perm filename F89.IN[LET,JMC] blob
sn#880612 filedate 1989-12-31 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00501 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00059 00002 ∂01-Oct-89 1340 donc@vaxa.isi.edu Re: looking for references about prototyping
C00061 00003 ∂01-Oct-89 1529 clm%zaphod.es.llnl.gov@lll-lcc.llnl.gov hello
C00066 00004 ∂02-Oct-89 1303 jutta@coyote.stanford.edu AI Division meeting
C00068 00005 ∂02-Oct-89 1328 ginsberg@Polya.Stanford.EDU nonmonotonic thesis problem?
C00070 00006 ∂02-Oct-89 1542 CLT closing escrow
C00071 00007 ∂02-Oct-89 1554 @Gang-of-Four.Stanford.EDU:arg@lucid.com new-qlisp
C00073 00008 ∂02-Oct-89 2129 qphysics-owner@neat.cs.toronto.edu An Interesting Puzzle
C00076 00009 ∂03-Oct-89 0314 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM the speed of life
C00082 00010 ∂03-Oct-89 0519 qphysics-owner@neat.cs.toronto.edu An Interesting Puzzle
C00084 00011 ∂03-Oct-89 0901 weening@Gang-of-Four.Stanford.EDU David Cyrluk
C00086 00012 ∂03-Oct-89 0942 qphysics-owner@neat.cs.toronto.edu Qualitative Reasoning Puzzle from QPHYSICS mailing list
C00093 00013 ∂03-Oct-89 1039 qphysics-owner@neat.cs.toronto.edu Re: Qualitative Reasoning Puzzle from QPHYSICS mailing list
C00099 00014 ∂03-Oct-89 1047 VAL Commonsense and Nonmonotonic Reasoning Seminar
C00101 00015 ∂03-Oct-89 1601 qphysics-owner@neat.cs.toronto.edu An Interesting Puzzle
C00105 00016 ∂03-Oct-89 1605 qphysics-owner@neat.cs.toronto.edu time fudge factor
C00108 00017 ∂03-Oct-89 2150 srh@flash.bellcore.com Searle and Connectionism
C00126 00018 ∂04-Oct-89 0243 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
C00146 00019 ∂04-Oct-89 1008 bhr@sumex-aim.stanford.edu CS300
C00150 00020 ∂04-Oct-89 1602 MPS
C00151 00021 ∂04-Oct-89 1638 srh@flash.bellcore.com Parallelism: Simulated and Real
C00159 00022 ∂06-Oct-89 0113 beeson@ucscd.UCSC.EDU an article you would want to see
C00161 00023 ∂06-Oct-89 0113 beeson@ucscd.UCSC.EDU said article
C00162 00024 ∂06-Oct-89 0535 menke@harvard.harvard.edu re: talk at harvard, october 2
C00164 00025 ∂06-Oct-89 0951 perlis@cs.UMD.EDU hi
C00168 00026 ∂06-Oct-89 1119 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU Elephant 2000
C00170 00027 ∂06-Oct-89 1159 rlg@ai.mit.edu Hertz Foundation Fellowship
C00172 00028 ∂06-Oct-89 1218 weening@Gang-of-Four.Stanford.EDU Software pipelining example
C00177 00029 ∂06-Oct-89 1244 srh@flash.bellcore.com Free-Floating Intentionality
C00197 00030 ∂06-Oct-89 1510 ME re: autologout
C00198 00031 ∂06-Oct-89 1532 harnad@clarity.Princeton.EDU Searle and Connectionism
C00216 00032 ∂06-Oct-89 1607 harnad@clarity.Princeton.EDU Dyer on Searle
C00264 00033 ∂06-Oct-89 1647 sl@van-bc.wimsey.bc.ca Re: Networks Considered Harmful - For Electronic Mail (fwd)
C00270 00034 ∂06-Oct-89 2027 srh@flash.bellcore.com re: Free-Floating Intentionality
C00272 00035 ∂08-Oct-89 0800 JMC
C00273 00036 ∂08-Oct-89 1051 boyer@CLI.COM Dinner
C00274 00037 ∂08-Oct-89 1146 nilsson@Tenaya.Stanford.EDU
C00282 00038 ∂08-Oct-89 1439 boyer@CLI.COM Dinner
C00283 00039 ∂09-Oct-89 0209 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
C00290 00040 ∂09-Oct-89 0932 @MCC.COM:EATON@MAXIMILLION.CP.MCC.COM last questions on arrangements for MCC visit
C00296 00041 ∂09-Oct-89 0932 @MCC.COM:EATON@MAXIMILLION.CP.MCC.COM list of attendees to MCC workshop, as of last week
C00302 00042 ∂09-Oct-89 0950 boyer@CLI.COM Dinner is on.
C00303 00043 ∂09-Oct-89 1419 arg@lucid.com bug fixed in new-qlisp
C00305 00044 ∂09-Oct-89 1625 @Score.Stanford.EDU,@ai.ai.mit.edu,@mc.lcs.mit.edu,@central.cis.upenn.edu:dale@linc.cis.upenn.edu CFP: Third Logical Biennial Chaika, Bulgaria
C00314 00045 ∂09-Oct-89 1630 MPS Keynote speaker
C00315 00046 ∂09-Oct-89 2026 H.HARPER@Macbeth.Stanford.EDU re: IN THE NEWS II
C00317 00047 ∂10-Oct-89 1156 SUSIE@MAXIMILLION.CP.MCC.COM Fall Associates Workshop
C00319 00048 ∂10-Oct-89 1202 ACW@ELEPHANT-BUTTE.SCRC.Symbolics.COM large periods
C00324 00049 ∂10-Oct-89 1327 0002814144@mcimail.com RE: test and ssx
C00326 00050 ∂10-Oct-89 1345 MPS
C00327 00051 ∂10-Oct-89 1353 betsy@russell.Stanford.EDU CSLI TINLunch
C00329 00052 ∂10-Oct-89 1413 betsy@russell.Stanford.EDU Re: date, title and abstract
C00331 00053 ∂10-Oct-89 1525 MPS size
C00332 00054 ∂10-Oct-89 1530 @IU.AI.SRI.COM,@peabody.teleos.com:leslie@teleos.com Procedural questions
C00335 00055 ∂10-Oct-89 1530 @IU.AI.SRI.COM,@peabody.teleos.com:leslie@teleos.com Meeting?
C00337 00056 ∂10-Oct-89 1632 MPS trip
C00338 00057 ∂10-Oct-89 2237 harnad@clarity.Princeton.EDU Verificationism and the Symbol Grounding Problem
C00358 00058 ∂11-Oct-89 0141 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
C00365 00059 ∂11-Oct-89 0645 Mailer failed mail returned
C00366 00060 ∂11-Oct-89 0856 kpeters@cdp.uucp Knott manuscript
C00368 00061 ∂11-Oct-89 1011 @Score.Stanford.EDU,@ai.ai.mit.edu,@MC.lcs.mit.edu:bundy%aipna.edinburgh.ac.uk@nsfnet-relay.ac.uk RA Post in AI Dept at Edinburgh
C00374 00062 ∂11-Oct-89 1204 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU Tuesday seminar
C00376 00063 ∂11-Oct-89 1357 weening@Gang-of-Four.Stanford.EDU Lucid work on Qlisp
C00380 00064 ∂11-Oct-89 1423 RWF re: Lyman talk, oct 26
C00381 00065 ∂11-Oct-89 1529 ACW@ELEPHANT-BUTTE.SCRC.Symbolics.COM large periods
C00391 00066 ∂11-Oct-89 1651 ginsberg@Polya.Stanford.EDU Review of Akawa's paper on presupposition
C00396 00067 ∂11-Oct-89 1733 nilsson@Tenaya.Stanford.EDU re: CIFE
C00398 00068 ∂11-Oct-89 2056 CLT CLT Itinerary San Diego / Ventura trip
C00399 00069 ∂12-Oct-89 0230 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
C00409 00070 ∂12-Oct-89 0602 CLT for sale
C00410 00071 ∂12-Oct-89 0607 CLT lathrop
C00412 00072 ∂12-Oct-89 1030 VAL testing the email connection
C00415 00073 ∂12-Oct-89 1052 nilsson@Tenaya.Stanford.EDU
C00423 00074 ∂12-Oct-89 1735 ACW@ELEPHANT-BUTTE.SCRC.Symbolics.COM large periods
C00426 00075 ∂12-Oct-89 1934 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM Speed comparison, 135 degree reflection, almost exponential periods
C00434 00076 ∂12-Oct-89 2211 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
C00437 00077 ∂13-Oct-89 1331 ACW@ELEPHANT-BUTTE.SCRC.Symbolics.COM large periods
C00440 00078 ∂13-Oct-89 1425 MLB@WHITE.SWW.Symbolics.COM composition contests
C00445 00079 ∂13-Oct-89 1539 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C00447 00080 ∂13-Oct-89 2035 underdog@Portia.stanford.edu employment
C00449 00081 ∂14-Oct-89 1930 uhlik@sun-valley.Stanford.EDU washer
C00451 00082 ∂14-Oct-89 1949 GOODMAN%uamis@rvax.ccit.arizona.edu sharing some praise
C00454 00083 ∂15-Oct-89 0814 beeson@ucscd.UCSC.EDU evolution
C00456 00084 ∂15-Oct-89 1118 feb6399@ultb.isc.rit.edu Question
C00459 00085 ∂15-Oct-89 1136 feb6399@ultb.isc.rit.edu one more thing
C00461 00086 ∂15-Oct-89 1746 0002814144@mcimail.com re: test and ssx
C00462 00087 ∂15-Oct-89 1941 pehoushe@Gang-of-Four.Stanford.EDU Primes with Qlambda
C00469 00088 ∂15-Oct-89 2246 rathmann@eclipse.stanford.edu Scheduling Orals
C00471 00089 ∂16-Oct-89 0800 JMC
C00472 00090 ∂16-Oct-89 0933 MPS phone
C00473 00091 ∂16-Oct-89 0934 CLT Monday, November 20, 1989
C00474 00092 ∂16-Oct-89 0942 CLT Visitor
C00476 00093 ∂16-Oct-89 1006 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU Re: Tuesday seminar
C00478 00094 ∂16-Oct-89 1037 helen@russell.Stanford.EDU re: Lunch Meeting
C00480 00095 ∂16-Oct-89 1045 ginsberg@Polya.Stanford.EDU re: Review of Akawa's paper on presupposition
C00482 00096 ∂16-Oct-89 1122 VAL DARPA visit on Nov. 20
C00483 00097 ∂16-Oct-89 1500 JJW Telnet from SAIL to Go4
C00485 00098 ∂16-Oct-89 1508 harnad@clarity.Princeton.EDU BBS Call for Commentators: Visual Field Specialization
C00491 00099 ∂16-Oct-89 1604 carol@lucid.com new new-qlisp
C00493 00100 ∂16-Oct-89 1634 ingrid@russell.Stanford.EDU IAP brochure
C00495 00101 ∂16-Oct-89 1705 carol@lucid.com new new-qlisp
C00497 00102 ∂16-Oct-89 1707 winograd@loire.stanford.edu December AI Qualifying Exam
C00504 00103 ∂16-Oct-89 1737 VAL Commonsense and Nonmonotonic Reasoning Seminar
C00506 00104 ∂16-Oct-89 1800 JMC
C00507 00105 ∂16-Oct-89 2001 JMC
C00508 00106 ∂17-Oct-89 0056 harnad@clarity.Princeton.EDU The Road to Beijing
C00518 00107 ∂17-Oct-89 0646 @Score.Stanford.EDU,@MCC.COM:greene@mcc.com Elephant 2000
C00520 00108 ∂17-Oct-89 0940 winograd@loire.stanford.edu elephant 2000
C00522 00109 ∂17-Oct-89 0947 AI.LENAT@MCC.COM Re: visit dates
C00524 00110 ∂17-Oct-89 1048 winograd@loire.stanford.edu re: elephant 2000
C00525 00111 ∂17-Oct-89 1400 JMC
C00526 00112 ∂17-Oct-89 1516 MPS Calls
C00527 00113 ∂17-Oct-89 1946 perlis@cs.UMD.EDU reference on context
C00529 00114 ∂18-Oct-89 1300 bledsoe@cs.utexas.edu Sorry I missed you
C00531 00115 ∂18-Oct-89 1429 Mailer re: quake
C00533 00116 ∂18-Oct-89 1607 Mailer events in Hungary
C00535 00117 ∂18-Oct-89 1840 @Score.Stanford.EDU:kirsh%cogsci@ucsd.edu aij
C00538 00118 ∂18-Oct-89 2027 CLT sarah
C00539 00119 ∂18-Oct-89 2042 Mailer re: events in Hungary
C00542 00120 ∂18-Oct-89 2259 ito@ito.ecei.tohoku.ac.jp Quake
C00544 00121 ∂18-Oct-89 2325 ME mail addresses
C00545 00122 ∂19-Oct-89 0912 MPS
C00546 00123 ∂19-Oct-89 1225 pehoushe@Gang-of-Four.Stanford.EDU Pat + meeting
C00548 00124 ∂19-Oct-89 1759 carol@lucid.com new new-qlisp
C00550 00125 ∂19-Oct-89 2015 shoham@hudson.Stanford.EDU Agent-Oriented Programming
C00558 00126 ∂19-Oct-89 2140 shoham@hudson.Stanford.EDU Re: Elephant
C00559 00127 ∂19-Oct-89 2154 mt@media-lab.media.mit.edu Elephant 2000
C00561 00128 ∂19-Oct-89 2223 mt@media-lab.media.mit.edu re: Elephant 2000
C00562 00129 ∂20-Oct-89 0700 ito@ito.ecei.tohoku.ac.jp R
C00565 00130 ∂20-Oct-89 0734 darden@cs.UMD.EDU ok?
C00567 00131 ∂20-Oct-89 1036 bledsoe@cs.utexas.edu re: Sorry I missed you
C00569 00132 ∂20-Oct-89 1053 lesperan@neat.cs.toronto.edu Elephant 2000
C00571 00133 ∂20-Oct-89 1320 nilsson@Tenaya.Stanford.EDU winter quarter course
C00575 00134 ∂20-Oct-89 1321 @IU.AI.SRI.COM,@sunset.ai.sri.com:felix@ai.sri.com re: Slingshot effect... (how does it work?)
C00577 00135 ∂20-Oct-89 1345 @Score.Stanford.EDU:james@cs.rochester.edu elephants
C00579 00136 ∂20-Oct-89 1421 james@cs.rochester.edu re: elephants
C00580 00137 ∂20-Oct-89 1441 ginsberg@Polya.Stanford.EDU Re: winter quarter course
C00582 00138 ∂20-Oct-89 1518 Mailer re: Slingshot effect... (how does it work?)
C00587 00139 ∂20-Oct-89 1529 VAL re: winter quarter course
C00589 00140 ∂20-Oct-89 1554 bledsoe@cs.utexas.edu re: Sorry I missed you
C00591 00141 ∂20-Oct-89 1715 Q4034%PUCC.BITNET@Forsythe.Stanford.EDU Tidal perturbations
C00594 00142 ∂21-Oct-89 0833 CLT driving lesson
C00595 00143 ∂21-Oct-89 1638 harnad@clarity.Princeton.EDU Sci. Am., you and Searle
C00621 00144 ∂21-Oct-89 2242 harnad@clarity.Princeton.EDU Two Ways to Refute Searle
C00627 00145 ∂22-Oct-89 0043 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM special (mental) cases
C00631 00146 ∂22-Oct-89 0158 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM tough nombres
C00635 00147 ∂22-Oct-89 0202 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM email address
C00637 00148 ∂22-Oct-89 0831 darden@cs.UMD.EDU re: ok?
C00639 00149 ∂22-Oct-89 1950 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C00641 00150 ∂22-Oct-89 1959 kirsh%cogsci@ucsd.edu re: aij
C00643 00151 ∂22-Oct-89 2211 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM An Ethiopian in the fuel supply.
C00646 00152 ∂23-Oct-89 0844 MPS
C00647 00153 ∂23-Oct-89 1000 JMC
C00648 00154 ∂23-Oct-89 1003 MPS Vote
C00649 00155 ∂23-Oct-89 1021 MPS books
C00650 00156 ∂23-Oct-89 1023 MPS lunch
C00651 00157 ∂23-Oct-89 1102 scales@Polya.Stanford.EDU CS 499
C00653 00158 ∂23-Oct-89 1107 scales@Polya.Stanford.EDU re: CS 499
C00654 00159 ∂23-Oct-89 1219 pedre@cs.utexas.edu Elephant 2000
C00656 00160 ∂23-Oct-89 1333 carol@lucid.com new new-qlisp
C00658 00161 ∂23-Oct-89 1540 davis@Polya.Stanford.EDU Lunch
C00659 00162 ∂23-Oct-89 1728 VAL Commonsense and Nonmonotonic Reasoning Seminar
C00660 00163 ∂23-Oct-89 1735 VAL Gelfond
C00661 00164 ∂23-Oct-89 1736 @IBM.COM:AMR@YKTVMH2 Note from AMR
C00675 00165 ∂24-Oct-89 0931 Mailer re: A Sense of Direction
C00680 00166 ∂24-Oct-89 0935 korf@CS.UCLA.EDU Re: Cindy Mason
C00682 00167 ∂24-Oct-89 0948 VAL Commonsense and Nonmonotonic Reasoning Seminar
C00683 00168 ∂24-Oct-89 1134 chandler@Polya.Stanford.EDU [Mail Delivery Subsystem <MAILER-DAEMON@tenaya.stanford.edu> :
C00687 00169 ∂24-Oct-89 1316 winograd@loire.stanford.edu re: elephant 2000
C00689 00170 ∂24-Oct-89 1533 Mailer re: A Sense of Direction
C00692 00171 ∂24-Oct-89 1541 bscott@Polya.Stanford.EDU Salary Research Charge
C00694 00172 ∂24-Oct-89 1624 pollack@cis.ohio-state.edu AAAI workshop grants
C00697 00173 ∂24-Oct-89 1650 Mailer re: A Sense of Direction
C00699 00174 ∂24-Oct-89 1922 paulf@bodega.Stanford.EDU Daily editorial
C00700 00175 ∂24-Oct-89 1956 paulf@bodega.Stanford.EDU re: Daily editorial
C00701 00176 ∂25-Oct-89 0054 roode@orc.olivetti.com [ je: From today's WSJ ]
C00706 00177 ∂25-Oct-89 0847 mkatz@sesame.Stanford.EDU Re: A Sense of Direction
C00709 00178 ∂25-Oct-89 0922 MPS
C00710 00179 ∂25-Oct-89 0934 Mailer re: A Sense of Direction
C00712 00180 ∂25-Oct-89 1009 CLT recursion
C00714 00181 ∂25-Oct-89 1239 katiyar@Polya.Stanford.EDU seminar next tuesday
C00716 00182 ∂25-Oct-89 1355 jones@Sunburn.Stanford.EDU Nonmonotonic Logic Course
C00718 00183 ∂25-Oct-89 1556 pehoushe@Gang-of-Four.Stanford.EDU QWAIT
C00720 00184 ∂25-Oct-89 1636 Mailer re: A Sense of Direction
C00724 00185 ∂25-Oct-89 1800 JMC
C00725 00186 ∂26-Oct-89 0524 LISTSERV@VM1.NoDak.EDU Message
C00726 00187 ∂26-Oct-89 1019 JCMA@REAGAN.AI.MIT.EDU Elephant 2000
C00728 00188 ∂26-Oct-89 1123 JCMA@REAGAN.AI.MIT.EDU re: Elephant 2000
C00730 00189 ∂26-Oct-89 1427 winograd@loire.stanford.edu December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
C00733 00190 ∂26-Oct-89 1753 S.SUMMER-RAIN@Macbeth.Stanford.EDU re: A Sense of Direction
C00736 00191 ∂26-Oct-89 2127 ma@src.dec.com Elephant paper?
C00738 00192 ∂26-Oct-89 2347 young@Neon.Stanford.EDU Elephant 2000 References?
C00740 00193 ∂27-Oct-89 0049 Mailer re: A Sense of Direction
C00742 00194 ∂27-Oct-89 0054 Mailer re: new earthquake concern
C00743 00195 ∂27-Oct-89 1720 Mailer re: A Sense of Direction
C00752 00196 ∂27-Oct-89 1725 bobrow.pa@Xerox.COM Elephant 2000
C00754 00197 ∂27-Oct-89 1734 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C00755 00198 ∂27-Oct-89 2000 JMC
C00756 00199 ∂28-Oct-89 0915 CLT driving
C00757 00200 ∂28-Oct-89 1138 christos%cs@ucsd.edu Re: ucsd visit
C00759 00201 ∂29-Oct-89 0900 JMC
C00760 00202 ∂30-Oct-89 1049 rlg@ai.mit.edu Hertz recommendation
C00762 00203 ∂30-Oct-89 1125 siegman@sierra.STANFORD.EDU Re: committee on administration
C00765 00204 ∂30-Oct-89 1436 bobrow.pa@Xerox.COM re: Elephant 2000
C00766 00205 ∂30-Oct-89 1753 woolf@venera.isi.edu verifying addresses
C00769 00206 ∂30-Oct-89 1901 VAL Commonsense and Nonmonotonic Reasoning Seminar
C00771 00207 ∂30-Oct-89 2019 jutta@coyote.stanford.edu reminder and agenda for AI Division meeting
C00774 00208 ∂31-Oct-89 0159 @loire.stanford.edu:marty@cis.Stanford.EDU December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
C00776 00209 ∂31-Oct-89 0536 CLT driving
C00777 00210 ∂31-Oct-89 0854 CLT driving
C00778 00211 ∂31-Oct-89 0908 katiyar@Neon.Stanford.EDU reminder
C00779 00212 ∂31-Oct-89 1324 @IBM.COM:LEORA@YKTVMH
C00781 00213 ∂31-Oct-89 1346 @loire.stanford.edu:hayes@kanga.parc.xerox.com December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
C00783 00214 ∂31-Oct-89 1553 sf@csli.Stanford.EDU
C00784 00215 ∂31-Oct-89 1553 VAL Gelfond
C00785 00216 ∂31-Oct-89 1621 sf@csli.Stanford.EDU Re: reply to message
C00787 00217 ∂31-Oct-89 1633 wab@sumex-aim.stanford.edu re: Busing
C00791 00218 ∂31-Oct-89 1718 hsu@Neon.Stanford.EDU Elephant 2000
C00792 00219 ∂31-Oct-89 2000 JMC
C00793 00220 ∂31-Oct-89 2000 JMC
C00794 00221 ∂01-Nov-89 0648 tom@Polya.Stanford.EDU Boise
C00796 00222 ∂01-Nov-89 0958 MPS
C00797 00223 ∂01-Nov-89 1201 JMC
C00798 00224 ∂01-Nov-89 1301 JMC
C00799 00225 ∂01-Nov-89 1624 paek@Neon.Stanford.EDU Elephant 2000 paper
C00800 00226 ∂01-Nov-89 1744 hayes@kanga.parc.xerox.com Vlad Dabija
C00802 00227 ∂02-Nov-89 0936 JMC
C00803 00228 ∂02-Nov-89 0948 hayes@kanga.parc.xerox.com Vlad Dabija
C00806 00229 ∂02-Nov-89 1219 MPS
C00807 00230 ∂02-Nov-89 1445 winograd@loire.stanford.edu AI QUAL SCHEDULED FOR WED. DECEMBER 6
C00810 00231 ∂02-Nov-89 1501 @IBM.COM:LEORA@YKTVMH book proposal
C00812 00232 ∂02-Nov-89 1507 MPS
C00813 00233 ∂03-Nov-89 1051 CLT qlisp
C00814 00234 ∂03-Nov-89 1501 CLT ∂02-Nov-89 1513 gilberts@Polya.Stanford.EDU Letters of Recommendation
C00816 00235 ∂03-Nov-89 1546 gumby@Gang-of-Four.Stanford.EDU Elephant 2000
C00817 00236 ∂03-Nov-89 1610 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU Reception for Matijasevitch and Girard
C00823 00237 ∂03-Nov-89 1610 stager@Polya.Stanford.EDU CS323 Winter Quarter
C00825 00238 ∂03-Nov-89 1632 MPS Room reservation
C00826 00239 ∂03-Nov-89 1735 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C00828 00240 ∂04-Nov-89 0121 B.BSK@Macbeth.Stanford.EDU re: One Week until Egg-Drop I
C00829 00241 ∂04-Nov-89 1037 beeson@ucscd.UCSC.EDU Re: Shrinking Lattice Polyhedra
C00834 00242 ∂04-Nov-89 1610 ACT Prancing Pony Bill
C00836 00243 ∂05-Nov-89 0029 ME NS
C00837 00244 ∂05-Nov-89 0939 ME AP
C00838 00245 ∂05-Nov-89 2259 alex@jessica.Stanford.EDU Thesis news AND questions
C00841 00246 ∂06-Nov-89 0515 cross@vax.darpa.mil Re: When
C00843 00247 ∂06-Nov-89 0617 @Score.Stanford.EDU,@MCC.COM:greene@mcc.com Elephant 2000
C00845 00248 ∂06-Nov-89 0825 stager@Polya.Stanford.EDU re: CS323 Winter Quarter
C00846 00249 ∂06-Nov-89 1047 perrault@ai.sri.com illocutionary vs perlocutionary
C00851 00250 ∂06-Nov-89 1219 MPS
C00852 00251 ∂06-Nov-89 1332 stager@Polya.Stanford.EDU re: CS323 Winter Quarter
C00854 00252 ∂06-Nov-89 1459 stager@Polya.Stanford.EDU re: CS323 Winter Quarter
C00855 00253 ∂06-Nov-89 1515 MPS
C00856 00254 ∂06-Nov-89 1636 COLLEEN@SUWATSON.stanford.edu Message from Arcady Blinov
C00858 00255 ∂06-Nov-89 1748 alex@jessica.Stanford.EDU re: Thesis news AND questions
C00859 00256 ∂06-Nov-89 1802 alex@jessica.Stanford.EDU re: Thesis news AND questions
C00861 00257 ∂06-Nov-89 2258 eaf@sumex-aim.stanford.edu [MAILER-DAEMON@sumex-aim.stanford.edu (Mail Delivery Subsystem) :
C00867 00258 ∂07-Nov-89 0717 beeson@ucscd.UCSC.EDU re: Shrinking Lattice Polyhedra
C00870 00259 ∂07-Nov-89 0827 MPS
C00871 00260 ∂07-Nov-89 1021 VAL Commonsense and Nonmonotonic Reasoning Seminar
C00874 00261 ∂07-Nov-89 1434 COLLEEN@SUWATSON.stanford.edu re: Message from Arcady Blinov
C00876 00262 ∂07-Nov-89 1435 pasquale%cs@ucsd.edu ucsd visit
C00878 00263 ∂08-Nov-89 0851 MPS Bureaucratic Meetings
C00879 00264 ∂08-Nov-89 1029 VAL Re: visit
C00881 00265 ∂08-Nov-89 1043 CLT qlisp
C00882 00266 ∂08-Nov-89 1055 Mailer re: The education president
C00891 00267 ∂08-Nov-89 1238 pasquale%cs@ucsd.edu re: ucsd visit
C00894 00268 ∂08-Nov-89 1347 POSTMASTER%McMaster.CA@Forsythe.Stanford.EDU Policies
C00897 00269 ∂08-Nov-89 1624 rick@hanauma.stanford.edu global warming seminar tonight 7:30 in Cubberly
C00898 00270 ∂08-Nov-89 1722 POSTMASTER%McMaster.CA@Forsythe.Stanford.EDU RE: re: Policies
C00902 00271 ∂09-Nov-89 0812 Mailer re: The education president
C00911 00272 ∂09-Nov-89 0913 beeson@ucscd.UCSC.EDU Elephant
C00913 00273 ∂09-Nov-89 0940 beeson@ucscd.UCSC.EDU cute little proof
C00915 00274 ∂09-Nov-89 1250 JJW SAIL DD monitors
C00916 00275 ∂09-Nov-89 1342 weening@Gang-of-Four.Stanford.EDU Re: SAIL DD monitors
C00918 00276 ∂09-Nov-89 1413 ME extra DD
C00919 00277 ∂09-Nov-89 1532 stager@Polya.Stanford.EDU re: CS323 Winter Quarter
C00921 00278 ∂10-Nov-89 0153 cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK E-mail address (cliff jones)
C00925 00279 ∂10-Nov-89 0907 bhr@sumex-aim.stanford.edu CS300 Winter Quarter
C00929 00280 ∂10-Nov-89 1119 paek@Neon.Stanford.EDU Nonmonotonic Reasoning class
C00931 00281 ∂10-Nov-89 1152 winograd@loire.stanford.edu AI QUAL SCHEDULE - ABOUT TO BE MADE OFFICIAL
C00934 00282 ∂10-Nov-89 1405 MPS phone call
C00935 00283 ∂10-Nov-89 1424 ME failed mail returned
C00937 00284 ∂10-Nov-89 1431 @loire.stanford.edu:hayes@kanga.parc.xerox.com AI QUAL SCHEDULE - ABOUT TO BE MADE OFFICIAL
C00939 00285 ∂10-Nov-89 1525 MPS
C00940 00286 ∂10-Nov-89 1548 VAL Notes on Cross's visit
C00941 00287 ∂11-Nov-89 2311 larson@unix.sri.com Re: The education president
C00945 00288 ∂12-Nov-89 1341 perlis@cs.UMD.EDU Putnam's Theorem
C00948 00289 ∂12-Nov-89 1414 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C00951 00290 ∂12-Nov-89 1649 gstuck@note.nsf.gov Re: electronic reviewing
C00953 00291 ∂13-Nov-89 0553 qphysics-owner@neat.cs.toronto.edu QP-Workshop 1990 - Call for papers
C00959 00292 ∂13-Nov-89 1036 korf@CS.UCLA.EDU Stanford Visit
C00961 00293 ∂13-Nov-89 1701 VAL Commonsense and nonmonotonic reasoning seminar - no meeting
C00962 00294 ∂14-Nov-89 0111 root@Gang-of-Four.Stanford.EDU
C00969 00295 ∂14-Nov-89 0800 JMC
C00970 00296 ∂14-Nov-89 1054 nilsson@Tenaya.Stanford.EDU reminder
C00976 00297 ∂14-Nov-89 1600 tajnai@Polya.Stanford.EDU AI Day on Campus, 6/7/90
C00980 00298 ∂14-Nov-89 1638 sanu@cs.utexas.edu elephant 2000
C00982 00299 ∂14-Nov-89 1649 sanu@cs.utexas.edu re: elephant 2000
C00984 00300 ∂14-Nov-89 1754 rdz@Gang-of-Four.Stanford.EDU MIT talk of possible interest
C00987 00301 ∂15-Nov-89 1003 chandler@Polya.Stanford.EDU Faculty Reports
C00990 00302 ∂15-Nov-89 1308 chandler@Polya.Stanford.EDU DARPA Visit Agenda
C00996 00303 ∂15-Nov-89 1359 kirsh%cogsci@ucsd.edu hewitt
C01063 00304 ∂15-Nov-89 1401 kirsh%cogsci@ucsd.edu part 2 hewitt
C01105 00305 ∂15-Nov-89 1817 jlh@vsop.Stanford.EDU Tuesday schedule for DARPA visit
C01108 00306 ∂15-Nov-89 2117 fuller@sierra.STANFORD.EDU Myrhvold@microsoft
C01110 00307 ∂16-Nov-89 0650 CLT show and tell
C01111 00308 ∂16-Nov-89 1058 pehoushe@Gang-of-Four.Stanford.EDU New low-cost extensions to Qlisp
C01114 00309 ∂16-Nov-89 1158 RPG@SAIL.Stanford.EDU New Qlisp Stuff
C01117 00310 ∂16-Nov-89 1206 iam@Gang-of-Four.Stanford.EDU New Qlisp Stuff
C01119 00311 ∂16-Nov-89 1252 RPG@SAIL.Stanford.EDU New Qlisp Stuff
C01122 00312 ∂16-Nov-89 1254 rpg@lucid.com New Qlisp Stuff
C01124 00313 ∂16-Nov-89 1336 RPG@SAIL.Stanford.EDU New Qlisp Stuff
C01126 00314 ∂16-Nov-89 1402 SJG (on TTY73, at TV-112 1402)
C01127 00315 ∂17-Nov-89 0817 Mailer Re: Fitzwater speaks
C01129 00316 ∂17-Nov-89 0859 MPS Faculty Report
C01130 00317 ∂17-Nov-89 0959 VAL Seminar at Berkeley
C01131 00318 ∂17-Nov-89 1141 ceb@might.Stanford.EDU possibility of a referee?
C01136 00319 ∂17-Nov-89 1533 VAL job hunting news
C01138 00320 ∂17-Nov-89 1545 MPS
C01139 00321 ∂17-Nov-89 1555 pehoushe@Gang-of-Four.Stanford.EDU The time of basic operations
C01142 00322 ∂17-Nov-89 1643 Mailer re: A Sense of Direction
C01145 00323 ∂17-Nov-89 1650 rpg@lucid.com The time of basic operations
C01150 00324 ∂17-Nov-89 1717 VAL Boehm's visit
C01151 00325 ∂17-Nov-89 1957 Mailer re: Berlin@Deutschland
C01153 00326 ∂17-Nov-89 2257 GLB
C01155 00327 ∂18-Nov-89 0600 CLT@SAIL.Stanford.EDU qlet or not qlet
C01157 00328 ∂18-Nov-89 0938 RPG@SAIL.Stanford.EDU re: qlet or not qlet
C01161 00329 ∂18-Nov-89 1000 JMC
C01162 00330 ∂18-Nov-89 1044 @SAIL.Stanford.EDU:jlm@lucid.com qlet or not qlet
C01167 00331 ∂18-Nov-89 1114 CLT@SAIL.Stanford.EDU qlet or not qlet
C01169 00332 ∂18-Nov-89 1406 @SAIL.Stanford.EDU:rpg@lucid.com qlet or not qlet
C01172 00333 ∂18-Nov-89 1938 pehoushe@Gang-of-Four.Stanford.EDU processes and functions
C01178 00334 ∂19-Nov-89 0847 RPG reply to message
C01179 00335 ∂19-Nov-89 1141 rpg@lucid.com processes and functions
C01184 00336 ∂19-Nov-89 1548 RPG@SAIL.Stanford.EDU Restrictions on Process Functions
C01187 00337 ∂19-Nov-89 1846 RPG Boehm
C01190 00338 ∂19-Nov-89 2148 rz@cs.cornell.edu Parallel Computer Algebra Workshop
C01195 00339 ∂20-Nov-89 0900 JMC
C01196 00340 ∂20-Nov-89 0938 pehoushe@Gang-of-Four.Stanford.EDU Implicit Closures or Process Functions
C01200 00341 ∂20-Nov-89 1008 rpg@lucid.com Implicit Closures or Process Functions
C01204 00342 ∂20-Nov-89 1151 pehoushe@Gang-of-Four.Stanford.EDU Implicit Closures or Process Functions
C01210 00343 ∂20-Nov-89 1543 rpg@lucid.com Implicit Closures or Process Functions
C01214 00344 ∂20-Nov-89 1543 rpg@lucid.com Round Table
C01215 00345 ∂20-Nov-89 1754 ME mail address
C01216 00346 ∂20-Nov-89 1843 VAL Commonsense and Nonmonotonic Reasoning Seminar
C01218 00347 ∂21-Nov-89 0155 weening@Gang-of-Four.Stanford.EDU
C01262 00348 ∂21-Nov-89 0634 bajcsy@central.cis.upenn.edu Hello
C01270 00349 ∂21-Nov-89 0733 beeson@ucscd.UCSC.EDU Re: Thanks for your paper.
C01272 00350 ∂21-Nov-89 0738 beeson@ucscd.UCSC.EDU learning
C01275 00351 ∂21-Nov-89 0901 Mailer re: silver lining to the quake
C01277 00352 ∂21-Nov-89 1005 MPS Expenses
C01278 00353 ∂21-Nov-89 1045 pehoushe@Gang-of-Four.Stanford.EDU qlet& and qlet
C01282 00354 ∂21-Nov-89 1114 VAL re: book
C01283 00355 ∂21-Nov-89 1901 VAL re: book
C01284 00356 ∂21-Nov-89 1914 ME SAIL
C01285 00357 ∂21-Nov-89 1926 cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK Lifschitz
C01287 00358 ∂21-Nov-89 2020 rpg@lucid.com qlet& and qlet
C01289 00359 ∂21-Nov-89 2048 ellis@src.dec.com Distribution of "su" newsgroups
C01291 00360 ∂22-Nov-89 0717 weening@Gang-of-Four.Stanford.EDU
C01293 00361 ∂24-Nov-89 0937 cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK Lifschitz
C01296 00362 ∂24-Nov-89 1415 harnad@clarity.Princeton.EDU Connectionist Learning/Representation: BBS Call for Commentators
C01302 00363 ∂24-Nov-89 1446 VAL
C01303 00364 ∂24-Nov-89 2039 ME second DD monitor in your office
C01306 00365 ∂25-Nov-89 1338 rpg@lucid.com BAA
C01309 00366 ∂25-Nov-89 2210 ginsberg@Polya.Stanford.EDU Penrose
C01311 00367 ∂26-Nov-89 0859 ginsberg@Polya.Stanford.EDU re: Penrose
C01313 00368 ∂26-Nov-89 1358 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C01315 00369 ∂26-Nov-89 1556 VAL Elephant
C01316 00370 ∂26-Nov-89 1737 VAL reply to message
C01317 00371 ∂27-Nov-89 0800 JMC
C01318 00372 ∂27-Nov-89 1107 chandler@Polya.Stanford.EDU Faculty reports....again
C01320 00373 ∂27-Nov-89 1216 ross%cs@ucsd.edu Distinguished Lecturer Expenses
C01322 00374 ∂27-Nov-89 1722 RPG CV
C01323 00375 ∂27-Nov-89 1744 winograd@loire.stanford.edu AI QUAL SCHEDULE
C01325 00376 ∂27-Nov-89 1745 VAL Commonsense and Nonmonotonic Reasoning Seminar
C01328 00377 ∂27-Nov-89 2312 golub@na-net.stanford.edu Re: Faculty reports....again
C01330 00378 ∂28-Nov-89 1005 tajnai@Polya.Stanford.EDU Daniel Scales
C01331 00379 ∂28-Nov-89 1044 tajnai@Polya.Stanford.EDU re: Daniel Scales
C01333 00380 ∂29-Nov-89 0915 MPS Vacation
C01334 00381 ∂29-Nov-89 1235 MPS
C01335 00382 ∂30-Nov-89 0824 MPS MCC
C01336 00383 ∂30-Nov-89 1059 qphysics-owner@neat.cs.toronto.edu Mailing List
C01339 00384 ∂30-Nov-89 1112 VAL re: Elephant programs as sentences
C01341 00385 ∂30-Nov-89 1125 GLB
C01346 00386 ∂30-Nov-89 1332 shoham@Hudson.Stanford.EDU agent-oriented programming
C01347 00387 ∂30-Nov-89 1455 winograd@loire.stanford.edu Locations and lunch for AI qual
C01350 00388 ∂01-Dec-89 1224 @Score.Stanford.EDU:kirsh%cogsci@ucsd.edu aij
C01353 00389 ∂01-Dec-89 1807 kirsh%cogsci@ucsd.edu re: aij
C01356 00390 ∂02-Dec-89 1527 kirsh%cogsci@ucsd.edu hewitt part1
C01409 00391 ∂02-Dec-89 1528 kirsh%cogsci@ucsd.edu hewitt part2
C01464 00392 ∂03-Dec-89 1057 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C01467 00393 ∂03-Dec-89 2355 mdbomber@Portia.stanford.edu re: Bozo Bush vs. Chinese students [was Re: Recent su.etc Statistics]
C01468 00394 ∂04-Dec-89 1000 JMC
C01469 00395 ∂04-Dec-89 1138 scherlis@vax.darpa.mil NSF-DARPA Formal Methods
C01475 00396 ∂04-Dec-89 1820 VAL Commonsense and Nonmonotonic Reasoning Seminar
C01477 00397 ∂04-Dec-89 1904 RWF re: World War II
C01478 00398 ∂04-Dec-89 2211 ACT Prancing Pony Bill
C01480 00399 ∂04-Dec-89 2256 harnad@clarity.Princeton.EDU Emperor's New Mind: BBS Call for Commentators
C01485 00400 ∂05-Dec-89 0630 CLT CLT itinerary
C01486 00401 ∂05-Dec-89 0647 CLT cate
C01487 00402 ∂05-Dec-89 0914 phil@ub.d.umn.edu Re: Philosophy and Artificial Intelligence
C01491 00403 ∂05-Dec-89 0915 MPS
C01492 00404 ∂05-Dec-89 0925 MPS
C01493 00405 ∂05-Dec-89 1000 JMC
C01494 00406 ∂05-Dec-89 1225 MPS
C01495 00407 ∂05-Dec-89 1300 JMC
C01496 00408 ∂05-Dec-89 1437 bscott@Polya.Stanford.EDU Your Questions
C01498 00409 ∂05-Dec-89 1448 lamport@src.dec.com Re: Dec connections
C01500 00410 ∂05-Dec-89 1719 peters@russell.Stanford.EDU [Takayasu Ito: Your Proposal]
C01503 00411 ∂05-Dec-89 1804 winograd@loire.stanford.edu REMINDER - AI QUAL TOMORROW (WED) MORNING AT 8:45
C01504 00412 ∂05-Dec-89 1951 kirsh%cogsci@ucsd.edu no comment
C01506 00413 ∂06-Dec-89 0836 rabin@harvard.harvard.edu Re: sabbatical
C01508 00414 ∂06-Dec-89 0900 JMC
C01509 00415 ∂06-Dec-89 1221 kirsh%cogsci@ucsd.edu re: no comment
C01511 00416 ∂06-Dec-89 1315 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU MTC qual reading list
C01513 00417 ∂06-Dec-89 1322 hemenway@Sunburn.Stanford.EDU Re: MTC qual reading list
C01515 00418 ∂06-Dec-89 1505 @Polya.Stanford.EDU:lincoln@ghoti.Stanford.EDU RE: MTC qual reading list
C01522 00419 ∂06-Dec-89 1633 @Polya.Stanford.EDU:lincoln@ghoti.Stanford.EDU RE: MTC qual reading list
C01524 00420 ∂06-Dec-89 1706 ortiz@itstd.sri.com Thanks
C01526 00421 ∂06-Dec-89 1936 ash@sumex-aim.stanford.edu Qual
C01528 00422 ∂06-Dec-89 2100 kirsh%cogsci@ucsd.edu re: no comment
C01530 00423 ∂07-Dec-89 0942 MPS FAX
C01532 00424 ∂07-Dec-89 1012 MPS
C01533 00425 ∂07-Dec-89 1100 MPS DMV
C01534 00426 ∂07-Dec-89 1356 VAL Manchester
C01541 00427 ∂07-Dec-89 1610 VAL re: Elephant programs as sentences
C01542 00428 ∂07-Dec-89 1719 bergman@Sunburn.Stanford.EDU proposals
C01544 00429 ∂08-Dec-89 0809 hewitt@ai.mit.edu aij
C01546 00430 ∂08-Dec-89 0835 MPS Fruitfly paper
C01547 00431 ∂08-Dec-89 0846 MPS
C01548 00432 ∂08-Dec-89 1043 shoham@Hudson.Stanford.EDU cheap pun
C01549 00433 ∂08-Dec-89 1309 MPS
C01550 00434 ∂08-Dec-89 1332 cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK Lifschitz
C01553 00435 ∂08-Dec-89 1441 VAL re: Lifschitz
C01554 00436 ∂08-Dec-89 1548 kirsh%cogsci@ucsd.edu hewitt
C01556 00437 ∂08-Dec-89 1809 scales@Gang-of-Four.Stanford.EDU OPS5 in QLISP
C01567 00438 ∂10-Dec-89 0946 hewitt@ai.mit.edu aij
C01570 00439 ∂10-Dec-89 1219 ME Here is the text of AIJ.TEX[1,CDR] (1417 lines)
C01571 00440 ∂10-Dec-89 1716 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C01573 00441 ∂11-Dec-89 0009 kirsh%cogsci@ucsd.edu [hewitt@ai.mit.edu: My AIJ paper]
C01634 00442 ∂11-Dec-89 0016 kirsh%cogsci@ucsd.edu 2 of Hewitt
C01687 00443 ∂11-Dec-89 1100 JMC
C01688 00444 ∂11-Dec-89 1250 siegman@sierra.STANFORD.EDU Re: Stanford costs and overhead
C01693 00445 ∂11-Dec-89 2000 JMC
C01694 00446 ∂12-Dec-89 1110 harnad@clarity.Princeton.EDU BBS Penrose Book Review
C01700 00447 ∂12-Dec-89 1146 VAL index
C01701 00448 ∂12-Dec-89 1252 harnad@clarity.Princeton.EDU re: BBS Penrose Book Review
C01703 00449 ∂12-Dec-89 1300 JMC
C01704 00450 ∂12-Dec-89 1307 scales@Neon.Stanford.EDU Re: cs499
C01706 00451 ∂12-Dec-89 1309 harnad@clarity.Princeton.EDU re: BBS Penrose Book Review
C01708 00452 ∂12-Dec-89 1613 weening@Gang-of-Four.Stanford.EDU Text needed for DARPA proposal
C01710 00453 ∂12-Dec-89 1754 PKR re: chapter 4
C01712 00454 ∂13-Dec-89 0939 MPS fruitfly
C01713 00455 ∂13-Dec-89 1012 @CUNYVM.CUNY.EDU:fraenkel@wisdom.weizmann.ac.il
C01715 00456 ∂13-Dec-89 1701 MPS Paper
C01716 00457 ∂14-Dec-89 0801 HK.RLS@Forsythe.Stanford.EDU memorandum
C01717 00458 ∂14-Dec-89 1304 CLT bing
C01718 00459 ∂14-Dec-89 1424 irvine@sumex-aim.stanford.edu Betty Scott
C01721 00460 ∂14-Dec-89 1438 stager@Sunburn.Stanford.EDU Winter TA
C01723 00461 ∂14-Dec-89 1731 @Sunburn.Stanford.EDU,@Polya.Stanford.EDU:nilsson@Tenaya.Stanford.EDU reappointments
C01727 00462 ∂15-Dec-89 0821 VAL lunch
C01728 00463 ∂15-Dec-89 0916 VAL re: lunch
C01729 00464 ∂15-Dec-89 1140 paek@Neon.Stanford.EDU re: Winter TA
C01731 00465 ∂15-Dec-89 1655 jolle@cs.psu.edu
C01738 00466 ∂17-Dec-89 0731 jf@research.att.com
C01740 00467 ∂17-Dec-89 1224 VAL
C01741 00468 ∂18-Dec-89 1213 VAL re: reply to message
C01742 00469 ∂18-Dec-89 1322 VAL Re: openings at Berkeley
C01747 00470 ∂18-Dec-89 1457 MPS Car
C01748 00471 ∂18-Dec-89 1610 VAL Towers
C01749 00472 ∂18-Dec-89 2000 JMC
C01750 00473 ∂19-Dec-89 1209 masahiko@sato.riec.tohoku.ac.jp NFS/JSPS proposal
C01753 00474 ∂19-Dec-89 1255 CLT NFS/JSPS proposal
C01755 00475 ∂19-Dec-89 1640 @IU.AI.SRI.COM,@sunset.ai.sri.com:konolige@ai.sri.com nonmon90 workshop
C01757 00476 ∂19-Dec-89 1659 weening@Gang-of-Four.Stanford.EDU Alliant visit
C01759 00477 ∂19-Dec-89 1915 @IU.AI.SRI.COM,@sunset.ai.sri.com:konolige@ai.sri.com re: nonmon90 workshop
C01765 00478 ∂20-Dec-89 1551 GLB
C01767 00479 ∂20-Dec-89 1553 GLB
C01773 00480 con-Dec-89 0827 beeson@ucscd.UCSC.EDU autism and common-sense reasoning
C01775 00481 ∂21-Dec-89 1047 MPS
C01776 00482 ∂21-Dec-89 1209 phil@ub.d.umn.edu Philosophy and Artificial Intelligence
C01780 00483 ∂21-Dec-89 1418 shankar@argon.csl.sri.com thesis
C01785 00484 ∂21-Dec-89 1423 beeson@ucscd.UCSC.EDU rationalizing what we are going to do anyway
C01788 00485 ∂22-Dec-89 1039 VAL
C01789 00486 ∂22-Dec-89 1110 Mailer re: Castro?
C01790 00487 ∂22-Dec-89 1145 MPS interview
C01791 00488 ∂22-Dec-89 1326 weening@Gang-of-Four.Stanford.EDU
C01796 00489 ∂22-Dec-89 1448 MPS MERRY OLE' CHRISTMAS
C01798 00490 ∂22-Dec-89 1535 shankar@argon.csl.sri.com Thanks!
C01800 00491 ∂26-Dec-89 0527 jsl%pres_res.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU E-mail
C01802 00492 ∂26-Dec-89 1056 jsl%casp1.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU Re: Final version of Networks considered harmful.
C01806 00493 ∂26-Dec-89 1213 jsl%casp1.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU Re: Final version of Networks considered harmful.
C01901 00494 ∂27-Dec-89 1033 eaf@sumex-aim.stanford.edu a merry christmas gift arrived
C01903 00495 ∂27-Dec-89 1334 bscott@Polya.Stanford.EDU Sabbatical Credit
C01906 00496 ∂27-Dec-89 1603 JMC wrong number?
C01907 00497 ∂28-Dec-89 1149 stager@Sunburn.Stanford.EDU CS499 grade sheet
C01909 00498 ∂28-Dec-89 1409 JMC dinner
C01910 00499 ∂28-Dec-89 1420 JMC cars
C01911 00500 ∂28-Dec-89 1508 JMC books
C01912 00501 ∂28-Dec-89 1529 JMC re: books
C01918 ENDMK
C⊗;
∂01-Oct-89 1340 donc@vaxa.isi.edu Re: looking for references about prototyping
Received: from vax.darpa.mil ([26.6.0.106]) by SAIL.Stanford.EDU with TCP; 1 Oct 89 13:40:25 PDT
Received: from vaxa.isi.edu by vax.darpa.mil (5.61/5.61+local)
id <AA10920>; Sun, 1 Oct 89 15:11:55 -0400
Posted-Date: Sun, 01 Oct 89 12:11:09 PDT
Message-Id: <8910011911.AA05128@vaxa.isi.edu>
Received: from LOCALHOST by vaxa.isi.edu (5.61/5.61)
id AA05128; Sun, 1 Oct 89 12:11:11 -0700
To: boyland@sequoia.Berkeley.EDU, cps@vax.darpa.mil
Subject: Re: looking for references about prototyping
Date: Sun, 01 Oct 89 12:11:09 PDT
From: Don Cohen <donc@vaxa.isi.edu>
As long as we're using this mailing list to spread references
(seems like a good initial use), I'll put in my 2 cents worth
for ap5, an extension to commonlisp which provides database-like
abstractions and facilities. Parts of it are described in
AAAI '86 and SigMod '89 (papers by Cohen); more documentation
is available too.
∂01-Oct-89 1529 clm%zaphod.es.llnl.gov@lll-lcc.llnl.gov hello
Received: from zaphod.es.llnl.gov ([128.115.4.30]) by SAIL.Stanford.EDU with TCP; 1 Oct 89 15:29:07 PDT
Received: by zaphod.es.llnl.gov (4.0/1.26)
id AA01192; Sun, 1 Oct 89 15:32:06 PDT
Date: Sun, 1 Oct 89 15:32:06 PDT
From: clm%zaphod.es.llnl.gov@lll-lcc.llnl.gov (Cynthia Mason 422-8911)
Message-Id: <8910012232.AA01192@zaphod.es.llnl.gov>
To: mccarthy@sail.stanford.edu
Subject: hello
Sorry I haven't been in touch for so long...
I left here Sept 9 for a workshop in Seattle, which was great..
the thesis work won a prize and all that, which was good news.
But from where I lie, it seems really insignificant...after
the workshop, I found out my fellowship had been extended for
another year. This also was good news. So, I decided to take a
long weekend, (Sept 15) and go hiking. Well...
I've been
flat on my back for most of the last week and a half or so...
The doctors can't seem to figure it out (duh...), although
listen to this...as a result of going up to Mt. Whitney, by car,
10,000 feet, I got joint pain, headache, EXTREME fatigue, cerebral
malfunction, (and the latest symptom is my entire leg went numb)
These are the same symptoms I had when I got bent..
You might recall, I got bent
in January, and had undergone decompression treatment twice already,
for this.
The docs seem to feel its been too long since the initial insult
to be related...that is, their theory is that the bubbles have
gone away by now. So, they're testing me for all kinds
of diseases and weirdnesses, all coming out negative.
It just seems so obvious to me...this has to be related to my previous
decompression sickness. But they have this notion its
been too long.
I've done enough reading myself to know the state of diving medicine
is in chaos...ie. they're constantly updating the dive tables,
there's some evidence the dive recompression tables themselves
are wrong, and we know there is a lot of variability between
individuals, and within an individual him/herself (Tuesday if
you dive, you'll get bent, but on Wednesday, you won't).
And
the doctors have no explanation for why I got bent in the first
place, but I did. So...here I sit, messed up again from going to
altitude. But I need to convince these guys it HASN"T been to
long for this to be a possibility.
If you know any doctors or diving research people, you might
ask them if they know of any current research about longevity of
the microbubbles, or usefulness of decompression treatment, etc.
I would be truly grateful....
Well, I guess I better go get horizontal now (sigh)...
(I'd rather be working on my thesis)
Bubbles
∂02-Oct-89 1303 jutta@coyote.stanford.edu AI Division meeting
Received: from coyote.stanford.edu by SAIL.Stanford.EDU with TCP; 2 Oct 89 13:03:38 PDT
Received: by coyote.stanford.edu; Mon, 2 Oct 89 13:02:27 PDT
Date: 2 Oct 1989 1302-PDT (Monday)
From: Jutta McCormick <jutta@coyote.stanford.edu>
To: binford@coyote.stanford.edu, feigenbaum@SUMEX-AIM.Stanford.EDU,
genesereth@polya.stanford.edu, jones@polya.stanford.edu,
ok@coyote.stanford.edu, jmc@sail.stanford.edu
Cc: latombe@coyote.stanford.edu, nilsson@tenaya.stanford.edu,
winograd@csli.stanford.edu, shoham@polya.stanford.edu,
jutta@coyote.stanford.edu
Subject: AI Division meeting
Jean-Claude Latombe is calling the first meeting of the Computer Science AI
Division for Wednesday, November 1, 1:30 - 3:00 p.m., in the Cedar Hall
conference room. One of the agenda items will be "preparation for the
Visiting Committee." Please send me (jutta@coyote) any additional items you
would like to have discussed. The final agenda will be distributed a few
days before the meeting.
Please let me know right away if you will be able to attend the meeting.
Jutta McCormick
∂02-Oct-89 1328 ginsberg@Polya.Stanford.EDU nonmonotonic thesis problem?
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 2 Oct 89 13:26:39 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA09408; Mon, 2 Oct 89 13:25:20 -0700
Date: Mon, 2 Oct 89 13:25:20 -0700
From: Matthew L. Ginsberg <ginsberg@Polya.Stanford.EDU>
Message-Id: <8910022025.AA09408@Polya.Stanford.EDU>
To: feed@Polya.Stanford.EDU, principia@Polya.Stanford.EDU
Subject: nonmonotonic thesis problem?
I've had an idea for choosing among competing extensions that I don't think
I'm going to have the time to investigate over the course of the next few
months. Anyone looking for a potential thesis topic?
Matt
∂02-Oct-89 1542 CLT closing escrow
we have a appointment to sign papers Tuesday Oct 10 at 4:30pm.
If this is not good let me know asap so we can rearrange.
∂02-Oct-89 1554 @Gang-of-Four.Stanford.EDU:arg@lucid.com new-qlisp
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 2 Oct 89 15:54:08 PDT
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU with SMTP (5.61/25-eef) id AA05859; Mon, 2 Oct 89 15:53:18 -0700
Received: from bhopal ([192.43.178.13]) by heavens-gate id AA10224g; Mon, 2 Oct 89 15:52:38 PDT
Received: by bhopal id AA11121g; Mon, 2 Oct 89 15:54:44 PDT
Date: Mon, 2 Oct 89 15:54:44 PDT
From: Ron Goldman <arg@lucid.com>
Message-Id: <8910022254.AA11121@bhopal>
To: qlisp@go4.stanford.edu
Subject: new-qlisp
If you are trying to use new-qlisp, you need to recompile your code. Generally
when we bring up a new version of qlisp/new-qlisp you need to recompile. In
particular new-qlisp now does a check on remaining stack space on entry to
each function; old code doesn't have this check.
Ron
∂02-Oct-89 2129 qphysics-owner@neat.cs.toronto.edu An Interesting Puzzle
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 2 Oct 89 21:28:53 PDT
Received: by neat.cs.toronto.edu id 2384; Tue, 3 Oct 89 00:29:22 EDT
Received: from nooksack.cs.washington.edu by neat.cs.toronto.edu with SMTP id 2277; Tue, 3 Oct 89 00:25:31 EDT
Received: by nooksack.cs.washington.edu (5.52.1/6.13)
id AA04373; Mon, 2 Oct 89 21:26:05 PDT
Date: Tue, 3 Oct 89 00:26:05 EDT
From: weld@nooksack.cs.washington.edu (Dan Weld)
Message-Id: <8910030426.AA04373@nooksack.cs.washington.edu>
To: qphysics@ai.toronto.edu
Subject: An Interesting Puzzle
Resent-From: qphysics-owner@cs.toronto.edu
Resent-To: qphysics@cs.toronto.edu
Resent-Reply-To: weld@nooksack.cs.washington.edu (Dan Weld)
Resent-Message-Id: <89Oct3.002922edt.2384@neat.cs.toronto.edu>
Resent-Date: Tue, 3 Oct 89 00:27:00 EDT
Assuming that most of the readers of this list are in the northern
hemisphere, most readers are lamenting the increasingly shorter days of
fall. But did everyone realize that the change in daylight hours is not
symmetric? On any given day the sun rises N minutes later and sets M
minutes earlier, but M is not necessarily equal to N!
Any good qualitative explanations for this mystery???
Dan
∂03-Oct-89 0314 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM the speed of life
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 3 Oct 89 03:14:49 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 426602; 3 Oct 89 06:06:25 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 109525; Tue 3-Oct-89 03:00:41 PDT
Date: Tue, 3 Oct 89 03:00 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: the speed of life
To: "acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
cc: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
math-fun@RUSSIAN.SPA.Symbolics.COM, "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <8909231839.AA05616@megalon.acad.com>
Message-ID: <19891003100034.5.RWG@TSUNAMI.SPA.Symbolics.COM>
Date: Sat, 23 Sep 89 11:39:41 PDT
From: acad!megalon!rudy@uunet.UU.NET (Rudy Rucker)
ok. I timed CA lab in 320 by 200 mode and we get .193 secs/gen.
Jeez, that's faster than my bitblt program!
In
the 80 by 100 mode it goes a lot faster, but I didn't think of putting
a generation counter. When you speak of your "bit blt" program is
the idea that you do boolean logic on whole planes of bits at once?
Yes. Workstations (playstations) generally have special microcode and
even datapaths for souping up boolean ops on bit rectangles. My life
program does about 32 bitblts/gen, which works out to about one machine
instruction per bit, regardless of rectangle size.
This suggests that your program must be doing several bits at once.
I've heard of an atari ca program that works this way.
Long after the SmallTalkers did it where the Future got Fumbled.
This scheme is attractive for its speed and continuous scalability. With
O(log(width)) bitblts, you can even rotate, transpose, or sort bit arrays
into odd/even columns.
I got some interesting papers yesterday from a guy called Per Bak. He
works at Brookhaven Nat Lab and his buzzword is Self-Organized Critical
Phenomena. He did some interesting statistical studies, one of which
was to take a died-out life board and then turn on one pixel and see
what happens. He defines the "total activity" to be the number of births and
deaths following such a perturbation. He gets some kind of fractal power
law. Have you heard of this stuff?
Sounds vaguely familiar, and vaguely bogus. I seem to recall that died-out
areas have a density of only about 2.8%, so he should be getting rather
sparse results. Yet on the other hand, for sufficiently large fields, there
should be no permanently died-out areas, and there should be a finite, but
small probability of a perturbation initiating infinitely many births and
deaths. Can a "died out area" contain a pulsar? A pentadecathlon? A gun
dumping into an eater? Tweak one bit in the eater, and boom.
Did we send you a copy of CA Lab?
Long ago. But we didn't have 386s at Symbolics then, so I had Weyhrauch
take it to IBUKI, and then never got over there.
∂03-Oct-89 0519 qphysics-owner@neat.cs.toronto.edu An Interesting Puzzle
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 3 Oct 89 05:19:23 PDT
Received: by neat.cs.toronto.edu id 2629; Tue, 3 Oct 89 08:18:55 EDT
Received: from Princeton.EDU by neat.cs.toronto.edu with SMTP id 2153; Tue, 3 Oct 89 08:13:38 EDT
Received: from notecnirp.Princeton.EDU by Princeton.EDU (5.58+++/2.22)
id AA20285; Tue, 3 Oct 89 08:13:09 EDT
Received: by notecnirp.Princeton.EDU (5.51/1.85)
id AA24279; Tue, 3 Oct 89 08:12:57 EDT
Date: Tue, 3 Oct 89 08:12:57 EDT
From: eps@Princeton.EDU (Elisha Sacks)
Message-Id: <8910031212.AA24279@notecnirp.Princeton.EDU>
To: weld@nooksack.cs.washington.edu
Cc: qphysics@cs.toronto.edu
In-Reply-To: Dan Weld's message of Tue, 3 Oct 89 00:26:55 EDT <89Oct3.002917edt.2347@neat.cs.toronto.edu>
Subject: An Interesting Puzzle
Resent-From: qphysics-owner@cs.toronto.edu
Resent-To: qphysics@cs.toronto.edu
Resent-Reply-To: eps@Princeton.EDU (Elisha Sacks)
Resent-Message-Id: <89Oct3.081855edt.2629@neat.cs.toronto.edu>
Resent-Date: Tue, 3 Oct 89 08:14:52 EDT
There are even days when N and M have different signs!
elisha.
∂03-Oct-89 0901 weening@Gang-of-Four.Stanford.EDU David Cyrluk
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 3 Oct 89 09:01:02 PDT
Received: by Gang-of-Four.Stanford.EDU (5.61/25-eef) id AA08321; Tue, 3 Oct 89 09:00:16 -0700
Date: Tue, 3 Oct 89 09:00:16 -0700
From: Joe Weening <weening@Gang-of-Four.Stanford.EDU>
Message-Id: <8910031600.AA08321@Gang-of-Four.Stanford.EDU>
To: jmc@sail, clt@sail, iam@sail, jk@sail
Subject: David Cyrluk
There is a new PhD student, David Cyrluk, who you may be interested in
talking to. He is interested in symbolic algebra, Groebner bases, and
theorem proving. (He worked at GE on these things for a few years.)
I hope we can offer to support him this year. I've put him on the
Qlisp mailing list, though he seems more interested in the algebraic
aspects than parallelism.
∂03-Oct-89 0942 qphysics-owner@neat.cs.toronto.edu Qualitative Reasoning Puzzle from QPHYSICS mailing list
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 3 Oct 89 09:42:24 PDT
Received: by neat.cs.toronto.edu id 2741; Tue, 3 Oct 89 12:42:51 EDT
Received: from cs.utexas.edu by neat.cs.toronto.edu with SMTP id 2594; Tue, 3 Oct 89 12:36:27 EDT
Received: from maltese.cs.utexas.edu by cs.utexas.edu (5.59/1.42)
id AA27054; Tue, 3 Oct 89 11:23:35 CDT
Date: Tue, 3 Oct 89 12:25:54 EDT
Posted-Date: Tue, 3 Oct 89 11:25:54 cdt
Message-Id: <8910031625.AA04123@maltese.cs.utexas.edu>
Received: by maltese.cs.utexas.edu (14.5/1.4-Client)
id AA04123; Tue, 3 Oct 89 11:25:54 cdt
From: hartman@cs.utexas.edu
Sender: hartman@cs.utexas.edu
To: qphysics@ai.toronto.edu, weld@nooksack.cs.washington.edu,
eps@Princeton.EDU
Cc: ajit@cs.utexas.edu, berleant@cs.utexas.edu, bert@cs.utexas.edu,
chiu@cs.utexas.edu, dave@mcc.com, dalle@iv1.cc.utexas.edu,
dvorak@cs.utexas.edu, farquhar@cs.utexas.edu, feng@cs.utexas.edu,
fouche@cs.utexas.edu, franke@mcc.com, woodwlee@cs.utexas.edu,
wylee@cs.utexas.edu, raman@cs.utexas.edu, rickel@cs.utexas.edu,
throop@cs.utexas.edu, kuipers@cs.utexas.edu, simmons@cs.utexas.edu,
porter@cs.utexas.edu, widman@cs.utexas.edu, bess@cs.utexas.edu,
byun@cs.utexas.edu, froom@cs.utexas.edu, dmpierce@cs.utexas.edu,
akira@cs.utexas.edu, wylee@cs.utexas.edu, fellow@cs.utexas.edu,
jc@cs.utexas.edu, hartman@cs.utexas.edu
In-Reply-To: kuipers@cs.utexas.edu's message of Tue, 3 Oct 89 09:12:55 CDT <8910031412.AA28180@ai.cs.utexas.edu>
Subject: Qualitative Reasoning Puzzle from QPHYSICS mailing list
Resent-From: qphysics-owner@cs.toronto.edu
Resent-To: qphysics@cs.toronto.edu
Resent-Reply-To: hartman@cs.utexas.edu
Resent-Message-Id: <89Oct3.124251edt.2741@neat.cs.toronto.edu>
Resent-Date: Tue, 3 Oct 89 12:38:50 EDT
>From: weld@nooksack.cs.washington.edu (Dan Weld)
>To: qphysics@ai.toronto.edu
>Subject: An Interesting Puzzle
>Assuming that most of the readers of this list are in the northern
>hemisphere, most readers are lamenting the increasingly shorter days of
>fall. But did everyone realize that the change in daylight hours is not
>symmetric? On any given day the sun rises N minutes later and sets M
>minutes earlier, but M is not necessarily equal to N!
>
>Any good qualitative explanations for this mystery???
>From: eps@Princeton.EDU (Elisha Sacks)
>There are even days when N and M have different signs!
>From: bess@cs.utexas.edu (Bess Sullivan)
>I personally think that at the beginning of the world N and M were the
>same, but the sun got confused when daylight savings time was invented
>and it has stayed that way ever since.
Bess is right!
The asymmetry is primarily due to the difference between clock and
celestial time. Noon in our (standard local) time isn't necessarily
the middle of the day. Difference is given by "equation of time"
which is sine-like fudge function to make clocks, calendars etc. come
out right. My "Graphical Timetable of the Heavens" (at home) gives
the function and more details.
There are also effects due to atmospheric refraction, latitude, etc.
Did anyone notice that day and night as given by sunrise and sunset
times weren't really equal on the equinox? Published sunrise and
sunset times are when the sun APPEARS to cross the horizon. However
the sun's apparent position is different from it's "true" position
because Earth's atmosphere bends its light, especially when sun is
near horizon and light is passing through lots of atmosphere. Does
the illusion make sunset more or less romantic?
Enjoy the sunsets and sunrises!
John Hartman
Dept. of Computer Sciences 512-471-9586 (o)
TAY 2.124 -472-3606 (h)
The University of Texas at Austin
Austin, TX 78712
∂03-Oct-89 1039 qphysics-owner@neat.cs.toronto.edu Re: Qualitative Reasoning Puzzle from QPHYSICS mailing list
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 3 Oct 89 10:39:10 PDT
Received: by neat.cs.toronto.edu id 2748; Tue, 3 Oct 89 13:40:00 EDT
Received: from cs.utexas.edu by neat.cs.toronto.edu with SMTP id 2596; Tue, 3 Oct 89 13:16:56 EDT
Received: from ratliff.cs.utexas.edu by cs.utexas.edu (5.59/1.42)
id AA05601; Tue, 3 Oct 89 12:03:43 CDT
Date: Tue, 3 Oct 89 13:03:09 EDT
From: jc@cs.utexas.edu
Posted-Date: Tue, 3 Oct 89 12:03:09 CDT
Message-Id: <8910031703.AA03556@ratliff.cs.utexas.edu>
Received: by ratliff.cs.utexas.edu (5.59/1.5-Client)
id AA03556; Tue, 3 Oct 89 12:03:09 CDT
X-Mailer: Mail User's Shell (6.5.6 6/30/89)
To: hartman@cs.utexas.edu, qphysics@ai.toronto.edu,
weld@nooksack.cs.washington.edu, eps@Princeton.EDU
Subject: Re: Qualitative Reasoning Puzzle from QPHYSICS mailing list
Cc: ajit@cs.utexas.edu, berleant@cs.utexas.edu, bert@cs.utexas.edu,
chiu@cs.utexas.edu, dave@mcc.com, dalle@iv1.cc.utexas.edu,
dvorak@cs.utexas.edu, farquhar@cs.utexas.edu, feng@cs.utexas.edu,
fouche@cs.utexas.edu, franke@mcc.com, woodwlee@cs.utexas.edu,
wylee@cs.utexas.edu, raman@cs.utexas.edu, rickel@cs.utexas.edu,
throop@cs.utexas.edu, kuipers@cs.utexas.edu, simmons@cs.utexas.edu,
porter@cs.utexas.edu, widman@cs.utexas.edu, bess@cs.utexas.edu,
byun@cs.utexas.edu, froom@cs.utexas.edu, dmpierce@cs.utexas.edu,
akira@cs.utexas.edu, wylee@cs.utexas.edu, fellow@cs.utexas.edu,
jc@cs.utexas.edu
Resent-From: qphysics-owner@cs.toronto.edu
Resent-To: qphysics@cs.toronto.edu
Resent-Reply-To: jc@cs.utexas.edu
Resent-Message-Id: <89Oct3.134000edt.2748@neat.cs.toronto.edu>
Resent-Date: Tue, 3 Oct 89 13:19:55 EDT
Actually, I am not sure that John is right. For the moment forget
about our standard calandar time and consider the earth simply
as a rotating body. We may define some function R as the time it
really takes for the earth to rotate (measured in seconds or whatever).
For our purposes R is a constant. Now, take the time now (call it
t1). Consider the difference between t1 and the sunrise today
and the sunset today (here I am talking about the actual time the
disk of the sun rises and sets). Just so we have lots of variables
we can call them trise1 and tset1. Now consider a time t2 = t1 + R.
Again we can define trise2 and tset2. One might expect that
trise1 - trise2 = tset1 = tset2 (i.e. the sunrise and sunset
have changed by the same amount). But this would, I think, be
false. Consider the time interval between sunrise1 and sunrise2,
and the interval between sunset1 and sunset2. These intervals
are about the same size so one might expect the earth's angle with
the sun to change by about the same amount. However, these intervals
occur at different times and so the change in the angles can in
fact be quite different (consider the case where sunrise1 and
sunrise2 stradle the equinox and so trise1 = trise2, but we
would not have tset1 = tset2. One could ever imagine cases where
the change in signs are not equal !).
Finally, three caveats:
1. I am making all this up so I may be wrong.
2. The effects John describes may overwelm the (small) effects I
am describing.
3. I leave it up to the qsim experts to try to translate all
this to qsim.
Jimi
∂03-Oct-89 1047 VAL Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
A SURVEY OF FORMAL NONMONOTONIC REASONING
Vladimir Lifschitz
Monday, October 16, 3:15pm
MJH 252
The Commonsense and Nonmonotonic Reasoning Seminar is a forum
in which we discuss current research and open problems related
to the logical foundations of AI, with a special emphasis on
commonsense reasoning and knowledge, formal nonmonotonic
reasoning, and foundations of logic programming.
This quarter, we are planning to include a few introductory
lectures on these subjects, for the benefit of those who are
new to this area of AI. The October 16 meeting will be the
first in this series. A basic knowledge of logic will be
assumed.
∂03-Oct-89 1601 qphysics-owner@neat.cs.toronto.edu An Interesting Puzzle
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 3 Oct 89 16:01:41 PDT
Received: by neat.cs.toronto.edu id 3109; Tue, 3 Oct 89 19:01:08 EDT
Received: from arisia.Xerox.COM by neat.cs.toronto.edu with SMTP id 2857; Tue, 3 Oct 89 18:54:47 EDT
Received: from kanga.parc.Xerox.COM by arisia.Xerox.COM with SMTP
(5.61+/IDA-1.2.8/gandalf) id AA16566; Tue, 3 Oct 89 11:18:29 -0700
Received: by kanga.parc.xerox.com
(5.61+/IDA-1.2.8/gandalf) id AA05583; Tue, 3 Oct 89 11:18:31 PDT
Message-Id: <8910031818.AA05583@kanga.parc.xerox.com>
Date: Tue, 3 Oct 89 14:18:31 EDT
From: Pat Hayes <hayes@parc.xerox.com>
To: weld@nooksack.cs.washington.edu
Cc: qphysics@cs.toronto.edu
In-Reply-To: Dan Weld's message of Tue, 3 Oct 89 00:27:18 EDT <89Oct3.003027edt.2618@neat.cs.toronto.edu>
Subject: An Interesting Puzzle
Resent-From: qphysics-owner@cs.toronto.edu
Resent-To: qphysics@cs.toronto.edu
Resent-Reply-To: Pat Hayes <hayes@parc.xerox.com>
Resent-Message-Id: <89Oct3.190108edt.3109@neat.cs.toronto.edu>
Resent-Date: Tue, 3 Oct 89 18:56:00 EDT
Why do you find it mysterious? That is, what makes you think they
should be the same? Maybe thats what needs the (qualitative?)
explanation. If you make a few sketches of the real geometry ( try a
view looking down on the globe from the solar systems vertical, so
that the north pole is to one side but the day-terminator is a
straight line: now imagine rotating the terminator ) it seems quite
clear that M and N are likely to be different.
Heres one that puzzled me for a long time, the harvest moon. Around
the spring solstice, the full moon rises and sets very steeply; but at
the fall solstice, it rises at a shallow angle to the horizon. This
used to help with late harvesting, and was taken as a sign of Gods
benevolence. But why is there any difference?
Pat
∂03-Oct-89 1605 qphysics-owner@neat.cs.toronto.edu time fudge factor
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 3 Oct 89 16:04:49 PDT
Received: by neat.cs.toronto.edu id 3152; Tue, 3 Oct 89 19:05:28 EDT
Received: from arisia.Xerox.COM by neat.cs.toronto.edu with SMTP id 3014; Tue, 3 Oct 89 18:58:11 EDT
Received: from kanga.parc.Xerox.COM by arisia.Xerox.COM with SMTP
(5.61+/IDA-1.2.8/gandalf) id AA16814; Tue, 3 Oct 89 11:35:09 -0700
Received: by kanga.parc.xerox.com
(5.61+/IDA-1.2.8/gandalf) id AA05994; Tue, 3 Oct 89 11:35:14 PDT
Message-Id: <8910031835.AA05994@kanga.parc.xerox.com>
Date: Tue, 3 Oct 89 14:35:14 EDT
From: Pat Hayes <hayes@parc.xerox.com>
To: hartman@cs.utexas.edu
Cc: qphysics@cs.toronto.edu
In-Reply-To: hartman@cs.utexas.edu's message of Tue, 3 Oct 89 12:39:22 EDT <89Oct3.124947edt.2757@neat.cs.toronto.edu>
Subject: time fudge factor
Resent-From: qphysics-owner@cs.toronto.edu
Resent-To: qphysics@cs.toronto.edu
Resent-Reply-To: Pat Hayes <hayes@parc.xerox.com>
Resent-Message-Id: <89Oct3.190528edt.3152@neat.cs.toronto.edu>
Resent-Date: Tue, 3 Oct 89 18:59:56 EDT
Quality old clocks have the equation of time fudge factor built in,
using a cam to move the minute hand around a little. Some
particularly fine ones ( now all in museums ) have two minute hands,
one showing celestial, one 'standard' time, which in England was
called 'railway time', giving a good idea of why it came into use.
Early railway timetables had to give arrival and departure times in
local time, which differed from city to city and were very confusing.
The railway companies introduced a standard time for their own use,
and of course straightforward ( so affordable ) clock mechanisms dont
pay attention to celestial abberations.
Pat
∂03-Oct-89 2150 srh@flash.bellcore.com Searle and Connectionism
Received: from flash.bellcore.com (petrus.bellcore.com) by SAIL.Stanford.EDU with TCP; 3 Oct 89 21:50:51 PDT
Received: by flash.bellcore.com (5.58/1.1)
id AA16221; Tue, 3 Oct 89 23:08:40 EDT
Date: Tue, 3 Oct 89 23:08:40 EDT
From: srh@flash.bellcore.com (stevan r harnad)
Message-Id: <8910040308.AA16221@flash.bellcore.com>
To: harnad@elbereth.rutgers.edu
Subject: Searle and Connectionism
[Apologies if this is received twice; mailer problems at princeton.]
To: John Searle
From: Stevan Harnad
John, before I circulate the Chinese Gym piece (which I have only now
gotten a chance to read) I want to go on record as urging you to
reconsider the argument before launching it.
This piece, in my opinion, would devalue the currency that you have
established with the original Chinese Room Argument. The remarkable
power of that Argument -- which I have now defended successfully
countless times -- is its built-in immunity to any "Systems Reply."
Whenever anyone says that YOU may not be understanding Chinese but the
"system" might, you point out that you are in fact performing ALL the
FUNCTIONS of the system! What FUNCTIONALIST can counter that? They are
committed to the irrelevance of the implementation of the function, so
they can't say that in the computer version the system thinks and in
the Searle version it does not, yet all there IS to the system in the
Searle version is you!
This, I think, is what has made the Argument so long-lived and the
object of so many obsessive and unsuccessful attempts to refute it.
Most of those attempts are various variants of the Systems Reply --
the persisting intuition that even if a part doesn't understand, the
whole could. In rebutting this on your behalf, I keep on reminding
people that in the Chinese Room there ARE no functioning parts apart
from you!
And as to the second most common rejoinder -- also
Systems-Reply-inspired -- that the brain has nonthinking parts too, and
works as a system -- I keep reminding people that you have nothing
against systems, parts, or even systems of which you may be a
nonthinking part! Logically speaking, your Argument applies only to
"systems" that allegedly understand, but ALL of whose functions you
yourself can perform without understanding. Pure symbol manipulating
systems of the kind produced by Strong AI are systems of the kind all
of whose functions you can perform without understanding, hence they do
not understand.
The coup de grace is when you point out that there is at least one
"system" that you DO believe understands, and that's the brain, and it
understands because it does have the causal powers necessary and
sufficient for understanding, whereas symbol systems do not. You are
not even opposed to artificial systems, as long as they have the
requisite causal powers, which, again, on the strength of the Chinese
Room Argument, symbol systems (and any other system all of whose
functions you yourself can perform without yourself understanding) do
not!
This network of points and counterpoints is, in my experience, the
enduring strength of the Chinese Room Argument against all comers to
date. But in the Chinese Gym you give up this built-in immunity to the
Systems Reply, leaving yourself expressing the same kind of
skepticism about the distributed "system" of un-understanding bodies
there as the god-fearing old lady expresses about both the computer AND
the brain!
You don't want to do that! I'm not saying that nets DO have the
requisite causal powers to understand; I'm just saying that so far you
don't have an Argument that deserves to stand in the same room as the
Chinese Room Argument to the effect that they don't. You have only the
old lady's skepticism about dumb parts. But that's just what the
Systems Reply has been waiting for! If you endorse this as an Argument
against connectionism, you'll not only lose there, you'll cast a shadow
on the rock-solid ground the original Chinese Room Argument stood on,
making it seem that you were making the weaker case -- vulnerable to the
Systems Reply -- all along.
There is a way out of this, I believe, and I hope you'll consider
using IT as your argument against connectionism instead: It is a fact
that despite all the talk about "brain-likeness" and "networks of
activation with neuron-like units," etc., all nets, and all their
accomplishments to date, have been COMPUTER SIMULATIONS of nets; these,
like all computer simulations, are all in reality of course just symbol
manipulating systems! Yet nets have been proffered as radical
alternatives to symbol systems.
Are they? Well, in a sense they are, because they could in principle
be implemented as real nets, which are radically different from symbol
systems, which can only be implemented as symbol systems (e.g., turing
machines or von-Neumann style digital computers, or their functional
equivalents, like a Chinese Army with a string of beads). In my view,
though, it is not the possibility of this radically different
implementation that is the real difference between the symbolic
approach and the connectionistic approach. They are actually just a
rival set of algorithms within the same basic framework, all of which
CAN be not only SIMULATED but also IMPLEMENTED purely symbolically.
Now THAT's the key to resurrecting the Chinese Room: As I argued in
"Minds, Machines and Searle," any function that can be completely
implemented purely symbolically is vulnerable to the Chinese Room
Argument. Unless the connectionists can give a reason why the
implementation of a net as a physical net rather than as a computer
simulation of a net is ESSENTIAL to its putative mental powers, you --
you alone, not a gym full of boys, which might, just might, have the
requisite system properties -- you alone will still be able to perform
all the functions of the net, just as you were able to perform all the
functions of the pure symbol manipulator ("Strong AI"), without
understanding!
And I don't think the connectionists will want to argue that there IS
this magical essential difference between the two implementations of
the same function because, after all, they too are functionalists, for
whom the function is implementation-independent. In any case, if they
DO want to say there's an essential difference, they will be stuck
waving their arms about the function's actually having to HAPPEN in
parallel in real time (which you can't ape, even hypothetically) rather
than serially, which seems not much better than saying that a function
must happen at a certain speed, rather than more slowly. I say
hand-waving, because this is not a FUNCTIONAL distinction: No
functional reason is given why one implementation of the function
should have different causal powers from the other. If the serial,
slower implementation delivers exactly the same goods, the burden is on
THEM to say why one should have a mind and the other should not.
Let me close with a plug for the symbol grounding problem: Unlike
the hand-waving for the "essential" role of speed or parallelism in
implementing a mind, the appeal to transducer function has a much
stronger, non-hand-waving rationale: By its nature transducer function
(1) cannot be performed by you,
(2) is nonsymbolic, and
(3) offers a natural link (via analog representations -- likewise
nonsymbolic -- and categorical representations, which CAN be
accomplished by a net, but only as a mindless component) to the world
of objects to which internal symbols might refer.
Hence transduction strikes at the heart of why pure symbol
manipulation may be ungrounded -- and lack intrinsic intentionality,
and be vulnerable to the Chinese Room Argument.
So I want to say that transduction of physical energy from the objects
in the world is not only an essentially nonsymbolic function, but it
is essential to the implementation of a mind. (The reason why
transduction cannot be just an add-on component module, as in the
standard Robot Reply, is, as you know, a more complicated story, and
more closely linked to my own specific, nonmodular, bottom-up grounding
theory, so I won't repeat it here. Suffice it to say that there is a
coherent rationale for transduction's essential role in the physical
substrate of mind, and hence for its immunity to the Chinese Room
Argument, whereas there is no such rationale for parallelism, speed,
or even the property of physical continuity.)
And note that nets will need transducers (and analog re-presentations of
them) too. We don't know which of the brain's functions are necessary
and sufficient for giving it the causal power to implement a mind and
which functions are irrelevant to it. The Chinese Room Argument shows
that (if the brain does symbol manipulation at all), symbolic function
is NOT SUFFICIENT for implementing a mind. The variant I described above
suggests that ditto is true for connectionist function: it's NOT
SUFFICIENT, for much the same reason. So all I would add is that
trandsucer function may be NECESSARY and even primary, with all the
rest grounded in it. Beyond that, at this stage of our understanding,
neither I nor you nor anyone else should claim to have given an
exhaustive list of the functions that are necessary and sufficient to
give a system the causal power to have a mind.
So I respectfully recommend that you jettison the Chinese Gym Argument
and instead deal with connectionism by turning the Chinese Room
Argument on its head, as follows. Suppose there are three rooms:
(1) In one there is a real Net (implemented as physical units, with
real physical links, real excitatory/inhibitory interconnections
real parallel distributed processing, real backpropping, etc.) that
could pass the Turing Test in Chinese (Chinese symbols in, Chinese
symbols out).
(2) In the second there is a computer simulation of (1) that likewise
passes the TT in Chinese.
(3) In the third is Searle, performing ALL the functions of (2),
likewise passing the Chinese TT (while still not understanding, of
course).
Now the connectionists have only two choices:
Either they must claim that all three understand Chinese (in which case
they are back up against the old Chinese Room Argument), or the
essentialists among them will have to claim that (1) understands but (2)
and (3) do not -- but without being able to give any functional reason
whatsoever why.
Now the foregoing is an ARGUMENT, in the spirit of the original
Chinese Room Argument, whereas the "Chinese Gym" is more like one of
those pale Hollywood sequels ("Chinese Room II"), trying to capture the
old glory (and its gate), but delivering only the style and not the
substance.
Here's a suggestion: If you want a sample of the kind of ironclad
defense I could successfully mount against the typical attempted
assaults on the Chinese Room -- but a defense that would definitely
FAIL in defense of the Chinese Gym, see the next message, which is
Dyer's critique of Searle, with my rebuttals.
If despite all this, however, you still want to launch the Chinese Gym
rather than the variant I suggested skyward, let me know, and I'll
start the count-down...
Cheers, Stevan
∂04-Oct-89 0243 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 4 Oct 89 02:43:42 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 426219; 2 Oct 89 10:21:58 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 109464; Sun 1-Oct-89 23:54:47 PDT
Date: Sun, 1 Oct 89 23:54 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: large periods
To: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, math-fun@RUSSIAN.SPA.Symbolics.COM,
"Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
cc: "hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19890915131149.0.RWG@TSUNAMI.SPA.Symbolics.COM>
Message-ID: <19891002065437.1.RWG@TSUNAMI.SPA.Symbolics.COM>
Character-Type-Mappings: (1 0 (NIL 0) (NIL NIL :SMALL) "TVFONT")
Fonts: CPTFONT, TVFONT
I scanned 1/3 of the possible glider impacts on the giant spark of
the unmuzzled p46 shuttle. It was cornucopious, logiccomponentwise.
The goodies include: 90 degree turn with the opposite phase parity
of the old one; 135 degree turn (:)!; 180 degree turn with sufficient
side-kick that just two shuttles form a delay-loop memory; one or two
sideslips; and, neatest of all, a configuration of two shuttles which
switches a lightweight spaceship beam on and off with successive gliders!
Date: Fri, 15 Sep 89 06:11 PDT
. . .
Ramblings about the size of oscillators: Life squanders
real estate. A given area counts through a minute fraction
of its potential states. So oscillators are large compared
to their periods. Can this be quantified?
How many states can a given area achieve? (Harder than the
Garden of Eden problem?) What is the largest area that can
be counted through all states? (= Minimum GoE -1 ? Could a
"grandorphan" be smaller?) Achieved by surrounding it with
an appropriate glider storm?
Clearly, above some size, maximum period as a function of area must
grow exponentially, because one could build a Turing type, but
finite "eons counter" only a fixed size increment over its memory.
What is the base of that exponential? I.e., the densest possible
memory.
With these new components, just two guns and five shuttles (half-guns)
suffice to make a shiftregister pseudorandom generator whose period
scales exponentially with the shuttle spacing. With an additional 90
degree shuttle, the pseudorandom lightweight beam escapes.
For those of you with small fonts or good software, here is a period
46*31*8191 = 11680366 in 140 by 167. (Generation number 2↑21.)
ε1 ...............................o
..............................oo
................oo...........ooo.o.........oo
................oo..........oo.............oo
.............................oo
..............................o
.
..............................o..........o
.............................oo........oo.o
............................oo........oo.ooo
.............................ooo.o.....o..o
..............................oo........oo
...............................o
.
.
.
.
..................................o
..................................oo
.................................o.o
.
..oo
..oo.......................................................oooooo.....oooo............oo
...........................................................oooo.......oo.o............oo
..............................................................oo.o.......o
................................................................oo.....oo
.
....................................o.o.........................oo.....oo
.......................................o......................oo.o.......o
......................oo...............o......................o.......oo.o
...o.....o.............oo...........o..o......................ooo.....oooo
..ooo...ooo...........o..............ooo
.oo.o...o.oo
.
.............................o
....o...o...................o.o
....o...o.............oo...o...o
......................oo...ooooo
..........................o.o.o.o.............................o
...........................o...o............................oo
...........o.................................................oo
...........oo..............o...o
..........o.o.............o.o.o.o
...........................ooooo
...........................o...o
............................o.o
.............................o
.
.
.
......................................o.o
.....................................o
.....................ooo.....ooo.....o
....................oo.oo...oo.oo....o..o
.......oo.oo........oo.o.o.o.o.oo....ooo
......o.....o........o...o.o...o
.....................o..o...o..o
.....o.......o........ooo...ooo
.....o..o.o..o
.....ooo...ooo
.
.......................................o........................................................ooo
.....................................oo.........................................................o
......................oo..............oo.........................................................o
......................oo
.
.
..............................ooo
.....oo.......................ooo
.....oo.................oo...ooooo
........................oo..oo...oo
............................oo...oo
.
....................................o.o................................................................................................ooo
...................................o...o.........................................................................................oo....ooo
............................oo...oo....o.........................................................................................oo...o...o
............................oo...oooo..o.............................................................................................o.....o
.............................ooooo...ooo..............................................................................................oo.oo
........oo....................ooo
.......o..o...................ooo.....................................................................................................oo.oo
.......o.......o.....................................................................................................................o.....o
.......o..o...o.o.....................................................................................................................o...o
.......oo.oo.oo.oo.....................................................................................................................ooo
...........o.o.........................................................................................................................ooo
..........oo.oo........ooo.....ooo
.......................ooo.....ooo
........................oo.....oo
..........................o...o
........................o..o.o..o
........ooo...ooo......o...o.o...o
.......o.........o......o..o.o..o...............................................................................................oo.......oo
.......o...o.o...o......ooo...ooo...............................................................................................o.oo...oo.o
........ooo...ooo...............................................................................................................o...o.o...o
.........o.....o................................................................................................................oo..o.o..oo
.................................................................................................................................ooo...ooo
..................................................................................................................................o.....o
........................oo............o.o
........................oo...........o
.....................................o
.....................................o..o
.....................................ooo
.................................................................................................................................oo
.................................................................................................................................oo
.
.
.
.
..............................................................................................................oo
..................o..........................................................................................o.o
...................o...........................................................................................o
.................ooo
.....ooo...ooo
.....ooo...ooo
....o.oo...oo.o
....o.oo...oo.o
.....oo.....oo
......o.....o
.
.
....................................o.o...........................................................ooo
.......oo.oo..................o........o............................................................o
......o.....o...............o.o........o...........................................................o
....ooooo.oo.oo..............oo.....o..o
...oo.ooo.oo.oo......................ooo
...ooo......o
....o.o
.....oo
.
.
.
.......................................................................................oo
.........................................o............................................o.o
..........................................o.............................................o
........................................ooo
.
.
.
.
.
...................................oo
....................................o
.................................ooo
.................................o.........................................ooo
.............................................................................o
............................................................................o
.
.
.
.
.
.
.
.
................................................................oo
...............................................................o.o
.................................................................o
......................................................................oo
.....................................................oo..............oo.oo......oo
.....................................................oo...............o..o......oo
......................................................................o..o
.......................................................................oo
.
......................................................ooo.ooo..........oo
.....................................................o.oo.oo.o........o..o
....................................................oo.......oo.......o..o
.....................................................o.oo.oo.o.......oo.oo
......................................................ooo.ooo.........oo
.
ε0This one computes a[n] = a[n-1] eqv a[n-18].
To release the spaceship beam out the bottom, delete the eater
("fish-hook") about 28 rows from the bottom.
(Turpitude: frotteur a pentadecathon against one of the disused
giant sparks to get an additional period factor of 15.)
My fancy program hits a paging wall at about 2 million, and thus must
cheat by irreversibly ripening the universe every a million or two.
The tightly spaced beam loops trapped between sidekickers can be
crossed, so that a region of space is packed with a startling
density of gliders traveling in all four directions without
colliding. This, then, gives us a lower bound of about 36 cells/bit
of eons memory, with the possibility of period 10↑100 or more fitting
on a screen.
∂04-Oct-89 1008 bhr@sumex-aim.stanford.edu CS300
Received: from sumex-aim.stanford.edu by SAIL.Stanford.EDU with TCP; 4 Oct 89 10:08:17 PDT
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AA07439; Wed, 4 Oct 89 10:10:39 PDT
Date: Wed, 4 Oct 1989 10:10:38 PDT
From: Barbara Hayes-Roth <bhr@sumex-aim.stanford.edu>
To: CAB@sail.stanford.edu, binford@coyote.stanford.edu,
cheriton@pescadero.stanford.edu, Gail@sol-margret.stanford.edu,
dill@amadeus.stanford.edu, feigenbaum@SUMEX-AIM.Stanford.EDU,
rwf@sail.stanford.edu, RPG@sail.stanford.edu,
Genesereth@score.stanford.edu, Ginsberg@polya.stanford.edu,
goldberg@polya.stanford.edu, golub@patience.stanford.edu,
ag@pepper.stanford.edu, halpern@ibm.com,
hayes-roth@SUMEX-AIM.Stanford.EDU, jlh@amadeus.stanford.edu,
Herriot@score.stanford.edu, Iwasaki@SUMEX-AIM.Stanford.EDU,
ARK@sail.stanford.edu, OK@coyote.stanford.edu, DEK@sail.stanford.edu,
lam@mojave.stanford.edu, latombe@coyote.stanford.edu,
val@sail.stanford.edu, linton@amadeus.stanford.edu,
zm@sail.stanford.edu, mayr@polya.stanford.edu, jmc@sail.stanford.edu,
ejm@sierra.stanford.edu, miller@kl.sri.com, jcm@polya.stanford.edu,
bmoore@ai.sri.com, nii@SUMEX-AIM.Stanford.EDU,
nilsson@tenaya.stanford.edu, oliger@pride.stanford.edu,
pratt@jeeves.stanford.edu, ALS@sail.stanford.edu,
shoham@score.stanford.edu, singh@score.stanford.edu,
shortliffe@SUMEX-AIM.Stanford.EDU, tobagi@sierra.stanford.edu,
ullman@score.stanford.edu, waldinger@ai.sri.com,
Wiederhold@SUMEX-AIM.Stanford.EDU, wilf@score.stanford.edu,
winograd@csli.stanford.edu
Cc: bhr@sumex-aim.stanford.edu
Subject: CS300
Message-Id: <CMM.0.88.623524238.bhr@sumex-aim.stanford.edu>
Hello -
We have a full schedule of speakers for CS300. Because we had many
more potential speakers than slots, we are considering extending
CS300 into winter quarter. More on that later.
Barbara Hayes-Roth
∂04-Oct-89 1602 MPS
Hi
Got a call from G. Campbell's wife today. You are interested
in the Mt Pelerin Meeting. As director she is extending you
an invitation to the Nov 27-30 in Christ Church, New Zealand
or next year in South America, Nov 11-16, 1990. Where does
Analise Anderson fit into this? She mentioned her in the
conversation.
Pat
∂04-Oct-89 1638 srh@flash.bellcore.com Parallelism: Simulated and Real
Received: from flash.bellcore.com (petrus.bellcore.com) by SAIL.Stanford.EDU with TCP; 4 Oct 89 16:38:15 PDT
Received: by flash.bellcore.com (5.58/1.1)
id AA17642; Wed, 4 Oct 89 19:06:02 EDT
Date: Wed, 4 Oct 89 19:06:02 EDT
From: srh@flash.bellcore.com (stevan r harnad)
Message-Id: <8910042306.AA17642@flash.bellcore.com>
To: palmer@cogsci.berkeley.edu
Subject: Parallelism: Simulated and Real
Simulated Vs. Real Parallelism
To: palmer@cogsci.berkeley.edu (Stephen E. Palmer)
Steve, you wrote:
> It seems to me that one can claim that there is an important and
> principled difference between the real connectionist net (in parallel
> hardware) and the simulated one (in a standard digital computer).
> It lies not at the "computational" level of the function computed,
> but at the algorithmic level of how the computation gets done. A
> parallel algorithm is not actually the same as a simulated parallel
> algorithm at the algorithmic level because the simulated one doesn't
> actually happen in parallel. Thus, an information processing type
> functionalist could counter what I take to be your claim that a
> functionalist would have to accept the same fate for the real net
> and the simulated net on the grounds that they differ in more than
> their mere implementation: they differ in the algorithm that is
> used to compute the function.
I know what the computational (software) and the implementational
(hardware) levels are, but I'm not sure what an algorithmic "level"
would be. (It seems to me that for Marr it was just a useful
conceptual or methodological level for the theorist in thinking
about a problem in AI.) I'm also not sure what an "information
processing type functionalist" is, but let me count some of the
ways you can be a "functionalist":
There are I/O functionalists, for whom I/O equivalence (Turing Test
Performance), irrespective of algorithm, is fine-grained enough:
Whatever passes the Chinese TT understands Chinese. These functionalists
should clearly be prepared to concede themselves refuted if Searle
performs all the internal functions of their candidate system and
yet fails to understand.
For "algorithmic" functionalists, it would presumably be
important that Searle should perform the same algorithm, and
if the "algorithm" is essentially nonsymbolic -- i.e., it cannot
be implemented as formal symbol manipulation on a Turing Machine
-- then Searle may not be able to perform it. Those I called
"essentialists" about parallelism (or about processing speed or
capacity, or [sub-NP] complexity or continuity) might fall in this
category -- but, as I mentioned, they would still owe us a functional
reason why two algorithms that gave exactly the same results should
differ so radically (one giving rise to a mind, the other not).
But would most connectionists really want to argue that a serially
simulated net was using a different ALGORITHM from a net implemented
in parallel? After all, Fodor & Pylyshyn have proposed that the
relevant functional level for the "cognitive" functionalist is that
of the "virtual" machine, and that the rest is just implementational
detail. You could write and execute a parallel process on a machine
that gave you virtual parallelism even though it was all implemented
below the virtual level as ordinary serial computation. The
"algorithms" of connectionism are presumably comprised of functions
like the unit interconnectivities, the generalized delta rule (back
prop), etc. If "algorithm" means anything at all, it's surely the
formal rule you program your computer to follow, not the
specific way the execution is implemented.
I guess it boils down to the fact that inasmuch as a parallel
computation is a FORMAL notion, it must be a hardware-independent
one. If a process requires the result of 30 other processes in
order to be computed, it can't matter formally whether the 30 others
are executed in series or in parallel, any more than it matters
whether they're executed quickly or slowly (within realizable
limits). And yet its this FORMAL level to which the functionalists
-- I call them symbolic functionalists -- are committed.
On the other hand, as I said, to the extent that connectionists
are NOT formal functionalists, they are free to claim that the
implementation does matter -- but then they have the burden of
showing how and why parallelism should be adequate for implementing
a mind, while its functionally equivalent serial implementation is
not. By way of contrast, there is no problem in accounting without
any hand-waving or mysterious essentialism for the radical functional
difference between my recommended candidate, transduction, and its
computer-simulated counterpart.
Stevan Harnad
∂06-Oct-89 0113 beeson@ucscd.UCSC.EDU an article you would want to see
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 6 Oct 89 01:12:15 PDT
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA03149; Fri, 6 Oct 89 01:12:59 -0700
Date: Fri, 6 Oct 89 01:12:59 -0700
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8910060812.AA03149@ucscd.UCSC.EDU>
To: jmc@sail.stanford.edu
Subject: an article you would want to see
New Scientist, 19 August, 1989, pp. 33-36.
"Recent theories in cognitive psychology allow us to understand that
emotions are not especially irrational. Rather, they are important in
the management of our goals and actions.
....emotions arise...when two different concerns clash, or when someone
else does something more or less than we expected....we need some mechanism
that can do three things. (1) ...handle interruptions and potential
interruptions...(2) change priorities and mangage...transition from one
activity to another...(3)...reprogram ourselves in the light of new
knowledge acquired as a result of [important and unanticipated] events...
∂06-Oct-89 0113 beeson@ucscd.UCSC.EDU said article
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 6 Oct 89 01:13:14 PDT
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA03157; Fri, 6 Oct 89 01:13:55 -0700
Date: Fri, 6 Oct 89 01:13:55 -0700
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8910060813.AA03157@ucscd.UCSC.EDU>
To: jmc@sail.stanford.edu
Subject: said article
If you have no easy access to New Scientist and wish to see it, I'll
make you a photocopy.
∂06-Oct-89 0535 menke@harvard.harvard.edu re: talk at harvard, october 2
Received: from harvard.harvard.edu by SAIL.Stanford.EDU with TCP; 6 Oct 89 05:35:13 PDT
Received: by harvard.harvard.edu (5.54/a0.25)
(for JMC@SAIL.Stanford.EDU) id AA05002; Fri, 6 Oct 89 08:38:14 EDT
Received: by endor.HARVARD.EDU; Fri, 6 Oct 89 08:38:11 EDT
Date: Fri, 6 Oct 89 08:38:11 EDT
From: menke@harvard.harvard.edu (Baiba Menke)
To: JMC@SAIL.Stanford.EDU
Subject: re: talk at harvard, october 2
Thank you. You gave the information in person to Rabin's
secretary who forwarded it to me.
We'll send the check along as soon as it is processed.
Baiba Menke
∂06-Oct-89 0951 perlis@cs.UMD.EDU hi
Received: from yoohoo.cs.UMD.EDU by SAIL.Stanford.EDU with TCP; 6 Oct 89 09:50:59 PDT
Received: by yoohoo.cs.UMD.EDU (5.61/UMIACS-0.9/04-05-88)
id AA01484; Fri, 6 Oct 89 12:57:42 -0400
Date: Fri, 6 Oct 89 12:57:42 -0400
From: perlis@cs.UMD.EDU (Don Perlis)
Message-Id: <8910061657.AA01484@yoohoo.cs.UMD.EDU>
To: jmc@sail.stanford.edu
Subject: hi
Cc: perlis@cs.UMD.EDU
Johm,
Let me again say how much I enjoyed your visit
here. It was also a treat for the students Terry and Michael
to meet with you.
As for the "meaning" issue, your comments about
(not) having names for words I think is a good point. What
I would tend to think about that is that even though there
may not be a specific reserved token for a given word, none
the less the agent (child) may have the ability to focus
attention on it as a thing. Just as we can mentally point to
(refer to, single out as of interest) a person whose name
we do not know, we can also think about an utterance made
in our presence (perhaps as "what was said" or simply "that")
even without explicit quotes. ANd this pointing ability in
turn seems to me to amount to an appearance/reality
disctinction capability. Quotation is just one possible (albeit
very powerful) way to formalize this.
I would maintain, then, that there is a kind of behavioral
watershed that occurs when the appearance/reality distinction
(ARD, as I call it) is available to a reasoner. I don't feel it
is worthwhile to argue over just how to use the word "mean", that
will all wash out in the end when we finally see what a full-blown
intelligence amounts to. What is important is to try to focus on
what (each individual scientist thinks) are the most revealing
aspects of rational behavior. This may be what you were getting
at in saying we can study whatever relation between inner processing
and outer world we want.
My essay "Intentioanlity and Defaults" (in the batch I gave you
when you were here) goes into some of these things a bit, tho
not in technical detail.
Best,
Don
∂06-Oct-89 1119 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU Elephant 2000
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Oct 89 11:19:36 PDT
Received: from iswim.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61/25-eef) id AA16811; Fri, 6 Oct 89 11:20:26 -0700
Received: by iswim.stanford.EDU (5.57/Ultrix3.0-C)
id AA00972; Fri, 6 Oct 89 11:13:48 PDT
Message-Id: <8910061813.AA00972@iswim.stanford.EDU>
To: John McCarthy <JMC@sail.stanford.edu>
Subject: Elephant 2000
Cc: katiyar@polya.Stanford.EDU
In-Reply-To: Your message of 06 Oct 89 10:39:00 -0700.
<1ph9GK@SAIL.Stanford.EDU>
From: John C. Mitchell <jcm@Polya.Stanford.EDU>
Reply-To: John C. Mitchell <jcm@Polya.Stanford.EDU>
Date: Fri, 06 Oct 89 11:13:47 PDT
Sender: jcm@iswim
Great. I was about to ask you whether you would be interested
in doing this. Meetings are Tues at 4. I think all are uncommitted
at this point (including next Tues), except one week before
Thanksgiving and the days that conflict with Departmental
Colloquia (I don't know the dates off hand, but I'll get
them from Anoop). Pick any one that is convenient for you.
John
∂06-Oct-89 1159 rlg@ai.mit.edu Hertz Foundation Fellowship
Received: from life.ai.mit.edu by SAIL.Stanford.EDU with TCP; 6 Oct 89 11:59:37 PDT
Received: from gracilis (gracilis.ai.mit.edu) by life.ai.mit.edu (4.1/AI-4.10) id AA08572; Fri, 6 Oct 89 15:00:37 EDT
From: rlg@ai.mit.edu (Bob Givan)
Received: by gracilis (4.0/AI-4.10) id AA00366; Fri, 6 Oct 89 15:00:24 EDT
Date: Fri, 6 Oct 89 15:00:24 EDT
Message-Id: <8910061900.AA00366@gracilis>
To: jmc@sail.stanford.edu
Subject: Hertz Foundation Fellowship
Professor McCarthy,
Would you be willing to write me a reference for the
Hertz Foundation Fellowship? I must apply this month,
and would appreciate your help. You could use the
two papers I gave you last weekend as additional fodder
on which to base your reference.
If you assent, I will forward the appropriate form to
you.
Hope your trip finished up successfully, including the talk
at Harvard (several people have asked me about Elephant 2000 (?)
since you explained it to me).
Thanx much,
Bob Givan
rlg@ai.mit.edu
∂06-Oct-89 1218 weening@Gang-of-Four.Stanford.EDU Software pipelining example
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Oct 89 12:18:43 PDT
Received: by Gang-of-Four.Stanford.EDU (5.61/25-eef) id AA07523; Fri, 6 Oct 89 12:17:57 -0700
Date: Fri, 6 Oct 89 12:17:57 -0700
From: Joe Weening <weening@Gang-of-Four.Stanford.EDU>
Message-Id: <8910061917.AA07523@Gang-of-Four.Stanford.EDU>
To: qlisp@Gang-of-Four.Stanford.EDU
Subject: Software pipelining example
I'm trying out an example of software pipelining based on a program
that I was shown this summer. It's a prime-finding sieve, and has
been used by various people as an example of coroutines, streams,
concurrency and other programming constructs, so you may have seen it
before, though I hadn't. Here is my Qlisp version:
(defun print-primes (limit)
(qwait
(counter 3 limit (filter 2))))
(defun counter (number limit proc)
(unless (> number limit)
(funcall proc number)
(counter (1+ number) limit proc)))
(defun filter (prime)
(print prime)
(let ((next-proc nil))
(function (qlambda t (number)
(unless (zerop (mod number prime))
(when (null next-proc)
(setq next-proc (filter number)))
(funcall next-proc number))))))
The top-level call establishes a limit on the size of primes to be
found, and runs the COUNTER function, which generates numbers up to
this limit and passes them to the first filter process, which filters
out multiples of 2. Each filter process is created by a call to
FILTER with a prime as an argument; it prints the prime and then
creates a process which is called with numbers, and discards multiples
of the prime. The first number that it doesn't discard must be a new
prime (because it has passed all of the previous prime filters), so it
calls FILTER to create a new filter for this number. It then passes
subsequent non-multiples of its prime to this filter. At any stage,
if P is the prime most recently found, the pipeline of processes will
look like:
+---------+ +-----------+ +-----------+ +-----------+
| counter | --> | filter(2) | --> | filter(3) | --> ... --> | filter(p) |
+---------+ +-----------+ +-----------+ +-----------+
The main thing that I don't like about this program is that each
filter process has to keep a small amount of state, namely whether it
has yet created the next process. In the form that I originally saw
it, processes communicated by sending and receiving messages from
channels, and an output channel for each process was created along
with that process. Then you could make a different process read from
that channel at some point, which would modify the pipeline without
changing the behavior of processes already running.
∂06-Oct-89 1244 srh@flash.bellcore.com Free-Floating Intentionality
Received: from flash.bellcore.com (petrus.bellcore.com) by SAIL.Stanford.EDU with TCP; 6 Oct 89 12:44:22 PDT
Received: by flash.bellcore.com (5.58/1.1)
id AA15625; Fri, 6 Oct 89 13:45:17 EDT
Date: Fri, 6 Oct 89 13:45:17 EDT
From: srh@flash.bellcore.com (stevan r harnad)
Message-Id: <8910061745.AA15625@flash.bellcore.com>
To: harnad@elbereth.rutgers.edu
Subject: Free-Floating Intentionality
Free-Floating Intentionality, or, "I've Got You On My Brain"
Dr Michael G Dyer <dyer@CS.UCLA.EDU> wrote:
> WHY is it that you are convinced by [Searle's] argument from
> introspection (i.e. that if one imagines doing ALL of the primitive
> functions of a system, that is capable of exhibiting overall
> intelligent/intentional behavior, and one does not feel like one is
> intentional, then [it follow that] there is no intentionality)?
Because I think that the only real difference between whether a system
really understands or only acts as if it understands but doesn't
understand is that there is something it's like to understand, and the
system that really understands has THAT, and the one that doesn't,
doesn't. For example, I think that there's something it's like to
understand English and something it's like to understand Chinese. I
have the former and not the latter, because I understand English and
not Chinese.
But before you hasten to declare that then you don't care about the
difference between (1) a system that "really" understands in this
introspective sense and (2) a system that only acts as if it understands,
notice that you're headed for trouble there, because even YOU wouldn't
be happy with (3) a system that just said "Yes, I understand" in response
to everything. Why? Because you don't believe it really understands,
even though it acts as if it does.
Now this was the motivation behind the original (talk-only) version of
the Turing Test: How to rule out (3)? Well, require that it not be so
easy to "see through" the system (as you would quickly see through the
repetitious routine of (3)): Require (2) to be so hard to see through
that you can't tell it apart from a real, understanding person from the
way it acts (verbally); then admit you have no rational grounds for
denying that it really understands. Well Searle has given you some
rational grounds: Even if you can't see through (2)'s performance from
the way it acts, Searle shows that (if he can himself execute ALL of
its functions without himself understanding) (2) doesn't really
understand because there's NOBODY HOME TO BE DOING THE UNDERSTANDING
(and the only one home, Searle, doesn't understand -- if he's to be
taken at his word, as I'm inclined to do; others have insisted on
seeing evidence of dual personality or "speaking in tongues" in all
this...).
The case to bear in mind is the hardest one: Searle has memorized all
the machine tables [n.b., not IMAGINED memorizing them, but ACTUALLY
memorized them; the thought experiment is imaginary, but the simulation
being imagined is not imaginary]; so everything's internalized; there are
no other parts, and Searle himself is all there is to the system.
Now you want to say that, despite Searle's denial, and despite the fact
that there's no one else in sight, there's some understanding going on
there. Well, let me ask you this then: WHO's doing the understanding?
And while you're at it, where is he (she?) after the lights go out,
when Searle's put away the Chinese toys and gone beddy-bye. I know that
one understanding system is there, sleeping away. Do you think there
are two? Since when? Since Searle memorized the tables? Talk about
special creation: An understanding system comes into being because a
person memorizes a bunch of meaningless symbol-manipulation rules...
> A system that has intelligence (intentionality or any other emergent
> phenomena) usually does NOT have any insight into the operations being
> formed that give rise to its emergent properties. If I had to really be
> aware of every sigmoidal function my neurons are executing I would have
> no time to be intelligent and my normal state of awareness would be
> totally bizarre. The act of doing is different from introspecting on
> the act of doing (since it is a different act of doing).
This is changing the subject. The question was only whether a system
that understands Chinese has enough "insight into its operations" to
know THAT it understands Chinese -- not HOW it understands Chinese, or
how its neurons understand Chinese, or anything else.
> In fact, the human experience of being intentional or aware (even self-
> aware) probably RELIES on humans NOT having access to the experience of
> what it's like to do all of the underlying cognitive and/or neural
> operations that give rise to the intentionality itself (as an emergent
> phenomenon). I've tried imagining myself doing all of the LISP
> functions involved in the BORIS NLP system and I can't imagine what
> it's like to be the BORIS system. I've tried imagining doing all the
> backprop operations of a language learning system and I can't imagine
> what it's like to have it's state of consciousness (if it has one).
Unless you're a panpsychist, you don't really believe EVERYTHING has a
mind. (And if you are, this discussion is pointless, because the
matter's settled, there's no need to pass the TT, and even (3)
understands.)
But if you're not a panpsychist then you probably better stop trying to
imagine what it's like to be BORIS or backprop because, until further
notice, it's probably safe to assume there's NOTHING it's like to be
BORIS or backprop or most other nonliving things: Nobody home. As for
your brain: You already know what it's like to be your whole brain,
and perhaps even your left hemisphere plus brainstem; but hold off on
what it's like to be lesser subsystems, because again, there's
probably nobody home. (And as I said, lack of introspective access to
the nature of the brain function underlying understanding is beside the
point.)
> Let's imagine every quantum/biochemical/bioelectrical/neural operation
> done by your brain and imagine Searle imagining himself doing every
> operation (e.g. he moves ions into place for neural spiking etc.). I'm
> sure he's not going to understand what he's doing or have any insight
> into the intentionality of the brain/mind he is creating (he's just
> following a giant set of simulation instructions). So what? Who EXPECTS
> him to understand the resulting state of mind? If I talk to Searle I
> get information about what it's like to simulate a brain with some
> (unknown) intentionality. If I talk directly to the mind residing on
> that Searle-created-brain, then I can have a conversation with that
> mind and it will describe it's state of mind. If the simulation is of
> YOUR brain, then it will give me one of your favorite arguments about
> why symbol grounding is important and why Searle's chinese box is
> right. (and perhaps even ultimately be persuaded by this counter
> argument :-)
[Repeat: We're not to imagine Searle imagining, we're to imagine him DOING.]
Until further notice (i.e., until someone successfully refutes
Searle's Chinese Room Argument), there's strong reason to doubt that pure
symbolic simulations of either the performance of a person with a mind
OR the functions of the brain have a mind. This should not be particularly
surprising because, as I pointed out in "Minds, Machines and Searle,"
lots of symbolic simulations fail to have the critical properties of
the thing they're simulating: Simulated furnaces don't heat. Simulated
airplanes don't fly. Why should simulated "minds" understand? A
simulated brain, after all, couldn't see, because simulated transducers
don't really transduce.
I like the idea of talking to the mind "on" the "Searle-created-brain,"
though, especially if it's my twin. Remind me again (because I'm
beginning to feel a certain familial interest), what becomes of him
when Searle stops thinking about simulating my brain and goes to sleep?
What's on my doppelganger's mind then, and where's his
"Searle-created-brain"...?
> You seem to be thinking that, just because Searle is the only
> functioning part, there is no one else to talk to BUT Searle, but
> that's clearly NOT true! For example, when we build a natural language
> understanding system, the only functioning parts in the computer are
> the program instruction interpreter hardware circuits, bus circuits,
> memory registers, etc. But I don't even consider trying to talk to any
> of THOSE. I type input TO THE SYSTEM I have built and talk TO THAT
> SYSTEM. Even if Searle is behind every bus and register operation, I
> can still sit down and have a conversation WITH THE SYSTEM, by feeding
> input TO THE SYSTEM. It's clearly NOT Searle who replies (any more than
> it is a single register replying, who may be busy flipping its
> flip-flops, etc). -- It so happens that I can stop and ALSO talk to
> Searle, but SO WHAT?
I seem powerless to communicate to you that it's not that anyone's
doubting the distinction between a system and a part of the system, or
that a system can have properties its parts don't have, or that
understanding systems (including the brain) can have non-understanding
parts, of which one non-understanding part might even be Searle. It's
still true that in the Chinese room Searle is both the part and the
whole system (if not, who/what is?), so HE's the only one whose word
I'm inclined to take as to whether or not he understands Chinese.
Now I realize you're going to rush to tell me that at the same time
he's denying it in English, in Chinese he's telling me he CAN
understand Chinese, to which I reply: Who's "he"? And where was he
last Tuesday, before Searle memorized the machine table? And how come
Searle doesn't know about him (he's both systems, isn't he?)? Or
perhaps Searle is speaking in tongues? Or has another personality.
What's his name? Shur-Lee? Ok, where were you Monday, Shur-Lee?
Can anybody corroborate that, Shur-Lee?
But this is all just silly sci-fi, as far as I'm concerned, designed to
milk intuitions about a counterfactual situation that's about as likely
to be possible as time-travel or telepathy. I give reasons in "Minds,
Machines and Searle" why no symbols-only program could ever pass the TT
in the first place, so Searle would never need go to such lengths to
show it had no mind -- and we need never confront paradoxes like the
above. The reason is the Symbol Grounding Problem, according to which
the system that can pass the (symbols-only) TT will first have to be
capable of passing the TTT (Total Turing Test), which will require
drawing on NONSYMBOLIC functions to implement robotic functions (like
transduction) that can only be simulated, but not implemented
symbolically. So Searle can't perform those functions either (and my
doppelganger can forever rest in peace).
> I don't know a single computer scientist or AI researcher who does
> anything more than laugh with incredulity at Searle's argument. The
> argument is not successful in CS/AI circles and every colleague, after
> hearing Searle give a talk, is always amazed that Searle seems to
> convince non-computer scientists (who have perhaps not had the
> experience of building layers and layers of virtual machines).
And I think Searle's Argument will go down in intellectual history as
having summarily lampooned the risible credulousness with which CS/AI
circles in the '70s and '80's were ready to gussy up extrapolations
of the performance feats of their toy symbol-crunchers with mentalistic
interpretations. (It's true that laymen have a naive blanket
incredulity about artificial minds; they will stand refuted when we
actually come up with one. But for now they're on the money despite
themselves. And as far as I'm concerned, pure symbol crunchers are out
of the running for good.)
Stevan Harnad
∂06-Oct-89 1510 ME re: autologout
To: CLT
CC: JMC
OK, autologout is disabled for you two. The reason it was happening
today is that a number of people are logged in on DDs and we have
a bunch of DD channels detached because of bad RAMs (which can be
fixed, I guess it's about time).
∂06-Oct-89 1532 harnad@clarity.Princeton.EDU Searle and Connectionism
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 6 Oct 89 15:32:41 PDT
Received: from cognito.Princeton.EDU by Princeton.EDU (5.58+++/2.22)
id AA24975; Fri, 6 Oct 89 18:14:33 EDT
Received: by cognito.Princeton.EDU (4.0/1.85)
id AA08696; Tue, 3 Oct 89 14:46:54 EDT
Date: Tue, 3 Oct 89 14:46:54 EDT
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8910031846.AA08696@cognito.Princeton.EDU>
To: harnad@elbereth.rutgers.edu
Subject: Searle and Connectionism
To: John Searle
From: Stevan Harnad
John, before I circulate the Chinese Gym piece (which I have only now
gotten a chance to read) I want to go on record as urging you to
reconsider the argument before launching it.
This piece, in my opinion, would devalue the currency that you have
established with the original Chinese Room Argument. The remarkable
power of that Argument -- which I have now defended successfully
countless times -- is its built-in immunity to any "Systems Reply."
Whenever anyone says that YOU may not be understanding Chinese but the
"system" might, you point out that you are in fact performing ALL the
FUNCTIONS of the system! What FUNCTIONALIST can counter that? They are
committed to the irrelevance of the implementation of the function, so
they can't say that in the computer version the system thinks and in
the Searle version it does not, yet all there IS to the system in the
Searle version is you!
This, I think, is what has made the Argument so long-lived and the
object of so many obsessive and unsuccessful attempts to refute it.
Most of those attempts are various variants of the Systems Reply --
the persisting intuition that even if a part doesn't understand, the
whole could. In rebutting this on your behalf, I keep on reminding
people that in the Chinese Room there ARE no functioning parts apart
from you!
And as to the second most common rejoinder -- also
Systems-Reply-inspired -- that the brain has nonthinking parts too, and
works as a system -- I keep reminding people that you have nothing
against systems, parts, or even systems of which you may be a
nonthinking part! Logically speaking, your Argument applies only to
"systems" that allegedly understand, but ALL of whose functions you
yourself can perform without understanding. Pure symbol manipulating
systems of the kind produced by Strong AI are systems of the kind all
of whose functions you can perform without understanding, hence they do
not understand.
The coup de grace is when you point out that there is at least one
"system" that you DO believe understands, and that's the brain, and it
understands because it does have the causal powers necessary and
sufficient for understanding, whereas symbol systems do not. You are
not even opposed to artificial systems, as long as they have the
requisite causal powers, which, again, on the strength of the Chinese
Room Argument, symbol systems (and any other system all of whose
functions you yourself can perform without yourself understanding) do
not!
This network of points and counterpoints is, in my experience, the
enduring strength of the Chinese Room Argument against all comers to
date. But in the Chinese Gym you give up this built-in immunity to the
functional reply, leaving yourself expressing the same kind of
skepticism about the distributed "system" of un-understanding bodies
there as the god-fearing old lady expresses about both the computer AND
the brain!
You don't want to do that! I'm not saying that nets DO have the
requisite causal powers to understand; I'm just saying that so far you
don't have an Argument that deserves to stand in the same room as the
Chinese Room Argument to the effect that they don't. You have only the
old lady's skepticism about dumb parts. But that's just what the
Systems Reply has been waiting for! If you endorse this as an Argument
against connectionism, you'll not only lose there, you'll cast a shadow
on the rock-solid ground the original Chinese Room Argument stood on,
making it seem that you were making the weaker case -- vulnerable to the
Systems reply -- all along.
There is a way out of this, I believe, and I hope you'll consider
using IT as your argument against connectionism instead: It is a fact
that despite all the talk about "brain-likeness" and "networks of
activation with neuron-like units," etc., all nets, and all their
accomplishments to date, have been COMPUTER SIMULATIONS of nets, which
are all in reality of course just symbol manipulating systems! Yet nets
have been proffered as radical alternatives to symbol systems.
Are they? Well, in a sense they are, because they could in principle
be implemented as real nets, which are radically different from symbol
systems, which can only be implemented as symbol systems (e.g., turing
machines or von-Neumann style digital computers, or their functional
equivalents, like a Chinese Army with a string of beads). In my view,
though, it is not the possibility of this radically different
implementation that is the real difference between the symbolic
approach and the connectionistic approach. They are actually just a
rival set of algorithms within the same basic framework, all of which
CAN be not only SIMULATED but also IMPLEMENTED purely symbolically.
Now THAT's the key to resurrecting the Chinese Room: As I argued in
"Minds, Machines and Searle," any function that can be completely
implemented purely symbolically is vulnerable to the Chinese Room
Argument. Unless the connectionists can give a reason why the
implementation of a net as a physical net rather than as a computer
simulation of a net is ESSENTIAL to its putative mental powers, you --
you alone, not a gym full of boys, which might, just might, have the
requisite system properties -- you alone will still be able to perform
all the functions of the net, just as you were to perform all the functions
of the pure symbol manipulator ("Strong AI"), without understanding!
And I don't think the connectionists will want to argue that there IS
this magical essential difference between the two implementations of
the same function because, after all, they too are functionalists, for
whom the function is implementation-independent. In any case, if they
DO want to say there's an essential difference, they will be stuck
waving their arms about the function's actually having to HAPPEN in
parallel in real time (which you can't ape, even hypothetically) rather
than serially, which seems not much better than saying that a function
must happen at a certain speed, rather than more slowly. I say
hand-waving, because this is not a FUNCTIONAL distinction: No
functional reason is given why one implementation of the function
should have different causal powers from the other. If the serial,
slower implementation delivers exactly the same goods, the burden is on
THEM to say why one should have a mind and the other should not.
Let me close with a plug for the symbol grounding problem: Unlike
the hand-waving for the "essential" role of speed or parallelness in
implementing a mind, the appeal to transducer function has a much
stronger, non-hand-waving rationale: By its nature transducer function
(1) cannot be performed by you,
(2) is nonsymbolic, and
(3) offers a natural link (via analog representations -- likewise
nonsymbolic -- and categorical representations, which CAN be
accomplished by a net, but only as a mindless component) to the world
of objects to which internal symbols might refer.
Hence transduction strikes at the heart of why pure symbol
manipulation may be ungrounded -- and lack intrinsic intentionality,
and be vulnerable to the Chinese Room Argument.
So I want to say that transduction of physical energy from the objects
in the world is not only an essentially nonsymbolic function, but it
is essential to the implementation of a mind. (The reason why
transduction cannot be just an add-on component module, as in the
standard Robot Reply, is, as you know, a more complicated story, and
more closely linked to my own specific, nonmodular, bottom-up grounding
theory, so I won't repeat it here. Suffice it to say that there is a
coherent rationale for transduction's essential role in the physical
substrate of mind, and hence for its immunity to the Chinese Room
Argument, whereas there is no such rationale for parallelness, speed,
or even the property of physical continuity.)
And note that nets will need transducers (and analog re-presentations of
them) too. We don't know which of the brain's functions are necessary
and sufficient for giving it the causal power to implement a mind and
which functions are irrelevant to it. The Chinese Room Argument shows
that (if the brain does symbol manipulation at all), symbolic function
is NOT SUFFICIENT for implementing a mind. The variant I described above
suggests that ditto is true for connectionist function: it's NOT
SUFFICIENT, for much the same reason. So all I would add is that
trandsucer function may be NECESSARY and even primary, with all the
rest grounded in it. Beyond that, at this stage of our understanding,
neither I nor you nor anyone else should claim to have given an
exhaustive list of the functions that are necessary and sufficient to
give a system the causal power to have a mind.
So I respectfully recommend that you jettison the Chinese Gym Argument
and instead deal with connectionism by turning the Chinese Room
Argument on its head, as follows. Suppose there are three rooms:
(1) In one there is a real Net (implemented as physical units, with
real physical links, real excitatory/inhibitory interconnections
connections, real parallel distributed processing, real backpropping,
etc.) that could pass the Turing Test in Chinese (Chinese symbols in,
Chinese symbols out).
(2) In the second there is a computer simulation of (1) that likewise
passes the TT in Chinese.
(3) In the third is Searle, performing ALL the functions of (2),
likewise passing the Chinese TT, while still not understanding, of
course.
Now the connectionists have only two choices:
Either they must claim that all three understand Chinese (in which case
they are back up against the old Chinese Room Argument), or the
essentialists among them will have to claim that (1) understands but (2)
and (3) do not -- but without being able to give any functional reason
whatsoever why.
Now the foregoing is an ARGUMENT, in the spirit of the original
Chinese Room Argument, whereas the "Chinese Gym" is more like one of
those pale Hollywood sequels ("Chinese Room II"), trying to capture the
old glory (and its gate), but delivering only the style and not the
substance.
Here's a suggestions: If you want a sample of the kind of ironclad of
defense I could successfully mount against the typical attempted
assaults on the Chinese Room -- but a defense that would definitely
FAIL in defense of the Chinese Gym, see the next message, which is
Dyer's critique of Searle, with my rebuttals.
If despite all this, however, you still want to launch the Chinese Gym
rather than the variant I suggested skyward, let me know, and I'll
start the count-down...
Cheers, Stevan
∂06-Oct-89 1607 harnad@clarity.Princeton.EDU Dyer on Searle
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 6 Oct 89 16:06:58 PDT
Received: from cognito.Princeton.EDU by Princeton.EDU (5.58+++/2.22)
id AA27749; Fri, 6 Oct 89 18:59:57 EDT
Received: by cognito.Princeton.EDU (4.0/1.85)
id AA08733; Tue, 3 Oct 89 14:57:19 EDT
Date: Tue, 3 Oct 89 14:57:19 EDT
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8910031857.AA08733@cognito.Princeton.EDU>
To: srh@elbereth.rutgers.edu
Subject: Dyer on Searle
[Here are Michael Dyer's Comments on Searle, which I promised to
circulate. Since they are long (the whole text is around 500 lines)
and quote extensively from me, I have restricted my own responses to
some interpolated lines here and there, indented, and enclosed in
square brackets, like this. SH]
From: Dr Michael G Dyer <dyer@CS.UCLA.EDU>
To: Stevan Harnad <harnad@clarity.Princeton.EDU>
Stevan, you wrote:
> You are apparently satisfied with the "Systems Reply." I definitely am
> not. Searle has repeatedly said that it's not that he doesn't believe
> intelligence arises from "component interactions," because he does
> believe that, for example, it arises from the component interactions in
> the brain. Nor is Searle committed to any particular kind of component
> type. He just holds (correctly, in my view) that his argument has
> shown that certain types of system (pure symbol systems, as it turns out,
> in my analysis) are the wrong types of system to exhibit intrinsic
> intentionality; their intentionality is all parasitic on our
> interpretations.
The problem with Searle's "intentionality" is that it is always
fundamentally "parasitic on our interpretations" -- i.e. YOUR having
intentionality (to ME) is parasitic on MY interpretation that bestows
intentionality to YOU. History is full of one (often distinct racial)
group denying (or reducing the amount of) intentionality "bestowed" on
another group.
[No, Searle's point has has nothing to do with vague social judgments;
his point is much simpler; The sentences in a book don't have intrinsic
meanings. They're just scratches on paper. Their meaning depends on our
interpretations. The meanings in our head do not. They are intrinsic. SH]
Maybe it depends on what you mean by "a pure symbol system". A computer
is a general purpose device. It can push both numbers and symbols
around since they are BOTH "symbols" (in the sense of patterns moved
about and transformed in systematic ways, dependent on other
patterns). There is a basic ambiguity concerning the use of the term
"symbol". I myself write papers about "connectionism versus symbolism",
but that discussion is at the level of what is the most natural
scientific language for describing cognition, and is NOT intended to
imply that somehow "connectionism" gives us some magic that cannot be
done on a von Neumann machine (or Turing machine).
> [You have to distinguish between a symbolic simulation and a causal
> implementation. Some physical processes, such as transduction, can be
> simulated by symbol manipulation but they cannot be implemented as
> symbol manipulation. Whether or not connectionism is vulnerable to
> Searle's Argument depends on whether or not there is any essential
> functional or causal difference between a symbolic simulation of a net
> and a nonsymbolic implementation. SH]
I can simulate any kind of neural system (to any level of detail) you
want with the 'symbols' of a von Neumann Machine (and Turing machine
ultimately, since a Turing machine can simulate a Von Neumann machine,
albeit incredibly more slowly). There's greater and greater cost as you
try to make the simulation more detailed (e.g. if you want quarks, then
I might need 1,000 supercomputers to simulate that level of detail for
a microsecond), but as long as Searle accepts that it is PATTERNS of
interaction, then the computer can do it, once we set up a
correspondence between the patterns that the computer can realize and
those patterns we are hypothesizing for whatever reality we are trying
to model.
[Except for symbol manipulation itself, a symbolic simulation of a
physical process is not the same as a causal implementation of that
process. A simulated transducer cannot do real transduction.
A simulated airplane cannot fly. Searle holds no brief against
interaction patterns in general, just purely symbolic ones. SH]
> {Dyer misses] the essence of Searle's Argument: The person is not just
> being a piece of a system, HE'S DOING EVERYTHING IN THE SYSTEM.
> Whatever functions you want to point to that a symbol-manipulator is
> performing in order to pass the Turing Test, Searle can simply do them
> all himself. Introspection is the right test for whether the mental state
> that is being attributed to the system he is simulating is really there
> under conditions where he is performing every last function the system
> is performing. That's why Searle says he doesn't see much left to hang
> a mind on in the walls and ceiling of the Chinese Room and the chalk on
> the blackboard. You're tilting at windmills if you want to insist that
> THAT system has a mind. It's not that Searle is claiming NO system can
> have a mind, even one of which he is himself a component. He's just
> claiming that THAT system -- the pure symbol-cruncher under discussion
> in the Chinese Room Argument -- has no mind.
Now we really have a CLEAR DISAGREEMENT! Imagine that we encounter a
strange form of intelligence from the planet XIMULA. Each XIMULAN is
highly intelligent (composes operas, etc.). However, one day humans
discover that the brains of the XIMULANs consist of societies of
miniature creatures, called the MINIXIANS. The MINIXIANS within the
brain of a given XIMULAN all work together, communicating and
performing many operations, to 'control' the behavior of that XIMULAN.
It turns out that you can talk to a single MINIXIAN (they're that
smart). It turns out that no MINIXIAN knows really what the XIMULAN is
doing (or thinking). Now we have two levels of "intentionality". You
can talk to the XIMULAN and he/she will tell you about his/her loves,
desires, thoughts, feelings, opera ideas, etc. But you can also talk to
a given MINIXIAN about what he does and why, etc.
[This example is irrelevant because Searle would not deny it. Reread
the last two sentences of the quoted passage before it. SH]
Put more simply: (1) I can use lots of smart components to make a dumb
device; e.g. I can make a simple hand-calculator out of a football
field of humans, all who are taught how and when to perform simple
bionary operations. No human in the football field need even know what
he/she is doing. (2) I can use lots of dumb components to make a smart
device -- that's the goal of both AI and connectionism.
If Searle can move around fast enough (to keep the time frames in synch
-- so he'd have to be VERY fast!) to perform every operation that's
involved in a given program execution, then no one will be able to tell
the difference. Let's assume this program's execution runs a
connectionist simulation which is looking at a manufactured visual
scene and answering questions about it. Now it's really Searle who is
implementing the underlying machine. Now if we ask the system what it
sees in the scene, it will answer. But if we ask Searle himself what he
experiences he will say "Whew! I'm running around here, flipping this
bit, flopping that bit, moving this electron down this data bus, ...
This is really hectic!" etc. The point is, Searle's introspection on
his own experience will give us NO INSIGHT into the intentionality of
the connectionist system (that is talking to us about what it is seeing
in the visual scene).
[The example has gotten too baroque. But if the bottom line is that
this is again a case of pure symbol-manipulation and question-answering,
Searle's Argument is again valid. If nonsymbolic functions such as
transduction, analog transformations or essential parallelism are
involved, Searle's Argument fails, as I've shown. SH]
Ever since I got into AI (and connectionism, since 1986) I have always
been amazed that my students and I can design a system that, at the I/O
level, will tell you, e.g. why Milton Friedman is against protectionism
(e.g. see Alvarado's 1989 PhD. or Alvarado, Dyer, Flowers in AAAI-86
proceedings or in Knowledge-Based Systems (in press)), but when you
open the program up, it's just the interaction of a bunch of dumb
parts. If you want to make Searle do lots of LISP CARs/CDRs/CONSs (or
lots of sums & sigmoidals), why should I judge the intentionality of
the resulting system just because Searle's personal experience is that
he's just doing lots of CDRs or sigmoidals? That would be like judging
YOUR intentionality by polling one of your neurons. If I did that, I
would conclude you have the same intentionality as that neuron.
[Searle's Argument, valid in my view, is an attempt to show
why those pure symbol manipulation systems, ALL of whose functions he
can himself perform, don't have intrinsic intentionality no matter
what you can read off the top. The Argument is not valid for every
possible system and every possible function. He CAN do all CARS/
CDRS, symbolic sums and symbolic sigmoid approximations. He could
not do "real" multiplication, if that meant a nonlinear physical
process. But then neither could a symbol-manipulator. It could
just do a symbolic approximation. For implementing a mind, this
kind of functional difference could be crucial. SH]
We really have to separate the business of (a) symbol grounding from
that of (b) at what LEVEL of system behavior we are going to look for
intentionality. It is the cheapest kind of trick for Searle to ask us
to imagine doing CDRs (or neuron spiking) and then ask us to judge the
intentionality of the system based on our intentionality as we do any on
(or all!) of these low-level tasks.
I will tell you what the simlutaneous "doing" of all of my neurons is
like -- it's ME. But just as a neuron can't tell you about ME, I cannot
tell you about what it's like to be a neuron. I suspect Searle realizes
that he's pulling "massive levels confusion" over on people, and if he
can bolster his argument by using symbol groundings, he will, but it's
a red herring.
[I must repeat, Searle has nothing against systems with dumb components,
e.g., the brain; his argument only works against pure symbol systems. SH]
> And he's simply right about that, according to me. But, as I show in my
> paper, it turns out that a very simple variant is immune to Searle's
> objection: One in which the candidate system's functions -- the ones
> all of which Searle is committed to performing himself, if his argument
> is to go through -- include certain nonsymbolic functions, in
> particular, sensory transduction. If, for example, the physical system
> that implements the mental state of seeing something -- that state we
> all know whether or not we're in just as directly as we all know
> whether or not we're understanding Chinese -- includes transducer
> function, then Searle cannot simulate it Chinese-Room style. For either
> he will have to get only the transducer's output (in which case it's no
> wonder he's not seeing, since he is not performing all the functions
> the system he is simulating is supposed to perform) or he will have to
> BE the transducer, in which case he will be seeing after all.
We're building a simulation in the UCLA AI lab that has a retina of
artificial neurons. This simulation is done on a connection machine,
but could be done on a von neumann machine. Given enough time, Searle
himself could simulate the entire system. The system learns to describe
the motion of a simulated visual object (i.e. neuron firing in a region
of a 1K x 1K array of neurons making up the retina and projecting to
other maps, which extract motion, direction, etc.).
[If your system uses a symbolically simulated retina, and symbolically
simulated visual objects, then Searle will indeed be able to simulate
it, and his Argument will apply, and he'll be right, because the system
won't really be seeing. But if the system uses for its retina real
transducers, that transduce real energy from real objects, then Searle
will not be able to perform this nonsymbolic function himself, and
hence whether the system could really see remains open. -- I would
personally still doubt it till it could pass the Total Turing Test for
some organism at least... SH]
I believe that the invariant representation (as a pattern) of, say, a
cat, should be formed from visual experience in seeing the shape of
cat, but that issue is not related to whether or not the whole process
can be run on a computer (it can). I will send you a review I wrote of
a talk Searle gave here at UCLA. His talk so infuriated me that I wrote
a critique of it, but have not sent it anywhere.
[Simulated transduction is not transduction. The whole process of
visual experience hence cannot be implemented on a computer, only
symbolically simulated. And, as Searle showed, pure symbol systems
don't have minds. SH]
> The reason Searle's argument works with understanding and fails with
> seeing is that the candidate system for understanding was purely
> symbolic, whereas the candidate system for sensing was not. It was this
> immunity of a seemingly trivial function (transduction) to Searle's
> Argument that confirmed my hunch that sensory grounding may be
> critical for modeling the mind. (And of course I believe that the place
> to ground understanding is in sensory function.)
Sensory grounding is critical for modeling certain ASPECTS of mind.
Grounding gives a greater richness to the symbols and allows the system
to reason about 3-D objects and motions, etc. in a way that AI systems
currently cannot. But we can smoothly weaken the amount of grounding.
E.g. imagine a robot/person with a lower-resolution retina, with lower
resolution of sensory membranes on the hands and in the joints. Or
imagine more and more "noise" in the distributed patterns that
reconstruct sensory experiences. But the fact that we can build systems
with symbolic systematicity properties that read fragments of
editorial text (although highly primitive) indicates that those kinds
of organizations capture ANOTHER ASPECT of mind, one that is only
loosely coupled to grounding symbols.
[In my view, sensory grounding is not just a matter of convenience or
greater "richness": it's essential for having a mind -- not for having
"aspects" of a mind, but for having a mind at all. Nor do I think it's
a matter of a sensory module you can simply add on to a symbol cruncher
module that captures another "aspect" of mind; a grounded system is
hybrid through and through. -- And I don't think having a mind is
captured "aspect by aspect." I think having a mind is all or none (and
that toy systems don't have any; only TTT-scale systems do). SH]
> And I think Searle would be agnostic about whether or not something had
> to be "alive" (by which I assume you mean a natural biological creature
> that is not dead) to have a mind. For example, I doubt Searle would
> have any problem with artificial neurons, as long as they had the same
> causal powers as natural neurons. But Searle's extracurricular beliefs
> are irrelevant. Insofar as they are based on the Chinese Room Argument,
> he has given strong reasons for doubting that mere symbol-crunchers
> ("Strong AI") can have minds. The rest is moot.
Searle must know that we can simulate artificial neurons on standard
computers (using the same LOAD, STORE and other register operations in
the machine that, by the way, are also used to simulate CDRs), so I
can't believe that he would accept artificial neurons, because he can
ask you to imagine being a neuron and then say "where's the
intentionality?" -- THAT's why I think it boils down to Searle needing
a real neuron (i.e. one that's alive). Then the whole Searlean argument
will repeat itself for Artificial Life (just as it has for Searle's
attack on Artificial Intelligence).
[Please reread the discussion of symbolic simulation versus causal
implementation of nonsymbolic functions, e.g., transduction.]
> I repeat: although Searle hasn't said so explicitly, I will: The only
> kind of artificial system that is vulnerable to his Chinese Room
> Argument is a pure symbol manipulating system. When it comes to
> nonsymbolic functions, all bets are off. Searle mentioned in passing
> that a system with a mind must have all the causal powers of a real
> brain, but I'm sure he would agree that many of the causal powers of
> the brain may be irrelevant to its having a mind (as Stephen Hawking's
> tragic case shows so dramatically), and that all he can really insist
> on are those causal powers of the brain that are necessary and
> sufficient for having a mind (which begins to sound a little
> tautological), and that any system, natural or artificial, that had
> those, would have a mind. Apart from that, Searle simply points out,
> correctly, that the kind of system he can simulate in the Chinese room
> (a pure symbol system) clearly does not have a mind, whereas the brain
> clearly does.
Searle never said very clearly just WHAT is in those books that the
person inside the chinese box (CBX) is using (in order for the CBX to
speak Chinese). Here's some content for those books: The books consist
of the circuitry (as drawings, or just as symbols and pointers,
whichever you prefer) of a connectionist system (with neurons as
detailed as you choose). Searle grabs the chinese characters and lays
them on a page that corresonds to the CBX's "retina". Searle then
traces (with many fingers, or over much time!) the values of "retinal
cells" where the characters overlapped. Searle needs lots of blank
pages for storing his intermediate calculations, since he has to
propagate the activation values (or firing pulse patterns, if you want
a more detailed level) to the next layer. Each page of the book has
different layers, with directions as to which pages to flip to and what
calculations to make. Some of the pages have millions/billions of
patterns and these represent memories. To simulate memory recall,
perhaps the CBX books direct him to do some tensor operations. In any
case, after many page flippings and MANY calculations, he arrives at a
pattern on a "motor page". This page explains how he should move his
hand to produce chinese characters on output. Searle won't understand
what he is doing, but the CBX will perhaps be responding (in chinese)
with a witty retort.
[All this symbol manipulation and symbol matching is exactly what
Searle can do, and his argument correctly shows that if that's all the
system does then it does not have a mind, no matter what comes out on
top. Note that connectionism is just as open to his critique as long as
there's no essential difference in causal power between a simulated and
an implemented net. Note, though, that in "Minds, Machines and Searle"
I give reasons why an ungrounded system would not be able to make all
the right words come out on top. That success is only conjectured,
after all. You may have to be able pass the TTT to pass the TT. SH]
> So the moral is simply that in the attempt isolate that subset of the
> brain's function (or any other possible mind-supporting system's
> function) that is necessary and sufficient for having a mind, a pure
> symbol crunching module turns out not to the right subset!
It turns out that scientists now carefully examine the CBX books and
arrive at the conclusion that a subset of the patterns on the CBX pages
are acting as patterns that are invariant with respect to the earlier
(sensory) pages. These invariant patterns have systematic relations to
one another. Some scientists decide to call these "symbols". Other
scientists figure out how to simulate the entire CBX book on a von
neumann machine. One crazy scientist builds a Turing machine to
actually simulate the von Neumann machine, and runs the CBX program on
it, and so claims that it is all "symbolic". Clearly, we have
terminological problems on our hands.
[See discussion of books, simulation and implementation, passim. SH]
I said: "I've never heard Searle talk about the importance of
grounding symbols in perception. If I'm wrong, please point me to
his relevant paper(s)."
> Neither have I. But *I've* talked (and written) about it. And whereas I
> first thought I was disagreeing with Searle, he assured me that we were
> largely in agreement. So there you have it.
I would be worried if I found Searle agreeing with ME! :-) Seriously,
it's nice that Searle recognizes the importance of grounding. But don't
confuse that with his (false) arguments on intentionality.
> The problem is that "Gothic Style," if it's anything at all, is an
> objective, observable property. (So is behavior under stress.) In
> principle we can all agree that any structure that has certain
> geometric properties, irrespective of any other properties, is Gothic.
> Unfortunately, this is not true of understanding (or seeing, or any of
> the other subjective properties of having a mind). For even if a system
> has all the objective, observable properties of a system that has a
> mind, it may not have a mind. And that can be true as a simple matter
> of fact, not just a metaphysical possibility, as Searle's Chinese Room
> Argument shows.
As you can tell by now, I am totally unconvinced by Searle, so
referring to his Chinese Box doesn't help me. Is mental function the
"style" of patterns, or the actual "substance" the patterns are made
of? Computer scientists are now building optical computers. It will be
the "style" of the photons that count for the information processing
level, even if it is the "substance" of the photons that allows one to
do things with greater parallelism and more rapidly.
[It's not a matter of style vs. substance but of pure symbol manipulation
versus nonsymbolic functions. SH]
> But here I actually part ways with Searle. Searle rejects all versions
> of the Turing Test, very much for the above reasons, whereas I reject
> only the traditional symbolic version of the Turing Test (symbols in,
> symbols out, and presumably just symbol-crunching in between),
> because of the Symbol Grounding Problem. The robotic version -- what
> I've called the Total Turing Test (TTT) -- is good enough for me,
> especially since that's all I can ever have to go on even with my
> fellow human beings. Appearances could be deceiving there too (as the
> "Other Minds Problem" reminds us), but as far as I'm concerned, Total
> Turing power is power enough to convince me.
Turing didn't consider direct sensory recognition and manipulation
tasks in his test (TT) because he wanted to focus on language-related
aspects of mind (and he didn't want people to see the computer and make
judgements based on prejudice). IF the computer in the TT does not have
something like images and perceptual reasoning, then it won't be able
to hold up it's side of a conversation about how things look and move,
or about imaginary visual objects. Turing did NOT disallow the use of
such conversations in his TT.
[I agree with this, in fact said it myself in "Minds, Machines and
Searle," but proponents of "Strong AI" don't. They think a repertoire
of symbol strings describing visual objects would be enough.]
It is possibly the case that one could have a computer that can discuss
and understand something about visual and other sensory experiences,
but cannot actually move and recognize objects (in the same way that
people who are learning, say, Spanish, can understand sentences they
cannot generate on their own). But the extent to which you can divorce
comprehension of words ABOUT the sensory world from actually patterns
stored AS A RESULT OF sensory experiences is an open empirical
question. I tend to agree with you that to give a system the ability to
pass the TT of talking about the world will require a system that has
had experience with the world (i.e. your TTT) (but mainly because
forming such representations by hand will be just too hard an
engineering feat -- we will want them formed automatically, from having
the robot interacting with the world directly).
[I agree that there's nothing sacred about real-time history and
learning; just the endstate, but I think that must must include
nonsymbolic representations and capacities in order to ground the
system. Real-time learning just seems the easiest and most natural
way to get there from here in most cases. SH]
> Besides, no one knows what the relevant causal powers of the brain are
> anyway, so that certainly can't be the way to settle the matter. The
> brain's equivalent of your clay's "behavior under stress" may or may
> not be one of the properties any system must have in order to have a
> mind. To put it another way, it's not clear how much of the TTT will
> have to include not only the body's behavior, but the brain's. Both,
> after all, are objective, observable properties. The more you capture,
> the better your grounds for confidence. I just think (total) bodily
> performance capacity is probably a sufficient constraint.
I agree in general, but I would grant civil rights to a robot that
HOBBLES and is BLIND, but can discuss sex, religion and politics,
before I give civil rights to an athletic, visually acute robot with
the mind of a dog.
[I'd want to protect anything that had a mind, but I happen to
be a vegetarian... SH]
> Searle's "elusive" intentionality is the only kind there is. It's the
> stuff the systems that really see or feel or understand [English or
> Chinese] have (and know at first hand that they have, because they know
> exactly what it feels like to see or feel or understand [English or
> Chinese]) and that systems that only act exactly as if they see or feel
> or understand [English or Chinese] don't have (but don't know they
> don't have, because there's nobody home). It's just that I don't happen
> to believe you can get a system to ACT (TTT) exactly as if it sees...
> etc., without its really having a mind.
Searle's "intentionality" is of the worst kind! It represents a
sophisticated form of prejudice. This prejudice is not based on gender,
culture, or race, but on a "neural machismo-ism". I would hate to be
relegated as a "non-intentional entity" by Searle just because, say, N
of my neurons are damaged, so someone replaces them with a microchip
that simulates their connectivity and firing patterns and then links
the chip up to neighboring, "real" neurons.
[This persistent misunderstanding of Searle does not advance matters:
I can only repeat: Searle would not deny the above as long as the
artificial parts had the right causal powers to keep sustaining your
mind. All he's claimed is that a system whose ONLY function is symbol
manipulation will not have a mind. SH]
> Pure symbol-crunchers, on the other hand, cannot by their very nature
> have ALL the requisite powers, because some of the powers are
> nonsymbolic -- and, according to my theory, the symbolic ones must be
> (nonmodularly) grounded in the nonsymbolic ones; so an autonomous
> symbol crunching mudule will have no mental powers at all.
We have two extremes of a spectrum: at one end we have lots of sensory
patterns and architecture for coordination and recognition tasks. At
the other end we have lots of sensory-invariant patterns, related to
one another in a logico-rational-linguistic manner. You're arguing, as
I see it, for moving the point on this scale from the logico-linguistic
extreme toward the middle and give richness to symbols. Others have
argued for moving from the sensory extreme to the middle and give
systematicity of abstract thought to distributed representations.
They're both going to meet in the middle somewhere.
[It's not a continuum. And it's not just a matter of the old
"bottom-up" approach meeting the "top-down" one in the middle: It's
bottom-up all the way. The shapes of symbol tokens in formal symbol
system are arbitrary. The shapes of objects in the world, and of the
features in their sensory projections, are nonarbitrary. Hence a
dedicated symbol system, whose functioning is constrained in this
second way, is no longer a formal symbol system at all: Its primitive
symbols are grounded by their resemblance and causal connection to the
real physical objects they pick out and stand for. SH]
Searle, however, is on another axis: one extreme is that it's "all
style", the other extreme is that it's "all substance". All AI and
connectionists that I know believe that it's "all style". Every time I
hear Searle, he seems to be saying it's "all substance".
[I don't think the style/substance distinction has any substance. One
can be a functionalist -- distinguishing structure from multiply
realizable function -- without being a symblic functionalist -- for
whom all function is symbolic function, forgetting that some
functions can only be simulated, but not implemented as symbol
manipulation. SH]
> From a TTT standpoint, AI's feats are all mere toys.
True, but AI has excelled at combinatorial, symbolic and complex
control architectures, while connectionist models have excelled at
sensory processing. The grounding problem addresses the issue of just
how the two approaches should be merged. Rather than a simple hybrid, I
think we both feel that symbols should be embedded in, and formed out
of, "connectoplasm" rather than symbols staying like ASCII and having
some kind of pointer to a sensory map.
[Dunno about "connectoplasm." For me sensory transduction, sensory
analogs and analog transformations are just as important; I would
use the nets to learn the features of sensory categories, thereby
connecting the category names -- the primitive symbols -- to the
objects they refer to. SH]
I'll send you my critique of Searle's UCLA talk.
-- Michael Dyer
∂06-Oct-89 1647 sl@van-bc.wimsey.bc.ca Re: Networks Considered Harmful - For Electronic Mail (fwd)
Received: from van-bc.wimsey.bc.ca by SAIL.Stanford.EDU with TCP; 6 Oct 89 16:47:29 PDT
Received: by van-bc.wimsey.bc.ca (/\=-/\ Smail3.1.17.5 #17.78 smail)
id <m0gD4BR-0000UmC@van-bc.wimsey.bc.ca>; Fri, 6 Oct 89 16:41 PDT
Message-Id: <m0gD4BR-0000UmC@van-bc.wimsey.bc.ca>
From: sl@wimsey.bc.ca (Stuart Lynne)
Subject: Re: Networks Considered Harmful - For Electronic Mail (fwd)
To: jmc@sail.stanford.edu
Date: Fri, 6 Oct 89 16:40:58 PDT
X-Mailer: ELM [version 2.2 PL11]
I read with great interest your recent comments in the telecom digest about
email and fax.
I've being working along similiar lines for the past few years. Starting with
a short stay at Sydney Developments locally where I clashed with their
philosophy about how and where X.400 mail systems should be implemented. I
did manage to demonstrate for them an X.400 to Fax gateway, but they showed
a distinct lack of interest. [They have since gone out of business.]
For the last year I've been implementing a Fax gateway under Unix. My
original impetus was simply that I wanted to extend my email so that I could
uses addresses like:
..!fax!1-800-555-5555!user
Of course my customers also want other facilities, but the original impetus
for me was an email extension.
> The main technical requirement is the development of a set of point-to
> point telephone mail protocols. Any of several existing network mail protocols
> could be adapted for the purpose. Presumably the same kinds of modems and
> dialers that are used for fax would be appropriate but would give better
> transmission speeds.
>
> Perhaps the organizationally simplest solution would be to get one or
> more of the various UNIX consortia to add a direct mail telephone protocol to
> UUCP. Such a protocol would allow mail to be addressed to a user-id at a
> telephone number. The computer would require a dialer and a modem with
> whatever characteristics were taken as standard and it would be well to use th
Are you aware of what Intel is offerring in their MS-DOS Fax product
(Connection Co-Processor). It allows an MS-DOS user to send and receive
fax's. And to send and receive EMail from another user equipped (of course)
with an Intel Connection Co-Processor.
We're just looking at adapting our Unix software to use their board. And one
of the options is to extend our current email to fax gateway to work with
their EMail connection.
I can see the following scenario. You are sending a mail message to a
mailing list. Various addresses on the list can be of the normal internet
format with others being a fax phone number. The message addressed to the
fax phone numbers are given to the fax gateway software. It dials the remote
fax. If it is a normal G3 type fax, the ASCII data is rendered to a bitmap
and delivered as a normal fax. If it is an extended machine (for example an
MS-DOS machine with Intel Co-processor, or a machine which adopts the Fax
EMail standard) the EMail is delivered directly.
There are other interesting developments in the G3 fax world that are
interesting as well. Error correction, higher speeds, higher resolution,
file transfer. It would appear that the G3 fax is capable of adapting new
functions. So hopefully if someone like Intel got behind an EMail standard
it could be adopted as a CCITT standard.
Anyway I would appreciate any further suggestions or comments you'd like to make
on this subject. I'd like to think that I'm in a position to actually try
and implement something along these lines.
--
Stuart.Lynne@wimsey.bc.ca uunet!van-bc!sl 604-937-7532(voice) 604-939-4768(fax)
∂06-Oct-89 2027 srh@flash.bellcore.com re: Free-Floating Intentionality
Received: from flash.bellcore.com (petrus.bellcore.com) by SAIL.Stanford.EDU with TCP; 6 Oct 89 20:27:37 PDT
Received: by flash.bellcore.com (5.58/1.1)
id AA14537; Fri, 6 Oct 89 23:29:17 EDT
Date: Fri, 6 Oct 89 23:29:17 EDT
From: srh@flash.bellcore.com (stevan r harnad)
Message-Id: <8910070329.AA14537@flash.bellcore.com>
To: JMC@sail.stanford.edu, srh@flash.bellcore.com
Subject: re: Free-Floating Intentionality
Cc: srh@flash.bellcore.com
Except that I'm not prepared to believe that a person
can get a case of dual personality just from memorizing
a set of meaningless symbols and rules for manipulating
them. Do you, really? I mean the shrinks are saying that
you need stronger stuff to cause such a radical change,
like early child abuse or drugs or something...
Stevan
∂08-Oct-89 0800 JMC
boyer, cindy, elliott
∂08-Oct-89 1051 boyer@CLI.COM Dinner
Received: from CLI.COM by SAIL.Stanford.EDU with TCP; 8 Oct 89 10:51:06 PDT
Received: by CLI.COM (4.0/1); Sun, 8 Oct 89 12:51:55 CDT
Date: Sun, 8 Oct 89 12:51:55 CDT
From: Robert S. Boyer <boyer@CLI.COM>
Message-Id: <8910081751.AA17624@CLI.COM>
To: jmc@sail.stanford.edu
Subject: Dinner
Reply-To: boyer@cli.com
You are invited to dinner at the Boyers Thursday evening.
∂08-Oct-89 1146 nilsson@Tenaya.Stanford.EDU
Received: from Tenaya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 8 Oct 89 11:46:12 PDT
Received: by Tenaya.Stanford.EDU (NeXT-0.8/25-eef) id AA04602; Sun, 8 Oct 89 11:44:20 PDT
Date: Sun, 8 Oct 89 11:44:20 PDT
From: Nils Nilsson <nilsson@tenaya.stanford.edu>
Message-Id: <8910081844.AA04602@Tenaya.Stanford.EDU>
To: feigenbaum@sumex-aim.stanford.edu, brown@sumex-aim.stanford.edu,
jmc@sail.stanford.edu, clt@sail.stanford.edu, val@sail.stanford.edu
latombe@cs, cannon@sierra, wiederhold@cs, luckham@sierra, hennessy@cs,
horowitz@sierra, plummer@sierra, vva@isl, nanni@mojave, cheriton@cs,
engelmore@sumex, rindfleisch@sumex, gibbons@sierra, levinthal@sierra,
goodman@sierra, linvill@sierra, losleben@cis, gupta@cs
Subject: Visit by Barry Boehm
BCC: nilsson
Barry Boehm, the new ISTO Director at DARPA and ex TRW person, will be
visiting Stanford on November 20 (Monday) and 21 (Tuesday). (The CIS
adcom meeting is on Nov. 20---a busy day.) Steve Cross, who is taking
over from Bob Simpson as a program director in ISTO, has suggested a
tentative schedule and has asked me to be the central Stanford point
in coordinating this schedule.
I'll attach Steve's suggested schedule below, but first some known
modifications and elaborations:
1) Barry would like to meet sometime during the visit with David
Cheriton (if we can bring him out of sabbatical for an hour). Steve
has suggested that Barry could meet David on Wednesday morning if
needed.
2) I have asked Steve whether or not Barry will be available for
dinner on either Monday or Tuesday evening.
3) Anoop Gupta is asking Barry if he could give a CS colloquium talk
on Tuesday afternoon at 4:15. We'll hear from Anoop on that soon.
4) I have invited Barry to come to our regular CS faculty lunch on
Tuesday to give us all a perspective on DARPA/ISTO these days.
5) I will be asking Jim Gibbons to help me welcome Barry to Stanford,
probably first thing Monday morning.
6) Steve says "Ideally the roundtable discussion at the end of
[Monday] would focus on AI and help shape Barry's vision about what
he should focus on during his tenure."
Here is the schedule proposed by Steve Cross. People should let me
know asap if there is anything about this schedule that won't work for
them. Also, we should treat the schedule as a suggestion. Perhaps
people close to the projects mentioned, and some of the senior people
on this distribution list, will propose major additions/deletions.
-------
Draft Agenda for Dr. Boehm's Site Visit to Stanford
Monday, 20 Nov 89
0830 - 0930 Overview, Dr. Nils Nilsson, Dept Head, Computer Science
0930 - 1200 Intelligent Systems for Science and Engineering Applications,
Dr. Ed Fegeinbaum (3 projects)
1200 - 1230 Lunch
1230 - 1315 Expert Systems - Multiprocessor Architectures, Dr. Harold Brown
1315 - 1430 Foundations of AI, Dr. John McCarthy
1430 - 1445 Break
1445 - 1600 Prof Jean-Claude Latombe, robotics
1600 - 1715 Prof Robert Cannon, robotics
1715 - 1830 Roundtable Discussion
Tuesday, 21 Nov 89
0830 - 0930 Data bases, Dr. Gio Weiderhold
0930 - 1100 Software, Dr. Luckham
1100 - 1230 Microsupercomputers, John Hennessey and Mark Horowitz
1230 - 1330 Lunch
1330 - 1430 Structured Process Flow, Jim Plummer
1430 - 1515 CAD for EEPROM, Abbas El Gammal
1515 - 1600 Logic Synthesis, Giovanni DeMicheli
-------
Presuming that something like this schedule will work, I propose that
the meetings on Monday morning, including a working box lunch, and up
through Harold Brown's talk, be at the KSL at 1101 Welch Road. To
save transportation time, I would recommend that I give my overview
talk there also. If we can get Jim Gibbons to come, perhaps we could
start at 8 am instead of at 8:30.
In order to see some robotics activity, I propose we travel to Cedar
after Brown's talk to hear from John McCarthy (and/or Vladimir, Carolyn)
and then hear about robotics. We can then take a brief walk to Durand
(if Bob Cannon agrees) to see the granite-table robots and then to
have the roundtable discussion, say, in Durand 450 with refreshments on
the patio outside of Durand 450. Bob, can you arrange for that? Having
the roundtable in Durand 450 would give us plenty of room to be able
to invite several people.
Then, on Tuesday, I propose we have the meetings in the CIS building.
Perhaps John Hennessy, with Paul Losleben's help, can make any necessary
arrangments (reserving CIS 101, etc.). Barry can walk to mjh for the
CS faculty lunch and then back to CIS. Then back to Psych 040 if Barry
agrees to give the cs colloquium.
Sound ok?
-Nils
∂08-Oct-89 1439 boyer@CLI.COM Dinner
Received: from CLI.COM by SAIL.Stanford.EDU with TCP; 8 Oct 89 14:39:34 PDT
Received: by CLI.COM (4.0/1); Sun, 8 Oct 89 16:40:10 CDT
Date: Sun, 8 Oct 89 16:40:10 CDT
From: Robert S. Boyer <boyer@CLI.COM>
Message-Id: <8910082140.AA18491@CLI.COM>
To: JMC@SAIL.Stanford.EDU
Subject: Dinner
Reply-To: boyer@cli.com
Wednesday would be an even better evening for a dinner at the Boyers.
Let us know if you can fit it in. I could pick you up at the airport.
∂09-Oct-89 0209 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 9 Oct 89 02:09:18 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 434132; 9 Oct 89 05:05:49 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 109722; Mon 9-Oct-89 02:01:49 PDT
Date: Mon, 9 Oct 89 02:01 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: large periods
To: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, math-fun@RUSSIAN.SPA.Symbolics.COM,
"Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
cc: "hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"michaelg@polya.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19891002065437.1.RWG@TSUNAMI.SPA.Symbolics.COM>
Message-ID: <19891009090130.4.RWG@TSUNAMI.SPA.Symbolics.COM>
Date: Sun, 1 Oct 89 23:54 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
. . .
With these new components, just two guns and five shuttles (half-guns)
suffice to make a shiftregister pseudorandom generator whose period
scales exponentially with the shuttle spacing. With an additional 90
degree shuttle, the pseudorandom lightweight beam escapes.
I figured out how to get a pseudorandom glider beam without the additional
shuttle. The smallest case (a[n] = a[n-1] eqv a[n-12]) fits (except for the
beam) in 128 by 140, and runs at a visually satisfying 12.8 hz in the clife
(bitblt) program. Period = 149730 = 46*7*15*31. Let me know if you want
the dots emailed, or a fax. It looks pretty neat after one period, at zoom
1. About 60 gliders fit on the page, forming a crazy Morse-code message.
The period of a[n] = a[n-1] eqv a[n-k] is an interesting function of k:
k period
2 l(2)
3 l(3)
4 l(4)
5 l(2,3)
6 l(6)
7 l(7)
8 l(2,6)
9 l(9)/l(3)
10 l(3,7)
11 l(2,9)
12 l(3,4,5)
13 l(5,8)
14 l(2,5,7)
15 l(15)
16 l(8,8)
17 l(2,3,12)/l(4)
18 l(5,13)
19 l(3,4,5,7)
20 l(2,5,13)
21 l(7,14)/l(2)
22 l(22)
23 l(2,8,13)
24 l(3,21)
25 l(6,8,11),
where l(a,b,...) = lcm(2↑a - 1,2↑b - 1, ...). Note the fudging at 8,
9, 17, 21, . . . , indicating my failure even to characterize this
function, let alone explain it. (Back in the 60s, Rollo Silver gave
me a Lincoln Labs paper by N. Zierler which probably explained all
this in terms of polynomials with coefficients mod 2.)
Another oddity: it is easy to write a loop to run this recurrence
backwards k steps/whack. But frontwards seems almost as hard as
a full convolution, capable of doubling the current number of steps.?
My fancy program hits a paging wall at about 2 million, and thus must
cheat by irreversibly ripening the universe every a million or two.
Then it hit a much harder wall at 4.8 million. I am seriously faked out.
The ripened universe at that point occupies only 20k (six-slot)
macrocells, yet the machine pages for days when trying to ripen a measly
2↑16. I'm beginning to think I tricked the ephemeral GC into some sort of
pessimal area growth, perhaps when a flip got postponed a couple of hours.
20k is a surprising compression of that pseudorandom stick of 100k
lightweights, considering how wretchedly many phases of period 46 there
are mod 2↑n, the cell sizes. Maybe compressibility by this method will
become a new randomness metric :)
I checked that the large period sequences are not the binary expansions
of relatively small rationals. (If they were, they'd be a lot cheaper
for conventional machines to generate.)
∂09-Oct-89 0932 @MCC.COM:EATON@MAXIMILLION.CP.MCC.COM last questions on arrangements for MCC visit
Received: from MCC.COM by SAIL.Stanford.EDU with TCP; 9 Oct 89 09:31:58 PDT
Received: from MAXIMILLION.CP.MCC.COM by MCC.COM with TCP/SMTP; Mon 9 Oct 89 11:32:58-CDT
Date: Mon 9 Oct 89 11:19:51-CDT
From: EATON@MAXIMILLION.CP.MCC.COM
To: jmc%sail.stanford.edu@mcc
Subject: last questions on arrangements for MCC visit
Dr. McCarthy,
Have you decided if you would like us to arrange dinner on
Wednesday might?
You will have all day Thursday to spend with Advanced Computing
Technologies researchers (unless you want to go to the
presentations at Packaging/Interconnect). I assume that Doug and
Mary are arranging this for you. The Science Fiction Writers panel
is from 3 - 5 Thursday. We will have a seat reserved for you.
The ACT presentations to the Associates and Shareholders are on
Friday and you are invited to them all.
I am forwarding a list of RSVPs to the workshop. We are getting
an unexpectedly good turnout this time, including a number of
people from companies that are thinking about joining MCC. You will
have an audience of about 100 for your dinner talk.
Please let me (eaton@mcc.com) or Susie (gifford@mcc.com) know if
there is anything we can do to help.
Sincerely,
Mark Eaton
∂09-Oct-89 0932 @MCC.COM:EATON@MAXIMILLION.CP.MCC.COM list of attendees to MCC workshop, as of last week
Received: from MCC.COM by SAIL.Stanford.EDU with TCP; 9 Oct 89 09:32:14 PDT
Received: from MAXIMILLION.CP.MCC.COM by MCC.COM with TCP/SMTP; Mon 9 Oct 89 11:33:03-CDT
Date: Mon 9 Oct 89 11:20:49-CDT
From: EATON@MAXIMILLION.CP.MCC.COM
To: jmc%sail.stanford.edu@mcc
Subject: list of attendees to MCC workshop, as of last week
From: PONY%"SUSIE@MAXIMILLION.CP.MCC.COM" 6-OCT-1989 16:28
To: EATON
Subj: [Resent-MM Mail] Fall Associates Workshop headcount
Date: Fri 6 Oct 89 16:27:43-CDT
From: Susie Gifford <SUSIE@MAXIMILLION.CP.MCC.COM>
Subject: Fall Associates Workshop headcount
To: eaton@MAXIMILLION.CP.MCC.COM, delfeld@MAXIMILLION.CP.MCC.COM,
maysonave@mcc.com, ohara@mcc.com, herrell@mcc.com, kroger@mcc.com,
wilson@MAXIMILLION.CP.MCC.COM, susie@MAXIMILLION.CP.MCC.COM,
allard@mcc.com, shepherd@mcc.com, hearn@mcc.com, whalen@mcc.com
Message-ID: <623712463.670000.SUSIE@MAXIMILLION.CP.MCC.COM>
Mail-System-Version: <VAX-MM(243)+TOPSLIB(135)+PONY(228)@MAXIMILLION.CP.MCC.COM>
Fall Associates Workshop RSVPs
------------------------------
Count as of 10/6/89: 92
SHAREHOLDERS: (10 companies represented)
------------
AMD
Ben Oliver
Larry Drake
Ben Bennett
Kamil Tueni
Bill Hughes
Naidu Golla
Joe Akins
Bellcore
Bill Mansfield
Steve Cheng
Steven Rohall
SBC Technology Resources (Bell RBOC)
Steve Dimmit (PI only)
Stephen Zvolner
John Tadlock
U. S. West Advanced Technologies (Bell RBOC)
Paul Bauer
NYNEX (Bell RBOC)
John Thomas
Boeing
K. C. Chang
GE
Constatine Neugebauer
Harris
Bruce Kraemer
Jay Muns
Ed Andre
Lockheed
Chin-Liang Chang
Mark Longley
Motorola
Maurice Karpman
John W. Stafford
Rockwell
John Mather
National Semiconductor
Court Skinner will send someone
Westinghouse
R. N. Horton
ASSOCIATES: (9 companies represented)
-----------
Allied
Robert T. Kane
Dell Computer
Byron Jarratt
Greg Stewart
Terry Parks
E-Systems
Samuel Hubbard
2 others
Paul O. Johnson (Garland Division)
Mr. J. J. Jackson (")
Mr. K. K. Agarwal (")
Mr. C. T. Brodnax (")
Ms. C. A. Davies (")
2 others (")
Randy Turnbull (ECI Division)
Bill Mahood (Melpar Division)
Bruce Dautrick (")
Tim Harvey (")
Keith Seawright (Greenville Division)
Mark White (")
Mark Przilas (")
4 others from Greenville Division
General Dynamics
Sylvia Darnall
Adrien Presley (Mr.)
2 from Corporate
Joe Van Fossen (Pomona Division)
Lew M. Jobe
LTV
Jim Billingsley
NSA
Jim Openshaw
SEI
Stan Rifkin
Tracor
Dennis Sustare
Dick Battin
Ed Smith
Dr. Ron Ham
United Technologies
Frank Bresnock
Petre Dan
2 others
PROSPECTS: (14 companies represented)
----------
ALCOA
Sam Bansal
Kal Buchavelky
Molex
Ivor Redmond
Kulicke & Soffa
George Berkin
Space Industries Inc.
James Calaway
Intel
Shuck Steidel
Compaq
Tim Magnuson
Chris Gintz
Keith Lucke
SRI, David Sarnoff Research Center
L. Onyshkevych
OACIS
Ted Lewis
SRC
John H. Kelly
ITT Defense
Nick Manzo
Steve Reiss
FMC
Dr. N. S. Sridharan
Olin Mesa
Scott Voss
Unisys
Dennis Dean
Dave Ray
Teradyne
Garry Gillette
Evans & Sutherland
Donald R. Mullen
MITRE
Chuck Howell
Andrea Weiss or Jim Antonisse
Ashton-Tate
Harry Wong
-------
∂09-Oct-89 0950 boyer@CLI.COM Dinner is on.
Received: from CLI.COM by SAIL.Stanford.EDU with TCP; 9 Oct 89 09:50:53 PDT
Received: by CLI.COM (4.0/1); Mon, 9 Oct 89 11:51:41 CDT
Date: Mon, 9 Oct 89 11:51:41 CDT
From: Robert S. Boyer <boyer@CLI.COM>
Message-Id: <8910091651.AA22988@CLI.COM>
To: JMC@SAIL.Stanford.EDU
Subject: Dinner is on.
Reply-To: boyer@cli.com
Good. I'll meet you at the Austin airport, this Wednesday, at 4:55pm
flight SW 957. We'll have dinner at the Boyers that evening.
∂09-Oct-89 1419 arg@lucid.com bug fixed in new-qlisp
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 9 Oct 89 14:19:29 PDT
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA16154; Mon, 9 Oct 89 14:18:28 -0700
Received: from bhopal ([192.43.178.13]) by heavens-gate id AA19493g; Mon, 9 Oct 89 14:16:47 PDT
Received: by bhopal id AA26003g; Mon, 9 Oct 89 14:18:11 PDT
Date: Mon, 9 Oct 89 14:18:11 PDT
From: Ron Goldman <arg@lucid.com>
Message-Id: <8910092118.AA26003@bhopal>
To: qlisp@go4.stanford.edu
Subject: bug fixed in new-qlisp
There was a bug with returning multiple values in the new stack code.
I just patched new-qlisp to fix it. Let me know if you encounter any
more problems.
Ron
∂09-Oct-89 1625 @Score.Stanford.EDU,@ai.ai.mit.edu,@mc.lcs.mit.edu,@central.cis.upenn.edu:dale@linc.cis.upenn.edu CFP: Third Logical Biennial; Chaika, Bulgaria
Received: from Score.Stanford.EDU by SAIL.Stanford.EDU with TCP; 9 Oct 89 16:25:49 PDT
Received: from lcs.mit.edu (MINTAKA.LCS.MIT.EDU.#Internet) by SCORE.STANFORD.EDU with TCP; Mon 9 Oct 89 16:26:14-PDT
Received: from ai.ai.mit.edu by mintaka.lcs.mit.edu id aa10486;
9 Oct 89 19:22 EDT
Received: from MC.LCS.MIT.EDU (CHAOS 3131) by AI.AI.MIT.EDU 9 Oct 89 19:21:37 EDT
Received: from lcs.mit.edu (CHAOS 15044) by MC.LCS.MIT.EDU 9 Oct 89 19:19:47 EDT
Received: from CENTRAL.CIS.UPENN.EDU by mintaka.lcs.mit.edu id aa10346;
9 Oct 89 19:12 EDT
Received: from LINC.CIS.UPENN.EDU by central.cis.upenn.edu
id AA21215; Mon, 9 Oct 89 19:12:01 -0400
Return-Path: <dale@linc.cis.upenn.edu>
Received: by linc.cis.upenn.edu
id AA03155; Mon, 9 Oct 89 19:11:52 EDT
Date: Mon, 9 Oct 89 19:11:52 EDT
From: Dale Miller <dale@linc.cis.upenn.edu>
Posted-Date: Mon, 9 Oct 89 19:11:52 EDT
Message-Id: <8910092311.AA03155@linc.cis.upenn.edu>
To: types@theory.lcs.mit.edu, theorem-provers@mc.lcs.mit.edu
Subject: CFP: Third Logical Biennial; Chaika, Bulgaria
FIRST ANNOUNCEMENT AND CALL FOR PAPERS
THIRD LOGICAL BIENNIAL
In Honour of S.C Kleene. June 6 - 15, 1990
Chaika near Varna; Bulgaria
Kleene'90 is the third edition of the Logical Biennial, succeeding
Goedel'86 and Heyting'88. Its scientific programme includes a Summer
School with invited lectures, a Conference with selected contributions,
discussions and informal seminars, and will focus on the following
TOPICS: Recursion Theory
Modal and Non Classical Logic
Proof Theory
Logic in Theoretical Computer Science
Applications of Logic to Mathematics
Kleene Session
The PROGRAMME COMMITTEE presently consists of
Johan van Benthem (Amsterdam) George Metakides (Brussels)
William Blok (Christchirch, NZ) Albert Meyer (Cambridge, MA)
George Boolos (Cambridge, MA) Dale Miller (Philadelphia)
Samuel Buss (La Jolla) Grigorii Mints (Tallinn)
Barry Cooper (Leeds) Yiannis Moschovakis (Los Angeles)
Kosta Dosen (Belgrade) Anil Nerode (Ithaca)
Leo Esakia (Tbilisi) Jeff Paris (Manchester)
Ronald Fagin (San Jose) Solomon Passy (Sofia)
Solomon Feferman (Stanford) Petio Petkov (Sofia)
Melvin Fitting (Bronx) Richard Shore (Ithaca)
George Gargov (Sofia) Theodore Slaman (Chicago)
Jean-Yves Girard (Paris) Craig Smorynski (Westmont)
Robert Goldblatt (Wellington) Ivan Soskov (Sofia)
Valentin Goranko (Sofia) Gaisi Takeuti (Urbana)
Lyubomir Ivanov/Chairman (Sofia) Tinko Tinchev (Sofia)
Alexander Kechris (Pasadena) John Tucker (Swansea)
Antonin Kucera (Prague) Vladimir Uspenski (Moscow)
Manuel Lerman (Storrs)
There will be about 25 INVITED SPEAKERS to be announced in the next
communication.
Submissions to the Conference from the above mentioned areas will be
evaluated by the Programme Committee for inclusion in the Proceedings.
Papers must contain original contributions, be clearly written and include
appropriate reference and comparison with related work. Five copies of a
draft full paper of no more than 15 double-spaced standard pages
accompanied by one-page camera-ready ABSTRACT should reach the PC Chairman
according to the following
DEADLINES November 15, 1989 - submission
March 15, 1990 - notification
May 20, 1990 - final version due
GRANTS
There will be a limited number of grants covering the local expenses
during the meeting. Applicants should send a Curriculum Vitae, list of
publications and a cover letter describing their request. Applications
should reach the OC Chairman prior to November 15, 1989.
ORGANIZING COMMITTEE Valentin Goranko, Lyubomir Ivanov,
Solomon Passy, Petio Petkov/Chairman
Please, direct requests for inclusion in the mailing list and all
correspondence regarding the Meeting to the Chairman at:
MAILING ADDRESS Kleene '90 Mathematic Faculty;
Blvd. Anton Ivanov 5; Sofia 1126; Bulgaria
CABLES kleene '90; mathemat faculty; sofia 1126; bulgaria
PHONES: (+359.2) 62561, exts. 598, 524
Dated: May 25, 1989
-----------------------------------------------
APPENDIX TO THE FIRST ANNOUNCEMENT
PROGRAMME COMMITTEE:
The Programme Committee of Kleene'90 has been joined in the meantime by
Michael Beeson, Gian-Carlo Rota and Gerald Sacks.
INVITED LECTURERS:
The following have accepted the invitation to give lectures:
Marat Arslanov, S. Artemov, George Boolos, Samuel Buss, Kosta Dosen, Leo
Esakia, Melvin Fitting, Jean-Yves Girard, Alexander Kechris, Jan Krajicek,
Antonin Kucera, Jim Lambek, Yiannis Moschovakis, Anil Nerode, Dana Scott,
Gaisi Takeuti, Alasdair Urquhart, Dimiter Vakarelov, Frank Veltman, Jeff
Zucker.
Due to various reason, few of them need more time to finally confirm their
participation. Several other invitees are yet to reply.
Dated: September 1, 1989
-----------------------------------------------
∂09-Oct-89 1630 MPS Keynote speaker
Judy Fortune, Seattle Pacific Univ (206) 281-3503
Would like you to be keynote speaker for knowledge based
systems seminar on April 17, 1990. They want to know what
your honorarium would be. It is being sponsored by Seattle Pacific U and
Boeing Computer Services.
∂09-Oct-89 2026 H.HARPER@Macbeth.Stanford.EDU re: IN THE NEWS II
Received: from Macbeth.Stanford.EDU by SAIL.Stanford.EDU with TCP; 9 Oct 89 20:26:11 PDT
Date: Mon 9 Oct 89 19:40:24-PDT
From: spinning wheel <H.HARPER@Macbeth.Stanford.EDU>
Subject: re: IN THE NEWS II
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <FiyuU@SAIL.Stanford.EDU>
Message-ID: <12532871697.27.H.HARPER@Macbeth.Stanford.EDU>
oh... From what I gather.. Berkeley had some instances of "racism" similar
to what happened here at Stanford. So, the result was a new policy...not
unlike the attempted revision of the SU Fundamemtal Stndard.
The new speech policy cites that verbal harrassment of others is
grounds for dismissal... (it reads much the same way that the SU revision
did). Un/fortunately, the speech policy was passed.. some students tried
to protest against it, but to no avail.
I'm sorry about the vaguity of the exact rules of the new Berkeley
"speech policy", but I don't have a copy of it with me.
-H.Harper
-------
∂10-Oct-89 1156 SUSIE@MAXIMILLION.CP.MCC.COM Fall Associates Workshop
Received: from MAXIMILLION.CP.MCC.COM by SAIL.Stanford.EDU with TCP; 10 Oct 89 11:56:32 PDT
Date: Tue 10 Oct 89 11:35:54-CDT
From: Susie Gifford <SUSIE@MAXIMILLION.CP.MCC.COM>
Subject: Fall Associates Workshop
To: jmc@sail.stanford.edu
Cc: eaton@MAXIMILLION.CP.MCC.COM, susie@MAXIMILLION.CP.MCC.COM
Message-ID: <624040554.960000.SUSIE@MAXIMILLION.CP.MCC.COM>
Mail-System-Version: <VAX-MM(243)+TOPSLIB(135)+PONY(228)@MAXIMILLION.CP.MCC.COM>
I have reserved a VIP Room for you at the Austin Stouffer Hotel
on Wednesday and Thursday nights. The address is 9721 Arboretum
Boulevard. Will the Boyers' pick you up at the airport? Would
you like a rental car to have? Or I would be happy to pick you
up. Let me know if there is anything I can do. We are excited
about having you here!
-------
∂10-Oct-89 1202 ACW@ELEPHANT-BUTTE.SCRC.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 10 Oct 89 12:02:34 PDT
Received: from YUKON.SCRC.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 434640; 10 Oct 89 13:17:06 EDT
Received: from WHIMBREL.SCRC.Symbolics.COM by YUKON.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 498604; Tue 10-Oct-89 13:17:53 EDT
Date: Tue, 10 Oct 89 13:16 EDT
From: Allan C. Wechsler <ACW@YUKON.SCRC.Symbolics.COM>
Subject: large periods
To: rwg@RUSSIAN.SPA.Symbolics.COM, "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
math-fun@RUSSIAN.SPA.Symbolics.COM, "Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
cc: "hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19891002065437.1.RWG@TSUNAMI.SPA.Symbolics.COM>
Message-ID: <19891010171652.5.ACW@WHIMBREL.SCRC.Symbolics.COM>
Suppose you have a Life oscillator. At any given generation it fits
inside some bounding rectangle. Look at the area of the bounding
rectangle over all generations in the cycle of the oscillator. Define
the "area" of the oscillator to be the maximum area achieved by the
bounding rectangle over a complete cycle of the oscillator.
The question has come up, "what is the largest period oscillator that
can be packed into a rectangle of area A?" We already know how to make
oscillators whose period scales linearly with A. Can we do better?
Gosper has tried to answer this by building shift-and-xor machines.
By juxtaposing linear oscillators of periods K, 2K, ... nK, we can build
an oscillator of period lcm(K,2K,...,nK) in area n(n-1)SK for some
constant S. I don't know how to derive an asymptotic formula for the
lcm. Is this exponential yet?
I know Gosper wants to forbid (on esthetic grounds) these lcm
oscillators where no individual bit has the desired period. But I also
know that if I presented such an oscillator to Gosper, he would be able
to cludge it so that there was a bit of at least the desired period.
Handwaving plan: have a halo of circulating gliders around the whole
structure. Each time the stream passes one end of a component
oscillator, have a "tap" into the oscillator that xors with the
surrounding glider stream.
∂10-Oct-89 1327 0002814144@mcimail.com RE: test and ssx
Received: from NRI.Reston.VA.US by SAIL.Stanford.EDU with TCP; 10 Oct 89 13:27:31 PDT
Date: Tue, 10 Oct 89 16:19 EDT
From: Jerry Pournelle <0002814144@mcimail.com>
To: John McCarthy <JMC@sail.stanford.edu>
Subject: RE: test and ssx
Message-Id: <20891010201902/0002814144NB3EM@mcimail.com>
Received. Thanks. Now let's see if you get this one back.
Does ARPA Net still exist, and is there a way to get on
through some local node, or must I call Stanford? I assume the JXP
account still exists although I haven't acces{ed it in YEARs.
Forgotten the password, too, alas.
Too damned much work to do.
∂10-Oct-89 1345 MPS
Richard Lyman returned your call 5-1496
∂10-Oct-89 1353 betsy@russell.Stanford.EDU CSLI TINLunch
Received: from russell.Stanford.EDU by SAIL.Stanford.EDU with TCP; 10 Oct 89 13:53:27 PDT
Received: by russell.Stanford.EDU (4.0/inc-1.0)
id AA15865; Tue, 10 Oct 89 13:55:03 PDT
Date: Tue 10 Oct 89 13:55:03-PDT
From: Betsy Macken <BETSY@CSLI.Stanford.EDU>
Subject: CSLI TINLunch
To: jmc@sail.stanford.edu
Cc: peters@CSLI.Stanford.EDU
Message-Id: <624056103.0.BETSY@CSLI.Stanford.EDU>
Mail-System-Version: <SUN-MM(242)+TOPSLIB(128)@CSLI.Stanford.EDU>
Stanley just told me that you had offered to give a talk on Elephant to
CSLI. We'd love to have you do that! We still have TINLunch
slots that I can switch around, so I could give you either
Oct. 26, Nov. 2, or Nov 9. Remember that TINLunch has been expanded
since we are having seminar series instead of individual talks at
the 2:15 seminar. So TINLunch would be a perfect place for you to
give this talk. Would any of those dates work for you?
Betsy
-------
∂10-Oct-89 1413 betsy@russell.Stanford.EDU Re: date, title and abstract
Received: from russell.Stanford.EDU by SAIL.Stanford.EDU with TCP; 10 Oct 89 14:12:38 PDT
Received: by russell.Stanford.EDU (4.0/inc-1.0)
id AA16054; Tue, 10 Oct 89 14:14:11 PDT
Date: Tue 10 Oct 89 14:14:10-PDT
From: Betsy Macken <BETSY@CSLI.Stanford.EDU>
Subject: Re: date, title and abstract
To: JMC@SAIL.Stanford.EDU
Cc: ingrid@CSLI.Stanford.EDU
Message-Id: <624057250.0.BETSY@CSLI.Stanford.EDU>
In-Reply-To: <1FjrES@SAIL.Stanford.EDU>
Mail-System-Version: <SUN-MM(242)+TOPSLIB(128)@CSLI.Stanford.EDU>
OK, let's confirm Oct. 26, at noon, in the Cordura conference room.
Thanks for offering and for the abstract.
Betsy
-------
∂10-Oct-89 1525 MPS size
The lady at the faculty club does not know the size of the
rooms, but each table in the small rooms is seated for 12 people.
∂10-Oct-89 1530 @IU.AI.SRI.COM,@peabody.teleos.com:leslie@teleos.com Procedural questions
Received: from IU.AI.SRI.COM by SAIL.Stanford.EDU with TCP; 10 Oct 89 15:30:39 PDT
Received: from peabody.teleos.com by IU.AI.SRI.COM via SMTP with TCP;
Tue, 10 Oct 89 15:30:44-PST
Received: by peabody.teleos.com (3.2/4.16) id AA13859 for
nilsson@tenaya.stanford.edu; Tue, 10 Oct 89 15:27:28 PDT
Date: Tue, 10 Oct 89 15:27:28 PDT
From: Leslie Kaelbling <leslie@teleos.com>
Message-Id: <8910102227.AA13859@peabody.teleos.com>
To: nilsson@tenaya.stanford.edu, jmc@sail.stanford.edu
Subject: Procedural questions
Reply-To: leslie%teleos.com@ai.sri.com
I have a couple of procedural questions that we didn't get to address this
morning:
- If the dissertation outline is reasonable then I'll start writing actual
chapters. What is the standard mechanism for having them read? I guess the
default is to just distribute things as I finish them and hope for the best.
- I would like to receive "Distinction in Teaching." In order to satisfy the
requirements I must teach or TA one more 200 level course. I would like to
teach or TA CS221 in the Spring, if possible. Also, no-one I have spoken to
(Sharon Hemenway, Betty Scott) has any idea what it means to have this
distinction, since it was created as a tax dodge and has never before been
awarded.
∂10-Oct-89 1530 @IU.AI.SRI.COM,@peabody.teleos.com:leslie@teleos.com Meeting?
Received: from IU.AI.SRI.COM by SAIL.Stanford.EDU with TCP; 10 Oct 89 15:30:44 PDT
Received: from peabody.teleos.com by IU.AI.SRI.COM via SMTP with TCP;
Tue, 10 Oct 89 15:31:16-PST
Received: by peabody.teleos.com (3.2/4.16) id AA13862 for
jmc@sail.stanford.edu; Tue, 10 Oct 89 15:28:07 PDT
Date: Tue, 10 Oct 89 15:28:07 PDT
From: Leslie Kaelbling <leslie@teleos.com>
Message-Id: <8910102228.AA13862@peabody.teleos.com>
To: jmc@sail.stanford.edu
Subject: Meeting?
Reply-To: leslie%teleos.com@ai.sri.com
Would you like to schedule an appointment for me to fill you in on
what I've been doing?
- Leslie
∂10-Oct-89 1632 MPS trip
In your travel folder there is a letter ( I think on top)
that confirms your hotel reservation, and if I remember something
about transportation when you arrive there. Have a good trip.
Pat
∂10-Oct-89 2237 harnad@clarity.Princeton.EDU Verificationism and the Symbol Grounding Problem
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 10 Oct 89 22:37:34 PDT
Received: from reason.Princeton.EDU by Princeton.EDU (5.58+++/2.22/mailrelay)
id AA27524; Wed, 11 Oct 89 01:31:08 EDT
Received: by reason.Princeton.EDU (4.0/1.86)
id AA06633; Wed, 11 Oct 89 01:28:30 EDT
Date: Wed, 11 Oct 89 01:28:30 EDT
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8910110528.AA06633@reason.Princeton.EDU>
To: harnad@elbereth.rutgers.edu
Subject: Verificationism and the Symbol Grounding Problem
Is Symbol Grounding Just Verificationism?
To: Michael Nitabach, Brain & Cognitive Sciences, MIT <miken@ai.mit.edu>
Mike, you wrote:
> for a symbol to be grounded is to possess a procedure for verifying on
> the basis of transducer outputs that its referent is now presenting
> itself to the organism. The only way that I see this as differing from
> standard procedural semantics is that you are postulating a new
> mechanism for obtaining these procedures--namely, as the emergent
> result of the operation of a connectionist architecture. How much
> evidence is there that *unsupervised* connectionist networks can and do
> learn appropriate input categorizations?
Some features of my model may resemble verificationist or procedural
semantics. I'm not sure, because I've never seen it put together quite
the way I propose. Verificationism is a philosophy, whereas what I'm
addressing is the question of how to generate actual performance:
discrimination, categorization and description of objects from their
sensory projections. I doubt that the British Associationists or
the logical positivists or the present-day verificationists have ever
made this their explicit modeling task, so we cannot judge empirically
whether or not they had any way to accomplish it. (The ironic thing is
that even my own proposal is at a level of abstraction that is quite
remote from actual implementation. Nevertheless, the constraint of
aiming at implementable, performance-generating mechanisms inevitably
gives rise to a different flavor of theorizing.)
I also think that, apart from proposing connectionism as the
feature-learner, I've in addition proposed an original complementary
role for iconic (analog) projections of the sensory surface and for
context-dependent category invariants in generating discrimination,
similarity judgment and identification performance. Finally, I don't
see why you would think I was particularly interested in UNsupervised
nets. Category learning is supervised learning par excellence,
constrained by feedback from the consequences of MIScategorization.
So it's the capacity of supervised nets that's at issue. (It may well
prove insufficient, though.)
> How likely do you think it is that... all primitive (in the sense of
> non-decomposable) symbols can be defined solely in terms of sensory
> properties? [or] that most symbols will turn out to be complex--built
> out of a small set of elementary, sensorily grounded, ones... a browse
> through the dictionary doesn't leave the impression that many words are
> defined in terms of a small subset of sensorily grounded ones...
> (example from Fodor... "The Present Status of the Innateness Controversy".)
I think it's not only likely, but necessary. But not being a
philosopher or a lexicographer, it's not "definitions" I'm looking for.
It's a categorization mechanism that is capable of discriminating,
naming and describing objects and states of affairs as we do. Since I
don't believe in magic, and I'm certainly not prepared to subscribe to
what I've dubbed the "Big Bang Theory of the Origin of Knowledge"
(whereas Fodor would), according to which our categories are
unlearnable from sensory data and hence must be innate, I conclude that
the basis for our successful sorting and labeling must be contained in
the input, otherwise we simply wouldn't be able to do it.
Fodor and many others (Wittgenstein among them) are believers in
"vanishing intersections." They think that the objects and states of
affairs that we name and describe DON'T HAVE the sensory invariants
that are sufficient to generate our successful categorization
performance (nor are they recursively grounded in the names of objects
and states of affairs that do). I think they do have them, for the
following reasons:
I don't think it has ever really been tested whether or not
category invariants exist, because no one has really tried yet.
Instead, theorists have simply entered the category hierarchy at an
arbitrary point, picked a category -- say, "games," or "goodness," or
"symbolist poetry" -- and declared: I can't define this in terms of
features that are necessary and sufficient for being a member (and
certainly not sensory ones!), therefore they do not exist. (This is
discussed in the last chapter of "Categorical Perception: The
Groundwork of Cognition," Cambridge University Press 1987, S. Harnad,
ed.)
Well, first of all, the ontological version of this question -- the
problem of defining what things really exist, in terms of their
essential properties -- is simply not the psychologist's or
neuroscientist's problem. We only have to explain HOW organisms
discriminate, identify and describe what they CAN discriminate,
identify and describe (doesn't that sound like a kind of
"verificational" problem?). And I think the null hypothesis here is
still that they learn to do this from input, initially sensory. There
is no "poverty of the stimulus" argument for concrete and abstract
categories, as there is for natural language syntax. No one has proved
or given evidence that they are unlearnable.
So I think it's more sensible to interpret the absence of invariants as
evidence of the fact that (1) invariants, like so much else in
cognition, are simply inaccessible to introspection, and hence will
have to be found the hard way, and that (2) the hard way (which is to
model the real-time, bottom-up course of category acquisition) has
simply not been explored yet!
Now I'm not saying there may not be some innate sensory categories;
their sensory invariants were "learned" by evolution (and some small
complement of them -- but only a very small one -- may even be like
"spandrels"). But whether innate or learned, the invariants must
exist, else we could not categorize successfully, as we in fact do. (Our
success, however, is provisional, approximate, and dependent on the
context of alternatives we've sampled to date; being always open to
revision, it is hardly a basis for a lapidary dictionary entry.)
I am not even claiming that the primitive sensory categories -- the
ones that are necessary to ground the system -- are fixed
and immutable. Not only are sensory categories, like all categories,
open to revision if the context of alternatives grows or changes, but a
grounded system is in principle "cognitively penetrable" through and
through: As long as the revision is able to percolate through the system
(all of it, if necessary), there's no obstacle to a bottom-level
category being revised on the basis of top-down information -- as long
as the symbols at the top are grounded! A system, once grounded, can
be forever revising its foundations -- pulling itself up by the
bootstraps, so to speak, as long as there's still some provisional
ground to lean on. And there always is; because once words get their
meanings on the basis of approximate and provisional sensory
invariants -- good enough to have picked out their referents
initially, at least -- then the rest of the adjustments can all be
done at the symbolic level, as long as they continue to use SOME
sensory grounding.
Saul Kripke (an essentialist philosopher) inadvertently gave me a good
example of this. He pointed out that if we baptized the term "gold"
on these rare, yellow-colored, shiny rocks, and started to use them in
trade, etc., and then we discovered that some of them were fool's
gold, which wasn't rare or precious (or we weren't prepared to
consider it such), then of course there would be no problem about
revising our features for gold so as to include gold and exclude fool's
gold (based on whatever features will reliably tell them apart). But,
Kripke asked, what if it turned out that we had unknowingly baptized "gold"
purely on the basis of samples of fool's gold, never yet even having
encountered real gold? Would that make any difference?
Now Kripke's concern was with essentialism, so his punchline was that,
no, it was real gold that we had intended all along. I'm not interested
in essentialism; but what I conclude is that we could and would
certainly keep on using the name "gold" for the real stuff, even if it
turned out that what we'd been using it on until that date had all been
the fake stuff. The reason is that, to an approximation, the real stuff
and the fake stuff share enough sensory invariants to ground the term
to begin with, and from that point on, the rest can be accomplished by
revision at the symbolic level (unless the sensory invariant that does
distinguish fool's gold from real gold has no name yet; no such problem
here; let's say it was the property of having atomic number 24 [?] in
the periodic table -- a perfectly well-grounded measurement, in
principle).
So that's why I say that a kind of bootstrapping and
top-down-revisability exists at all levels of the system. The sensory
invariants, as well as the category system as a whole, are just
approximations to reality -- picking things out successfully (as
dictated by feedback from the consequences of MIScategorization)
on the basis of the sample of confusable alternative inputs sampled to
date. No eternal "definitions" are etched in stone; just the
wherewithal to keep discriminating, identifying and describing things
as successfully as we do, based on the input we've encountered so
far.
> Attempts in the Philosophy of Science to ground scientific theories
> in a small vocabulary of operational, sensory based, terms have failed
> miserably (e.g. logical positivism)... an organism's
> conceptual armament [is] a very complex "theory" of the world...
As I said, this is not philosophy of science, it's a model for our
categorization performance capacity. And "primitives" are always
revisable; they're just needed to get the symbolic level off
the ground in the first place. And "theories" don't just float freely
in the head, because of the symbol grounding problem. Their elementary
terms must be connected to their (provisional) referents by some
(provisionally) reliable sensory invariants. Otherwise your "theory
driven" organism is just in Searle's Chinese room of meaningless
symbols.
> I don't see how it is possible for a primitive symbol to "resemble" its
> referent, in the clear sense you have developed, while still playing
> the appropriate syntactic role in mental computation. Either an
> alternate mechanism for grounding is required, or we must develop some
> theory of cognitive processing which isn't based on the notion of
> syntactic computation over mental representations.
As I state at the end of the paper, I think the latter is true: Since
the kind of (1) bottom-up, (2) hybrid (symbolic/nonsymbolic) system
required by my grounding proposal is (3) a dedicated symbol system,
it's no longer clear whether it's really a symbol system at all, for
the reason you mention (which I likewise spell out at the end of the
paper): A pure symbol system manipulates symbol tokens on the basis of
syntactic rules operating purely on the their (arbitrary) shapes.
That's supposed to be the only constraint. In a hybrid system grounded
along the lines I propose, there is a second constraint, namely, the
connection between the elementary symbols and the analog sensory
projections and invariance-detectors that pick out their referents
on the basis of their (nonarbitrary) shapes. With this extra
constraint, all bets about formal symbol systems are off until the
formal properties of such dedicated devices are investigated more
deeply.
Stevan Harnad
∂11-Oct-89 0141 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 11 Oct 89 01:41:05 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 434930; 11 Oct 89 03:07:54 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 109840; Wed 11-Oct-89 00:03:45 PDT
Date: Wed, 11 Oct 89 00:02 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: large periods
To: ACW@YUKON.SCRC.Symbolics.COM
cc: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, math-fun@RUSSIAN.SPA.Symbolics.COM,
"Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19891010171652.5.ACW@WHIMBREL.SCRC.Symbolics.COM>
Message-ID: <19891011070248.5.RWG@TSUNAMI.SPA.Symbolics.COM>
Date: Tue, 10 Oct 89 13:16 EDT
From: Allan C. Wechsler <ACW@YUKON.SCRC.Symbolics.COM>
Suppose you have a Life oscillator. At any given generation it fits
inside some bounding rectangle. Look at the area of the bounding
rectangle over all generations in the cycle of the oscillator. Define
the "area" of the oscillator to be the maximum area achieved by the
bounding rectangle over a complete cycle of the oscillator.
The question has come up, "what is the largest period oscillator that
can be packed into a rectangle of area A?" We already know how to make
oscillators whose period scales linearly with A. Can we do better?
Isn't this what I was asking in my "Life squanders real estate" msg?:
Clearly, above some size, maximum period as a function of area must
grow exponentially, because one could build a Turing type, but
finite "eons counter" only a fixed size increment over its memory.
What is the base of that exponential? I.e., the densest possible
memory.
Gosper has tried to answer this by building shift-and-xor machines.
By juxtaposing linear oscillators of periods K, 2K, ... nK, we can build
an oscillator of period lcm(K,2K,...,nK) in area n(n-1)SK for some
constant S. I don't know how to derive an asymptotic formula for the
lcm. Is this exponential yet?
Well, yes, but on 5 Oct, I redistributed Hickerson's 3 Oct msg, including
It occurred to me that there's a rather trivial way to get 'exponential'
growth of period as a function of area: Independent glider shuttles
of periods 1*60, 3*60, ..., (2n+1)*60 have an lcm period that's something
like exp(sqrt(area)).
And on 1 Oct, I repeated the "real estate" quote, while announcing
With these new components, just two guns and five shuttles (half-guns)
suffice to make a shiftregister pseudorandom generator whose period
scales exponentially with the shuttle spacing.
Are you leery of shiftregister sequences because some of them fall far short
of max period? That period table I sent yesterday suggested most lengths
would be close to max, although I plead gross ignorance of the facts.
I know Gosper wants to forbid (on esthetic grounds) these lcm
oscillators where no individual bit has the desired period. But I also
know that if I presented such an oscillator to Gosper, he would be able
to cludge it so that there was a bit of at least the desired period.
Handwaving plan: have a halo of circulating gliders around the whole
structure. Each time the stream passes one end of a component
oscillator, have a "tap" into the oscillator that xors with the
surrounding glider stream.
That would be hard in the general case. But I still don't see what problem
it solves that a shiftregister with a "spacefilling" glider loop doesn't.
Or, if necessary, equip a loop with a genuine binary counter.
Irrelevant challenge to those who have read this far: What is the smallest
initial population that can grow by exactly 1 each gen?
∂11-Oct-89 0645 Mailer failed mail returned
The following message was undeliverable to the address(es) below
because the destination Host Name(s) are Unknown:
elliott%slacvm.bitnet@forsysthe
------- Begin undelivered message: -------
11-Oct-89 0644 JMC Lyman talk, oct 26
To: elliott%slacvm.bitnet@forsysthe, RWF@SAIL.Stanford.EDU
He will tell us his opinions at noon in the
Faculty Club on that day. Invite others. There
won't be printed publicity.
------- End undelivered message -------
∂11-Oct-89 0856 kpeters@cdp.uucp Knott manuscript
Received: from arisia.Xerox.COM by SAIL.Stanford.EDU with TCP; 11 Oct 89 08:55:54 PDT
Received: from cdp.UUCP by arisia.Xerox.COM with UUCP
(5.61+/IDA-1.2.8/gandalf) id AA04320; Wed, 11 Oct 89 08:35:20 -0700
Message-Id: <8910111535.AA04320@arisia.Xerox.COM>
Date: Wed, 11 Oct 89 08:35:20 -0700
From: <kpeters@cdp.uucp>
To: jmc@sail.stanford.edu
Subject: Knott manuscript
Cc: kpeters@arisia.Xerox.COM
Dear Professor McCarthy:
I am emailing you merely to check if you have had a chance to look
at the Knott manuscript on Interpreting LISP, which we sent you in
mid-September. We would like to be able to get back to Knott with
comments, and feel it is important to get your opinion on the proposal.
Please respond at your earliest convenience.
Best regards,
Alice Peters
Academic Press
P.S. Our answerback usually doesn't work. Please contact me via
cdp!kpeters@labrea.stanford.edu
∂11-Oct-89 1011 @Score.Stanford.EDU,@ai.ai.mit.edu,@MC.lcs.mit.edu:bundy%aipna.edinburgh.ac.uk@nsfnet-relay.ac.uk RA Post in AI Dept at Edinburgh
Received: from Score.Stanford.EDU by SAIL.Stanford.EDU with TCP; 11 Oct 89 10:11:25 PDT
Received: from lcs.mit.edu (MINTAKA.LCS.MIT.EDU.#Internet) by SCORE.STANFORD.EDU with TCP; Wed 11 Oct 89 10:12:35-PDT
Received: from ai.ai.mit.edu by mintaka.lcs.mit.edu id aa10866;
11 Oct 89 10:54 EDT
Received: from MC.LCS.MIT.EDU (CHAOS 3131) by AI.AI.MIT.EDU 11 Oct 89 10:54:48 EDT
Received: from lcs.mit.edu (CHAOS 15044) by MC.LCS.MIT.EDU 11 Oct 89 10:52:48 EDT
Received: from NSFNET-RELAY.AC.UK by mintaka.lcs.mit.edu id aa10801;
11 Oct 89 10:49 EDT
Received: from aipna.edinburgh.ac.uk by NSFnet-Relay.AC.UK
via Janet with NIFTP id ab06214; 11 Oct 89 13:53 BST
Date: Wed, 11 Oct 89 13:19:20 BST
Message-Id: <11959.8910111219@etive.aipna.ed.ac.uk>
From: Alan Bundy <bundy%aipna.edinburgh.ac.uk@nsfnet-relay.ac.uk>
Subject: RA Post in AI Dept at Edinburgh
To: otten%swivax.uucp@nsfnet-relay.ac.uk, ailist@kl.sri.com,
jws%ibm-b.rutherford.ac.uk@nsfnet-relay.ac.uk,
arpanet-bboards%mc.lcs.mit.edu@nsfnet-relay.ac.uk,
comp-ai@ucbvax.berkeley.edu,
mod-ki <@uunet.uu.net,@unido.uucp:mod-ki@gmdzi>,
humanist%utoronto.bitnet@nsfnet-relay.ac.uk, nnsc@nnsc.nsf.net,
fox@vtcs1.cs.vt.edu, fj-ai%etl.jp@nsfnet-relay.ac.uk,
nl-kr@cs.rochester.edu, theorem-provers%mc.lcs.mit.edu@nsfnet-relay.ac.uk,
lfcs-interest%lfcs.edinburgh.ac.uk@nsfnet-relay.ac.uk,
messages%aipna.edinburgh.ac.uk@nsfnet-relay.ac.uk,
msgs%itspna.ed.ac.uk@nsfnet-relay.ac.uk,
msgs%eusip.edinburgh.ac.uk@nsfnet-relay.ac.uk,
msgs%epistemi.edinburgh.ac.uk@nsfnet-relay.ac.uk
Department of Artificial Intelligence
University of Edinburgh
RESEARCH FELLOW
(Mathematical Reasoning)
Applications are invited for an SERC supported post, tenable from 1st
January 1990, or on a mutually agreed date. Appointment will be to
September 30th 1991, initially, but with a possibility of renewal.
The research is to develop proof plans, a technique for guiding the
search for a proof in automatic theorem proving. The main application
is to the automatic synthesis, verification and transformation of
logic programs using constructive logic. The project is led by
Professor Alan Bundy and Dr Alan Smaill.
Candidates should possess a PhD or have equivalent research or
industrial experience. Knowledge of logic is essential and knowledge
of artificial intelligence, formal methods in software engineering or
logic programming would be an advantage. Salary is on the AR1A scale
in the range 10,458-16,665 pounds p.a., according to age,
qualifications and experience.
Applicants should send a CV and the names of two referees to:
Prof. Alan Bundy.
Department of Artificial Intelligence,
University of Edinburgh,
80 South Bridge,
Edinburgh,
EH1 1HN.
as soon as possible. The closing date for applications
is 14th November
1990. Further details may be obtained from Prof. Bundy (at the above
address or email to bundy@uk.ac.edinburgh or bundy@rutgers.edu)
quoting reference number 5717/E.
∂11-Oct-89 1204 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU Tuesday seminar
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 11 Oct 89 12:04:40 PDT
Received: from iswim.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61/25-eef) id AA06149; Wed, 11 Oct 89 12:04:09 -0700
Received: by iswim.stanford.EDU (5.57/Ultrix3.0-C)
id AA02051; Wed, 11 Oct 89 12:07:24 PDT
Message-Id: <8910111907.AA02051@iswim.stanford.EDU>
To: jmc@sail.Stanford.EDU
Subject: Tuesday seminar
Cc: katiyar@neon.Stanford.EDU
From: John C. Mitchell <jcm@Polya.Stanford.EDU>
Reply-To: John C. Mitchell <jcm@Polya.Stanford.EDU>
Date: Wed, 11 Oct 89 12:07:13 PDT
Sender: jcm@iswim
Would you be willing to talk about Elephant 2000 either
next Tues, Oct 17, or Tues Oct 31 at 4 PM? There is a
departmental colloquium Oct 24, and another on Nov 21.
At this point, I could arrange any other Tues that suits
you, but the schedule is going to get a little less flexible
soon. I am especially interested in hearing what you have
to say, so I hope you can make it.
John
∂11-Oct-89 1357 weening@Gang-of-Four.Stanford.EDU Lucid work on Qlisp
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 11 Oct 89 13:57:41 PDT
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA23674; Wed, 11 Oct 89 13:56:57 -0700
Date: Wed, 11 Oct 89 13:56:57 -0700
From: weening@Gang-of-Four.Stanford.EDU (Joe Weening)
Message-Id: <8910112056.AA23674@Gang-of-Four.Stanford.EDU>
To: jmc@sail.Stanford.EDU, clt@sail.Stanford.EDU
Subject: Lucid work on Qlisp
The following exchange of messages is somewhat disturbing:
Date: Tue, 10 Oct 89 10:04:12 -0700
>From: weening@Gang-of-Four.Stanford.EDU (Joe Weening)
To: arg
Subject: Another proposal
Would it be hard to do the following: before and after each GC, call
the function that asks the operating system for the accumulated CPU
time, and then have a user-settable flag that allows this to be
subtracted from the times reported by TIME and QTIME, or report it as
a separate number. This would let us determine if a program is
getting good speedup in non-garbage collection activities.
Date: Tue, 10 Oct 89 14:17:25 PDT
>From: Ron Goldman <arg@lucid.com>
To: weening@go4.stanford.edu
Subject: re: GC times
Joe - Separating out the time taken by GC is certainly a good idea. If and
when we have some time (like in several months) we might be able to
get around to it.
Ron
To: Ron Goldman <arg@lucid.com>
Subject: Re: GC times
Date: Tue, 10 Oct 89 22:05:23 PDT
>From: Joe Weening <weening>
This is really discouraging. Do even minor changes to the system have
to take this long?
Joe
Date: Wed, 11 Oct 89 13:12:04 PDT
>From: Ron Goldman <arg@lucid.com>
To: weening@Gang-of-Four.Stanford.EDU
Subject: manpower
Joe - 'Fraid so. We have essentially zero time budgeted for work on the
Alliant per se. As the Encore work progresses I hope to use the
version of Qlisp on the Alliant as a place to try out new things
(e.g. dataflow for future checks, heavy-weight futures, pmi-functions).
When we get a Qlisp on the Encore and add new features hopefully we'll
be able to do some retrofitting of the Alliant version. The main
problem is a big lack of manpower; we're already behind on the Encore
work.
Ron
∂11-Oct-89 1423 RWF re: Lyman talk, oct 26
[In reply to message rcvd 11-Oct-89 06:44-PT.]
I'll be there.
∂11-Oct-89 1529 ACW@ELEPHANT-BUTTE.SCRC.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 11 Oct 89 15:29:48 PDT
Received: from YUKON.SCRC.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 435260; 11 Oct 89 18:27:06 EDT
Received: from WHIMBREL.SCRC.Symbolics.COM by YUKON.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 499158; Wed 11-Oct-89 18:28:03 EDT
Date: Wed, 11 Oct 89 18:26 EDT
From: Allan C. Wechsler <ACW@YUKON.SCRC.Symbolics.COM>
Subject: large periods
To: rwg@RUSSIAN.SPA.Symbolics.COM, ACW@YUKON.SCRC.Symbolics.COM
cc: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, math-fun@RUSSIAN.SPA.Symbolics.COM,
"Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19891011070248.5.RWG@TSUNAMI.SPA.Symbolics.COM>
Message-ID: <19891011222651.7.ACW@WHIMBREL.SCRC.Symbolics.COM>
Date: Wed, 11 Oct 89 00:02 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Date: Tue, 10 Oct 89 13:16 EDT
From: Allan C. Wechsler <ACW@YUKON.SCRC.Symbolics.COM>
Suppose you have a Life oscillator. At any given generation it fits
inside some bounding rectangle. Look at the area of the bounding
rectangle over all generations in the cycle of the oscillator. Define
the "area" of the oscillator to be the maximum area achieved by the
bounding rectangle over a complete cycle of the oscillator.
The question has come up, "what is the largest period oscillator that
can be packed into a rectangle of area A?" We already know how to make
oscillators whose period scales linearly with A. Can we do better?
Isn't this what I was asking in my "Life squanders real estate" msg?:
Yes. I was just restating the background, prehaps to make it clear to
myself.
Clearly, above some size, maximum period as a function of area must
grow exponentially, because one could build a Turing type, but
finite "eons counter" only a fixed size increment over its memory.
What is the base of that exponential? I.e., the densest possible
memory.
Somehow I missed that paragraph in the original posting.
Gosper has tried to answer this by building shift-and-xor machines.
By juxtaposing linear oscillators of periods K, 2K, ... nK, we can build
an oscillator of period lcm(K,2K,...,nK) in area n(n-1)SK for some
constant S. I don't know how to derive an asymptotic formula for the
lcm. Is this exponential yet?
Well, yes,
How can you tell? And since we can do this with technology we already
own, can you tell me the base of the exponential? Just so we'll have a
lower bound to challenge us?
but on 5 Oct, I redistributed Hickerson's 3 Oct msg, including
It occurred to me that there's a rather trivial way to get 'exponential'
growth of period as a function of area: Independent glider shuttles
of periods 1*60, 3*60, ..., (2n+1)*60 have an lcm period that's something
like exp(sqrt(area)).
I want to see Dean's math. And e↑(a↑(1/2)) << k↑a no matter how small k
is. So if Dean's math is right, we haven't achieved exponentiality yet,
and we don't have a lower bound for the information capacity of life
space.
And on 1 Oct, I repeated the "real estate" quote, while announcing
With these new components, just two guns and five shuttles (half-guns)
suffice to make a shiftregister pseudorandom generator whose period
scales exponentially with the shuttle spacing.
Are you leery of shiftregister sequences because some of them fall far short
of max period? That period table I sent yesterday suggested most lengths
would be close to max, although I plead gross ignorance of the facts.
No. I'm confident that we could prove the existence of an infinite
sequence of devices with period > k↑A. Then the information capacity of
life is > log[2]k bits/cell. I would just like to see an actual numeric
lower bound. It sounds like you now possess enough data to compute it.
I know Gosper wants to forbid (on esthetic grounds) these lcm
oscillators where no individual bit has the desired period. But I also
know that if I presented such an oscillator to Gosper, he would be able
to cludge it so that there was a bit of at least the desired period.
Handwaving plan: have a halo of circulating gliders around the whole
structure. Each time the stream passes one end of a component
oscillator, have a "tap" into the oscillator that xors with the
surrounding glider stream.
That would be hard in the general case. But I still don't see what problem
it solves that a shiftregister with a "spacefilling" glider loop doesn't.
Or, if necessary, equip a loop with a genuine binary counter.
Irrelevant challenge to those who have read this far: What is the smallest
initial population that can grow by exactly 1 each gen?
Cool.
My own irrelevant challenge:
I want to know if anyone has been working on lightspeed wires. These
would be stable conduits that could conduct pulses at the speed of
light. The pulses could look something like
...ooooooooooooooo...
...o o o o o o o o...
which moves upward at the speed of light. (It alternates between the
phase shown and another with longer cilia.)
We "just" need to develop stable retaining walls and (the very hard part)
a way to make the pulses turn corners.
∂11-Oct-89 1651 ginsberg@Polya.Stanford.EDU Review of Akawa's paper on presupposition
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 11 Oct 89 16:51:01 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA28710; Wed, 11 Oct 89 16:51:51 -0700
Date: Wed, 11 Oct 89 16:51:51 -0700
From: Matthew L. Ginsberg <ginsberg@Polya.Stanford.EDU>
Message-Id: <8910112351.AA28710@Polya.Stanford.EDU>
To: bobrow.pa@Xerox.COM
Subject: Review of Akawa's paper on presupposition
Cc: ginsberg@Polya.Stanford.EDU, jmc@sail
This may be the worst paper I have ever read.
For a start, the English is terrible. I have done my best to
understand the points that the author is making, but in many cases,
was simply unable to make much sense of the run-on sentences and
peculiar phrasing. I understand that it is often difficult for
non-native speakers to write clear English prose, but the English in
this particular paper is absolutely awful.
The technical content of the paper, as far as I can tell, is on a par
with the language. The very definition on which the paper is founded
-- that on the top of p.4 -- is senseless. The definition of #K says
that it contains any sentence consistent with K, where consistency is
"as usually understood in nonmonotonic logic."
This is circular in the extreme; for a start, the set of such
consistent sentences is often itself inconsistent, and the different
approaches to nonmonotonic logic disagree *completely* about what the
sentences consistent with a given base theory are.
Given that the basic definition doesn't make any sense, I couldn't get
much out of the rest of the paper. But any time I did understand
anything, I was disappointed. For example:
The equation in the middle of p.7, professing to solve the Yale
shooting problem, in fact is simply encoding the right answer and
proceeding from there. In the next paragraph, it appears that Akawa
is sanctioning chronological ignorance as a solution; this is
something that was abandoned by most of the nonmonotonic community
quite some time ago as a bad idea.
On p.8, Akawa tries to draw philosophical conclusions from the fact
that brave approaches to default reasoning can come to believe
contradictions, somehow connecting the issue with minimal model
semantics. This is at odds with my understanding of the situation (so
much so that I can't even figure out what Akawa is trying to say).
On p.9, it appears that the # operation may not be unique. The
original definition certainly appeared to be unique, to the extent I
understood it.
Later on p.9, it is implied that the important difference between
McCarthy's two papers on circumscription was the introduction of the
ab predicate. Not only is the ab predicate an idea of Levesque's
(I believe), but the differences between McCarthy's two papers
are principally elsewhere.
All of Section 6 appears to say nothing.
∂11-Oct-89 1733 nilsson@Tenaya.Stanford.EDU re: CIFE
Received: from Tenaya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 11 Oct 89 17:33:36 PDT
Received: by Tenaya.Stanford.EDU (NeXT-0.8/25-eef) id AA01291; Wed, 11 Oct 89 17:31:46 PDT
Date: Wed, 11 Oct 89 17:31:46 PDT
From: Nils Nilsson <nilsson@tenaya.stanford.edu>
Message-Id: <8910120031.AA01291@Tenaya.Stanford.EDU>
To: John McCarthy <JMC@sail.stanford.edu>
Subject: re: CIFE
Cc: nilsson@Tenaya.Stanford.EDU
Yes, I think the CIFE people would be interested
in Elephant and in CBCL. I have been trying to interest
one of my graduate students in related matters without
success so far, so I don't have a student to suggest.
Perhaps you don't have to identify a particular student
in the proposal---leaving that for later.
∂11-Oct-89 2056 CLT CLT Itinerary San Diego / Ventura trip
Thursday Oct 12
Dana Inn and Marina
1710 W. Mission Blvd.
San Diego CA
619-222-6440
Fri Oct 13 - Sat Oct 14
Vagabond Inn
756 E Thompson Blvd
805-648-5371
∂12-Oct-89 0230 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 12 Oct 89 02:30:43 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 435431; 12 Oct 89 05:27:02 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 109886; Thu 12-Oct-89 02:14:46 PDT
Date: Thu, 12 Oct 89 02:14 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: large periods
To: ACW@YUKON.SCRC.Symbolics.COM
cc: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, math-fun@RUSSIAN.SPA.Symbolics.COM,
"Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19891011222651.7.ACW@WHIMBREL.SCRC.Symbolics.COM>
Message-ID: <19891012091408.7.RWG@TSUNAMI.SPA.Symbolics.COM>
Date: Wed, 11 Oct 89 18:26 EDT
From: Allan C. Wechsler <ACW@YUKON.SCRC.Symbolics.COM>
. . .
[ACW] By juxtaposing linear oscillators of periods K, 2K, ... nK, we can build
an oscillator of period lcm(K,2K,...,nK) in area n(n-1)SK for some
constant S. I don't know how to derive an asymptotic formula for the
lcm. Is this exponential yet?
[rwg] Well, yes,
How can you tell? And since we can do this with technology we already
own, can you tell me the base of the exponential? Just so we'll have a
lower bound to challenge us?
It's probably in Hardy & Wright (mine has disappeared), certainly in Hardy's
"Ramanujan" (neq "Collected Papers"), p27+. Crudely, all you need is the
prime number thm: pi(x) ~~ (ln x)/x, nth prime ~~ n ln n. Neglecting
squares, cubes, etc., we have
lcm(1,2,...n) = prod k ln k ~~ e↑n, by Stirling's formula.
1<k<(ln n)/n
Plot the log. What a strange place to find e.
(Actually, I think this e↑n can be derived by "elementary" arguments, and
the prime number theorem deduced by equivalence.)
[rwg] but on 5 Oct, I redistributed Hickerson's 3 Oct msg, including
It occurred to me that there's a rather trivial way to get 'exponential'
growth of period as a function of area: Independent glider shuttles
of periods 1*60, 3*60, ..., (2n+1)*60 have an lcm period that's something
like exp(sqrt(area)).
I want to see Dean's math. And e↑(a↑(1/2)) << k↑a no matter how small k
is.
Good point. LCMs are less efficient. (Yay!)
So if Dean's math is right, we haven't achieved exponentiality yet,
and we don't have a lower bound for the information capacity of life
space.
What you mean "we", white man? Again from 1 Oct:
The tightly spaced beam loops trapped between sidekickers can be
crossed, so that a region of space is packed with a startling
density of gliders traveling in all four directions without
colliding. This, then, gives us a lower bound of about 36 cells/bit
of eons memory, with the possibility of period 10↑100 or more fitting
on a screen.
I.e., just splicing this giant delay line into a shift register will
give an exponential(a) period eons counter.
. . .
Are you leery of shiftregister sequences because some of them fall far short
of max period? That period table I sent yesterday suggested most lengths
would be close to max, although I plead gross ignorance of the facts.
No. I'm confident that we could prove the existence of an infinite
sequence of devices with period > k↑A. Then the information capacity of
life is > log[2]k bits/cell. I would just like to see an actual numeric
lower bound. It sounds like you now possess enough data to compute it.
By the above, ~ 1/36. (The compact form of the sidekicker I planned to use
intrudes by one cell into the previous path, but I think I see a(n inelegant)
fix.) Gliders can operate as densely as 1/20, but I'm not sure we have the
requisite focusing machinery.
Gliders are unlikely to be optimal. One could imagine a radix 4 delay-line
featuring all three weights of spaceship, but this would probably cost more
than twice the real estate.
Gliders and spaceships are vacuum tube technology. Higher speeds and
densities await the discovery of a semiconductive (nonexplosive) lattice.
Irrelevant challenge to those who have read this far: What is the smallest
initial population that can grow by exactly 1 each gen?
Cool.
Schick has a four ship puffer (~ 50 dots) which drops traffic lights (12 dots)
every 24 ticks. Two of these out of phase would do the job, modulo period 6
ripple (low amplitude). Hickerson's arsenal of p6s should yield one or two
that would reduce this to p2 or 3 ripple, which would be trivial to finish
off. I'd guess ~130 for the initial total.
My own irrelevant challenge:
I want to know if anyone has been working on lightspeed wires.
Fiber optics.
These
would be stable conduits that could conduct pulses at the speed of
light. The pulses could look something like
...ooooooooooooooo...
...o o o o o o o o...
which moves upward at the speed of light. (It alternates between the
phase shown and another with longer cilia.)
We "just" need to develop stable retaining walls and (the very hard part)
a way to make the pulses turn corners.
Childishly unrealistic optimism. Unless . . . Say, Dean?
∂12-Oct-89 0602 CLT for sale
To: su-bboards@SAIL.Stanford.EDU
CC: JMC@SAIL.Stanford.EDU
bed -- kingsize mattress, box springs, frame
Sears imperial (top line) made by Simmons
excellent condition, about 8 years old
$150 (or best offer)
washer -- Maytag, 5button, about 15 years old works well.
$50 (or best offer)
Contact
John McCarthy jmc@sail.stanford.edu 732-4430(ofc), 857-0672(home)
Carolyn Talcott clt@sail.stanford.edu 732-0937(ofc), 857-9029(home)
(clt will be away til 16 Oct)
∂12-Oct-89 0607 CLT lathrop
Bob wants to have the asbestos people come in as soon as escrow
closes. This means things need to be cleared out. I have
sent a for sale msg for the bed/washer to su-market (and bboard by
mistake). If it isn't sold by Monday we can either move the mattress
to the bathroom or just call good will.
Here is a list of things that Alcon needs to do to finish up.
File notice of completion
Clear rubble from yard
Cut of reinforcement bars where wing wall was removed
Reconnect shed to electricity
Driveway lights (one at each side of porch) are missing
Clean garbage disposal (where is plug?)
Louver windows have 3 missing slats two are there, one is broken
Alcon Contractors
652-0311
858-1202 is number at 846 Lathrop.
∂12-Oct-89 1030 VAL testing the email connection
∂12-Oct-89 0456 @NSFnet-Relay.AC.UK,@research1.computer-science.manchester.ac.uk:john@research7.computer-science.manchester.ac.uk testing the email connection
Received: from NSFnet-Relay.AC.UK by SAIL.Stanford.EDU with TCP; 12 Oct 89 04:56:45 PDT
Received: from research1.computer-science.manchester.ac.uk
by NSFnet-Relay.AC.UK via Janet with NIFTP id aa03400;
12 Oct 89 11:27 BST
Received: from r7f.r7.cs.man.ac.uk by r1.cs.man.ac.uk; Thu, 12 Oct 89 11:34:33 BST
From: John Gurd <john%research7.computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK>
Date: Thu, 12 Oct 89 11:33:53 BST
Message-Id: <11283.8910121033@r7f.r7.cs.man.ac.uk>
To: VAL@sail.stanford.edu
In-Reply-To: Vladimir Lifschitz's message of 02 Oct 89 1309 PDT <1zfqTp@SAIL.Stanford.EDU>
Subject: testing the email connection
Vladimir,
I'm sorry to take so long to reply to your
last message, but I have been away for a few days.
I have spoken to the Vice-Chancellor. He says
he is almost ready to write to you, but that he is waiting
for some additional information that he expects to receive
during the next week. I did not get the impression that
there was a substantial problem involved. I have asked him
to let me know the moment he sends a letter, and I will
mail you about it immediately.
Thanks for the information on your wife's
interests - I have been told of a number of other teaching
possibilities and will let you know the position when I
have followed these up.
Best wishes,
John.
∂12-Oct-89 1052 nilsson@Tenaya.Stanford.EDU
Received: from Tenaya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 12 Oct 89 10:51:49 PDT
Received: by Tenaya.Stanford.EDU (NeXT-0.8/25-eef) id AA01618; Thu, 12 Oct 89 10:47:21 PDT
Date: Thu, 12 Oct 89 10:47:21 PDT
From: Nils Nilsson <nilsson@tenaya.stanford.edu>
Message-Id: <8910121747.AA01618@Tenaya.Stanford.EDU>
To: abbas@isl.stanford.edu, losleben@cis.stanford.edu,
plummer@sierra.stanford.edu, nanni@galileo.stanford.edu,
jlh@vsop.stanford.edu, horowitz@mojave.stanford.edu, toole@darpa.mil,
cross@darpa.mil, gibbons@sierra.stanford.edu,
levinthal@sierra.stanford.edu, feigenbaum@sumex-aim.stanford.edu,
engelmore@sumex-aim.stanford.edu, goodman@sierra.stanford.edu,
linvill@cis.stanford.edu, jmc@sail.stanford.edu
jct@sail, val@sail, wiederhold@sail, ullman@cs, latombe@cs, binford@cs, cannon@sierra, wiederhold@cs, luckham@sierra, cheriton@cs, gupta@cs, jutta@coyote, bhr@sumex, khatib@cs
Subject: Visit by Barry Boehm
cc: nilsson
Folks,
As I mentioned in a previous message (I have heard that some of you did not
get copies of that message), Barry Boehm, the new Director of DARPA/ISTO, will
be visiting Stanford on Monday and Tuesday, November 20 and 21, 1989. Barry
was formerly at TRW. Steve Cross, a program director at DARPA/ISTO, has asked
me to help arrange Barry's schedule while at Stanford. Steve sent a suggested
schedule that has now been modified a bit. We should note that the CIS ad com
meeting is scheduled for Monday, November 20. I hope that those of us who will
be participating in both meetings will be able to do some shuttling back and forth
and time-sharing.
Please confirm that you can participate at the scheduled time and place. Due to
the distributed nature of Stanford projects, I suggest that we divide the visit into
"sections" with a section leader who will keep the participants in that section on
schedule and who will handle any last minute schedule changes within his section.
Session 1; leader Bob Engelmore
Monday, November 20, 1989
Knowledge Systems Lab
1101 Welch Road, Palo Alto
8:00 - 9:30 Welcome and Overview of Stanford EE and CS Departments
Dean James Gibbons, Joe Goodman, Nils Nilsson
9:30- noon Review of KSL projects
Bob Engelmore and KSL staff
noon - 1:15 Discussion and working lunch at KSL followed by travel to Cedar Hall
Session 2; leader Jean-Claude Latombe
Monday, November 20, 1989
Robotics Laboratory
Cedar Hall
1:15 - 2:30 Foundations of AI and Related Projects
John McCarthy, Vladimir Lifschitz, and Carolyn Talcott
2:30 - 5:00 Stanford Robotics Projects
Jean-Claude Latombe, Tom Binford, and Bob Cannon
Session 3; leader Bob Cannon
Monday, November 20, 1989
Durand Building
5:00 - 5:30 Tour of Aero/Astro Robotics Lab
5:30 - 6:30 Roundtable Discussion (Durand Conference Room 450)
[Steve suggests that this roundtable focus on AI-related topics;
so all of the robotics and AI people are invited.]
7:00 - Possible small group for dinner? (Nils to follow up and arrange.)
Sessions 4 & 5; leaders John Hennessy and Jim Plummer
Tuesday, November 21, 1989
CIS Building
Conference Room 101
8:30 - 9:15 Database Research
Gio Wiederhold
9:15 - 10:00 Parallel Computation
Jeff Ullman
10:00 - 5:00 [[[Steve had listed the following people and topics in his
suggested agenda. I think John H. and Jim P. should review these and decide
how best to structure, label, and schedule them:
Software, David Luckham
Microsupercomputers, John Hennessy and Mark Horowitz
Structured Process Flow, Jim Plummer
CAD for EEPROM, Abbas El Gammal
Logic Synthesis, Giovanni DiMicheli
Distributed Systems, David Cheriton
the agenda for these topics can spill over into Wednesday morning if needed.
Note that the CS/CSL faculty have a regularly scheduled faculty lunch, with
sandwiches, etc. provided, on Tuesdays. We would like to invite Barry to come to
that lunch on Tuesday and perhaps informally provide some prospective about his
views on DARPA-funded research, long-range goals, etc. This would also give
Barry a chance to see the CSD facilities in Margaret Jacks Hall.
]]]
Bob Engelmore: Can you arrange for the KSL Conference room, o/h projectors, coffee,
and lunch on Monday morning?
Jutta and Jean-Claude: Can you do the same for Cedar facilities on Monday afternoon?
Bob Cannon: Can you do the same for Durand and Durand 450 on late Monday
afternoon?
John Hennessy: Can you be the contact person for arrangements for Tuesday?
Any questions, comments, suggestions, etc. can be addressed to the various
section leaders and/or me.
Thanks,
-Nils
∂12-Oct-89 1735 ACW@ELEPHANT-BUTTE.SCRC.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 12 Oct 89 17:35:07 PDT
Received: from YUKON.SCRC.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 435764; 12 Oct 89 17:42:02 EDT
Received: from WHIMBREL.SCRC.Symbolics.COM by YUKON.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 499674; Thu 12-Oct-89 17:37:20 EDT
Date: Thu, 12 Oct 89 17:35 EDT
From: Allan C. Wechsler <ACW@YUKON.SCRC.Symbolics.COM>
Subject: large periods
To: rwg@RUSSIAN.SPA.Symbolics.COM, ACW@YUKON.SCRC.Symbolics.COM
cc: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, math-fun@RUSSIAN.SPA.Symbolics.COM,
"Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19891012091408.7.RWG@TSUNAMI.SPA.Symbolics.COM>
Message-ID: <19891012213553.8.ACW@WHIMBREL.SCRC.Symbolics.COM>
I'm not sure if you answered this. What if you just use the
prime-period (times a constant) shuttles? Then the area would be
proportional to only the sum of the primes less than n, not to n↑2. Is
this enough gain to improve the disjoint-shuttle scheme to
exponentiality?
∂12-Oct-89 1934 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM Speed comparison, 135 degree reflection, almost exponential periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 12 Oct 89 19:33:57 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 435911; 12 Oct 89 22:33:32 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 109924; Thu 12-Oct-89 19:29:39 PDT
Date: Thu, 12 Oct 89 19:29 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: Speed comparison, 135 degree reflection, almost exponential periods
To: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: The message of 3 Oct 89 15:43 PDT from HUL%PSUVM.BITNET@CORNELLC.cit.cornell.edu
Supersedes: <19891006060624.3.RWG@TSUNAMI.SPA.Symbolics.COM>
Comments: Retransmission of failed mail.
Message-ID: <19891013022916.9.RWG@TSUNAMI.SPA.Symbolics.COM>
Date: Tue, 3 Oct 89 18:43 EDT
From: "Dean Hickerson" <HUL%PSUVM.BITNET@CORNELLC.cit.cornell.edu>
I tried out your period 11680366 pattern in my Apple II program; it runs at
about 9.1 generations per second.
That is embarrassing both because it is nearly twice as fast as my
bitblt pgm, and because the period turns out to be 46*7*15*31*127 =
19015710. I dropped a bit (in the worst sense) when I counted the states.
To see that it is really computing a[n] = a[n-1] eqv a[n-19], uncork the
spaceships and observe the bit sequence after the initial 0 burst is
10101010101010101011... . The first doubled bit being 1 indicates an odd
number of state bits in the shift register.
At that rate, it should finish a complete
cycle in about two weeks. Oh well, at least it's fast enough to see how
it works.
I should have warned people to leave extra space at the bottom if they
uncork the ships, since lightweights hitting a boundary clamped to 0
instantly bounce back as gliders.
The 135 degree turn is amazing!
It sure changes the economics of glider logic. Look for gunless fanout
next.
Since the centinal is based on the p46
shuttle, can it also do the 135?
No. The centinal is extremely weird. One could equally say that the
p54 is based on the p46. More properly, the thing that looks like a
p46 simply isn't in those contexts. It gets killed and reincarnated
via a totally diferent mechanism from the p46 shuttle.
If so, can your period 900+200n gun be
simplified?
Yes, but not with 135 turners. First of all, the 900+200n needs middleweights,
not lightweights. Second of all, a mere pentadecathlon can gliderize the bonus
block a la the centinal. So, by choosing an oddly even multiple of 15, half
the spaceships (and thus half the machinery) can be discarded, giving something
perhaps as small as 270+60n.
I experimented with the centinal to see if I could replace the
2 eaters at one end by a block, but it didn't work; the block is fussier
about its exact location than the eaters. Have you found any viable
centinal variants?
I'd never dream of looking. The 135 bounce needs the giant spark off an
unmuzzled (actually, unsilencered) p46. The p100 muzzle miraculously
withstands two different hits, the harder of which stuns the shuttle into
losing four ticks and two cell positions. The only way to get p100
giant sparks would be to delete and reconstruct eaters and block every 100.
It occurred to me that there's a rather trivial way to get 'exponential'
growth of period as a function of area: Independent glider shuttles
of periods 1*60, 3*60, ..., (2n+1)*60 have an lcm period that's something
like exp(sqrt(area)).
Neat. Are odds better than primes?
This also works for my 1-dimensional question:
There's a 3 state cellular automaton with max period growing like
exp(sqrt(diameter)). If we want exp(diameter), 4 states apparently
suffice: I recently received some papers on 1-D CAs from Harold
McIntosh; one mentions a 4 state CA which simulates a binary counter.
I'm not sure how it works yet; it's on an IBM PC disk which he also sent,
but which I haven't tried yet.
Your crazy program might be able to find a binary counter in Life.
∂12-Oct-89 2211 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 12 Oct 89 22:11:41 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 435946; 13 Oct 89 01:08:33 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 109928; Thu 12-Oct-89 22:04:34 PDT
Date: Thu, 12 Oct 89 22:03 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: large periods
To: ACW@YUKON.SCRC.Symbolics.COM
cc: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, math-fun@RUSSIAN.SPA.Symbolics.COM,
"Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19891012213553.8.ACW@WHIMBREL.SCRC.Symbolics.COM>
Message-ID: <19891013050359.1.RWG@TSUNAMI.SPA.Symbolics.COM>
Date: Thu, 12 Oct 89 17:35 EDT
From: Allan C. Wechsler <ACW@YUKON.SCRC.Symbolics.COM>
I'm not sure if you answered this. What if you just use the
prime-period (times a constant) shuttles? Then the area would be
proportional to only the sum of the primes less than n, not to n↑2. Is
this enough gain to improve the disjoint-shuttle scheme to
exponentiality?
I think the point you are missing is that the n-times slower oscillators
are n times bigger. O(n) of them each averaging O(n) area consumes O(n↑2).
∂13-Oct-89 1331 ACW@ELEPHANT-BUTTE.SCRC.Symbolics.COM large periods
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 13 Oct 89 13:31:15 PDT
Received: from YUKON.SCRC.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 436196; 13 Oct 89 16:05:15 EDT
Received: from WHIMBREL.SCRC.Symbolics.COM by YUKON.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 500051; Fri 13-Oct-89 16:06:23 EDT
Date: Fri, 13 Oct 89 16:04 EDT
From: Allan C. Wechsler <ACW@YUKON.SCRC.Symbolics.COM>
Subject: large periods
To: rwg@RUSSIAN.SPA.Symbolics.COM, ACW@YUKON.SCRC.Symbolics.COM
cc: "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, math-fun@RUSSIAN.SPA.Symbolics.COM,
"Rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "acad!megalon!rudy@uunet.UU.NET"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rem@suwatson.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"hpm@rover.ri.cmu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19891013050359.1.RWG@TSUNAMI.SPA.Symbolics.COM>
Message-ID: <19891013200449.0.ACW@WHIMBREL.SCRC.Symbolics.COM>
Date: Thu, 12 Oct 89 22:03 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
I think the point you are missing is that the n-times slower oscillators
are n times bigger. O(n) of them each averaging O(n) area consumes O(n↑2).
But there aren't O(n) of them -- there are Numberofprimeslessthan(N) of
them. That is definitely less than O(n). I suspect that the average
area is less than O(n) as well.
∂13-Oct-89 1425 MLB@WHITE.SWW.Symbolics.COM composition contests
Received: from WHITE.SWW.Symbolics.COM ([128.81.57.24]) by SAIL.Stanford.EDU with TCP; 13 Oct 89 14:25:00 PDT
Received: from CHROME.SWW.Symbolics.COM by WHITE.SWW.Symbolics.COM via CHAOS with CHAOS-MAIL id 25778; Fri 13-Oct-89 14:21:55 PDT
Date: Fri, 13 Oct 89 14:20 PDT
From: mArQ lE bRuN <MLB@WHITE.SWW.Symbolics.COM>
Subject: composition contests
To: BIL%CCRMA-F4@SAIL.Stanford.EDU
cc: BRG%CCRMA-F4@SAIL.Stanford.EDU, RSK%CCRMA-F4@SAIL.Stanford.EDU, lbj-next%CCRMA-F4@SAIL.Stanford.EDU,
rwg@WHITE.SWW.Symbolics.COM, les@sail.stanford.edu, jmc@sail.stanford.edu,
moravec@rover.ri.cmu.edu
In-Reply-To: The message of 12 Oct 89 10:19 PDT from Bill Schottstaedt <BIL%CCRMA-F4@SAIL.Stanford.EDU>
Message-ID: <19891013212036.1.MLB@CHROME.SWW.Symbolics.COM>
[bunch o' SAIL alumni (yes, -i, damnit!) added]
Date: 12 Oct 89 1019 PDT
From: Bill Schottstaedt <BIL%CCRMA-F4@SAIL.Stanford.EDU>
I was just looking at the bulletin board of contests and other performance
opportunities for composers and it seemed that every contest deliberately
excluded white males. There's at least a dozen for women only, and one
for Afro-Americans only (Detroit Symphony). I wonder if this is legal
in the case of groups like the Detroit Symphony. Then just to irritate
me, I got a contest announcement for Illinois residents only -- why did
they mail it to my address in Menlo Park!
As I said in another message, they only do it to Illinois you. (Obviously because you ain't
got rhythm.) On the other hand, it's better than being invited to enter a DEcomposition
contest, right?
Did you happen to see Les Earnest's article a few months ago in the Proc. ACM (I think that's
where I saw it)? He had a whole thing about how on forms that requested that he specify race
he started writing in (in the Other blank) "Mongrel". This results in a few requests for
clarification (maybe, didn't he mean "Mongoloid"?) and also, occasionally, even some
intelligent dialog. I might be able to find my copy to send you.
So, enter a piece as a Mongrel composer. Demand your Mongrel rights! I can just see the
bumper stickers now: "I Listen To Mongrel Music"!
PS: I've had a similar problem with the "Religious Preference (Optional)" question on forms.
I always wanted to put down "Scientist", but I was afraid that might get parsed as "Christian
Scientist" or "Scientologist". But a phrase that came up in an offhand discussion with Hans
saved me by providing a great alternative: nowadays I like to put down "Ardent Mechanist"!
∂13-Oct-89 1539 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
A SURVEY OF FORMAL NONMONOTONIC REASONING
Vladimir Lifschitz
Monday, October 16, 3:15pm
MJH 252
The Commonsense and Nonmonotonic Reasoning Seminar is a forum
in which we discuss current research and open problems related
to the logical foundations of AI, with a special emphasis on
commonsense reasoning and knowledge, formal nonmonotonic
reasoning, and foundations of logic programming.
This quarter, we are planning to include a few introductory
lectures on these subjects, for the benefit of those who are
new to this area of AI. The October 16 meeting will be the
first in this series. A basic knowledge of logic will be
assumed.
∂13-Oct-89 2035 underdog@Portia.stanford.edu employment
Received: from Portia.stanford.edu by SAIL.Stanford.EDU with TCP; 13 Oct 89 20:35:03 PDT
Received: by Portia.stanford.edu (5.59/25-eef) id AA22608; Fri, 13 Oct 89 20:36:14 PDT
Date: Fri, 13 Oct 89 20:36:14 PDT
From: Dwight Joe <underdog@Portia.stanford.edu>
Message-Id: <8910140336.AA22608@Portia.stanford.edu>
To: jmc@sail.stanford.edu
Subject: employment
Professor McCarthy:
Do you know if there are any part-time intellectual
(i.e. NOT secretarial and NOT gopher-type) jobs available
in CS? I cannot work for more than 8 hours per week.
Thanks,
---Dwight
∂14-Oct-89 1930 uhlik@sun-valley.Stanford.EDU washer
Received: from sun-valley.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Oct 89 19:30:11 PDT
Received: from kirkwood.Stanford.EDU ([201.1.100.2]) by sun-valley.Stanford.EDU (4.0/inc-1.0)
id AA13679; Sat, 14 Oct 89 19:37:17 PDT
Received: by kirkwood.Stanford.EDU (4.0/inc-1.0)
id AA19401; Sat, 14 Oct 89 19:37:11 PDT
Date: Sat, 14 Oct 89 19:37:11 PDT
From: uhlik@sun-valley.stanford.edu
Message-Id: <8910150237.AA19401@kirkwood.Stanford.EDU>
To: jmc@sail.stanford.edu
Subject: washer
I would like to see the washer. I am moving to my new house on October 29
and at that time, I will need a washer and dryer. If you could hold the
washer until then, I might be interested (of course I would like to see it
first).
-- Chris Uhlik uhlik@sun-valley.Stanford.EDU
328-6438 (home) 723-3608 (work)
∂14-Oct-89 1949 GOODMAN%uamis@rvax.ccit.arizona.edu sharing some praise
Received: from Forsythe.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Oct 89 19:49:01 PDT
Received: by Forsythe.Stanford.EDU; Sat, 14 Oct 89 19:49:39 PDT
Date: Sat, 14 Oct 89 19:43 MST
From: GOODMAN%uamis@rvax.ccit.arizona.edu
Subject: sharing some praise
To: DUANE.ADAMS@C.CS.CMU.EDU, BLUMENTHAL@A.ISI.EDU, DONGARRA@ANL-MCS.ARPA,
GANNON%RDVAX.DEC%DECWRL@decwrl.dec.com, JAHIR@ATHENA.MIT.EDU,
HEARN@RAND-UNIX.ARPA, JLH@SIERRA.STANFORD.EDU, JMC@SAIL.STANFORD.EDU,
KNEMEYER@A.ISI.EDU, MCHENRY@GUVAX.BITNET, OUSTER@GINGER.BERKELEY.EDU,
Ralston@MCC.COM, CWEISSMAN@DOCKMASTER.ARPA
X-VMS-To: @NASOLD
Just to share a few kind words that came to us via people I bumped into
last week:
Joe Traub and Mary Shaw had many good words for our effort (both are on the
CSTB). Herb Simon came up to me and talked about it (very positively) for
about 15 minutes. At least 2 people told me that State told memebrs of the
board that our report was the best value/cost effort they had ever gotten
out of a contract report (no surprise to any of us, but it's nice to know
that State is at least telling people that).
Just about everyone with more than a passing familiarity with our book, and
a serious interest in the subject, believe we were instrumental in various
ways in helping to being about the recent major changes in PC controls.
∂15-Oct-89 0814 beeson@ucscd.UCSC.EDU evolution
Received: from ucscd.UCSC.EDU ([128.114.129.2]) by SAIL.Stanford.EDU with TCP; 15 Oct 89 08:14:38 PDT
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA28612; Sun, 15 Oct 89 08:15:27 -0700
Date: Sun, 15 Oct 89 08:15:27 -0700
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8910151515.AA28612@ucscd.UCSC.EDU>
To: jmc@sail.stanford.edu
Subject: evolution
is the mutation frequency really random? Somebody named Cairns
decided it had never been tested, and put E. Coli which had been
mutated not to be able to digest sugar in a sucrose-rich environment,
and measured the mutation frequency, finding it far above random.
The results were checked and other similar experiments performed.
Now apparently the biologists are in an uproar about it. I read this
in a relatively old New Scientist and then misplaced the article, but
if I can find it again I'll give you the reference.
∂15-Oct-89 1118 feb6399@ultb.isc.rit.edu Question
Received: from ultb.isc.rit.edu by SAIL.Stanford.EDU with TCP; 15 Oct 89 11:17:53 PDT
Received: by ultb.isc.rit.edu (5.57/5.17)
id AA10426; Sun, 15 Oct 89 14:23:21 EDT
Date: Sun, 15 Oct 89 14:23:21 EDT
From: feb6399@ultb.isc.rit.edu (F.E. Barrus )
Message-Id: <8910151823.AA10426@ultb.isc.rit.edu>
To: jmc@sail.stanford.edu
Subject: Question
I hope I'm not taking up too much of your time, but could you answer
a question for me?
I'm currently working on developing a pre-compiler and interpreter for
an object oriented language for simulating real-world (and imaginary)
environments. (for interactive fiction, etc.)
I also need to develop a sophisticated English parser for entering
commands, and an easy way to modify and manipulate the "world" after
it is first compiled.
I've written languages of this type and parsers before, but was never
fully satisfied with their capibilities.
Question: Could you recommend to me the best language (in your opinion)
to base my language on? Or could you tell me where I could get files
with programs similar to what I'm developing? Mainly I just need ideas
on how to implement a lot of this. I'm writing the pre-compiler and
interpreter in C, by the way.
Thanks a lot for any assistance you can give.
- Frank Barrus (Comp. Sci major at R.I.T.)
∂15-Oct-89 1136 feb6399@ultb.isc.rit.edu one more thing
Received: from ultb.isc.rit.edu by SAIL.Stanford.EDU with TCP; 15 Oct 89 11:36:06 PDT
Received: by ultb.isc.rit.edu (5.57/5.17)
id AA11030; Sun, 15 Oct 89 14:40:59 EDT
Date: Sun, 15 Oct 89 14:40:59 EDT
From: feb6399@ultb.isc.rit.edu (F.E. Barrus )
Message-Id: <8910151840.AA11030@ultb.isc.rit.edu>
To: jmc@sail.stanford.edu
Subject: one more thing
Oh, I forgot to add that I also want to be able to make the computer
controlled characters in the simulated world capable of "learning" from
what they do, so that they all have goals, and are capable of doing whatever
is necessary to accomplish these goals, and will interact with the
human-controlled characters as well.
∂15-Oct-89 1746 0002814144@mcimail.com re: test and ssx
Received: from NRI.Reston.VA.US by SAIL.Stanford.EDU with TCP; 15 Oct 89 17:45:52 PDT
Date: Sun, 15 Oct 89 20:36 EDT
From: Jerry Pournelle <0002814144@mcimail.com>
To: John McCarthy <JMC@sail.stanford.edu>
Subject: re: test and ssx
Message-Id: <42891016003624/0002814144NB3EM@mcimail.com>
Oh. pointless to rea]store account then.
I am here Sunday night and Monday night at Rickey Hyatt Housein Palos
alto with lunch p
speech at BYTE thingy tomorrow at Rickey.
this communicationcnahhel does seem to work.
Best
j
∂15-Oct-89 1941 pehoushe@Gang-of-Four.Stanford.EDU Primes with Qlambda
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 15 Oct 89 19:40:54 PDT
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA09488; Sun, 15 Oct 89 19:39:57 -0700
Date: Sun, 15 Oct 89 19:39:57 -0700
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8910160239.AA09488@Gang-of-Four.Stanford.EDU>
To: qlisp@Gang-of-Four.Stanford.EDU
Subject: Primes with Qlambda
I ran some experiments using a version of Joe's primes sieve. I
modified it so that it incremented a counter instead of printed the
prime. This avoids IO bottlenecks and harmful bugs.
NEW-QLISP
In the new-qlisp version, I had some difficulty getting the program to
run without hanging for moderate values of N. It usually hung after a
GC. However, all times printed below were obtained without a GC
taking place. Times are in milleseconds.
N Parallel Serial
50 127 48
100 212 115
200 436 337
400 939 910
800 2528 2726
1600 7425 8459
New-Qlisp actually got a small amount of speed-up on N=800 and 1600.
Caveat: these times may not be very reproducible, but the program is
at the end of this message.
Alt-NEW-QLISP
In my latest version, which builds on an old new-qlisp, using a macro
version of qlambda called nqlambda, I got the following times (program
at end of message):
N Parallel Serial #Qlambda Spawns
50 8 13 167
100 14 27 435
200 30 71 1324
400 69 186 3639
800 182 541 10964
1600 505 1687 34338
32000 83522 292024 5975297
64000 290597 1020797 20754300
In the 64000 run, there were 20754300 spawns by nqlambda functions in
parallel mode (No GC, due to efficient recycling of structures). For
program correctness, I don't think any dynamic spawning is permitted.
(i.e. The numbers have to walk through the qlambda pipeline in
increasing order.)
There weren't any spawns when running in serial, although there were
associated function calls. The speed-up approaches a limit determined
by the cost of spawning a qlambda call. With 8 processors it's about
3.5, with 16 processors it would be about 7.
One reason for the large difference in serial times is the new
features in New-Qlisp.
If anyone wants to see my version of qlambda, I'd be glad to show it
off.
-Dan
:: begin NEW-QLISP program
(defglobalvar *prime-array* nil)
(defun primes (limit)
(setf *prime-array* (make-array
*number-of-processors*
:initial-element 0))
;; the sieve
(qwait (counter 3 limit (filter 2)))
;; how many primes were there?
(let ((sum 0))
(dotimes (i *number-of-processors*)
(incf sum (aref *prime-array* i)))
sum))
(defun counter (number limit proc)
(do ((i number (1+ i)))
((> i limit))
(funcall proc i)))
(defun filter (prime &aux (next-proc nil))
(incf (svref *prime-array* (get-processor-number)))
#'(qlambda t (number)
(unless (zerop (mod number prime))
(when (null next-proc)
(setq next-proc (filter number)))
(funcall next-proc number))))
;;; end NEW-QLISP program
;;; **************************
;;; begin Alt-NEW-QLISP
(defglobalvar *prime-array* nil)
(defun primes (limit)
(setf *prime-array* (make-array
*number-of-processors*
:initial-element 0))
(counter 3 limit (filter 2))
(let ((sum 0))
(dotimes (i *number-of-processors*)
(incf sum (aref *prime-array* i)))
sum))
(defun counter (number limit proc)
(do ((i number (1+ i)))
((> i limit))
(funcall proc i)))
(defun filter (prime &aux (next-proc nil))
(incf (svref *prime-array* (get-processor-number)))
(nqlambda t (number)
(unless (zerop (mod number prime))
(when (null next-proc)
(setq next-proc (filter number)))
(funcall next-proc number))))
;;; end Alt-NEW-QLISP program
∂15-Oct-89 2246 rathmann@eclipse.stanford.edu Scheduling Orals
Received: from eclipse.stanford.edu by SAIL.Stanford.EDU with TCP; 15 Oct 89 22:46:14 PDT
Received: by eclipse.stanford.edu (4.0/inc-1.0)
id AA00119; Sun, 15 Oct 89 22:44:40 PDT
Date: Sun, 15 Oct 89 22:44:40 PDT
From: rathmann@eclipse.stanford.edu (Peter K. Rathmann)
Message-Id: <8910160544.AA00119@eclipse.stanford.edu>
To: gio@sumex.stanford.edu, jmc@sail.stanford.edu, shoham@polya.stanford.edu
Cc: rathmann@eclipse.stanford.edu
Subject: Scheduling Orals
Dear Committee Members,
With luck, you may have had occassion over the last week to review my
thesis draft. Based upon that, and if you feel I am ready, I would
like to schedule my orals, preferably before the Thanksgiving holiday.
Possible dates, at least four weeks away, include November 13th, 21st,
or 22nd. Thanks, I look forward to hearing from you at you earliest
convenience.
-Peter
∂16-Oct-89 0800 JMC
cate about form
∂16-Oct-89 0933 MPS phone
Terry Taug;ner, Information International
213 390-8611
Please call
∂16-Oct-89 0934 CLT Monday, November 20, 1989
I don't suppose there is another possible hour.
Who will take Timothy to school?
∂16-Oct-89 0942 CLT Visitor
As you may recall, we discussed with Gigina the possibilty
of one of her students visiting and working with Vladimir
and possibly Matt. He needs a formal letter of invitiation
in order to the get Italian grant to come. Below is my attempt.
If this is ok, I'll have pat put it on paper, else send me a
revision.
-------------------------------------------------------
Marco Schaerf
School of computer science
McGill University
3480 University
Montreal Quebec
Canada H3A 2A7
Dear Mr. Schaerf,
I am pleased to invite you to visit our research group in
the Stanford Computer Science Department to work in the
area of Artificial Intelligence for the
period January through June 1990.
Sincerely
John McCarthy
∂16-Oct-89 1006 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU Re: Tuesday seminar
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Oct 89 10:06:05 PDT
Received: from iswim.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61/25-eef) id AA10677; Mon, 16 Oct 89 10:07:04 -0700
Received: by iswim.stanford.EDU (5.57/Ultrix3.0-C)
id AA02448; Mon, 16 Oct 89 10:08:20 PDT
Message-Id: <8910161708.AA02448@iswim.stanford.EDU>
To: John McCarthy <JMC@sail.stanford.edu>
Subject: Re: Tuesday seminar
Cc: katiyar@polya.Stanford.EDU
In-Reply-To: Your message of 16 Oct 89 00:28:00 -0700.
<6m0JJ@SAIL.Stanford.EDU>
From: John C. Mitchell <jcm@Polya.Stanford.EDU>
Reply-To: John C. Mitchell <jcm@Polya.Stanford.EDU>
Date: Mon, 16 Oct 89 10:08:19 PDT
Sender: jcm@iswim
I am scheduled to speak tomorrow, but would gladly
put my talk off until the 31-st. So take your pick and
we'll get the announcement right away. Would you prefer the
17-th (tomorrow) or the 31-st?
John
∂16-Oct-89 1037 helen@russell.Stanford.EDU re: Lunch Meeting
Received: from russell.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Oct 89 10:36:59 PDT
Received: by russell.Stanford.EDU (4.0/inc-1.0)
id AA17955; Mon, 16 Oct 89 10:38:45 PDT
Date: Mon 16 Oct 89 10:38:44-PDT
From: Helen Nissenbaum <HELEN@CSLI.Stanford.EDU>
Subject: re: Lunch Meeting
To: JMC@SAIL.Stanford.EDU
Message-Id: <624562724.0.HELEN@CSLI.Stanford.EDU>
In-Reply-To: <12m9vs@SAIL.Stanford.EDU>
Mail-System-Version: <SUN-MM(242)+TOPSLIB(128)@CSLI.Stanford.EDU>
Cordura Hall Conference Room. We'll start at about 11:30 and go till 1.
--Helen
-------
∂16-Oct-89 1045 ginsberg@Polya.Stanford.EDU re: Review of Akawa's paper on presupposition
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Oct 89 10:45:27 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA13253; Mon, 16 Oct 89 10:46:26 -0700
Date: Mon, 16 Oct 89 10:46:26 -0700
From: Matthew L. Ginsberg <ginsberg@Polya.Stanford.EDU>
Message-Id: <8910161746.AA13253@Polya.Stanford.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: re: Review of Akawa's paper on presupposition
The review has already gone in; I heard back from Bobrow
with an apology (!) for subjecting you/us to it. He said
that he usually tries to weed out stuff this bad ...
My attitude when reviewing stuff that is clearly junk
depends on what the underlying attitude of the authors
appears to be. In Akawa's case, I had the impression
that he was just stringing together stuff at random,
not thinking at all, and had no real interest in thinking
at all. I couldn't think of a reason to soften the
review ...
Hoping that he's not some sort of an acquaintance of yours,
Matt
∂16-Oct-89 1122 VAL DARPA visit on Nov. 20
Do we want to stay there for Latombe's presentation (which will follow
ours)? If yes, I'll cancel the seminar that day.
∂16-Oct-89 1500 JJW Telnet from SAIL to Go4
My character macro is the following:
ESC 6 CR: s o u r c e <SP> . d m w a i t s <CR> βt
The meta-t command needs to be the last thing, because it puts Telnet
into transparent mode which stops further processing of the macro.
On Gang-of-Four, the file .dmwaits in my home directory contains:
set term=dmwaits
stty 2400
echo -n '↑]↑↑'
where the text in the quotes is actually the ascii characters ctrl-]
and ctrl-↑. This sends the sequence to put the terminal in roll mode,
which SAIL usually has turned off.
The overall procedure is:
1. Telnet Go4.
2. Log in.
3. At the shell prompt, type <esc> 6 <cr> to SAIL (or use the function
key on the terminal).
4. Release the SAIL lock key.
∂16-Oct-89 1508 harnad@clarity.Princeton.EDU BBS Call for Commentators: Visual Field Specialization
Received: from Princeton.EDU (indy.Princeton.EDU) by SAIL.Stanford.EDU with TCP; 16 Oct 89 15:08:21 PDT
Received: from reason.Princeton.EDU by Princeton.EDU (5.58+++/2.22/mailrelay)
id AA14555; Mon, 16 Oct 89 00:47:08 EDT
Received: by reason.Princeton.EDU (4.0/1.86)
id AA00463; Mon, 16 Oct 89 00:39:28 EDT
Date: Mon, 16 Oct 89 00:39:28 EDT
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8910160439.AA00463@reason.Princeton.EDU>
To: harnad@elbereth.rutgers.edu
Subject: BBS Call for Commentators: Visual Field Specialization
Below is the abstract of a forthcoming target article to appear in
Behavioral and Brain Sciences (BBS), an international,
interdisciplinary journal that provides Open Peer Commentary on important
and controversial current research in the biobehavioral and cognitive
sciences. Commentators must be current BBS Associates or nominated by a
current BBS Associate. To be considered as a commentator on this article,
to suggest other appropriate commentators, or for information about how
to become a BBS Associate, please send email to:
harnad@confidence.princeton.edu or write to:
BBS, 20 Nassau Street, #240, Princeton NJ 08542 [tel: 609-921-7771]
____________________________________________________________________
Functional Specialization in the Lower and Upper Visual Fields in Man:
Its Ecological Origins and Neurophysiological Implications
by Fred H. Previc
Crew Technology Division
USAF School of Aerospace Medicine
Brooks AFB, TX 78235-5301
Abstract: Functional specialization in the lower and upper visual
fields in man is reviewed and interpreted with respect to the origins
of the primate visual system. Many of the processing differences
between the vertical hemifields are related to the distinction between
near (personal) and far (extrapersonal) space, which are biased towards
the lower and upper visual fields respectively. It is hypothesized that
a significant enhancement of these functional specializations occurred
in conjunction with the emergence of forelimb manipulative skills and
fruit-eating, in which the fruit or distal object is searched and
attended to in central vision while the reaching motion of the hand and
other related manipulations are monitored in the proximal lower visual
field. Objects in far vision are searched and recognized primarily
using linear/local perceptual mechanisms, whereas nonlinear/global
processing is required in the lower visual field in order to perceive
the optically degraded and diplopic images contained in near vision.
The functional differences between the lower and upper visual fields
are correlated with their disproportionate representations in the
dorsal vs. ventral divisions of visual association cortex,
respectively, and in the magnocellular and parvocellular pathways that
project to them. The division between near and far visual functions may
also have contributed to the transformations of the lateral geniculate
nucleus, superior colliculus, and frontal visual areas which occurred
during the evolution of primate vision.
∂16-Oct-89 1604 carol@lucid.com new new-qlisp
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Oct 89 16:04:34 PDT
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA12342; Mon, 16 Oct 89 16:03:34 -0700
Received: from vesuvius ([192.9.200.11]) by heavens-gate id AA21495g; Mon, 16 Oct 89 15:58:00 PDT
Received: by vesuvius id AA01109g; Mon, 16 Oct 89 15:56:48 PDT
Date: Mon, 16 Oct 89 15:56:48 PDT
From: Carol Sexton <carol@lucid.com>
Message-Id: <8910162256.AA01109@vesuvius>
To: qlisp@go4.stanford.edu
Subject: new new-qlisp
I'm planning on installing a new new-qlisp this afternoon or evening.
(There are people currently running new-qlisp so I can't install it at the
moment.) This new lisp contains a couple of bug fixes which showed up when
running Joe's prime sieve program. One bug was a race condition involving
the creation of futures for qlambda processes and the other was a problem
with the processor register getting corrupted when the keyword parsing
routine was interrupted.
Carol
∂16-Oct-89 1634 ingrid@russell.Stanford.EDU IAP brochure
Received: from russell.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Oct 89 16:34:01 PDT
Received: from localhost by russell.Stanford.EDU (4.0/inc-1.0)
id AA21427; Mon, 16 Oct 89 16:35:42 PDT
Message-Id: <8910162335.AA21427@russell.Stanford.EDU>
To: jmc@sail.stanford.edu
Subject: IAP brochure
Date: Mon, 16 Oct 89 16:35:39 PDT
From: ingrid@russell.Stanford.EDU
John,
I'm having our IAP brochure reprinted and need an update on your
research interests.
Last year's brochure listed the following interests for you:
- formalizing commen-sense knowledge and reasoning, especially
nonmonotonic reasoning
Would you like to add or subtract anything from the above? Let me
know by Friday, 20 October, please.
Ingrid
∂16-Oct-89 1705 carol@lucid.com new new-qlisp
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Oct 89 17:05:01 PDT
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA12566; Mon, 16 Oct 89 17:03:49 -0700
Received: from vesuvius ([192.9.200.11]) by heavens-gate id AA21820g; Mon, 16 Oct 89 17:00:49 PDT
Received: by vesuvius id AA01225g; Mon, 16 Oct 89 16:59:49 PDT
Date: Mon, 16 Oct 89 16:59:49 PDT
From: Carol Sexton <carol@lucid.com>
Message-Id: <8910162359.AA01225@vesuvius>
To: qlisp@go4.stanford.edu
Subject: new new-qlisp
I've installed the new new-qlisp. The previous new-qlisp
is currently in /lucid/bin/new-qlisp.9oct89. If the
the new new-qlisp seems healthy, I'll delete
new-qlisp.9oct89 later this week.
Carol
∂16-Oct-89 1707 winograd@loire.stanford.edu December AI Qualifying Exam
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 16 Oct 89 17:07:03 PDT
Received: by loire.stanford.edu (5.59/25-eef) id AA08130; Mon, 16 Oct 89 17:05:42 PDT
Date: Mon, 16 Oct 89 17:05:42 PDT
Message-Id: <8910170005.AA08130@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: ai-qual-faculty@loire.stanford.edu, phd@cs
Subject: December AI Qualifying Exam
Students wishing to take the AI qual this quarter, as described below, should
send Sharon Hemenway the following information by next Monday, October 23:
1) Name
2) Depth area
3) Times during the week of December 4-8 when you COULD NOT take the exam.
Based on the responses we will choose a single day (or two if there are too many
people for one day) and set the exam. The next exam will be in the Spring.
If you have any questions, check with Sharon or with me. --t
------------------------------------------------------------------------------
Preliminary Announcement
Qualifying Examination in Artificial Intelligence
Computer Science Department
Stanford University
December, 1989
1. PURPOSE
The purpose of the AI qualifying examination is to determine whether a
student has sufficient breadth of knowledge in AI generally, and depth
of knowledge in some subtopic of AI, to undertake research toward his
or her doctoral dissertation.
2. FORMAT
The format is a ninety-minute oral examination before a committee of
three members, consisting of faculty or staff from Stanford and/or
senior researchers from other AI laboratories. The chair of each
committee will be a member of the Stanford faculty. Approximately one
hour will be spent on the breadth examination and 30 minutes on the
depth area. These relative durations are only approximate, and the
order of the two parts is up to the chair. The committee may want to
excuse the candidate for a few minutes before the end of the exam to
confer amongst themselves about the progress of the exam and to
determine how the remaining time can best be used.
3. GRADING
The committee has final authority on the outcome of the examination.
There are four possible outcomes:
Unconditional Pass. This outcome is appropriate if it is clear
that the student has satisfactory depth and breadth.
Conditional Pass. This outcome is appropriate if the student has
displayed adequate depth and substantial breadth but has clearly
identifiable weaknesses that can be remedied by directed work, such as
the completion of a specific course or research project. The
committee should prescribe the scope of the work, the time period
during which it must be completed, and whom is to certify completion.
Continuation. This outcome is appropriate if there is strong evidence
that the student has adequate depth and breadth but there is still
substantial doubt. If this outcome is selected, the student must meet
with the committee again. The committee should make clear to the
student the nature of its doubts and fix a time for the continuation
of the examination.
Deferral. This outcome is appropriate if it is clear that gaps in the
student's knowledge cannot be remedied by limited correctives. The
candidate is required to take the examination again.
All committees will meet at the end of the day to calibrate their
grading of the exams. Results of the exams will be available through
the committee chairs by 6:00 p.m. on May 26.
4. SYLLABUS
The syllabus for this AI Qual is the reading list updated last year by
Karen Myers and Ramin Zabih. Copies will be given to all examiners
shortly. Students should be familiar with basic concepts and
techniques in all of the areas listed and should know where to find
additional information.
In addition, students should be familiar with the specialized topic or
depth area in AI they have declared.
∂16-Oct-89 1737 VAL Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
A CONNOTATIVE TREATMENT OF CIRCUMSCRIPTION
Jun Arima
ICOT Research Center, Japan
Monday, October 23, 3:15pm
MJH 252
Circumscription proposed by McCarthy is one of the most promising
formalizations of nonmonotonic aspects of commonsense reasoning.
It has several versions. However, they are all proposed for the
denotative minimization of predicates, that is, circumscription
minimizes the extension of predicates. Regarding such treatment,
this paper considers three problems: absence of abnormal things, a
limitation on equality and inconsistency of circumscription. This
paper proposes a solution for them by presenting a connotative
treatment of circumscription. This treatment is based on the idea
of circumscribing predicates connotatively, that is, minimizing
the set of names denoting objects which satisfy a certain
predicate. Besides, this treatment provides a form of the unique
names hypothesis for a general database which allows the use of
equality and functions.
∂16-Oct-89 1800 JMC
Leora
∂16-Oct-89 2001 JMC
suppes re Kaelbling
∂17-Oct-89 0056 harnad@clarity.Princeton.EDU The Road to Beijing
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 17 Oct 89 00:56:13 PDT
Received: from reason.Princeton.EDU by Princeton.EDU (5.58+++/2.22/mailrelay)
id AA05561; Tue, 17 Oct 89 01:02:47 EDT
Received: by reason.Princeton.EDU (4.0/1.86)
id AA00976; Tue, 17 Oct 89 00:57:34 EDT
Date: Tue, 17 Oct 89 00:57:34 EDT
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8910170457.AA00976@reason.Princeton.EDU>
To: smoliar@vaxa.isi.edu
Subject: The Road to Beijing
J.H. Hexter and the Road to Beijing
To: Steve Smoliar
From: Stevan Harnad
Steve, you ask:
> What can a memory be OF, if not an experience?
It can be of how to do something, what something means, what something
looks/sounds like (in general, not on a particular occasion), etc.
> I, too, know that I understand English and I don't understand Chinese;
> but what do I know about YOU... all I know about your understanding
> of Chinese is what you have told me, which is neither better nor worse
> than a computer which is programmed to respond "Yes" when I type in "Do
> you understand me?"
Oh, my! Do things ever manage to get garbled up! Whether what I have
told you is better or worse than what a computer tells you is what is
AT ISSUE here. You can't trot it out as a first principle. And you
know a whale of a lot more about me than the Turing Test [TT] (especially
now that we've met): You know I'm a human being, made out of the same
wetware as you. So you have much better antecedent grounds for trusting
me than your VAX if I say I understand English and not Chinese. Ditto
for Searle. And ditto for Searle when he's doing everything the VAX is
doing when you think it's understanding Chinese, and he is telling you
he DOESN'T understand Chinese, and that the Chinese symbols he's
manipulating for you are just gibberish to him.
> In order for me to form any knowledgeable opinion of your understanding
> of Chinese, I would probably have to observe you in a Chinese environment.
Now this is no longer just the TT (symbols in, symbols out); it's the
Total Turing Test [TTT], calling for full sensorimotor interaction with
the world. And for a system to pass that, it must have nonsymbolic
functions, transduction at the least. Transduction can be simulated,
but not implemented symbolically. Hence it is immune to Searle's
Chinese Room Argument. That's what symbol grounding is about. I've
repeated this many times before. You should give some sign of having
understood it rather than repeating the step that preceded it every
time...
> does Diana Rigg know if you understand Chinese?... she is an actress
> and therefore used to dealing with people playing roles in which they
> say things which are not true about themselves.
Indeed she is. And she presumably also knows what it's like to play a
role in a language she doesn't understand...
> So, if she REALLY had to know if you understood Chinese... She
> would.... put you in an environment in which your knowledge of Chinese
> could be observed... In other words she would end up testing you
> exactly the same way we test the man in Searle's room and use those
> interactions to say whether or not you understood Chinese.
This mixing up of issues is getting to be too much for me: Searle's
room is just the TT (symbols in, symbols out), NOT the TTT. That's the
basic point of the symbol grounding problem. It NECESSARILY takes more
to pass the TTT than the TT (in particular, it REQUIRES nonsymbolic
functions). And, as I've said over and over, I think even the
conjecture that the TT could be successfully passed by a pure symbol
cruncher is counterfactual, again because of the symbol grounding
problem. I'm sure it would require full TTT power to pass the TT, and
that in turn would be immune to the Chinese Room Argument.
> I think the reason many of us in AI do not take Searle's argument very
> seriously is that... it really does not tell us anything we want to
> know. It certainly does not tell us very much about how intelligent
> agents get on in the world by interacting with other intelligent
> agents. Ultimately, it seems to be yet another attempt to demonstrate
> that there is something in being human which rises above the capability
> of any machine. Minsky, on the other hand, wishes to assure us that
> there is no stigma in being perceived as a machine, since machines have
> the power to behave with as much complexity as humans.. if not more
> so. Is this not a healthier attitude to assume if one is interested in
> the workings of the mind?
It's not a matter of "attitude" (which is ALL Minsky is giving us any
more). Searle doesn't show us which way is right, but he sure does show
us which way is wrong: Pure symbol manipulation. Stereotyping him as a
Luddite is just obfuscation. Searle repeats over and over that he
knows perfectly well we're machines; it's a question of what KIND of
machine we are. And one kind we're NOT is pure symbol crunchers. And
that's good to know. I feel it's time for me to trot out my J.H.
Hexter quote again (always at this same juncture):
in an academic generation a little overaddicted to "politesse," it
may be worth saying that violent destruction is not necessarily
worthless and futile. Even though it leaves doubt about the right
road for London, it helps if someone rips up, however violently, a
`To London' sign on the Dover cliffs pointing south...
Steve, I'm afraid I can't discuss this further until I hear a new
argument, or at least some sign of someone's having understood SOME
portion of the points I've been making.
Stevan Harnad
∂17-Oct-89 0646 @Score.Stanford.EDU,@MCC.COM:greene@mcc.com Elephant 2000
Received: from Score.Stanford.EDU by SAIL.Stanford.EDU with TCP; 17 Oct 89 06:46:09 PDT
Received: from MCC.COM by SCORE.STANFORD.EDU with TCP; Tue 17 Oct 89 06:47:15-PDT
Received: from luke.aca.mcc.com by MCC.COM with TCP/SMTP; Tue 17 Oct 89 08:47:09-CDT
Date: Tue, 17 Oct 89 08:46:30 CDT
From: greene@mcc.com (Kevin Greene)
Posted-Date: Tue, 17 Oct 89 08:46:30 CDT
Message-Id: <8910171346.AA14881@luke.aca.mcc.com>
Received: by luke.aca.mcc.com (3.2/ACTv4.1i)
id AA14881; Tue, 17 Oct 89 08:46:30 CDT
To: jmc@sail.stanford.edu
Subject: Elephant 2000
I was unable to attend your recent talk at UT on Elephant 2000. I would very
much appreciate receiving any write-up you might have on the subject. Thanks
in advance for your reply.
--Kevin
Kevin J Greene--MCC, 3500 W. Balcones Center Dr, Austin, TX, 78759
greene@mcc.com, 512-338-3471
∂17-Oct-89 0940 winograd@loire.stanford.edu elephant 2000
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 17 Oct 89 09:40:26 PDT
Received: by loire.stanford.edu (5.59/25-eef) id AA09037; Tue, 17 Oct 89 09:41:04 PDT
Date: Tue, 17 Oct 89 09:41:04 PDT
Message-Id: <8910171641.AA09037@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: jmc@sail
Subject: elephant 2000
John,
Two different people in the last couple of days have asked me about
Elephant 2000, assuming that since it is being done by a colleague of
mine at Stanford, and uses speech acts (about which I have made a fuss
at times) that I must know all about it. It sounds vaguely like things
we may have discussed long ago, but I don't know much. Do you have
written papers, or could we chat sometime? It sounds interesting.
--t
∂17-Oct-89 0947 AI.LENAT@MCC.COM Re: visit dates
Received: from MCC.COM by SAIL.Stanford.EDU with TCP; 17 Oct 89 09:47:08 PDT
Date: Tue 17 Oct 89 11:46:21-CDT
From: Doug Lenat <AI.LENAT@MCC.COM>
Subject: Re: visit dates
To: JMC@SAIL.Stanford.EDU
cc: AI.LENAT@MCC.COM
In-Reply-To: <tmW0Z@SAIL.Stanford.EDU>
Message-ID: <12534860704.42.AI.LENAT@MCC.COM>
Yes, Nov. 16 and 17 will be good. We will reactivate your contract
-- are the old terms acceptable (I don't know the binding for "old
terms", such as rate of pay, travelling expense reimbursement, etc.,
but my default action is to just say "reactivate the thing")? Let's
also try to get in a day of consulting during our upcoming trip to
Ca. (Oct. 30-Nov. 2) if possible. We would very much like to get
most of your consulting time for the coming year, or as much as you
can spare -- probably about half here in Texas and half in Palo Alto.
CYC has now advanced (in several dimensions) to the point where you
can make significant progress, I believe, both to it and using it.
Regards,
Doug
-------
∂17-Oct-89 1048 winograd@loire.stanford.edu re: elephant 2000
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 17 Oct 89 10:48:49 PDT
Received: by loire.stanford.edu (5.59/25-eef) id AA09412; Tue, 17 Oct 89 10:49:29 PDT
Date: Tue, 17 Oct 89 10:49:29 PDT
Message-Id: <8910171749.AA09412@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: JMC@SAIL.Stanford.EDU
Subject: re: elephant 2000
John,
I'll be out of town Thursday-Sun. If I get a draft tomorrow, I can
read it on the plane and we can talk next week. --t
∂17-Oct-89 1400 JMC
expenses
∂17-Oct-89 1516 MPS Calls
The only person I was unable to reach was Gelpi, A. Phil.
There is no Stanford number for him and 858--3923 is
always busy.
∂17-Oct-89 1946 perlis@cs.UMD.EDU reference on context
Received: from yoohoo.cs.UMD.EDU by SAIL.Stanford.EDU with TCP; 17 Oct 89 19:46:17 PDT
Received: by yoohoo.cs.UMD.EDU (5.61/UMIACS-0.9/04-05-88)
id AA13077; Tue, 17 Oct 89 22:53:13 -0400
Date: Tue, 17 Oct 89 22:53:13 -0400
From: perlis@cs.UMD.EDU (Don Perlis)
Message-Id: <8910180253.AA13077@yoohoo.cs.UMD.EDU>
To: jmc@sail.stanford.edu
Subject: reference on context
John, you might be interested in the OMEGA system developed
by Attardi & Simi. One basic reference is their paper in IJCAI-81;
another is Simi&Motta in the book "Meta-Level Architectures and
Reflection, edited by P. Maes and D. Nardi, North-Holland (1988).
I think they are trying to capture at least some aspects of what
you call contexts.
Don
∂18-Oct-89 1300 bledsoe@cs.utexas.edu Sorry I missed you
Received: from cs.utexas.edu by SAIL.Stanford.EDU with TCP; 18 Oct 89 12:59:56 PDT
Received: from alfalfa.cs.utexas.edu by cs.utexas.edu (5.59/1.43)
id AA01897; Wed, 18 Oct 89 15:00:58 CDT
Posted-Date: Wed, 18 Oct 1989 15:00:44 CDT
Received: by alfalfa.cs.utexas.edu (5.59/1.4-Client)
id AA01527; Wed, 18 Oct 89 15:00:46 CDT
Date: Wed, 18 Oct 1989 15:00:44 CDT
From: Woody Bledsoe <bledsoe@cs.utexas.edu>
To: jmc@sail.stanford.edu
Subject: Sorry I missed you
Message-Id: <CMM.0.86.624744044.bledsoe@alfalfa.cs.utexas.edu>
John,
I'm sorry that I missed seeing you when you were here, and especially
unhappy that I did not hear you speak about you new language.
It could be really important in the long run.
Woody
∂18-Oct-89 1429 Mailer re: quake
Received: from akbar.cac.washington.edu by SAIL.Stanford.EDU with TCP; 18 Oct 89 14:29:07 PDT
Received: from tomobiki-cho.cac.washington.edu by akbar.cac.washington.edu
(5.61/UW-NDC Revision: 2.6 ) id AA29205; Wed, 18 Oct 89 14:30:13 -0700
Date: Wed, 18 Oct 1989 14:19:46 PDT
From: Mark Crispin <MRC@CAC.Washington.EDU>
Sender: mrc@Tomobiki-Cho.CAC.Washington.EDU
Subject: re: quake
To: John McCarthy <JMC@sail.stanford.edu>
Cc: pkarp@ncbi.nlm.nih.gov, su-etc@sail.stanford.edu
In-Reply-To: <fnrJX@SAIL.Stanford.EDU>
Message-Id: <MailManager.624748786.4084.mrc@Tomobiki-Cho.CAC.Washington.EDU>
JMC, thanks for the update. KRON's coverage was broadcast nationwide.
From everything I've heard, it could have been a lot worse.
The question is; what this "The Big One" (= the SF Bay Area is "safe" for
the rest of our lifetimes) or is "The Big One" yet to come? I've heard
conflicting comments about that, suggesting that nobody really knows.
Seattle is overdue for a quake, plus we have volcanoes up here. We've
been watching the goings-on in the Bay Area for when we get ours...
-------
∂18-Oct-89 1607 Mailer events in Hungary
Received: from hanna.cac.washington.edu by SAIL.Stanford.EDU with TCP; 18 Oct 89 16:07:03 PDT
Received: from tomobiki-cho.cac.washington.edu by hanna.cac.washington.edu
(5.61/UW-NDC Revision: 2.6 ) id AA23459; Wed, 18 Oct 89 16:08:04 -0700
Date: Wed, 18 Oct 1989 16:02:16 PDT
From: Mark Crispin <mrc@SUMEX-AIM.Stanford.EDU>
Subject: events in Hungary
To: JMC@SAIL.Stanford.EDU, su-etc@SAIL.Stanford.EDU
Message-Id: <MS-C.624754936.1103527590.mrc@SUMEX-AIM.Stanford.EDU>
I haven't heard any comments from JMC about the events in Hungary, and
how that meshes with his theories that communist governments never terminate
voluntarily. In Poland, Solidarity staged a kind of coup, but this cannot be
said of what's happening in Hungary.
-------
∂18-Oct-89 1840 @Score.Stanford.EDU:kirsh%cogsci@ucsd.edu aij
Received: from Score.Stanford.EDU by SAIL.Stanford.EDU with TCP; 18 Oct 89 18:40:13 PDT
Received: from ucsd.edu by SCORE.STANFORD.EDU with TCP; Wed 18 Oct 89 17:24:17-PDT
Received: from cogsci.ucsd.edu by ucsd.edu; id AA10969
sendmail 5.61/UCSD-2.0-sun via SMTP
Wed, 18 Oct 89 17:24:26 -0700 for nilsson@score.stanford.edu
Received: from cogsci-sun.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA07399 for rosenblo@veneva.isi.edu; Wed, 18 Oct 89 17:26:55 PDT
Date: Wed, 18 Oct 89 17:24 PDT
From: David Kirsh <kirsh%cogsci-sun@ucsd.edu>
Subject: aij
To: bobrow@xerox.com, nilsson@score.stanford.edu, birnbaum@ils.nwu.edu,
lenat@mcc.com, feigenbaum@SUMEX-AIM.Stanford.EDU,
briansmith.pa@xerox.com, hewitt@ai.ai.mit.edu, jmc@score.stanford.edu,
brooks@ai.ai.mit.edu, rosenblo@veneva.isi.edu, norman@cogsci.ucsd.edu
Message-Id: <19891019002436.4.KIRSH@cogsci-sun.ucsd.edu>
Gentlemen,
It's been a long time coming but now we are approaching closure on the special
issue of the AIJ on the Foundations of AI. The final version of the
manuscript will be sent off to Danny Bobrow on Dec 1st. Those of you who
still have revisions to make MUST submit your final draft to me by Nov 25th.
A version of that special edition will appear later as a book with the MIT
Press. Right now we are considering using whatever royalties arise from the
sales of the book to create a fund for a future workshop on the foundations.
Please let me know if you dislike this idea.
I'd like to thank you all for your efforts.
-- David
∂18-Oct-89 2027 CLT sarah
called, wondered if you had heard whether Granny and Joe
were ok in Watsonville. Call her if you hear anything,
but no need to call otherwise. She may be down this weekend.
∂18-Oct-89 2042 Mailer re: events in Hungary
Received: from hanna.cac.washington.edu by SAIL.Stanford.EDU with TCP; 18 Oct 89 20:42:11 PDT
Received: from tomobiki-cho.cac.washington.edu by hanna.cac.washington.edu
(5.61/UW-NDC Revision: 2.6 ) id AA24238; Wed, 18 Oct 89 20:43:18 -0700
Date: Wed, 18 Oct 1989 20:30:36 PDT
From: Mark Crispin <mrc@SUMEX-AIM.Stanford.EDU>
Subject: re: events in Hungary
To: John McCarthy <JMC@sail.stanford.edu>
Cc: mrc@sumex-aim.stanford.edu, su-etc@sail.stanford.edu
In-Reply-To: <Nnufm@SAIL.Stanford.EDU>
Message-Id: <MS-C.624771036.1103527590.mrc@SUMEX-AIM.Stanford.EDU>
The next question is what's going to happen afterwards?
We can assume that Poland and Hungary, upon completing the transition into
free countries, are still going to keep those elements of the communist system
that are popular -- specifically, the social support programs. The sheer cost
of these systems are bound to guarantee instability in these countries, and
fertile grounds for demagogues.
What about the US? It can't be long before we start reaping the harvest of
the deficit and the total neglect of education here. The religious and
pseudo-scientific lunacy that has popped up in the past couple of decades is
not encouraging.
Will we end up with the worldwide power centers in Berlin and Tokyo?
-------
∂18-Oct-89 2259 ito@ito.ecei.tohoku.ac.jp Quake
Received: from RELAY.CS.NET by SAIL.Stanford.EDU with TCP; 18 Oct 89 22:59:33 PDT
Received: from [192.41.197.3] by RELAY.CS.NET id aa22989; 19 Oct 89 1:00 EDT
Received: from ccut.cc.u-tokyo.ac.jp by relay.cc.u-tokyo.ac.jp (5.61/2.3W-relay)
id AA06006; Thu, 19 Oct 89 15:00:32 +0900
Received: by ccut.cc.u-tokyo.ac.jp (5.61/6.4J.6-ut1.67)
id AA14471; Thu, 19 Oct 89 15:03:33 +0900
Received: from maia.ito.ecei.tohoku.ac.jp by hirose.cc.tohoku.ac.jp (3.2/6.4J.6-Oct17)
id AA22077; Thu, 19 Oct 89 12:47:17 JST
Received: by maia.ito.ecei.tohoku.ac.jp (4.0/6.4J.6)
id AA06405; Thu, 19 Oct 89 12:46:06 JST
Received: by aoba.ito.ecei.tohoku.ac.jp (4.0/6.4J.6)
id AA10861; Thu, 19 Oct 89 12:45:49 JST
Received: by ito.ito.ecei.tohoku.ac.jp (3.2/6.4J.6)
id AA22888; Thu, 19 Oct 89 12:37:31+0900
Date: Thu, 19 Oct 89 12:37:31+0900
From: Takayasu ITO <ito@ito.ecei.tohoku.ac.jp>
Return-Path: <ito@ito.ecei.tohoku.ac.jp>
Message-Id: <8910190237.AA22888@ito.ito.ecei.tohoku.ac.jp>
To: JMC%sail.stanford.edu%relay.cs.net@u-tokyo.ac.jp
Subject: Quake
How was the quake at your place? Are you and your family OK?
Takayasu Ito
∂18-Oct-89 2325 ME mail addresses
∂18-Oct-89 2306 JMC Is the following address likely to work?
ito@ito.ecei.tohoku.ac.jp
ME - Yes, the host name looks OK (the domain system recognizes it).
You can check this sort of thing out with the host command like this:
host ito.ecei.tohoku.ac.jp
If it doesn't say "not a registered domain name", you're in business.
Or you can simply mail to such an address. The mailer will quickly
let you know the same negative response if that's the case.
∂19-Oct-89 0912 MPS
Nancy Osmond, McMillan will be here next Monday, and would
like to make an appointment with you. 415-726-6910.
She is the science editor
∂19-Oct-89 1225 pehoushe@Gang-of-Four.Stanford.EDU Pat + meeting
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 19 Oct 89 12:25:05 PDT
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA06536; Thu, 19 Oct 89 12:24:14 -0700
Date: Thu, 19 Oct 89 12:24:14 -0700
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8910191924.AA06536@Gang-of-Four.Stanford.EDU>
To: jmc@sail.Stanford.EDU
Subject: Pat + meeting
Pat had a mild anxiety attack, so I took her home.
About your meeting: (I assume you know what this is about; I do not.)
She couldn't get Gelpi (858-3923) or Cohn 3-0661, 323-7983.
∂19-Oct-89 1759 carol@lucid.com new new-qlisp
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 19 Oct 89 17:59:27 PDT
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA07798; Thu, 19 Oct 89 17:58:36 -0700
Received: from vesuvius ([192.9.200.11]) by heavens-gate id AA10135g; Thu, 19 Oct 89 17:57:54 PDT
Received: by vesuvius id AA02415g; Thu, 19 Oct 89 17:59:53 PDT
Date: Thu, 19 Oct 89 17:59:53 PDT
From: Carol Sexton <carol@lucid.com>
Message-Id: <8910200059.AA02415@vesuvius>
To: qlisp@go4.stanford.edu
Subject: new new-qlisp
I'm about to install a new new-qlisp which
has a bug fix for a race condition involving
gc and deferred interrupts. Also execute-for-effect now
returns its argument instead of returning either t or nil.
Carol
∂19-Oct-89 2015 shoham@hudson.Stanford.EDU Agent-Oriented Programming
Received: from hudson.Stanford.EDU by SAIL.Stanford.EDU with TCP; 19 Oct 89 20:15:11 PDT
Received: by hudson.Stanford.EDU (5.61/25-eef) id AA02974; Thu, 19 Oct 89 20:16:14 -0700
Date: Thu, 19 Oct 1989 20:16:12 PDT
From: Yoav Shoham <shoham@hudson.Stanford.EDU>
To: jmc@sail
Subject: Agent-Oriented Programming
Message-Id: <CMM.0.88.624856572.shoham@hudson.Stanford.EDU>
John,
I saw the notice of your upcoming CSLI talk. I was struck by the similarity
to something I and a couple of students are doing, which I've been
calling Agent-Oriented Programming. I'll be talking about it on Thu, Nov 3,
at 2:15 at CSLI (the rational agents seminar). I'm busy writing up the
concept right now. I hope to have something ready by the end of next week.
Below is the intro section. I'd be delighted to get anything you have written
up, and your reaction.
Yoav
\title{Agent-Oriented Programming \\
machines, environments and formal agents}
\section{Introduction}
This manuscript touches on issues that are subject of much current
research, issues that include the relationship between a machine and its
environment, and the notion of agenthood. Many of the ideas here intersect
and interact with ideas others have had. For the sake of continuity, however,
I will delay placing this work in the context of other work until at the end.
The term `agents' is used a lot these days. This is true in AI, but also
outside it, for example
in connection with data bases and manufacturing automation.
Although very popular, the term has been
used in such diverse ways that it has become almost meaningless without
reference to a particular notion of agenthood. Some notions
are primarily intuitive, others quite formal. Some are very
austere, defining an agent essentially as a Turing-like machine, and others
ascribe to agents sensory-motor and natural language capabilities.
In this document we propose viewing `formal agents' as formal versions
of human agents,
possessing formal versions of knowledge and beliefs, desires
and goals, capabilities, and so on.
The result is a computational framework which we
call {\it agent-oriented programming (AOP)}.
The name is not accidental, as AOP can be viewed as an extension of the
{\it object-oriented programming (OOP)\/} paradigm.
We mean the latter in the spirit of
Hewitt's original Actors formalism, rather than in the more technological
sense in which it is used nowadays. Intuitively, whereas OOP proposes viewing
a computational system as made up of modules that are able to communicate
with one another and which have individual ways of handling in-coming
messages, AOP expands the picture by allowing the modules to
possess knowledge and beliefs about one another, to have
goals and desires, and possibly other similar notions.
This is one perspective on AOP, as a generalization of OOP. Another slightly
more formal perspective is as a generalization of epistemic logics, which
have been
used a fair amount in AI and distributed computation in recent years.
These logics describe the behavior of machines in terms of
notions such as knowledge and belief. These mentalistic-sounding notions are
actually given very precise computational meanings,
and are used not only to prove properties of distributed
systems, but to design them as well. A typical rule in such a
`knowledge-based' system is ``if processor A does not \underline{know}
that processor
B has received his message, then processor A will not send another message.''
AOP expands these logics by augmenting
them with formal notions of goals, desires, capabilities, and possibly others.
A typical rule in the resulting framework would be ``if agent A
\underline{knows} that agent B \underline{intends}
to do something agent A
does not \underline{want} done, A will \underline{request}
that B change his intention.'' In addition, temporal information is included
to anchor knowledge, desire and so on in particular points in time.
`Intensional' terms such as knowledge, beliefs, goals and so on are used
in a curious sense in the formal AI community.
On the one hand, the definitions (e.g., of knowledge)
come nowhere close
to capturing the full linguistic meanings. On the other hand,
the intuitions about these formal notions do indeed derive from the
everyday, commonsense meanings of the words. What is curious is that, despite
the disparity, the everyday intuition has proven a good guide to
employing the formal notions, when done in moderation.
We will return to OOP and logics of intensionality. For a moment, though,
let us forget about the concept of agenthood altogether, and concentrate
instead on models of computation.
∂19-Oct-89 2140 shoham@hudson.Stanford.EDU Re: Elephant
Received: from hudson.Stanford.EDU by SAIL.Stanford.EDU with TCP; 19 Oct 89 21:40:09 PDT
Received: by hudson.Stanford.EDU (5.61/25-eef) id AA03172; Thu, 19 Oct 89 21:41:13 -0700
Date: Thu, 19 Oct 1989 21:41:11 PDT
From: Yoav Shoham <shoham@hudson.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: Re: Elephant
In-Reply-To: Your message of 19 Oct 89 2027 PDT
Message-Id: <CMM.0.88.624861671.shoham@hudson.Stanford.EDU>
I'm at home, so tomorrow will be fine.
Thanks
Yoav
∂19-Oct-89 2154 mt@media-lab.media.mit.edu Elephant 2000
Received: from media-lab (media-lab.media.mit.edu) by SAIL.Stanford.EDU with TCP; 19 Oct 89 21:54:15 PDT
Received: by media-lab (5.57/4.8) id AA01034; Fri, 20 Oct 89 00:55:19 EDT
Date: Fri, 20 Oct 89 00:52 EDT
From: Michael Travers <mt@media-lab.media.mit.edu>
Subject: Elephant 2000
To: jmc@sail.stanford.edu
Message-Id: <19891020045207.2.MT@OUROBOROS.MEDIA.MIT.EDU>
I saw an announcement for a talk you are giving on "Elephant 2000". I
like the name, and the description of the language also sounds
fascinating. I'm presently working on a survey paper about different
ideas about the relation of knowledge and action in AI, and this work
sounds very relevant. Since I can't attend, I wonder if you have
anything written on this language and if you would be so kind as to send
me a copy? Thanks very much,
Michael Travers
MIT Media Lab E15-494
20 Ames St
Cambridge MA 02139
∂19-Oct-89 2223 mt@media-lab.media.mit.edu re: Elephant 2000
Received: from media-lab (media-lab.media.mit.edu) by SAIL.Stanford.EDU with TCP; 19 Oct 89 22:23:32 PDT
Received: by media-lab (5.57/4.8) id AA01342; Fri, 20 Oct 89 01:24:38 EDT
Date: Fri, 20 Oct 89 01:21 EDT
From: Michael Travers <mt@media-lab.media.mit.edu>
Subject: re: Elephant 2000
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <bncNf@SAIL.Stanford.EDU>
Message-Id: <19891020052127.6.MT@OUROBOROS.MEDIA.MIT.EDU>
OK, thanks. The paper I'm working on is for my PhD general exams, and
not presently destined for publication, if that makes a difference.
∂20-Oct-89 0700 ito@ito.ecei.tohoku.ac.jp R
Received: from RELAY.CS.NET by SAIL.Stanford.EDU with TCP; 20 Oct 89 07:00:16 PDT
Received: from [192.41.197.3] by RELAY.CS.NET id aa17816; 20 Oct 89 9:01 EDT
Received: from ccut.cc.u-tokyo.ac.jp by relay.cc.u-tokyo.ac.jp (5.61/2.3W-relay)
id AA21793; Fri, 20 Oct 89 23:00:05 +0900
Received: by ccut.cc.u-tokyo.ac.jp (5.61/6.4J.6-ut1.67)
id AA29903; Fri, 20 Oct 89 22:59:44 +0900
Received: from maia.ito.ecei.tohoku.ac.jp by hirose.cc.tohoku.ac.jp (3.2/6.4J.6-Oct17)
id AA01268; Fri, 20 Oct 89 18:47:08 JST
Received: by maia.ito.ecei.tohoku.ac.jp (4.0/6.4J.6)
id AA01147; Fri, 20 Oct 89 18:45:39 JST
Received: by aoba.ito.ecei.tohoku.ac.jp (4.0/6.4J.6)
id AA04214; Fri, 20 Oct 89 18:45:36 JST
Received: by ito.ito.ecei.tohoku.ac.jp (3.2/6.4J.6)
id AA24966; Fri, 20 Oct 89 18:27:36+0900
Date: Fri, 20 Oct 89 18:27:36+0900
From: Takayasu ITO <ito@ito.ecei.tohoku.ac.jp>
Return-Path: <ito@ito.ecei.tohoku.ac.jp>
Message-Id: <8910200827.AA24966@ito.ito.ecei.tohoku.ac.jp>
To: JMC%SAIL.Stanford.EDU%relay.cs.net@u-tokyo.ac.jp
Subject: R
Thanks for your mail on earthquake and for informing me that you and your familyare fine.
I am going to visit Stanford from October 30,31 and November 1.
I will try to call you or your wife at office/home during my stay at Palo Alto.
Best regards,
Takayasu ITO
∂20-Oct-89 0734 darden@cs.UMD.EDU ok?
Received: from mimsy.UMD.EDU by SAIL.Stanford.EDU with TCP; 20 Oct 89 07:34:46 PDT
Received: by mimsy.UMD.EDU (5.61/UMIACS-0.9/04-05-88)
id AA18652; Fri, 20 Oct 89 10:33:06 -0400
Date: Fri, 20 Oct 89 10:33:06 -0400
From: darden@cs.UMD.EDU (Prof. Lindley Darden)
Message-Id: <8910201433.AA18652@mimsy.UMD.EDU>
To: jmc@sail.stanford.edu
Subject: ok?
John,
How are you? Where were you when the earthquake
occurred? Is everything at Stanford ok?
sorry I wan't able to see you when you were in
Washington. My class and I had a lively dinner
at a Mexican restuarant. The group is an uneasy
alliance between historians of science, who want
the detailed particulars on each person and event,
and the philosophers, who are searching for general
patterns. One of the more amusing remarks of the
evening was a philosopher, who said he was not interested
in historical details, but studien "meta-anything."
Take care,
Lindley
∂20-Oct-89 1036 bledsoe@cs.utexas.edu re: Sorry I missed you
Received: from cs.utexas.edu by SAIL.Stanford.EDU with TCP; 20 Oct 89 10:35:56 PDT
Received: from alfalfa.cs.utexas.edu by cs.utexas.edu (5.59/1.43)
id AA00411; Fri, 20 Oct 89 12:37:03 CDT
Posted-Date: Fri, 20 Oct 1989 12:36:42 CDT
Received: by alfalfa.cs.utexas.edu (5.59/1.4-Client)
id AA02464; Fri, 20 Oct 89 12:36:43 CDT
Date: Fri, 20 Oct 1989 12:36:42 CDT
From: Woody Bledsoe <bledsoe@cs.utexas.edu>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: Sorry I missed you
In-Reply-To: Your message of 18 Oct 89 1304 PDT
Message-Id: <CMM.0.86.624908202.bledsoe@alfalfa.cs.utexas.edu>
John,
Maybe we can have dinner the evening of Nov 16, if you will be free then.
Or lunch Nov 16 (Thursday). I could also have lunch Friday but my class
ends at 12:00. We could invite Bob Boyer, Doug Lenat, etc if you like.
Woody
∂20-Oct-89 1053 lesperan@neat.cs.toronto.edu Elephant 2000
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 20 Oct 89 10:53:06 PDT
Received: by neat.cs.toronto.edu id 3340; Fri, 20 Oct 89 13:53:47 EDT
From: Yves Lesperance <lesperan@ai.toronto.edu>
To: jmc@sail.stanford.edu
Subject: Elephant 2000
Message-Id: <89Oct20.135347edt.3340@neat.cs.toronto.edu>
Date: Fri, 20 Oct 89 13:53:44 EDT
Hi! I saw an abstract about Elephant 2000 in the CSLI Calendar.
If you could send me a paper or reference on this work, I would
really appreciate. Thanks.
Yves Lesperance
Dept. of Computer Science
University of Toronto
Toronto, Ont. CANADA M5S 1A4
∂20-Oct-89 1320 nilsson@Tenaya.Stanford.EDU winter quarter course
Received: from Tenaya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 20 Oct 89 13:20:07 PDT
Received: by Tenaya.Stanford.EDU (NeXT-0.8/25-eef) id AA07585; Fri, 20 Oct 89 13:17:57 PDT
Date: Fri, 20 Oct 89 13:17:57 PDT
From: Nils Nilsson <nilsson@tenaya.stanford.edu>
Message-Id: <8910202017.AA07585@Tenaya.Stanford.EDU>
To: jmc@sail.stanford.edu, val@sail.stanford.edu, ginsberg@polya.stanford.edu
Subject: winter quarter course
Cc: nilsson@Tenaya.Stanford.EDU, floyd@polya.stanford.edu,
jones@polya.stanford.edu
Folks,
Kurt Konolige asked me recently about teaching the
following course. I told him to get in touch with our
curriculum chair, Bob Floyd. Perhaps Bob is already
asking you people for your opinions because the
course may overlap courses that you are teaching
or plan to teach. Any responses should be directed
to Bob Floyd.
-Nils
-----
Begin forwarded message:
Date: Tue, 17 Oct 89 14:41:55 PDT
From: Kurt Konolige <konolige@ai.sri.com>
To: floyd@polya.stanford.edu
Subject: winter quarter course
Cc: jones@polya.stanford.edu, nilsson@polya.stanford.edu
Bob, I was thinking of teaching a course in Nonmonotonic Logic
this coming winter, and talked the idea over with Nils. He told
me to send a course description to you, and hopefully it wouldn't
be too late to schedule it. It's meant to be a grad seminar.
I've appended the course description.
--kk
-----------------------
Nonmonotonic Logic: Theory and Applications
This course is a unified approach to the current nonmonotonic logic
subfield in AI. The concentration will be on the formalization of
defeasible reasoning in autoepistemic logic. Characterization of
other logical systems, such as default logic, circumscription, and
truth maintenance systems, will be presented within this framework.
Applications to be discussed include inheritance nets with exceptions,
reasoning about action and change, and reasoning about the beliefs and
intentions of agents.
3 units.
Requirements: Competence in first-order logic. No previous experience
with nonmonotonic or modal logics is required.
Instructor: Kurt Konolige
Center for the Study of Language and Information
SRI International
∂20-Oct-89 1321 @IU.AI.SRI.COM,@sunset.ai.sri.com:felix@ai.sri.com re: Slingshot effect... (how does it work?)
Received: from IU.AI.SRI.COM by SAIL.Stanford.EDU with TCP; 20 Oct 89 13:21:02 PDT
Received: from sunset.ai.sri.com by IU.AI.SRI.COM via SMTP with TCP;
Thu, 19 Oct 89 23:09:38-PST
Received: from morro.ai.sri.com by sunset.ai.sri.com (4.0/4.16) id
AA19699 for JMC@SAIL.Stanford.EDU; Thu, 19 Oct 89 23:10:13 PDT
Received: by morro.ai.sri.com (4.0/4.16) id AA00477 for
JMC@SAIL.Stanford.EDU; Thu, 19 Oct 89 23:10:12 PDT
Date: Thu, 19 Oct 89 23:10:12 PDT
From: Francois Felix INGRAND <felix@ai.sri.com>
Message-Id: <8910200610.AA00477@morro.ai.sri.com>
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <bnc8S@SAIL.Stanford.EDU>
Subject: re: Slingshot effect... (how does it work?)
Organization: SRI International, Menlo Park, CA
Phone: (415) 859 5584
John McCarthy writes:
>[In reply to message sent 17 Oct 89 20:59:28 GMT.]
>
>Is "Read my Lisp... No new syntax" yours? If not, whose?
>
I think it is mine. I got the idea during the campaign last year after
reading the signature of a french friend which was saying something
like:
"Carmen lips are more exciting than common lisp" Bizet
Felix
∂20-Oct-89 1345 @Score.Stanford.EDU:james@cs.rochester.edu elephants
Received: from Score.Stanford.EDU by SAIL.Stanford.EDU with TCP; 20 Oct 89 13:45:28 PDT
Received: from cayuga.cs.rochester.edu by SCORE.STANFORD.EDU with TCP; Fri 20 Oct 89 13:04:11-PDT
Received: from slate.cs.rochester.edu by cayuga.cs.rochester.edu (5.59/o) id AA17023; Fri, 20 Oct 89 14:14:25 EDT
Received: by slate.cs.rochester.edu (4.0/o) id AA00997; Fri, 20 Oct 89 14:17:53 EDT
Date: Fri, 20 Oct 89 14:17:53 EDT
From: james@cs.rochester.edu
Message-Id: <8910201817.AA00997@slate.cs.rochester.edu>
To: mccarthy@score.stanford.edu
Subject: elephants
John
I read with great interest the description of your
TINlunch talk next week. I've been working on some similar
issues recently and would be very interested to get a copy
of a paper, or draft, or whatever. Do you have something
you can send me??
James Allen
Dept Computer Science
Univ Rochester
Rochester, NY 14627
∂20-Oct-89 1421 james@cs.rochester.edu re: elephants
Received: from cayuga.cs.rochester.edu by SAIL.Stanford.EDU with TCP; 20 Oct 89 14:21:08 PDT
Received: from slate.cs.rochester.edu by cayuga.cs.rochester.edu (5.59/o) id AA19178; Fri, 20 Oct 89 17:18:36 EDT
Received: by slate.cs.rochester.edu (4.0/o) id AA01291; Fri, 20 Oct 89 17:22:03 EDT
Date: Fri, 20 Oct 89 17:22:03 EDT
From: james@cs.rochester.edu
Message-Id: <8910202122.AA01291@slate.cs.rochester.edu>
To: JMC@SAIL.Stanford.EDU
Subject: re: elephants
thanks
james
∂20-Oct-89 1441 ginsberg@Polya.Stanford.EDU Re: winter quarter course
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 20 Oct 89 14:41:09 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA18420; Fri, 20 Oct 89 14:41:48 -0700
Date: Fri, 20 Oct 89 14:41:48 -0700
From: Matthew L. Ginsberg <ginsberg@Polya.Stanford.EDU>
Message-Id: <8910202141.AA18420@Polya.Stanford.EDU>
To: ginsberg@polya.stanford.edu, jmc@sail.stanford.edu, val@sail.stanford.edu
Subject: Re: winter quarter course
Cc: floyd@polya.stanford.edu, nilsson@Tenaya.Stanford.EDU
Konolige's proposed course sounds very much like a survey of his
own work; I am in general opposed to such stuff. A course in nonmon
should cover all of the major formalisms and applications, as opposed
to one person's views.
Matt
∂20-Oct-89 1518 Mailer re: Slingshot effect... (how does it work?)
Received: from IU.AI.SRI.COM by SAIL.Stanford.EDU with TCP; 20 Oct 89 15:18:22 PDT
Received: from sunset.ai.sri.com by IU.AI.SRI.COM via SMTP with TCP;
Fri, 20 Oct 89 15:17:19-PST
Received: by sunset.ai.sri.com (4.0/4.16) id AA14427 for
su-etc@SAIL.Stanford.EDU; Fri, 20 Oct 89 15:17:41 PDT
Date: Fri, 20 Oct 89 15:17:41 PDT
From: Francois Felix INGRAND <felix@ai.sri.com>
Message-Id: <8910202217.AA14427@sunset.ai.sri.com>
To: JMC@SAIL.Stanford.EDU
Cc: felix@Warbucks.ai.sri.com, su-etc@SAIL.Stanford.EDU
In-Reply-To: <jnc49@SAIL.Stanford.EDU>
Subject: re: Slingshot effect... (how does it work?)
Organization: SRI International, Menlo Park, CA
Phone: (415) 859 5584
John McCarthy writes:
>[In reply to message from felix@AI.SRI.COM sent 17 Oct 89 20:59:28 GMT.]
>
>The spacecraft really can get a boost in velocity relative to the sun
>and not just a change in direction. If one takes a co-ordinate system
>in which the sun is at rest, then one can say that the planet contributes
>part of its energy to the spacecraft.
>
>The simplest case to compute is one in which the planet has velocity
>v1 in orbit and the spacecraft has velocity v2 in the same orbit
>but in the opposite direction, i.e. just missing a head on collision.
>If the spacecraft could come close enough to the planet, it could
>have its velocity relative to the planet reversed. This is
>approximately equivalent to saying that the spacecraft has an
>elastic collision with the planet. In that case the velocity
>of the spacecraft relative to the planet is initially v1+v2,
>and after the collision is -(v1+v2). This means that its velocity
>relative to the sun is now 2v1+v2 in the opposite direction, i.e.
>its velocity is increased by twice the orbital velocity of the
>planet, just as the velocity of a ping-pong ball leaving the
>a paddle is reversed and increased by twice the velocity of the
>paddle.
John,
I am not convince at all by your explanation. They do not play ping
pong with billion dollars satellite or spacecraft. At no time you have
contact (even with the atmosphere) and thus you cannot apply the
"conservation de la quantite de mouvement (in french)" rule.
Anyway, this morning I found a very convincing example (at least which
satisfy my intuition).
In a sun reference frame, imagine an immobile object in the plan of
the earth orbit, almost on the orbit itself, but a little bit
further (to avoid collision).
It appears to me quite obviously that when the earth will passe by the
object (which we will considered much lighter than the earth), the
object will be subject to the gravitation force and thus will be
accelerated.
My point is that there is no way that this object will still stand
still after the earth has passed nearby him (even if it fell on earth
it is moving...)
That's it, this prove that we have given some kinetic energy to the
object. Now, this can indeed be applied to boost satellite.
Felix
∂20-Oct-89 1529 VAL re: winter quarter course
To: ginsberg@Polya.Stanford.EDU
CC: JMC@SAIL.Stanford.EDU, floyd@Polya.Stanford.EDU,
nilsson@TENAYA.STANFORD.EDU
[In reply to message from ginsberg@Polya.Stanford.EDU sent Fri, 20 Oct 89 14:41:48 -0700.]
> A course in nonmon
> should cover all of the major formalisms and applications, as opposed
> to one person's views.
CS323, that JMC will be teaching, is supposed to cover all major formalisms, but,
I expect, it will be mainly about circumscription. Maybe it's not a bad idea to
offer in addition another course, that is supposed to cover all major formalisms
also, but will be mainly about a.e. logic...
--Vladimir
∂20-Oct-89 1554 bledsoe@cs.utexas.edu re: Sorry I missed you
Received: from cs.utexas.edu by SAIL.Stanford.EDU with TCP; 20 Oct 89 15:54:29 PDT
Received: from alfalfa.cs.utexas.edu by cs.utexas.edu (5.59/1.43)
id AA02034; Fri, 20 Oct 89 17:55:36 CDT
Posted-Date: Fri, 20 Oct 1989 17:55:15 CDT
Received: by alfalfa.cs.utexas.edu (5.59/1.4-Client)
id AA02629; Fri, 20 Oct 89 17:55:17 CDT
Date: Fri, 20 Oct 1989 17:55:15 CDT
From: Woody Bledsoe <bledsoe@cs.utexas.edu>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: Sorry I missed you
In-Reply-To: Your message of 20 Oct 89 1524 PDT
Message-Id: <CMM.0.86.624927315.bledsoe@alfalfa.cs.utexas.edu>
Good. I will call you at MCC that day to arrange the details. Or feel
free to call me.
∂20-Oct-89 1715 Q4034%PUCC.BITNET@Forsythe.Stanford.EDU Tidal perturbations
Received: from Forsythe.Stanford.EDU by SAIL.Stanford.EDU with TCP; 20 Oct 89 17:14:58 PDT
Received: by Forsythe.Stanford.EDU; Fri, 20 Oct 89 17:15:36 PDT
Received: by PUCC (Mailer R2.04) id 1076; Fri, 20 Oct 89 20:14:52 EDT
Date: Fri, 20 Oct 89 19:57:12 EDT
From: John Nash <Q4034%PUCC.BITNET@Forsythe.Stanford.EDU>
Subject: Tidal perturbations
To: John McCarthy <jmc@SAIL.Stanford.EDU>
Dear McCarthy,
Moonkwake!!
I think that the recent "Loma Prieta" earthquake of about 5 P.M.
Pacific Time on last Tuesday is possibly a good example for the
theory of the association of earthquakes with times of high tide.
On T.V. there had been repeated news stories last week-end about
how the unusual astronomical line-up of the moon and the sun, both
acting to induce tides on the earth, was feared to cause great co-
astal damage on the South Carolina coast in the wake of hurricane
Hugo. The super-high tides were expected for last Monday.
But what happened was that you got a quake in California. And
also the chinese got a quake west of Beijing.
Well, of course the tidal effects of the astronomical allign-
ment operate quite globally to affect circumstances on the earth.
This is in contrast to the hurricanes and the meteorological lows
which have more of a local character in relation to the earth's
surface. I wonder what you think about these opinions.
J. Nash, Q4034
Q4034@PUCC.Princeton.EDU
∂21-Oct-89 0833 CLT driving lesson
Please try to give Hazel a lesson today.
∂21-Oct-89 1638 harnad@clarity.Princeton.EDU Sci. Am., you and Searle
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 21 Oct 89 16:37:48 PDT
Received: from reason.Princeton.EDU by Princeton.EDU (5.58+++/2.22/mailrelay)
id AA21355; Sat, 21 Oct 89 19:30:46 EDT
Received: by reason.Princeton.EDU (4.0/1.87)
id AA02757; Sat, 21 Oct 89 19:26:57 EDT
Date: Sat, 21 Oct 89 19:26:57 EDT
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8910212326.AA02757@reason.Princeton.EDU>
To: pschurchland@ucsd.edu
Subject: Sci. Am., you and Searle
To: P & P Churchland
From: Stevan Harnad
THINKING THE UNTHINKABLE, OR, RUN THAT BY ME AGAIN, FASTER
Hi Pattie and Paul:
Thanks for sending me your Scientific American draft. I've seen
Searle's companion draft too. Here are a few comments:
(1) Unfortunately, in suggesting that Searle is the one who is begging
the question or assuming the truth of the hypothesis that syntax alone
can consititute semantics you seem to have the logic reversed: In fact,
Searle's the one who's TESTING that hypothesis and answering that question;
and the Chinese-Room thought-experiment shows that the hypothesis fails
the test and the answer is no! It is the proponents of the "systems
reply" -- which merely amounts to a reiteration of the hypothesis in
the face of Searle's negative evidence -- who are begging the question.
By the way, in endorsing the systems reply in principle, as you do
(apparently only because of its counterintuitiveness, and the fact that
other counterintuitive things have, in the history of science, turned
out to be correct after all), you leave out Searle's very apt RESPONSE
to the counterintuitive idea that the "system" consisting of him plus
the room and its contents might still be understanding even if he
himself is not: He memorizes the rules, and henceforth he IS all there
is to the system, yet still he doesn't understand Chinese. (And I hope
you won't rejoin with the naive hackers' multiple-personality gambit at
this point, which is CLEARLY wanting to save the original hypothesis at
any counterfactual price: There is no reason whatsoever to believe that
simply memorizing a bunch of symbols and symbol manipulation rules and
then executing them is one of the etiologies of multiple personality
disorder!)
As to the speed factor: Yes, that is one last holdout, if it is in
fact true that Searle could never pass the Chinese TT in real time. But
that's at the price of being prepared to believe that the difference
between having and not having a mind is purely a function of speed!
The phenomenon of phase transitions in physics notwithstanding, that
sounds like a fast one, too fast for me to swallow, at any rate.
Besides, once he's memorized the rules (PLEASE don't parallel the speed
argument with a capacity argument too!), it's not clear that Searle
could not manage a good bit of the symbol manipulation in real time
anyway. The whole point of this exercise, after all, is to show that
thinking can't be just symbol manipulation -- at ANY speed.
I don't know about you, but I've never been at all happy with
naive hackers' claims that all there is to mind is the usual stuff, but
(1) faster, (2) bigger, and (3) more "complex." I think the real
punchline's going to turn out to be a good bit more substantive than
this hand-waving about just (a lot) more of the same...
(2) All your examples about the groundlessness of prior skepticism
in the face of physical theories of sound, light and life were (perhaps
unknowingly) parasitic on subjectivity. Only now, in mind-modeling, is
the same old problem finally being confronted on its home turf. But All
prior bets are off, since those were all away-games. The buck, as Tom
Nagel notes, stops with qualia. I'll show this specifically with the
example below.
(3) It is ironic that your example of light = oscillating
electromagnetic radiation should also hinge on speed (frequency). You
say that Searle, in a parallel "simulation," would be waving the magnet
much too slowly, and would then unjustly proclaim "Luminous room, my
foot, Mr. Maxwell. It's pitch black in here!" But here you see how all
these supposedly analogous forms of skepticism are actually parasitic on
subjectivity (with shades of Locke's primary and secondary qualities):
Because of course the only thing missing is the VISIBILITY of light at
the slow frequency. It made perfect sense, and could have been pointed
out all along, that, if fast electromagnetic oscillations really are
light, then it might only be visible to the eye in some of its
frequency ranges, and invisible but detectable by other instruments in
other frequency ranges.
That story is perfectly tenable, and in no way analogous to Searle's
Argument, because it is objective: It's not "what it's like to see
light" (a subjective, "secondary" quality) that the Maxwellian
equation of light with EM radiation is trying to explain, it's the
objective physical property that, among other things, happens to be the
normal cause of the subjective quality of seeing light. The mistake
the sceptics were making is clearly NOT the same as Searle's. They
were denying an objective-to-objective equation: One set of objective
physical properties (among them the power to cause us to see light) was
simply being shown to be the same as another set of objective physical
properties. No one was trying to equate THE SUBJECTIVE QUALITY OF LIGHT
ITSELF with something objective. (Not until lately, that is.)
So, whereas concerns about subjectivity might indeed have been the
source of the earlier scepticism, all that scepticism was simply
misplaced. It was much ado about nothing. Ditto with sound and life:
Subjectivity, though lurking in each case, was really never at issue.
As Nagel puts it, one set of appearances was simply being replaced by
(or eliminated in favor of) another, in the new view, however
surprising the new appearances might have appeared. But no one was
really trying to replace APPEARANCES THEMSELVES by something else, by
the stuff of which all appearances would then allegedly be made: THAT
would have been a harder nut to crack.
But that's the one modern mind-modeling is up against, and Nagel is
right that this is another ball game altogether (my "home-game" analogy
was an understatement -- and the metaphors are as mixed as nuts by
now...). So no analogies carry over. It's not that the rules have
changed. It's just that nothing remotely like this has ever come up
before. So, in particular, you are NOT entitled to help yourself to the
speed analogy in trying to refute Searle's Chinese Room Argument.
Because whereas it would have been no problem at all for Maxwell to
"bite the bullet" and claim that very slow EM oscillation was still
light, only it wasn't visible, one CANNOT say that very slow
symbol-manipulation is still thinking only it's... what?
"Unthinkable?" You took the words out of my mouth.
(4) Your point about the immunity of parallel processes to the Chinese
Room Argument (unlike similar points about speed, capacity or
complexity) has somewhat more prima facie force because it really is
based on something Searle can't take over all by himself, the way he
could with symbol manipulation. On the face of it, Searle couldn't BE
the parallel system that was passing the TT in Chinese in the same way
he could BE the serial symbol system, so he could not take the next
step and show that he would not be understanding Chinese if he were
(and hence that neither would the system he was duplicating).
This is why I suggested to Searle that his "Chinese Gym" Argument fails
to have the force of his original Chinese Room Argument, and is indeed
vulnerable to a "systems reply." It's also why I suggested the
"three-room" argument to Searle (which he has now expanded to four
rooms in his Scientific American piece, somewhat unnecessarily, in my
view, to accommodate the Chinese Gym too), which is completely in the
spirit of the original Chinese Room Argument and puts the burden of
evidence or argument on the essential parallelist, where it belongs.
Here is the critical excerpt from my comments on an earlier draft by
Searle:
> So I respectfully recommend that you jettison the Chinese Gym Argument
> and instead deal with connectionism by turning the Chinese Room
> Argument on its head, as follows. Suppose there are three rooms:
>
> (1) In one there is a real Net (implemented as physical units, with
> real physical links, real excitatory/inhibitory interconnections
> real parallel distributed processing, real backpropping, etc.) that
> could pass the Turing Test in Chinese (Chinese symbols in, Chinese
> symbols out).
>
> (2) In the second there is a computer simulation of (1) that likewise
> passes the TT in Chinese.
>
> (3) In the third is Searle, performing ALL the functions of (2),
> likewise passing the Chinese TT (while still not understanding, of
> course).
>
> Now the connectionists have only two choices:
>
> Either they must claim that all three understand Chinese (in which case
> they are back up against the old Chinese Room Argument), or the
> essentialists among them will have to claim that (1) understands but (2)
> and (3) do not -- but without being able to give any functional reason
> whatsoever why.
So this is what parallelism is up against. I also went on to query the
Connectionists on this, as follows (and received multiple replies, most
along the lines of the 1st two, which I include below):
> From: Stevan Harnad
> To: connectionists@cs.cmu.edu
> Subject: Parallelism, Real vs. Simulated: A Query
>
> "I have a simple question: What capabilities of PDP systems do and
> do not depend on the net's actually being implemented in parallel,
> rather than just being serially simulated? Is it only speed and
> capacity parameters, or something more?"
> ----------------------------------------------------------------
>
> (1) From: skrzypek@CS.UCLA.EDU (Dr. Josef Skrzypek)
> Cc: connectionists@cs.cmu.edu
>
> Good (and dangerous) question. Applicable to Neural Nets in general
> and not only to PDP.
>
> It appears that you can simulate anything that you wish. In principle
> you trade computation in space for computation in time. If you can
> make your time-slices small enough and complete all of the necessary
> computation within each slice there seem to be no reason to have
> neural networks. In reality, simulation of synchronized, temporal
> events taking place in a 3D network that allows for feedback pathways
> is rather cumbersome.
>
> (2) From: Michael Witbrock <mjw@cs.cmu.edu>
>
> I believe that none of the properties depend on parallel implementation.
>
> There is a proof of the formal equivalence of continuous and discrete
> finite state automata, which I believe could be transformed to prove the
> formal equivalence of parallel and serially simulated pdp models.
Except for some equivocal stuff on "asynchronous" vs "synchronous"
processes, about which some claimed one thing and others claimed the
opposite, most respondents agreed that the parallel and serial
implementations were equivalent. Hence it is not true, as you write,
that parallel systems are "not threatened by [Searle's] Chinese Room
argument." They are, although someone may still come up with a
plausible reason why, although the computational difference is
nonexistent, the implementational difference is an essential one.
And that may, logically speaking, turn out to be (one of the)
answer(s) to the question of which of the "causal powers" of the brain
are actually relevant (and necessary/sufficient) for producing a mind.
I think Searle's Argument (and my Symbol Grounding Problem) have
effectively put pure symbol manipulation out of contention. I don't
think "the same, only faster, bigger, or more complex" holds much hope
either. And parallelism stands a chance only if someone can show what
it is about its implementation in that form, rather than in fast serial
symbolic form, is critical. My own favored candidate for the "relevant"
property, however, namely, sensory grounding, and sensory transduction
in particular, has the virtue of not only being, like parallelism,
invulnerable to the Chinese Argument (as I showed in "Minds, Machines
and Searle"), but also being a natural candidate for a solution to the
Symbol Grounding Problem, thereby, unlike paralellism, wearing the
reason WHY it's critical on its sleeve, so to speak.
(5) Finally, you write "We, and Searle, reject the Turing Test as a
sufficient condition for conscious intelligence." In this I must disagree
with both (or rather, all three) of you: The logic goes like this. So
far, only pure symbol crunching has been disqualified as a candidate
for being the sufficient condition for having a mind. But don't forget
that it was only a conjecture (and in my mind always a counterfactual
one) that the standard (language-only) Turing Test (only symbols in,
and symbols out), the TT, could be successfully passed by a pure symbol
cruncher. Searle's argument shows that IF the TT could be passed by
symbol crunching alone, THEN, because of the Chinese Room Argument, it
would not have a mind, and hence the TT is to be rejected.
Another possibility remains, however, which is that it is impossible to
successfully pass the TT with symbol crunching alone. The truth may
instead be that any candidate that could pass the TT would already have
to have and draw upon the causal power to pass the TTT, the Total
Turing Test, which includes all of our robotic, sensorimotor capacities
in the real world of objects. Now the TTT necessarily depends on
transduction, which is naturally and transparently immune to Searle's
Chinese Room Argument. Hence there is no reason to reject the TTT
(indeed, I would argue, there's no alternative to the TTT, which,
perhaps expanded to include neural function -- the "TTTT"? -- is simply
equivalent to empiricism!). And if a necessary condition for passing the
TT is the causal power to pass the TTT, then there's really no reason
left for rejecting the TT either.
Best wishes,
Stevan Harnad
∂21-Oct-89 2242 harnad@clarity.Princeton.EDU Two Ways to Refute Searle
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 21 Oct 89 22:42:32 PDT
Received: from cognito.Princeton.EDU by Princeton.EDU (5.58+++/2.22/mailrelay)
id AA25810; Sun, 22 Oct 89 01:35:45 EDT
Received: by cognito.Princeton.EDU (4.0/1.87)
id AA10501; Sun, 22 Oct 89 01:35:34 EDT
Date: Sun, 22 Oct 89 01:35:34 EDT
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8910220535.AA10501@cognito.Princeton.EDU>
To: harnad@elbereth.rutgers.edu
Subject: Two Ways to Refute Searle
[Recipients of the Searle/symbol-grounding discussions: If you want
your name removed from this discussion list, please let me know. -- SH]
TWO WAYS TO REFUTE SEARLE
John McCarthy <JMC@SAIL.Stanford.EDU> writes:
> To people in AI, obtuse philosophers like Searle and Harnad are too
> long winded and obscure to be worth trying to decipher. This is
> especially true, because it seems they aren't trying to raise any
> practical difficulties that would have to be overcome in order to make
> useful systems at any level of performance. They merely say that no
> level of performance would count as intelligent.
It might be a good idea to try to decipher the arguments anyway, just
in case there is something practical there, such as (1) how pure symbol
manipulation cannot be grounded in the objects it can be interpreted as
representing, and hence cannot be the right model for mental activity,
which IS grounded in the objects it represents; and (2) how successful
performance on the TTT would count as intelligent, but because of (1),
pure symbol manipulation will be unable to achieve it.
[Ceterum sentio: I am not a philosopher but a psychologist.]
> If we avoid these distractions for now, there will be more hope of an
> AI program during their lifetimes that will overwhelm their arguments.
> It should be capable of following the most obscure and lengthy
> arguments and refuting them at any desired length, from a paragraph to
> a three volume treatise.
To write a symbol manipulating program that can attempt to counter my
arguments, as AI researchers have attempted to counter them, i.e.,
unsuccessfully, is surely feasible, but would only amount to yet
another arbitrary toy fragment of the TT (not to be confused with the
TTT). I'll settle for either a system that can pass the TTT (which, as
I showed in "Minds, Machines, and Searle," would be automatically
immune to the Chinese Room Argument) -- or just a successful refutation
by a human being.
> Maybe for debugging it, we will also need a program capable
> of generating the arguments.
For that, the arguments will have to be deciphered and understood
first.
> However, perhaps I have them wrong. Do Searle and Harnad claim to have
> discovered any difficulty in achieving any particular behavioral
> performance? If so, what is the simplest thing they imagine we can't
> program computers to do that humans can do?
Arbitrary performance fragments won't do, because we're not just
looking for more clever toys, but the real thing; nothing less than TTT
performance will do the job.
Stevan Harnad
∂22-Oct-89 0043 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM special (mental) cases
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 22 Oct 89 00:43:40 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 438731; 22 Oct 89 03:40:53 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 110653; Sun 22-Oct-89 00:39:00 PDT
Date: Sun, 22 Oct 89 00:40 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: special (mental) cases
To: math-fun@RUSSIAN.SPA.Symbolics.COM
cc: "wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "hul@psuvm.psu.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
Message-ID: <19891022074030.6.RWG@TSUNAMI.SPA.Symbolics.COM>
of "nonlocal derangement". These two are nearly logically
equivalent:
m
==== a
\ n
> -------------- = 0,
/ m
==== /===\ a
n = 0 ! ! n
! ! (1 - --)
! ! a
k = 0 k
k # n
m>=1, and
m m
==== a ====
\ n \
> -------------- = > a .
/ m / n
==== /===\ a ====
n = 0 ! ! k n = 0
! ! (1 - --)
! ! a
k = 0 n
k # n
They might also follow from partial fraction identities in Knuth,
Vol 1, which I don't have handy :-(.
∂22-Oct-89 0158 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM tough nombres
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 22 Oct 89 01:58:26 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 438737; 22 Oct 89 04:55:43 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 110657; Sun 22-Oct-89 01:53:40 PDT
Date: Sun, 22 Oct 89 01:55 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: tough nombres
To: math-fun@RUSSIAN.SPA.Symbolics.COM
cc: "wri-tech@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "rcs@la.tis.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM,
"rwf@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
In-Reply-To: <19890724094644.7.RWG@TSUNAMI.SPA.Symbolics.COM>
Message-ID: <19891022085510.9.RWG@TSUNAMI.SPA.Symbolics.COM>
Date: Mon, 24 Jul 89 02:46 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Responding to an email query, I guessed that the term "Mertens's
constant" denoted
limit sum 1/p - ln ln x ~~ .26152?
x -> inf primes
p < x
based on Hardy & Wright, p 351.
Note the minor miracle that e is the only base of the
outer log for which the limit can be finite.
This constant is a bear to compute. I don't trust anything
past the .26, although for x = prime(2↑n), n = 16, 17, 18,
it's respectively .26157, .2615291, and .2615271.
DEK physmailed me a pointer to how to find .261497212747643...
with stone age calculus (ahem): purple Greene and Knuth, pp 66-7.
Thus it behooves Mathematica to have MertensM and KhinchinK and ...
(be-`hoove: v.t. To waist magically from the ass down.)
∂22-Oct-89 0202 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM email address
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 22 Oct 89 02:02:53 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 438739; 22 Oct 89 05:02:47 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 110659; Sun 22-Oct-89 02:00:58 PDT
Date: Sun, 22 Oct 89 02:02 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: email address
To: "dek@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
Message-ID: <19891022090228.0.RWG@TSUNAMI.SPA.Symbolics.COM>
Most reliable is
Reply-to: rwg@yukon.scrc.symbolics.com
(although now that Petti's quitting...)
∂22-Oct-89 0831 darden@cs.UMD.EDU re: ok?
Received: from mimsy.UMD.EDU by SAIL.Stanford.EDU with TCP; 22 Oct 89 08:31:01 PDT
Received: by mimsy.UMD.EDU (5.61/UMIACS-0.9/04-05-88)
id AA01569; Sun, 22 Oct 89 11:29:19 -0400
Date: Sun, 22 Oct 89 11:29:19 -0400
From: darden@cs.UMD.EDU (Prof. Lindley Darden)
Message-Id: <8910221529.AA01569@mimsy.UMD.EDU>
To: JMC@SAIL.Stanford.EDU, darden@CS.UMD.EDU
Subject: re: ok?
Hi John,
Glad to hear that you are fine.
I'm busy working on an NSF proposal that, if successful, would
fund me to visit at the Ohio State AI Lab in 1990-91 academic year.
I am playing with the idea that anomaly resolution within a scientific
theory (that is, change the theory so the data point is not longer
anomalous) is like diagnostic reasoning (find the failing part and
fix it). So, I want to see if B. Chandrasekaran's Functional Representation
System can be used to represent genetic theory and guide localization
of anomalies. Do you know Chandra? He has a very philosophical mind
for an electrical engineer.
Take care,
Lindley
∂22-Oct-89 1950 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
A CONNOTATIVE TREATMENT OF CIRCUMSCRIPTION
Jun Arima
ICOT Research Center, Japan
Monday, October 23, 3:15pm
MJH 252
Circumscription proposed by McCarthy is one of the most promising
formalizations of nonmonotonic aspects of commonsense reasoning.
It has several versions. However, they are all proposed for the
denotative minimization of predicates, that is, circumscription
minimizes the extension of predicates. Regarding such treatment,
this paper considers three problems: absence of abnormal things, a
limitation on equality and inconsistency of circumscription. This
paper proposes a solution for them by presenting a connotative
treatment of circumscription. This treatment is based on the idea
of circumscribing predicates connotatively, that is, minimizing
the set of names denoting objects which satisfy a certain
predicate. Besides, this treatment provides a form of the unique
names hypothesis for a general database which allows the use of
equality and functions.
∂22-Oct-89 1959 kirsh%cogsci@ucsd.edu re: aij
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 22 Oct 89 19:59:11 PDT
Received: from cogsci.ucsd.edu by ucsd.edu; id AA24022
sendmail 5.61/UCSD-2.0-sun via SMTP
Sun, 22 Oct 89 17:00:39 -0700 for JMC@sail.stanford.edu
Received: from cogsci-sun.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA01162 for JMC@sail.stanford.edu; Sun, 22 Oct 89 17:03:24 PDT
Date: Sun, 22 Oct 89 17:01 PDT
From: David Kirsh <kirsh%cogsci-sun@ucsd.edu>
Subject: re: aij
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <fnvc8@SAIL.Stanford.EDU>
Message-Id: <19891023000124.4.KIRSH@cogsci-sun.ucsd.edu>
John,
I have a brief and highly informal piece from you. But Carl had told me that
he had sent you his newer piece for you to comment properly on, and that you
would have a reply soon. Am I misinformed?
Trust all is well in other matters, your talk on a speech act programming
language certainly looks exciting.
-- David
∂22-Oct-89 2211 @ELEPHANT-BUTTE.SCRC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM An Ethiopian in the fuel supply.
Received: from ELEPHANT-BUTTE.SCRC.Symbolics.COM by SAIL.Stanford.EDU with TCP; 22 Oct 89 22:11:24 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 438814; 23 Oct 89 01:09:55 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 110675; Sun 22-Oct-89 22:08:08 PDT
Date: Sun, 22 Oct 89 22:09 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: An Ethiopian in the fuel supply.
To: "swolf@wri.com"@ELEPHANT-BUTTE.SCRC.Symbolics.COM
cc: rwg@RUSSIAN.SPA.Symbolics.COM
Message-ID: <19891023050937.4.RWG@TSUNAMI.SPA.Symbolics.COM>
bcc: "jmc@sail.stanford.edu"@ELEPHANT-BUTTE.SCRC.Symbolics.COM, mlb@RUSSIAN.SPA.Symbolics.COM
My guess is your Ethiopian inquirer was fronting for the
Soviets. As for dealing with them directly, I am reminded
of the remark that "the Capitalists will sell us the rope ...".
What do they need more than technology? They've got gold
mines and oil wells. Tell them the alternative to hard
cash is involvement with the State Department. Maybe
they'll trade Mathematica for more internal freedoms.
∂23-Oct-89 0844 MPS
∂20-Oct-89 1956 JMC
To: MPS
Is Dantzig dinner at the Faculty Club?
yes it is
∂23-Oct-89 1000 JMC
expenses
∂23-Oct-89 1003 MPS Vote
Hi,
Joyce needs to know how you voted on the Oussama Khatib issue asap.
Pat
∂23-Oct-89 1021 MPS books
Please do not forget to bring in the books due in the library
∂23-Oct-89 1023 MPS lunch
Hi,
I will need to know about how many for lunch Wednesday.
I should order late today.
I have a menu at work you can glance at to see what you think
everyone wuld like.
Pat
∂23-Oct-89 1102 scales@Polya.Stanford.EDU CS 499
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 23 Oct 89 11:02:45 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA06611; Mon, 23 Oct 89 11:02:15 -0700
Date: Mon, 23 Oct 89 11:02:15 -0700
From: Daniel J. Scales <scales@Polya.Stanford.EDU>
Message-Id: <8910231802.AA06611@Polya.Stanford.EDU>
To: jmc@sail
Subject: CS 499
Prof. McCarthy,
Since I am doing an RAship with the QLISP group this quarter, I have signed
up for 3 units of CS 499 with you as the advisor, in order to fill out my 9
units. Is that okay? At the time that I had to submit my study list, you
were away, so I just signed up and planned to ask you when you got back.
Dan Scales
∂23-Oct-89 1107 scales@Polya.Stanford.EDU re: CS 499
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 23 Oct 89 11:07:07 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA07060; Mon, 23 Oct 89 11:06:45 -0700
Date: Mon, 23 Oct 89 11:06:45 -0700
From: Daniel J. Scales <scales@Polya.Stanford.EDU>
Message-Id: <8910231806.AA07060@Polya.Stanford.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: re: CS 499
Yes, I'm working closely with Joe and Dan P.
Dan
∂23-Oct-89 1219 pedre@cs.utexas.edu Elephant 2000
Received: from cs.utexas.edu by SAIL.Stanford.EDU with TCP; 23 Oct 89 12:19:50 PDT
Received: from mothra.cs.utexas.edu by cs.utexas.edu (5.59/1.43)
id AA03309; Mon, 23 Oct 89 14:19:25 CDT
Posted-Date: Mon, 23 Oct 89 14:19:09 CDT
Message-Id: <8910231919.AA12326@mothra.cs.utexas.edu>
Received: by mothra.cs.utexas.edu (5.59/1.4-Client)
id AA12326; Mon, 23 Oct 89 14:19:15 CDT
From: pedre@cs.utexas.edu (Dino Pedreschi)
Date: Mon, 23 Oct 89 14:19:09 CDT
X-Mailer: Mail User's Shell (6.5.6 6/30/89)
To: jmc@sail.stanford.edu
Subject: Elephant 2000
Cc: pedre@cs.utexas.edu
Dear Prof. McCarthy:
I am a researcher from the univ. of Pisa - Italy, currently visiting the
Comp. Sci. Dept of the Univ. of Texas at Austin. I missed your talk about
Elephant 2000, and Bob Boyer told me that you should have a draft paper
about that. The abstract of your talk made me pretty curious, so I would
appreciate very much receiving some literature on the subject, even in a
preliminary stage.
Thank you very much anyway for your attention.
Dino Pedreschi
Department of Computer Sciences
University of Texas at Austin
Austin, Texas 78712 - 1188
e_mail: pedre@cs.utexas.edu
ph: (512) 471-9530
∂23-Oct-89 1333 carol@lucid.com new new-qlisp
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 23 Oct 89 13:33:23 PDT
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA21660; Mon, 23 Oct 89 13:31:07 -0700
Received: from vesuvius ([192.9.200.11]) by heavens-gate id AA01348g; Mon, 23 Oct 89 13:30:26 PDT
Received: by vesuvius id AA06886g; Mon, 23 Oct 89 13:32:28 PDT
Date: Mon, 23 Oct 89 13:32:28 PDT
From: Carol Sexton <carol@lucid.com>
Message-Id: <8910232032.AA06886@vesuvius>
To: qlisp@go4.stanford.edu
Subject: new new-qlisp
I finally installed the new new-qlisp that
I sent mail about last Friday.
Carol
∂23-Oct-89 1540 davis@Polya.Stanford.EDU Lunch
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 23 Oct 89 15:40:05 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA07921; Mon, 23 Oct 89 13:25:36 -0700
Date: Mon, 23 Oct 1989 13:25:33 PDT
From: "Thea E. Davis" <davis@Polya.Stanford.EDU>
To: jmc@sail
Subject: Lunch
Message-Id: <CMM.0.88.625177533.davis@Polya.Stanford.EDU>
The lunch scheduled for Thurs the 26th has been changed to Weds the 25th.
Prof Courtney from Classics will not be able to attend because of the change.
∂23-Oct-89 1728 VAL Commonsense and Nonmonotonic Reasoning Seminar
This is how I propose to advertize your talk. I this ok, or would you like
to include a more substantial abstract?
FORMALIZING COMMONSENSE KNOWLEDGE
John McCarthy
Monday, October 30, 3:15pm
MJH 252
This will be the second in our series of introductory lectures on
commonsense and nonmonotonic reasoning.
∂23-Oct-89 1735 VAL Gelfond
Gelfond may be able to come on Nov. 13 to give a talk about his new work
on a.e. logic if can buy him a ticket. Can we?
∂23-Oct-89 1736 @IBM.COM:AMR@YKTVMH2 Note from AMR
Received: from IBM.COM (ALMADEN.IBM.COM) by SAIL.Stanford.EDU with TCP; 23 Oct 89 17:36:49 PDT
Received: from YKTVMH2 by IBM.COM (IBM VM SMTP R1.2.1MX) with BSMTP id 9798; Mon, 23 Oct 89 17:36:03 PDT
Date: Mon, 23 Oct 89 20:28:28 EDT
From: AMR@IBM.COM
To: jmc@sail.stanford.edu
Subject: Note from AMR
Thought you might be interested in the following, sent to Searle and
Harnad:
------------------------------- Referenced Note ---------------------------
Date: 20 October 1989, 15:53:14 EDT
From: AMR at YKTVMH2 (Alexis Manaster-Ramer)
To: srh at flash.bellcore.com, searle at berkeley.cogsci.edu, amr
If I may mix in (I have been in touch just recently with Stevan,
who has just given me John's e-mail address):
I am thinking of writing a paper of my own focusing on the idea that
traditional AI people as well as its critics have been mistaken
in identifying the computational approach with the propositions that
any two entities whose I/O behavior is equivalent (indistinguishable
by the Turing Test) are computationally equivalent. Much of the theory
of computation (formal language theory and complexity theory) rely on
distinctions between classes of programs (grammars, machines) which are
TT-equivalent but which are distinct with respect to some feature of
their structure. The argument that Searle intends with his Chinese Room
is, of course, that no conventional program or Turing machine could
in principle be intelligent, conscious, and all the other good things.
The correct computational position is NOT the old-fashioned "strong AI"
position, but rather it holds that just because a program or TM or grammar
passes the TT, it does NOT mean that it possesses these properties, simply
because these properties might attach to some structural properties of the
biological machines that the conventional programs, TMs, and grammars do
not happen to share.
To be more specific, the theory of complexity focuses a lot of attention
on the distinction (which is NOT discernable by the TT) between deterministic
and nondeterministic machines, while formal language theory (like its sometimes
distant and indirect offshoots in theoretical linguistics) recognizes that
two grammars which generate the same language may differ in strong generative
capacity (i.e. the derivations of the sentences of the language), and this
notion (which is well-defined for only some classes of grammars, to be sure)
allows us to say quite precisely and formally why a grammar A is not a good
model of the language faculty for some language and a grammar B is (even though
both generate the language).
It remains to specify (a) what kinds of differences may count as differences
of strong generative capacity and (b) how it is
possible to deduce which of two weakly equivalent systems we are dealing
with (and a fortiori to tell which is say conscious and which is not).
It seems to me that the answer to (a) is that any differences at all can
conceivably be relevant. Thus, complexity theory has focused on certain
rather abstract ones, notably determinism and nondeterminism, but given
two physically realized systems, e.g. a PC and an AT (even if they are
running what is conventionally considered to be the same program), we
may insist that for some purpose we need to pay attention to the distinct
ions inherent in their use of different chips. By the same reasoning, if
we need to, we can get even lower into the chemical and physical structure
of the two systems. Thus, it would in principle be entirely consistent
with the computational view of things if we discovered that some particular
property of living beings depended on their being made up of protein etc.,
for example.
At the same time a good computationalist must believe that, regardless,
there exist TMs, grammars, programs which have the exact same I/O behavior
as any given biological system (assuming that biological systems are
ultimately digital; if not, we would need a model of analog computation)
and yet he would also realize how extremely unlikely it is that we could
write such a TM, program, or grammar in the absence of detailed information
about how the biological system really functions (i.e., about its strong
generative capacity, if you will). Of course, those who accept the Chinese
Room argument are assuming ex hypothesi that such a TM etc. exists, since
otherwise the argument is pointless, so this should not be controversial
position.
Finally, I think that it is possible to reason (though never with
certitude) about differences between entities with respect to strong
generative capacity, by focusing not on individual entities (or sets
of weakly equivalent ones), but by studying whole families of gradually
more and more different ones, trying to see what modifications in the
entities' structures produce what modifications in I/O behavior, trying
to understand the historical origin of the structures, etc. Thus, for
the biological systems, we compare different people, different species,
look at the effects of injuries and pathologies, etc., and we can
try to replicate the relevant effects with artificial systems.I think
that some of the existing work in various disciplines incl. linguistics
is in fact aiming at this kind of result.
Having said all this, I would just like to stress
that the traditional AI rhetoric is not in fact consistent with
what the theory of computation teaches us, that from what I have
said it follows that it might be true (as a factual matter) that
only say a biological system could be conscious but also that this
cannot be shown by a purely formal argument (like the Chinese Room
one), and that (although it is highly unlikely that we could ever
write a program which could pass the Turing Test by focusing just
on getting the correct I/O behavior without understaning the structural
principles on which human beings are put together) it is nevertheless
true that if such programs existed in any reasonable domain (say
translation from one NL to another), they would be tremendously useful
and would be considered quite correctly to be a vindication of the
years of efforts that have been invested in AI. It seems to me that
it is this last idea which makes it virtually impossible for people
actively in AI to accept the various arguments against them, since
the real objection (and the one that is NOT illuminated by the
Chinese Room argument and so on) is that the likelihood of attaining
such a program by hacking rather than doing empirical science ofthe
relevant sort is of the same order of magnitude as that of getting
such a program by typing at random. In linguistics, I think this
idea is reasonably clear: that you cannot hope to write any kind of
a grammar without some notion of the underlying structural principles,
but of course linguistics is not AI.
(For Stevan: I don't see the force of your argument about transduction,
since every physically realized computational system(as opposed to
a program or TM or grammar written on a piece of paper) has transduction
of a rather simple sort and behaves as it does in virtue of the laws of
physics (i.e. causally). But I still have not gotten your paper on this
subject. I am just judging by the remarks you make on the subject in
the e-mail to others that you have so kindly shared with me).
I hope all this makes sense, and I welcome (indeed plead for) comments.
Alexis Manaster Ramer
Visiting Scientist, IBM TJW Research Center
Associate Professor of Computer Science, Wayne State University
Ph. D. Linguistics, Univ. of Chicago, 1981
(Just so you guys know something about me.)
P.S. Has either of you seen Maudlin's argument about the
impossibility of machine consciousness in the J. of Philosophy.
I think the construction is incorrect (and cannot be fixed),
but what interests me is precisely that it amounts to an
attempt to show that two strongly equivalent systems may differ
with respect to such a property as consciousness. The flaw,
of course, is the two systems he compares are not strongly
equivalent, but as far as I know that is the only argument in
the literature which (without knowing it, I think) tries to raise
the issue.
∂24-Oct-89 0931 Mailer re: A Sense of Direction
Received: from akbar.cac.washington.edu by SAIL.Stanford.EDU with TCP; 24 Oct 89 09:31:31 PDT
Received: from tomobiki-cho.cac.washington.edu by akbar.cac.washington.edu
(5.61/UW-NDC Revision: 2.6 ) id AA27344; Tue, 24 Oct 89 09:31:03 -0700
Date: Tue, 24 Oct 1989 9:14:22 PDT
From: Mark Crispin <MRC@CAC.Washington.EDU>
Sender: mrc@Tomobiki-Cho.CAC.Washington.EDU
Subject: re: A Sense of Direction
To: John McCarthy <JMC@sail.stanford.edu>
Cc: RIC@rml2.sri.com, comments@kl.sri.com, su-etc@sail.stanford.edu
In-Reply-To: <70$N9@SAIL.Stanford.EDU>
Message-Id: <MailManager.625248862.7590.mrc@Tomobiki-Cho.CAC.Washington.EDU>
In <70$N9@SAIL.Stanford.EDU>, John McCarthy writes:
>The theory that the underclass is caused by defects in the
>educational system is interesting. Newhouse says,
>
>"... bloated deficits, an unstable dollar, and a deplorable
>public-education plant, the last reflected in a large and growing
>underclass of alienated, underemployed, and unemployed people."
>
>This suggests the underclass could be eliminated by fixing the
>educational system.
Seattle may test this theory. The so-called "Save Our Schools" city
proposition (Prop 34), currently slated to pass by a landslide, will have the
effect of ending school busing as a means to achieve integration in Seattle.
Rather than banning busing, it instead offers the board of education a huge
chunk of money that it can claim only if busing is abandoned. It also creates
magnet programs in the schools in disadvantaged areas to encourage voluntary
enrollment in those schools. Since the board of education is going to have to
go to the voters to ask for more money next year they will have a very
difficult time doing so if they pass this money up...
Of course, it took an absolute disaster to make this come about. In Seattle,
it was the so-called "controlled choice" busing scheme used this year, which
had the effect of not enrolling a fair number of kids in schools until several
weeks after school started because the city hadn't decided where to bus them
to yet! Many kids are bused 45 minutes each way each day.
Needless to say, the neo-liberals are up in arms over this. Purportedly,
allowing kids to go to their neighborhood schools would "re-segregate" the
schools. It doesn't matter to them that the group of black kids they bus to a
mostly-white school all end up in the same class, and vice versa...
Curiously, many black parents are strong supporters of Prop. 34. They feel
that their children's education is being hurt by long bus rides, not to
mention being concerned about the long periods where their children are
inadequately supervised and subject to bad influences.
-------
∂24-Oct-89 0935 korf@CS.UCLA.EDU Re: Cindy Mason
Received: from shemp.cs.ucla.edu by SAIL.Stanford.EDU with TCP; 24 Oct 89 09:35:15 PDT
Return-Path: <korf@CS.UCLA.EDU>
Received: from Denali.CS.UCLA.EDU by shemp.cs.ucla.edu (Sendmail 5.61a/2.22)
id AA00442; Tue, 24 Oct 89 09:34:57 -0700
Message-Id: <8910241634.AA00442@shemp.cs.ucla.edu>
Received: by denali; Tue, 24 Oct 89 09:44:58 pdt
Date: Tue, 24 Oct 89 09:44:58 pdt
From: Richard E Korf <korf@CS.UCLA.EDU>
To: JMC@SAIL.Stanford.EDU
In-Reply-To: John McCarthy's message of 19 Oct 89 2031 PDT <tnb2j@SAIL.Stanford.EDU>
Subject: Re: Cindy Mason
John,
Thanks for asking. Cindy is still having problems, most likely related to
decompression sickness. She is down here with me for now. We're playing musical
doctors trying to get a definitive diagnosis and treatment.
I have to put off any travel plans until we sort this out and she gets better.
Sorry about that.
-rich
∂24-Oct-89 0948 VAL Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
FORMALIZING COMMONSENSE KNOWLEDGE
John McCarthy
Monday, October 30, 3:15pm
MJH 252
This will be the second in our series of introductory lectures on
commonsense and nonmonotonic reasoning.
∂24-Oct-89 1134 chandler@Polya.Stanford.EDU [Mail Delivery Subsystem <MAILER-DAEMON@tenaya.stanford.edu> :
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 24 Oct 89 11:34:36 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA10891; Tue, 24 Oct 89 11:33:46 -0700
Date: Tue, 24 Oct 1989 11:33:45 PDT
From: "Joyce R. Chandler" <chandler@Polya.Stanford.EDU>
To: binford@coyote, cheriton@pescadero, feigenbaum@sumex-aim, rwf@sail,
genesereth@polya, golub@patience, guibas@dec.com, jlh@amadeus,
dek@sail, latombe@coyote, zm@sail, jmc@sail, ejm@sierra, oliger@pride,
pratt@Polya.Stanford.EDU, ullman@Polya.Stanford.EDU,
wiederhold@sumex-aim, winograd@csli
Subject: [Mail Delivery Subsystem <MAILER-DAEMON@tenaya.stanford.edu> :
Returned mail: User unknown ]
Message-Id: <CMM.0.88.625257225.chandler@Polya.Stanford.EDU>
Return-Path: <MAILER-DAEMON@Tenaya.Stanford.EDU>
Received: from Tenaya.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61/25-eef) id AA10307; Tue, 24 Oct 89 11:24:49 -0700
Received: by Tenaya.Stanford.EDU (NeXT-0.8/25-eef) id AA09459; Tue, 24 Oct 89 11:21:38 PDT
Date: Tue, 24 Oct 89 11:21:38 PDT
From: Mail Delivery Subsystem <MAILER-DAEMON@tenaya.stanford.edu>
Message-Id: <8910241821.AA09459@Tenaya.Stanford.EDU>
Subject: Returned mail: User unknown
To: nilsson@Tenaya.Stanford.EDU
----- Transcript of session follows -----
Connected to polya.Stanford.EDU:
>>> RCPT To:<tenured@polya.stanford.edu>
<<< 550 <tenured@polya.stanford.edu>... User unknown
550 tenured@cs... User unknown
----- Unsent message follows -----
Received: by Tenaya.Stanford.EDU (NeXT-0.8/25-eef) id AA09457; Tue, 24 Oct 89 11:21:38 PDT
Date: Tue, 24 Oct 89 11:21:38 PDT
From: Nils Nilsson <nilsson@tenaya.stanford.edu>
Full-Name: Nils Nilsson
Message-Id: <8910241821.AA09457@Tenaya.Stanford.EDU>
To: tenured@polya.stanford.edu
Subject: appt. needed
I have been asked by Elizabeth Traugott, Dean
of Graduate Studies, to appoint a "Departmental
Minority Graduate Student Officer." Quoting from
the request, "duties would include meeting with minority
graduate students in your department and
understanding their concerns, serving as liaison with the
department graduate admissions committee, and as advocate for multicultural issues in the department."
Any volunteers?
-Nils
∂24-Oct-89 1316 winograd@loire.stanford.edu re: elephant 2000
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 24 Oct 89 13:16:05 PDT
Received: by loire.stanford.edu (5.59/25-eef) id AA18979; Tue, 24 Oct 89 13:15:26 PDT
Date: Tue, 24 Oct 89 13:15:26 PDT
Message-Id: <8910242015.AA18979@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: JMC@SAIL.Stanford.EDU
Subject: re: elephant 2000
John,
I won't be able to get to the tinlunch on Thursday since I have a dentist
appointment. I am putting copies in your box of a couple of papers by
Searle that are relevant to the logic of speech acts. Also, you might
be interested in the book (which I don't have a copy of):
Searle, John R. and Vanderveken, Daniel, FOUNDATIONS OF ILLOCUTIONARY
LOGIC (Cambridge ; Cambridge University Press, 1985)
LOCATION: Green Stacks P95.55.S42 1985
--t
∂24-Oct-89 1533 Mailer re: A Sense of Direction
Received: from akbar.cac.washington.edu by SAIL.Stanford.EDU with TCP; 24 Oct 89 15:33:14 PDT
Received: from tomobiki-cho.cac.washington.edu by akbar.cac.washington.edu
(5.61/UW-NDC Revision: 2.6 ) id AA02410; Tue, 24 Oct 89 15:33:02 -0700
Date: Tue, 24 Oct 1989 15:11:41 PDT
From: Mark Crispin <MRC@CAC.Washington.EDU>
Sender: mrc@Tomobiki-Cho.CAC.Washington.EDU
Subject: re: A Sense of Direction
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Cc: RIC@RML2.SRI.COM, comments@KL.SRI.COM, su-etc@SAIL.Stanford.EDU
In-Reply-To: <o1svx@SAIL.Stanford.EDU>
Message-Id: <MailManager.625270301.7866.mrc@Tomobiki-Cho.CAC.Washington.EDU>
In <o1svx@SAIL.Stanford.EDU>, John McCarthy writes:
>The liberals still have the courts.
This was a concern when Prop. 34 was written. That's why it is written the
way it is -- the carrot is a chunk of money iff busing is axed, the stick is
if the voters are pissed off they can vote down the schools' request for money
the next election.
The board of education tried to get Prop. 34 taken off the ballot as
unconstitutional, but the courts refused. They said they couldn't see
anything that justified it not being voted on. Now, if the board takes it to
the court after it succeeds, then they're going to be in the position of
refusing money.
-------
∂24-Oct-89 1541 bscott@Polya.Stanford.EDU Salary Research Charge
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 24 Oct 89 15:41:34 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA09276; Tue, 24 Oct 89 15:41:13 -0700
Date: Tue, 24 Oct 1989 15:41:11 PDT
From: Betty Scott <bscott@Polya.Stanford.EDU>
To: JMC@Sail
Cc: BScott@Polya.Stanford.EDU
Subject: Salary Research Charge
Message-Id: <CMM.0.88.625272071.bscott@Polya.Stanford.EDU>
John, Nils said he talked with you about charging some of your academic year
salary to research accounts, since you are the P.I. on several. He asked
me to check with you again to see whether you may have changed your mind,
and do want to charge some percentage to research.
Thank you.
Betty
∂24-Oct-89 1624 pollack@cis.ohio-state.edu AAAI workshop grants
Received: from cheops.cis.ohio-state.edu by SAIL.Stanford.EDU with TCP; 24 Oct 89 16:23:21 PDT
Received: from toto.cis.ohio-state.edu by cheops.cis.ohio-state.edu (5.61/4.891004)
id AA18014; Tue, 24 Oct 89 19:23:05 -0400
Received: by toto.cis.ohio-state.edu (4.0/3.890428)
id AA08280; Tue, 24 Oct 89 19:23:03 EDT
Date: Tue, 24 Oct 89 19:23:03 EDT
From: Jordan B Pollack <pollack@cis.ohio-state.edu>
Message-Id: <8910242323.AA08280@toto.cis.ohio-state.edu>
To: jmc@sail.stanford.edu
Cc: pollack@cis.ohio-state.edu
Subject: AAAI workshop grants
Reply-To: pollack@cis.ohio-state.edu
Professor McCarthy,
We last met at Carl Hewitt's "Foundation of AI" workshop at MIT.
I write to find out if you are still coordinating small AAAI grants
to support conferences and workshops (with deliverables consisting
of a small report in AIMag)?
You co-funded (along with with ONR) a workshop I organized on AI &
Connectionism, April 1987, which was quite successful (Report in AIMag
Winter 88); I am collaborating on a broad cross-disciplinary workshop
(Columbus, late spring) on "Musical Cognition", and am organizing a
track on Musical Knowledge Representation.
Please let me know if its OK to send you a brief proposal on it, or
who I should write to instead.
Jordan Pollack
∂24-Oct-89 1650 Mailer re: A Sense of Direction
Received: from rml2.sri.com by SAIL.Stanford.EDU with TCP; 24 Oct 89 16:50:20 PDT
Date: Tue 24 Oct 89 16:44:43-PST
From: Ric Steinberger <RIC@RML2.SRI.COM>
Subject: re: A Sense of Direction
To: JMC@SAIL.STANFORD.EDU
Cc: MRC@CAC.WASHINGTON.EDU, RIC@RML2.SRI.COM, comments@KL.SRI.COM,
su-etc@SAIL.STANFORD.EDU
Message-ID: <625275884.0.RIC@RML2.SRI.COM>
In-Reply-To: <o1svx@SAIL.Stanford.EDU>
Mail-System-Version: <VAX-MM(246)+TOPSLIB(136)@RML2.SRI.COM>
> From: John McCarthy <JMC@SAIL.Stanford.EDU>
> The liberals still have the courts.
That's very interesting considering that Mr. Reagan has personally appointed
a large number of federal judges as well as Supreme Court Justices.
And "having the courts" is somewhat inflamatory. No one, no party, as I
understand our system of government, should ever "have" the courts.
Certainly, recent Supreme Court decisions regarding abortion rights,
affirmative action, labor laws, and the role of the federal government
do not indicate any liberal bias. Or, am I misunderstanding what
is meant by liberal?
-ric steinberger
-------
∂24-Oct-89 1922 paulf@bodega.Stanford.EDU Daily editorial
Received: from bodega.Stanford.EDU by SAIL.Stanford.EDU with TCP; 24 Oct 89 19:21:50 PDT
Received: by bodega.Stanford.EDU (5.57/Ultrix3.0-C)
id AA06604; Tue, 24 Oct 89 19:10:18 PDT
Date: Tue, 24 Oct 89 19:10:18 PDT
From: paulf@bodega.stanford.edu (Paul Flaherty)
Message-Id: <8910250210.AA06604@bodega.Stanford.EDU>
To: jmc@sail.stanford.edu
Subject: Daily editorial
Well, I see that the Lamely has once again decided to equate "oppositon to
affirmative action" to "racism". Care for a reply?
-=paulf
∂24-Oct-89 1956 paulf@bodega.Stanford.EDU re: Daily editorial
Received: from bodega.Stanford.EDU by SAIL.Stanford.EDU with TCP; 24 Oct 89 19:56:54 PDT
Received: by bodega.Stanford.EDU (5.57/Ultrix3.0-C)
id AA06816; Tue, 24 Oct 89 19:45:30 PDT
Date: Tue, 24 Oct 89 19:45:30 PDT
From: paulf@bodega.stanford.edu (Paul Flaherty)
Message-Id: <8910250245.AA06816@bodega.Stanford.EDU>
To: JMC@sail.stanford.edu
Subject: re: Daily editorial
The Daily is windier than most... ;-)
-=paulf
∂25-Oct-89 0054 roode@orc.olivetti.com [ je: From today's WSJ ]
Received: from orc.olivetti.com ([129.189.192.2]) by SAIL.Stanford.EDU with TCP; 25 Oct 89 00:54:23 PDT
Received: by orc.olivetti.com (4.0/SMI-3.2)
id AA13026; Wed, 25 Oct 89 00:54:06 PDT
Date: Wed, 25 Oct 1989 0:54:05 PDT
From: David Roode <roode@orc.olivetti.com>
To: jmc@sail.stanford.edu
Subject: [ je: From today's WSJ ]
Message-Id: <CMM.0.87.625305245.roode@Pisa>
Moonlighting?
---------------
Return-Path: <jed>
Received: from LOCALHOST by Thun.ORC.Olivetti.Com (4.0/SMI-3.2)
id AA00894; Mon, 23 Oct 89 17:41:57 PDT
Message-Id: <8910240041.AA00894@Thun.ORC.Olivetti.Com>
To: tech, ravasio@condor
Cc: cam-orl!mvw
Subject: From today's WSJ
X-Fax-Delivery-Window:
Date: Mon, 23 Oct 89 17:41:55 -0700
From: jed
Hewlett-Packard's Task Broker Assigns Jobs to Computers
Palo Alto, CA -- Hewlett-Packard will announce today a software program
that allows computers in a network to speed up computing tasks by
sending the tasks to each other.
Called Task Broker, the program acts something like an auctioneer among
a group of computers wired together. If a machine has a big computing
task, Task Broker asks other computers on the network for "bids" on the
job. It then determines which machine is free to do the task most
quickly and sends the task to that machine.
Hewlett-Packard claims that the software allows a network to run three
times as many tasks as conventional networks and will run each task twice as fast.
The new Hewlett-Packard program, said analyst John McCarthy at
Forrester Research Inc., a computer-market research company, "is a key
building block as people move to this new model of distributed
processing." In today's computer networks, some machines oftern sit
idle while others are overtaxed. With the Hewlett-Packard program, he
said, "You get more bang for the buck you've spent on computers."
The program, which will be shipped in January 1990, runs on the Unix
operating system. Hewlett-Packard will charge $5000 for a license
covering 10 users. The program now works on all Hewlett-Packard and
Apollo workstations and on computers made by Multiflow Computers Inc.
of Branford, Conn. Hewlett-Packard said it will sell versions next
year that run on Sun Microsystems Inc. and Digital Equipment Corp. machines.
The Task Broker differs from other programs that spread computing tasks
around a network. A previously available program, called Network
Computing System, developed by Hewlett-Packard's Apollo division, for
instance, takes a task and splits it up into parts, divvying up those
parts to several computers in a network for simultaneous processing.
But programs in individual computers must be revised to work with that system.
Applications won't have to be rewritten to work with Task Broker,
Hewlett-Packard said, and the user of a computer won't be able to tell
that another machine is doing the work. The Task Broker, "turns that
network into -- as far as the user is concerned -- one giant computer,"
said Bill Kay, general manager of Hewlett-Packard's workstation group.
∂25-Oct-89 0847 mkatz@sesame.Stanford.EDU Re: A Sense of Direction
Received: from sesame.Stanford.EDU by SAIL.Stanford.EDU with TCP; 25 Oct 89 08:47:11 PDT
Received: by sesame.Stanford.EDU (5.57/Ultrix2.4-C)
id AA16772; Wed, 25 Oct 89 08:44:59 PDT
Date: Wed, 25 Oct 89 08:44:59 PDT
From: mkatz@sesame.stanford.edu (Morris Katz)
Message-Id: <8910251544.AA16772@sesame.Stanford.EDU>
To: JMC@sail.stanford.edu (John McCarthy)
Subject: Re: A Sense of Direction
Newsgroups: su.etc
In-Reply-To: <o1svx@SAIL.Stanford.EDU>
Organization: Stanford University
In article <o1svx@SAIL.Stanford.EDU> you write:
>[In reply to message from MRC@CAC.Washington.EDU sent Tue, 24 Oct 1989 9:14:22 PDT.]
>
>The liberals still have the courts.
In light of the current split on the Supreme Court and the number of Reagan
appointees to the federal bench, I would be interested in seeing you justify
this statement. Nearly 3/4 of the federal bench has been appointed by
Republicans. Please feel free to tell me that Republicans are not
conservatives in the traditional sense of the word and that you are refering to
the traditional meaning of liberal. I can't determine how you intended to use
the word 'liberal' which has become a loaded term.
-- Morry Katz
--
-------------------------------------------------------------------------------
Morry Katz
katz@polya.stanford.edu
-------------------------------------------------------------------------------
∂25-Oct-89 0922 MPS
D. Lyman is ill and will be unable to attend today
∂25-Oct-89 0934 Mailer re: A Sense of Direction
Received: from rml2.sri.com by SAIL.Stanford.EDU with TCP; 25 Oct 89 09:34:12 PDT
Date: Wed 25 Oct 89 09:28:26-PST
From: Ric Steinberger <RIC@RML2.SRI.COM>
Subject: re: A Sense of Direction
To: JMC@SAIL.STANFORD.EDU
Cc: RIC@RML2.SRI.COM, MRC@CAC.WASHINGTON.EDU, comments@KL.SRI.COM,
su-etc@SAIL.STANFORD.EDU
Message-ID: <625336106.290000.RIC@RML2.SRI.COM>
In-Reply-To: <u1ufy@SAIL.Stanford.EDU>
Mail-System-Version: <VAX-MM(246)+TOPSLIB(136)@RML2.SRI.COM>
I suppose it wouldn't hurt for someone considered a liberal to state that
I have always been against bussing school children as a remedy for almost
anything. The money spent for bussing could be better used improving
the schools themselves: buying books, upgrading facilities, teacher training,
etc.
However while I don't personally support bussing to achieve integration
"targets", I have heard some positive statements made about the results.
In general, though, I believe it's a major waste of the children's time,
and a waste of money that is needed for more important educational
objectives.
Can I join the young Republicans now?
-ric steinberger %-)
-------
∂25-Oct-89 1009 CLT recursion
Here are a few references that come to mind:
Abelson, H. and G. J. Sussman
[1985]
Structure and interpretation of computer programs,
(The MIT Press, McGraw-Hill Book Company)
Burge, W. H.
[1975]
Recursive programming techniques,
(Addison-Wesley)
Henderson, P.
[1980]
Functional programming: Application and implementation,
(Prentice-Hall)
Boyer, R. S. and J. S. Moore
[1979]
A Computational Logic
(Academic Press, New York).
Friedman and Felleisen
[1987]
The Little Lisper
(The MIT Press, McGraw-Hill Book Company)
Manna and Waldinger - latest volumes
(I don't know the exact refs, also you might ask Waldinger if he
knows of more recent things).
∂25-Oct-89 1239 katiyar@Polya.Stanford.EDU seminar next tuesday
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 25 Oct 89 12:39:16 PDT
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA08126; Wed, 25 Oct 89 12:38:57 -0700
Date: Wed, 25 Oct 1989 12:38:55 PDT
From: Dinesh Katiyar <katiyar@Polya.Stanford.EDU>
To: jmc@sail
Subject: seminar next tuesday
Message-Id: <CMM.0.88.625347535.katiyar@Polya.Stanford.EDU>
john,
could you please send me an abstract (if you have one) of the
talk on ELEPHANT 2000 that you are giving next tuesday so that
i can mail it to the relevant mailing list ?
-dinesh
∂25-Oct-89 1355 jones@Sunburn.Stanford.EDU Nonmonotonic Logic Course
Received: from Sunburn.Stanford.EDU by SAIL.Stanford.EDU with TCP; 25 Oct 89 13:55:07 PDT
Received: by Sunburn.Stanford.EDU (5.61/25-eef) id AA26022; Wed, 25 Oct 89 13:55:26 -0700
Date: Wed, 25 Oct 1989 13:55:22 PDT
From: "H. Roy Jones" <jones@Sunburn.Stanford.EDU>
To: konolige@ai.sir.com
Cc: rwf@sail, jmc@sail, nilsson@Sunburn.Stanford.EDU, nilsson@cs
Subject: Nonmonotonic Logic Course
Message-Id: <CMM.0.88.625352122.jones@Sunburn.Stanford.EDU>
Kurt,
I haven't heard anything officially from the Curriculum committee about your
proposal yet. However, there is some concern that it overlaps considerably
with John McCarthy's course on nonmonotonic reasoning, CS323, which is also
scheduled for winter quarter. If you get a chance you might look at the
course description for 323 or talk to John and let me know what you think.
In any case, I'll keep you posted.
Roy Jones
723-3642
∂25-Oct-89 1556 pehoushe@Gang-of-Four.Stanford.EDU QWAIT
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 25 Oct 89 15:56:12 PDT
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA00400; Wed, 25 Oct 89 15:53:51 -0700
Date: Wed, 25 Oct 89 15:53:51 -0700
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8910252253.AA00400@Gang-of-Four.Stanford.EDU>
To: qlisp@Gang-of-Four.Stanford.EDU
Subject: QWAIT
QWAIT now works in the Stanford Qlisp system. If you just want to
wait, use (spawn t form), if you want the value, use (future form).
You can nest qwaits, they'll do the right thing. This could
significantly help OPS5; but as a general waiting scheme, QWAIT leaves
alot to be desired.
-dan
PS. Some of you know that, to use this system you need to run
/u/pehoushe/bin/new-qlisp from unix, and then (load "/qlisp/nqn-load").
The file /qlisp/nqn-fib.lisp contains lots of hairy fib examples,
although it doesn't excersize qwait or nqlambda.
∂25-Oct-89 1636 Mailer re: A Sense of Direction
To: RIC@RML2.SRI.COM, JMC@SAIL.Stanford.EDU
CC: MRC@CAC.WASHINGTON.EDU, su-etc@SAIL.Stanford.EDU
From: Robert W Floyd <RWF@SAIL.Stanford.EDU>
[In reply to message from RIC@RML2.SRI.COM sent Wed 25 Oct 89 09:28:26-PST.]
No, Ric, you can't join the Young Republicans yet.
First you have to figure out that your disappointment
about busing, though widespread among the public, has
no effect on practice because busing was ordered by the
courts rather than the legislatures. After 26 years,
it is time for the country to be able to reassess
busing in public debate and decide whether it is good
public policy. We are now busing the children of the
original busees, and I see no way to avoid busing
their children short of constitutional amendment.
Some years of my childhood I spent more than two hours
a day on buses, for purely logistic reasons, and I
think I lost some of my childhood that way. I hate
to see us doing the same thing forever, whatever the
noble intent.
I don't know how to avoid this kind of problem. The
courts need power to enforce their orders and to
overrule legislatures, for well known reasons, but
they fantasize that they know what will be good for
us, and they impose it without appeal. Then they
call it precedent so it can't be changed. Maybe the
recent purge of the CA Supreme Court will also set
a precedent. The founders in their wisdom set up
a system where right turns on red could be tried on
a small scale first. There is no proviision for that
kind of federalism in the courts. What if the Supreme
court had set up several regional integration plans
(including benign neglect) and watched their effects for
ten or twenty years (their long terms of office permit that)
before telling the whole country what was good for us?
Brown v Board of Educaation at the time looked like
a decision that could only be made by the courts.
Yet within ten years, even conservative legislators
like Dirkson were calling for numerous civil rights laws
"whose time had come". Moral: be skeptical of arguments
that only the courts can solve a problem.
∂25-Oct-89 1800 JMC
cacm proofs
∂26-Oct-89 0524 LISTSERV@VM1.NoDak.EDU Message
Received: from VM1.NoDak.EDU by SAIL.Stanford.EDU with TCP; 26 Oct 89 05:24:49 PDT
Received: from NDSUVM1.BITNET by VM1.NoDak.EDU (IBM VM SMTP R1.2.1MX) with BSMTP id 0847; Thu, 26 Oct 89 07:24:22 CDT
Received: by NDSUVM1 (Mailer R2.03B) id 0791; Thu, 26 Oct 89 07:24:19 CDT
Date: Thu, 26 Oct 89 07:24:02 CDT
From: Revised List Processor (1.6c) <LISTSERV@VM1.NoDak.EDU>
Subject: Message
To: JMC@SAIL.STANFORD.EDU
Your mailing for THEORYNT has been submitted to the list editor:
THEORYNT@YKTVMZ.
∂26-Oct-89 1019 JCMA@REAGAN.AI.MIT.EDU Elephant 2000
Received: from REAGAN.AI.MIT.EDU by SAIL.Stanford.EDU with TCP; 26 Oct 89 10:19:05 PDT
Received: from MORRISON.AI.MIT.EDU by REAGAN.AI.MIT.EDU via INTERNET with SMTP id 271108; 26 Oct 89 13:19:56 EDT
Date: Thu, 26 Oct 89 13:20 EDT
From: John C. Mallery <JCMA@REAGAN.AI.MIT.EDU>
Subject: Elephant 2000
To: jmc@sail.stanford.edu
Message-ID: <19891026172009.0.JCMA@MORRISON.AI.MIT.EDU>
Please send me a copy of the paper on this.
John Mallery
Artificial Intelligence Laboratory, NE43-797
Massachusetts Institute of Technology
545 Technology Square
Cambridge, Massachusetts 02139-4301
(617)253-5966
∂26-Oct-89 1123 JCMA@REAGAN.AI.MIT.EDU re: Elephant 2000
Received: from REAGAN.AI.MIT.EDU by SAIL.Stanford.EDU with TCP; 26 Oct 89 11:23:40 PDT
Received: from MORRISON.AI.MIT.EDU by REAGAN.AI.MIT.EDU via INTERNET with SMTP id 271139; 26 Oct 89 14:24:31 EDT
Date: Thu, 26 Oct 89 14:24 EDT
From: John C. Mallery <JCMA@REAGAN.AI.MIT.EDU>
Subject: re: Elephant 2000
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <1G29qH@SAIL.Stanford.EDU>
Message-ID: <19891026182440.1.JCMA@MORRISON.AI.MIT.EDU>
Date: 26 Oct 89 1026 PDT
From: John McCarthy <JMC@SAIL.Stanford.EDU>
[In reply to message sent Thu, 26 Oct 89 13:20 EDT.]
You're on the list.
Thanks.
∂26-Oct-89 1427 winograd@loire.stanford.edu December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 26 Oct 89 14:27:33 PDT
Received: by loire.stanford.edu (5.59/25-eef) id AA22666; Thu, 26 Oct 89 14:25:21 PDT
Date: Thu, 26 Oct 89 14:25:21 PDT
Message-Id: <8910262125.AA22666@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: ai-qual-faculty@loire.stanford.edu
Subject: December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
To the AI faculty (and associates):
We are scheduling the AI qual for the week of Dec. 4-8. We have not
yet set a specific day, and I would like to know when you are available.
Classes are still meeting that week. Please specify which half-days
(morning 9-12:15, afternoon 1:15-4:30) you ARE available and I will
try to work out a schedule. Also, let me know which of the following
depth areas you would be willing to serve as the depth examiner (each
committee has one person designated as chair, and one as depth examiner.
You don't do much different from anyone else except that you declare
yourself as being competetent to judge students in that area and presumably
ask more questions in it).
knowledge representation
temporal reasoning
logics (nomonotonic)
Please answer by next Tuesday (Oct. 31) so we can let the students know
a definite date.
Thanks. --t
∂26-Oct-89 1753 S.SUMMER-RAIN@Macbeth.Stanford.EDU re: A Sense of Direction
Received: from Macbeth.Stanford.EDU by SAIL.Stanford.EDU with TCP; 26 Oct 89 17:53:29 PDT
Date: Thu 26 Oct 89 17:50:18-PDT
From: William Brown, Jr. <S.SUMMER-RAIN@Macbeth.Stanford.EDU>
Subject: re: A Sense of Direction
To: JMC@SAIL.Stanford.EDU
cc: S.SUMMER-RAIN@Macbeth.Stanford.EDU, wab@SUMEX-AIM.Stanford.EDU
In-Reply-To: <70$N9@SAIL.Stanford.EDU>
Message-ID: <12537308100.25.S.SUMMER-RAIN@Macbeth.Stanford.EDU>
Re: whether Newhouse would perform said experiment.
Well, John, I have not been on the 'board for several months, but
I've obviously decided to step into this one.
Newhouse's political position is irrelevant to the
validity of his argument. It is a known fact that students from
richer schools (i.e. schools with more money per student) do much
better than students from poorer schools, especially when
cost-of-living adjustments are made. This does not PROVE that
investing more money into the poorer school systems will produce
better students, however, that is most likely the end result.
Newhouse was pointing out that we do not invest in a great deal
of our students - wich leaves them very little to give back. This
leaves us with the state of not really being in competition with
anyone BUT ourselves. Our true political enemies, at least the
openly hostile ones, are much too small militarily to be a real
threat to us. It is our internal social structure which is
ailing, and which we must fix to remain relatively viable in the
world market.
--Bill
-------
∂26-Oct-89 2127 ma@src.dec.com Elephant paper?
Received: from decwrl.dec.com by SAIL.Stanford.EDU with TCP; 26 Oct 89 21:15:52 PDT
Received: by decwrl.dec.com; id AA24872; Thu, 26 Oct 89 21:15:39 -0700
Received: by jumbo; id AA10787; Thu, 26 Oct 89 21:15:37 PDT
From: ma@src.dec.com (Martin Abadi)
Message-Id: <8910270415.AA10787@jumbo>
Date: 26 Oct 1989 2115-PDT (Thursday)
To: jmc@sail.stanford.edu
X-Folder-Carbon:
Subject: Elephant paper?
Hello,
I wonder if you have anything written up about the new Elephant.
It seems quite interesting---and bigger than the one you defined
a few years ago. If so, could you please have the write-up sent
to me (address below)? I did not make your talk today and I'll be
East next week, alas.
Thanks.
Martin
M. Abadi
DEC SRC
130 Lytton Ave.
Palo Alto, CA 94301
∂26-Oct-89 2347 young@Neon.Stanford.EDU Elephant 2000 References?
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 26 Oct 89 23:47:07 PDT
Received: by Neon.Stanford.EDU (5.61/25-eef) id AA21137; Thu, 26 Oct 89 23:45:45 -0700
Date: Thu, 26 Oct 1989 23:45:44 PDT
From: "R. Michael Young" <young@Neon.Stanford.EDU>
To: jmc@sail
Cc: young@Neon.Stanford.EDU
Subject: Elephant 2000 References?
Message-Id: <CMM.0.88.625473944.young@Neon.Stanford.EDU>
Professor McCarthy,
After hearing your talk today over at CSLI, I was wondering if I could get a
copy of any paper/extended abstract that you might have on the Elephant 2000
language or on the Speech Acts approach to programming. Were you refering
today to a paper which was published somewhere?
-Michael
∂27-Oct-89 0049 Mailer re: A Sense of Direction
To: JMC@SAIL.Stanford.EDU, su-etc@SAIL.Stanford.EDU
From: Les Earnest <LES@SAIL.Stanford.EDU>
In response to my earlier posting, John McCarthy asks:
> Do you think that raising the pay of teachers to that of auto mechanics -
> if indeed it is lower, would attract the same kind of people who
> were teachers in the 1930s and 1940s?
The world has changed a lot since the '30s and '40s. Until a time machine
is invented, it will not be possible to attract the same kind of people
who taught then. If what John really wants to know is whether raising
teachers' salaries will improve the quality of typical teachers, I would
say that it is necessary but not sufficient -- it will also be necessary
to improve the public esteem of the profession and wait a generation or
so for the teacher population to "turn over."
-Les Earnest
∂27-Oct-89 0054 Mailer re: new earthquake concern
To: JMC@SAIL.Stanford.EDU, su-etc@SAIL.Stanford.EDU
From: Les Earnest <LES@SAIL.Stanford.EDU>
Regarding the reported relative plate movement in the Loma Prieta quake,
John McCarthy asks:
> Did surveying give the two meter result this time?
I believe so; that is the only practical way to do it.
∂27-Oct-89 1720 Mailer re: A Sense of Direction
To: LES@SAIL.Stanford.EDU, JMC@SAIL.Stanford.EDU,
su-etc@SAIL.Stanford.EDU
From: Robert W Floyd <RWF@SAIL.Stanford.EDU>
[In reply to message from LES rcvd 27-Oct-89 00:49-PT.]
There were studies 15 or 20 years ago of the characteristics
of schools that produced good academic performance. I think
I saw the results in a book "Inequality" by Riesmann and Jenks.
Most of the obvious variables, including expenditures, were
insignificant after controlling for the advantages of the
student population. To get a good education, go to school
with bright, well brought up children of bright, well brought up
parents. This conforms with my observation that the way to
have a great graduate program is to have a program with such
a ggreat reputation that it attracts great students. When the
professors ignore them, they then educate each other very well.
This is why high reputations of grad schools are well deserved.
In California, expenditures on schools were horrendously
unequal until the late 70s. The Serrano decision, which
makes interesting reading, held that this violated equal
protection of the laws. So far as I know, the state now
provides most of the support for the schools (used to be
local prop taxes). The Gilroy - LAHills example shows
that money throwing doesn't fix the problem, though I
think the kids should have abundant resources whether
they are learning or not. The Catholic schools nationally
spend less money than the public schools and pay teachers
less. They produce better educated kids, I regret to
say. What they do works in the worst neighborhoods.
In the south side of Chicago, black non-Catholic
parents who want their kids to learn send their kids
to the Catholic schools.
I think it likely that kids will underachieve in a
society where they are not in the continual company
of a directly concerned adult. Sorry if the resulting
conclusion sounds anti-femiinist and anti-poor, but here it
comes. If you want a healthy, well educated, productive
kid, plan to have a parent stay home until junior high.
In practice that means mama. Most people can't afford that
because of high housing prices resulting from high land
prices, at least around here. So move already. Live
in Fresno or Pittsburgh or Houston, or better yet in
a small town in Indiana or Iowa. Listen to music on
records, order books by phone from a big city, rent
videotapes of foreign movies.
An alternative is to postpone childbearing for a long
time. Find a mate, then work for ten years putting
20% of income into stocks, bonds, real estate, etc.
Make that 30%, that would otherwise be going into
orthodontists bills and such. Then one parent goes
into nesting mode with the resources to do it right.
Meanwhile the parents have learned how to use the
resources effectively that will keep them from being
groun down by the stresses of childrearing.
Either way, divorce is out. Hardly any couple can
aford that along with the expenses of raising good
kids. Severely dysfunctional families should split
at all costs, but in marginal cases families should
stay nominally married and conserve resources. Daddy
can have a mistress, and Mama can have a tennis coach,
it's a lot cheaper than divorce. Greece and Spain and
France and Italy have been that way for ages, so
people can live with it.
I'm not proposing legislating any of this, except in
the sense of tax laws along the lines of IRAs on a
larger scale to defer taxation on saving, whether in
general or earmarked for child support. I just wish I
had adopted such a policy. And had a like-minded mate.
Another pattern we can still see in other countries,
and in some animal species like the canids, is to
leave childbearing to the dominant members of the tribe
with the others playing a support role. Among the wild
dogs of Africa, this is implemented by draconian policies
like killing pups of the non-dominant females. It is
probably hard to implement among us in a way the ACLU
would like. It might be a good strategy, though, even
for the non-dominant to get their genes into the next
generation.
I don't expect many of my readers to buy these ideas.
But in planning a life, we should take into account
not only the deliberate costs, but also the risks
of other costs like the adolescent child who can't
grow up because the only grownups she can pattern
herself after are the ones on TV, and they behave
like kids. When the hidden costs are added in, it
might not be a high price for the childbearing generation
to pay.
∂27-Oct-89 1725 bobrow.pa@Xerox.COM Elephant 2000
Received: from Xerox.COM by SAIL.Stanford.EDU with TCP; 27 Oct 89 17:25:29 PDT
Received: from Semillon.ms by ArpaGateway.ms ; 27 OCT 89 17:25:20 PDT
Date: Fri, 27 Oct 89 17:20 PDT
From: bobrow.pa@Xerox.COM
Subject: Elephant 2000
To: jmc@sail.stanford.edu
Message-ID: <19891028002030.1.BOBROW@BULLWINKLE.parc.xerox.com>
Line-fold: no
John,
I saw an announcement about a discussion of your new language, but couldn't
come. Is there a paper? If so, could you send me a copy? Are you willing to
give a talk at PARC?
thanks
danny
-------
∂27-Oct-89 1734 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
FORMALIZING COMMONSENSE KNOWLEDGE
John McCarthy
Monday, October 30, 3:15pm
MJH 252
This will be the second in our series of introductory lectures on
commonsense and nonmonotonic reasoning.
∂27-Oct-89 2000 JMC
Michael Stanford article
∂28-Oct-89 0915 CLT driving
Please see that Hazel gets a lesson today.
∂28-Oct-89 1138 christos%cs@ucsd.edu Re: ucsd visit
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 28 Oct 89 11:38:22 PDT
Received: from helios.ucsd.edu by ucsd.edu; id AA06431
sendmail 5.61/UCSD-2.0-sun via SMTP
Sat, 28 Oct 89 11:38:09 -0700 for JMC@sail.stanford.edu
Received: by helios.UCSD.EDU (3.2/UCSDPSEUDO.2)
id AA19866 for delivery to JMC@sail.stanford.edu; Sat, 28 Oct 89 11:37:36 PDT
Date: Sat, 28 Oct 89 11:37:36 PDT
From: christos%cs@ucsd.edu (Christos Papadimitriou)
Message-Id: <8910281837.AA19866@helios.UCSD.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: Re: ucsd visit
Cc: kay@cs.UCSD.EDU, pasquale@cs.UCSD.EDU
John:
It is Monday, Nov. 13th. I shall be out of town until the 10th, but my
colleague Joe Pasquale will be in touch with you concerning your trip.
We are all looking forward to your visit!
---Christos
∂29-Oct-89 0900 JMC
Susie about lunch
∂30-Oct-89 1049 rlg@ai.mit.edu Hertz recommendation
Received: from life.ai.mit.edu by SAIL.Stanford.EDU with TCP; 30 Oct 89 10:49:32 PST
Received: from gracilis (gracilis.ai.mit.edu) by life.ai.mit.edu (4.0/AI-4.10) id AA09047; Mon, 30 Oct 89 13:49:28 EST
From: rlg@ai.mit.edu (Bob Givan)
Received: by gracilis (4.0/AI-4.10) id AA00888; Mon, 30 Oct 89 13:49:14 EST
Date: Mon, 30 Oct 89 13:49:14 EST
Message-Id: <8910301849.AA00888@gracilis>
To: jmc@sail.stanford.edu, val@sail.stanford.edu
Subject: Hertz recommendation
hello,
in light of the earthquake especially, and
just generally, I'd like to check that the
Hertz recommendation forms I sent to you
2 weeks ago were received and have been
sent out. (the deadline is Nov. 1)
please let me know...
hope you survived the earthquake well...
thanks,
Bob Givan
rlg@ai.mit.edu
∂30-Oct-89 1125 siegman@sierra.STANFORD.EDU Re: committee on administration
Received: from sierra.STANFORD.EDU by SAIL.Stanford.EDU with TCP; 30 Oct 89 11:25:12 PST
Received: by sierra.STANFORD.EDU (3.2/4.7); Mon, 30 Oct 89 11:24:50 PST
Date: Mon, 30 Oct 89 11:24:50 PST
From: siegman@sierra.STANFORD.EDU (Anthony E. Siegman)
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: Re: committee on administration
In-Reply-To: Your message of 30 Oct 89 1025 PST
Message-Id: <CMM.0.88.625778688.siegman@>
I've retired to "private life" for a few years -- partly as an explicit
protest, refusing to help out with administration until situation
improves -- but will pass your message along.
∂30-Oct-89 1436 bobrow.pa@Xerox.COM re: Elephant 2000
Received: from Xerox.COM by SAIL.Stanford.EDU with TCP; 30 Oct 89 14:36:03 PST
Received: from Semillon.ms by ArpaGateway.ms ; 30 OCT 89 14:32:37 PST
Date: Mon, 30 Oct 89 14:27 PST
From: bobrow.pa@Xerox.COM
Subject: re: Elephant 2000
To: JMC@SAIL.Stanford.EDU
cc: bobrow.pa@Xerox.COM
In-Reply-To: <K2#O8@SAIL.Stanford.EDU>
Message-ID: <19891030222730.2.BOBROW@BULLWINKLE.parc.xerox.com>
Line-fold: no
Great. I look forward to seeing the paper, and will ask you agian about a talk in
January.
-------
∂30-Oct-89 1753 woolf@venera.isi.edu verifying addresses
Received: from vax.darpa.mil by SAIL.Stanford.EDU with TCP; 30 Oct 89 17:53:47 PST
Received: from venera.isi.edu by vax.darpa.mil (5.61/5.61+local)
id <AA29333>; Mon, 30 Oct 89 20:51:19 -0500
Posted-Date: Mon, 30 Oct 89 17:51:10 PST
Message-Id: <8910310151.AA21532@venera.isi.edu>
Received: from nzl.isi.edu by venera.isi.edu (5.61/5.61+local)
id <AA21532>; Mon, 30 Oct 89 17:51:25 -0800
To: cps@venera.isi.edu
Subject: verifying addresses
Cc: vgordon@venera.isi.edu, kemp@venera.isi.edu
Reply-To: woolf@venera.isi.edu
Date: Mon, 30 Oct 89 17:51:10 PST
From: Suzanne Woolf <woolf@venera.isi.edu>
This message is an attempt to verify addresses in the cps mailing list
on vax.darpa.mil. We have had complaints from people who have tried
to send mail to this list that some of the addresses for some of the
users on the list don't work. We have resolved as many of these as
possible, and would like to verify the ones we have repaired and
correct any left that may be wrong.
Please disregard this message unless it was delivered to you at an
address other than the one you prefer. If you would like to correct
the email address we have for you, please send the request to
cps-request@vax.darpa.mil.
Thanks very much for your patience.
Suzanne Woolf
Network Services, ISI
∂30-Oct-89 1901 VAL Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU, su-events@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
AUTOEPISTEMIC MODAL LOGICS
Grigory Shvarts
Program Systems Institute
of the USSR Academy of Sciences
Pereslavl-Zalessky, USSR
Monday, November 6, 3:15pm
MJH 252
A modal approach to nonmonotonic reasoning was proposed by Drew
McDermott and Jon Doyle in 1980-82. Almost immediately some
disadvantages of that approach were pointed out. Robert Moore
(1983) proposed his autoepistemic logic, which overcomes these
difficulties. Later, some authors (Kurt Konolige, Paul Morris
and others) found peculiarities of different kinds in Moore's
logic and proposed rather complicated solutions to these problems.
A careful mathematical analysis of Moore's and McDermott's
approaches shows that Moore's logic is merely a special case
of McDermott's logic, at least formally. The problems that arose
in Moore's logic may find a simple and uniform solution by
coming back to McDermott's original concept.
∂30-Oct-89 2019 jutta@coyote.stanford.edu reminder and agenda for AI Division meeting
Received: from coyote.stanford.edu by SAIL.Stanford.EDU with TCP; 30 Oct 89 20:19:14 PST
Received: by coyote.stanford.edu; Mon, 30 Oct 89 14:25:21 PST
Date: 30 Oct 1989 1425-PST (Monday)
From: Jutta McCormick <jutta@coyote.stanford.edu>
To: latombe@coyote.stanford.edu, binford@coyote.stanford.edu,
feigenbaum@SUMEX-AIM.Stanford.EDU, genesereth@polya.stanford.edu,
jones@polya.stanford.edu, jmc@sail.stanford.edu
Cc: ok@coyote.stanford.edu, nilsson@tenaya.stanford.edu,
shoham@polya.stanford.edu, winograd@csli.stanford.edu
Subject: reminder and agenda for AI Division meeting
AI Division Meeting
Wednesday, November 1, 1:30 - 3:00 p.m.
Cedar Hall Conference Room
Agenda
1. Preparation for Visiting Committee
2. Organization of the Division
3. AI Course Offerings
4. Need for Additional AI Faculty
5. Consulting AI Faculty
6. Space Allocation
7. Division Budget
8. Other Issues
∂31-Oct-89 0159 @loire.stanford.edu:marty@cis.Stanford.EDU December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 31 Oct 89 01:59:44 PST
Received: from CIS.Stanford.EDU by loire.stanford.edu (5.59/25-eef) id AA29905; Tue, 31 Oct 89 01:57:42 PDT
Received: by cis.Stanford.EDU (5.59/inc-1.0)
id AA13840; Tue, 31 Oct 89 01:55:03 PDT
Date: Tue, 31 Oct 89 01:55:03 PDT
From: marty@cis.Stanford.EDU (Marty Tenenbaum)
Message-Id: <8910310955.AA13840@cis.Stanford.EDU>
To: Winograd@csli.stanford.edu
Cc: ai-qual-faculty@loire.stanford.edu
In-Reply-To: Terry Winograd's message of Thu, 26 Oct 89 14:25:21 PDT <8910262125.AA22666@loire.stanford.edu>
Subject: December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
I would be available for 1 afternoon, monday, thursday, or friday.
JMT.
∂31-Oct-89 0536 CLT driving
I would like you to take the responsibility of seeing that
hazel gets driving practice. If you aren't willing to take
the time then hire someone to do it.
Also until Hazel has a liscence and 2 months of solid experience,
I would like you to take responsibility of getting Timothy to school
on Wednesday and for taking and picking up on Friday. If you
can't do it on some given day, the its your responsibility to
find someone who can. That is only half of the time and I think
its time you did your share not only of driving but arranging.
∂31-Oct-89 0854 CLT driving
Thanks much
∂31-Oct-89 0908 katiyar@Neon.Stanford.EDU reminder
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 31 Oct 89 09:08:35 PST
Received: by Neon.Stanford.EDU (5.61/25-eef) id AA02032; Tue, 31 Oct 89 09:08:06 -0800
Date: Tue, 31 Oct 1989 9:08:05 PST
From: Dinesh Katiyar <katiyar@Neon.Stanford.EDU>
To: jmc@sail
Subject: reminder
Message-Id: <CMM.0.88.625856885.katiyar@Neon.Stanford.EDU>
just to remind you, sir, that you are speaking on elephant 2000
this afternoon (at 4 in room 301, MJH)
-dinesh
∂31-Oct-89 1324 @IBM.COM:LEORA@YKTVMH
Received: from IBM.COM (ALMADEN.IBM.COM) by SAIL.Stanford.EDU with TCP; 31 Oct 89 13:24:39 PST
Received: from YKTVMH by IBM.COM (IBM VM SMTP R1.2.1MX) with BSMTP id 4297; Tue, 31 Oct 89 13:24:34 PST
Date: Tue, 31 Oct 89 16:23:59 EST
From: LEORA@IBM.COM
To: jmc@sail.stanford.edu
John,
I received your reminder about the book proposal. I managed to get
in touch with the person at Brown responsible for restoring my files,
and she said I would have them by tomorrow afternoon at the latest.
So, I will send it to you then. By the way, the file is in troff.
Is troff available at Stanford, or should I delete the troff commands,
merge lines, etc. to make things more readable?
Leora
∂31-Oct-89 1346 @loire.stanford.edu:hayes@kanga.parc.xerox.com December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 31 Oct 89 13:46:06 PST
Received: from arisia.Xerox.COM by loire.stanford.edu (5.59/25-eef) id AA00626; Tue, 31 Oct 89 13:42:09 PDT
Received: from kanga.parc.Xerox.COM by arisia.Xerox.COM with SMTP
(5.61+/IDA-1.2.8/gandalf) id AA10950; Tue, 31 Oct 89 13:40:33 -0800
Received: by kanga.parc.xerox.com
(5.61+/IDA-1.2.8/gandalf) id AA09800; Tue, 31 Oct 89 13:42:02 PST
Message-Id: <8910312142.AA09800@kanga.parc.xerox.com>
Date: Tue, 31 Oct 89 13:42:02 PST
From: Pat Hayes <hayes@parc.xerox.com>
To: Winograd@csli.stanford.edu
Cc: ai-qual-faculty@loire.stanford.edu
In-Reply-To: Terry Winograd's message of Thu, 26 Oct 89 14:25:21 PDT <8910262125.AA22666@loire.stanford.edu>
Subject: December AI qual - AVAILABILITY DATES NEEDED FROM YOU BY TUESDAY
I can make most any half-day after the 5 December. I would be willing
to go into depth on KR or temporal reasoning.
Pat
∂31-Oct-89 1553 sf@csli.Stanford.EDU
Received: from csli.Stanford.EDU by SAIL.Stanford.EDU with TCP; 31 Oct 89 15:53:20 PST
Received: by csli.Stanford.EDU (4.0/inc-1.0)
id AA05876; Tue, 31 Oct 89 15:53:52 PST
Date: Tue 31 Oct 89 15:53:52-PST
From: Sol Feferman <SF@CSLI.Stanford.EDU>
To: JMC@SAIL.Stanford.EDU
Message-Id: <625881232.0.SF@CSLI.Stanford.EDU>
In-Reply-To: <w4uPg@SAIL.Stanford.EDU>
Mail-System-Version: <SUN-MM(242)+TOPSLIB(128)@CSLI.Stanford.EDU>
Yes, i wanted to invite him too. Is Friday noon OK with you? It seems
best for Pat Suppes.
-------
∂31-Oct-89 1553 VAL Gelfond
Gelfond is going to be here on Nov. 13, and I invited him to give a
seminar. Can we pay him a honorarium?
∂31-Oct-89 1621 sf@csli.Stanford.EDU Re: reply to message
Received: from csli.Stanford.EDU by SAIL.Stanford.EDU with TCP; 31 Oct 89 16:21:47 PST
Received: by csli.Stanford.EDU (4.0/inc-1.0)
id AA07209; Tue, 31 Oct 89 16:22:21 PST
Date: Tue 31 Oct 89 16:22:20-PST
From: Sol Feferman <SF@CSLI.Stanford.EDU>
Subject: Re: reply to message
To: JMC@SAIL.Stanford.EDU
Message-Id: <625882940.0.SF@CSLI.Stanford.EDU>
In-Reply-To: <1S4z4J@SAIL.Stanford.EDU>
Mail-System-Version: <SUN-MM(242)+TOPSLIB(128)@CSLI.Stanford.EDU>
Great! I'll send a separate msg to Vladimir. Let's meet at the
Faculty Club at noon. I'll reserve.
-------
∂31-Oct-89 1633 wab@sumex-aim.stanford.edu re: Busing
Received: from sumex-aim.stanford.edu by SAIL.Stanford.EDU with TCP; 31 Oct 89 16:33:24 PST
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AA28020; Tue, 31 Oct 89 16:35:08 PST
Date: Tue, 31 Oct 1989 16:35:05 PST
From: "William A. Brown" <wab@sumex-aim.stanford.edu>
To: John McCarthy <JMC@sail.stanford.edu>
Cc: wab@sumex-aim.stanford.edu, su-etc@sumex-aim.stanford.edu
Subject: re: Busing
In-Reply-To: Your message of 29 Oct 89 2024 PST
Message-Id: <CMM.0.88.625883705.wab@sumex-aim.stanford.edu>
John, You wrote:
My own opinion is that some equalization of resources is ok, but
I think it won't solve the problem where there is a culture that
opposes the educational system.
I would also be against inhibition of volunteer parent activity,
although it's certainly ok to try to match it with volunteer
activity aimed at districts with difficulties.
-----------------
I don't think any CULTURE opposes education. I do think that several
ENVIRONMENTS may strongly hinder it. (Of course, one could argue that "punk
culture" opposes education, "the drug culture" opposes education, etc. This
is true. I am referring to racial cultures.) To combat these influences,
money, yes money, needs to be spent to insure that all children have the same
quality if educational resources. This will most likely mean putting police
in some schools, as I'm sure is already done. This also means putting
computers in schools that don't have them. This may also mean providing after
school security for teachers who wish to work after hours in an inner city
school. The list goes on.
As for parental participation, of course it should be encouraged. I really
don't understand why parents don't do more checking up on teachers. I have
gone to class with my younger brother for just that reason. At least one of
instructors completely failed in my book. I almost fell alseep in her very,
very, unstimulating class. Whether or not it should be the teachers'
responsibility to teach students, it must now be the parents', because the
public (and most of the private) school sytem has completely failed.
My experience was highly unusual in that I went to a public school that was
one of the best in the nation. On Friday nights the hottest spot was at home
- on your modem. You get the picture. Very nerdy. Reminds me of Stanford. :)
(Come to think of it, I'm STILL on my modem...)
--B.
∂31-Oct-89 1718 hsu@Neon.Stanford.EDU Elephant 2000
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 31 Oct 89 17:17:59 PST
Received: by Neon.Stanford.EDU (5.61/25-eef) id AA24616; Tue, 31 Oct 89 17:17:41 -0800
Date: Tue, 31 Oct 1989 17:17:40 PST
From: Yung-jen Hsu <hsu@Neon.Stanford.EDU>
To: jmc@sail
Subject: Elephant 2000
Message-Id: <CMM.0.88.625886260.hsu@Neon.Stanford.EDU>
I'd like to have a copy of the paper when it's available. Please send me
a message, and I'll come pick it up. Thanks in advance.
Jane
∂31-Oct-89 2000 JMC
check for sarah and stuff for irs
∂31-Oct-89 2000 JMC
Sarah check
∂01-Nov-89 0648 tom@Polya.Stanford.EDU Boise
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 1 Nov 89 06:48:08 PST
Received: by Polya.Stanford.EDU (5.61/25-eef) id AA09463; Wed, 1 Nov 89 06:47:47 -0800
Date: Wed, 1 Nov 89 06:47:47 -0800
From: Tom Dienstbier <tom@Polya.Stanford.EDU>
Message-Id: <8911011447.AA09463@Polya.Stanford.EDU>
To: JMC@SAIL.Stanford.EDU
In-Reply-To: John McCarthy's message of 31 Oct 89 2100 PST <G4bMa@SAIL.Stanford.EDU>
Subject: Boise
As you can see John it doesn't get much use. It cost us 20K per year
just for HP maintenance and depreciation. I belive it goes through no
more than 2-3 boxes per year. Not enough to cover its cost.
Any ideas for a good replacement?
tom
∂01-Nov-89 0958 MPS
Chris Anderson, The Scientist Newspaper, DC, 202-857-0355
He would like to talk to you about the historical funding of AI (funding
agencies) and how things have changed over the years. I asked for
a complimentary copy of the Newspaper, thinking you might be interested.
Pat
∂01-Nov-89 1201 JMC
letter to N.
∂01-Nov-89 1301 JMC
letter about Moravec book
∂01-Nov-89 1624 paek@Neon.Stanford.EDU Elephant 2000 paper
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 1 Nov 89 16:24:36 PST
Received: by Neon.Stanford.EDU (5.61/25-eef) id AA24425; Wed, 1 Nov 89 16:24:18 -0800
Date: Wed, 1 Nov 1989 16:24:14 PST
From: Eunok Paek <paek@Neon.Stanford.EDU>
To: jmc@sail
Subject: Elephant 2000 paper
Message-Id: <CMM.0.88.625969454.paek@Neon.Stanford.EDU>
I would like a copy of Elephant 2000 paper when it's available.
Thanks a lot.
-Eunok Paek
∂01-Nov-89 1744 hayes@kanga.parc.xerox.com Vlad Dabija
Received: from arisia.Xerox.COM by SAIL.Stanford.EDU with TCP; 1 Nov 89 17:43:53 PST
Received: from kanga.parc.Xerox.COM by arisia.Xerox.COM with SMTP
(5.61+/IDA-1.2.8/gandalf) id AA10512; Wed, 1 Nov 89 17:41:42 -0800
Received: by kanga.parc.xerox.com
(5.61+/IDA-1.2.8/gandalf) id AA06938; Wed, 1 Nov 89 17:43:15 PST
Message-Id: <8911020143.AA06938@kanga.parc.xerox.com>
Date: Wed, 1 Nov 89 17:43:15 PST
From: Pat Hayes <hayes@parc.xerox.com>
To: JMC@sail.stanford.edu
Cc: hayes@parc.xerox.com
Subject: Vlad Dabija
John,
you may recall me mentioning this Romanian guy, a potential student
who wants to talk to people. He will be here on MOnday 6th: can I
tell him to come and talk with you soon after lunch, say at 1.30 ? If
a different time is better ( has to be before 4 ) let me know asap.
Thanks.
Pat
∂02-Nov-89 0936 JMC
doug about contract
∂02-Nov-89 0948 hayes@kanga.parc.xerox.com Vlad Dabija
Received: from arisia.Xerox.COM by SAIL.Stanford.EDU with TCP; 2 Nov 89 09:48:26 PST
Received: from kanga.parc.Xerox.COM by arisia.Xerox.COM with SMTP
(5.61+/IDA-1.2.8/gandalf) id AA18642; Thu, 2 Nov 89 08:48:18 -0800
Received: by kanga.parc.xerox.com
(5.61+/IDA-1.2.8/gandalf) id AA20100; Thu, 2 Nov 89 08:49:53 PST
Message-Id: <8911021649.AA20100@kanga.parc.xerox.com>
Date: Thu, 2 Nov 89 08:49:53 PST
From: Pat Hayes <hayes@parc.xerox.com>
To: JMC@SAIL.Stanford.EDU
In-Reply-To: John McCarthy's message of 01 Nov 89 1845 PST <135vOV@SAIL.Stanford.EDU>
Subject: Vlad Dabija
Thanks, OK 1.30 In your office in the quad ( which I hope wasnt too
badly shookup ), I assume.
To recommend? Well, I dont know him personally but was impressed by
his CV. He got top honors everywhere in Romania and did a lot of AI
hacking all by himself, and seems to combine intelligence with energy.
Hes been working in LA for a year or so but is superkeen to get a
western PhD in AI. He is busy putting himself through the appropriate
graduate-school-entrance exams and has tiptop results so far, I
understand, in everything except English.( although his spokjen
English seems fine ). If he were American he would have got in touch
with you himself, but he was obvioulsy raised in a society in which
one cannot approach Famous Men unless one has a letter of
introduction: rather like 17th century England. He came to me with a
cover message ( and request to help him ) from someone back east.
Dont take it that I recommend him personally, only that I commend him
to your attention, and am acting as a sort of lubricant to enable him
to move around Stanford and talk to people.
Pat
∂02-Nov-89 1219 MPS
Hurd - 854-1901
∂02-Nov-89 1445 winograd@loire.stanford.edu AI QUAL SCHEDULED FOR WED. DECEMBER 6
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 2 Nov 89 14:45:39 PST
Received: by loire.stanford.edu (5.59/25-eef) id AA03773; Thu, 2 Nov 89 14:43:19 PDT
Date: Thu, 2 Nov 89 14:43:19 PDT
Message-Id: <8911022243.AA03773@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: ai-qual@loire.stanford.edu
Subject: AI QUAL SCHEDULED FOR WED. DECEMBER 6
Based on responses from students and faculty, we have scheduled the AI
Qual for Wednesday morning, December 6. There will be two exam times:
8:45-10:15 and 10:30-12:00. The faculty will meet at noon to discuss
grading, and results will be made available by the end of the day.
The students taking the exam and their depth areas are:
Becky Thomas: bthomas@neon knowledge representation
David Ash: ash@sumex-aim logic and planning
Fangzhen Lin: lin@neon nonmonotonic logic
Ramanathan Guha: guha@sumex-aim knowledge representation
STUDENTS: Let me know if you have a strong preference for the earlier
or later slot. I already know that David can't do the later one.
FACULTY: We have generally tried not to have the examining committee
include an advisor or someone who is working closely with the
student. Let me know if there are any such constraints for you of
which I should be aware in assigning the committees. Also, we will
provide lunch at the noon discussion.
Please get this information to me by next Wednesday and I will determine
the committees and send out final plans.
Thanks. --t
p.s. To faculty who told me that you are not available on that day:
You are receiving this message for your information and are about to be
taken off the distribution list so you won't be bothered further.
∂02-Nov-89 1501 @IBM.COM:LEORA@YKTVMH book proposal
Received: from IBM.COM (ALMADEN.IBM.COM) by SAIL.Stanford.EDU with TCP; 2 Nov 89 15:01:26 PST
Received: from YKTVMH by IBM.COM (IBM VM SMTP R1.2.1MX) with BSMTP id 4003; Thu, 02 Nov 89 15:01:19 PST
Date: Thu, 2 Nov 89 18:01:03 EST
From: LEORA@IBM.COM
To: jmc@sail.stanford.edu
Subject: book proposal
John,
I'm enclosing the book proposal and list of papers. I've made some
minor adjustments since I did this last year, and the troff isn't
working here properly, so you may get some bugs when you try to troff
it. Hope it won't be too bad. The macro package I use is ms.
I'd very much appreciate any comments you have on this. Looking forward
to speaking to you,
Leora
[moved it to leora.boo[f89,jmc]
∂02-Nov-89 1507 MPS
L. Morgenstern called. 212-781-6539. She said
you could call her at home.
∂03-Nov-89 1051 CLT qlisp
To: RPG@SAIL.Stanford.EDU, JMC@SAIL.Stanford.EDU,
weening@Gang-of-Four.Stanford.EDU
We should meet soon to decide
what should become of the qlisp
project. What about next Wednesday
just before the CPL meeting (1ish)?
∂03-Nov-89 1501 CLT ∂02-Nov-89 1513 gilberts@Polya.Stanford.EDU Letters of Recommendation
Date: Thu, 2 Nov 1989 15:12:50 PST
From: Edie Gilbertson <gilberts@Polya.Stanford.EDU>
To: CLT@Sail
Subject: Letters of Recommendation
Carolyn, The provost's office has requested three letters of
recommendation for your promotion, one from Nils and two from
outside. I'll ask Nils for one. Would you like to ask two
other people for one each? The are to be mailed to:
Attn: Mary Hanrahan
Cypress Hall, C-wing
cc: Rita Hernandez
Terman, 201
cc: Yvette Sloan
CSD
Please let me know if you would like to ask two associates for
recommendations, or if I should ask Nils to suggest people.
Thanks, -Edie
∂03-Nov-89 1546 gumby@Gang-of-Four.Stanford.EDU Elephant 2000
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 3 Nov 89 15:45:55 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA17495; Fri, 3 Nov 89 15:47:23 -0800
Date: Fri, 3 Nov 89 15:47:23 -0800
From: gumby@Gang-of-Four.Stanford.EDU (David Vinayak Wallace)
Message-Id: <8911032347.AA17495@Gang-of-Four.Stanford.EDU>
To: jmc@sail.Stanford.EDU
Subject: Elephant 2000
When you have a paper ready could you send me a copy please?
Best address is David Wallace, Box 8874, Stanford 94309.
Thanks
d
∂03-Nov-89 1610 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU Reception for Matijasevitch and Girard
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 3 Nov 89 16:10:11 PST
Received: from iswim.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61+IDA/25-eef) id AA17179; Fri, 3 Nov 89 16:09:33 -0800
Received: by iswim.stanford.EDU (5.57/Ultrix3.0-C)
id AA01622; Fri, 3 Nov 89 16:09:21 PST
Message-Id: <8911040009.AA01622@iswim.stanford.EDU>
To: lifschitz@Polya.Stanford.EDU, jmc@sail.Stanford.EDU, clt@sail.Stanford.EDU,
barwise@csli.Stanford.EDU, pratt@Polya.Stanford.EDU,
sf@csli.Stanford.EDU, shankar@csl.sri.com, meseguer@csl.sri.com,
narciso@csl.sri.com, andre@csli.Stanford.EDU,
etchemendy@csli.Stanford.EDU, kolaitis@gaia.ucsc.edu, ma@src.dec.com,
luca@src.dec.com, dek@sail.Stanford.EDU, shoham@Polya.Stanford.EDU,
israel@csli.Stanford.EDU, vardi@ibm.com, halpern@ibm.com,
iam@sail.Stanford.EDU, nilsson@Polya.Stanford.EDU,
zm@sail.Stanford.EDU, Pat%IMSSS@score.Stanford.EDU
Subject: Reception for Matijasevitch and Girard
From: John C. Mitchell <jcm@polya.stanford.edu>
Reply-To: John C. Mitchell <jcm@Polya.Stanford.EDU>
Date: Fri, 03 Nov 89 16:09:19 PST
Sender: jcm@iswim.Stanford.EDU
A reception in honor of distinguished visitors
Prof. Yuri Matijasevitch Prof. Jean-Yves Girard
Steklov. Inst. of Mathematics Universite Paris VII
will be held Tuesday evening, Nov 7, beginning at 7 PM. Wine, cheese,
cold cuts, salads, a few desserts and soft drinks will be served.
Spouses and significant others are welcome. The reception will be held
at the home of Vaughan and Margot Pratt, 2215 Old Page Mill Road
(which might better be referred to as 2215 Gerth Lane, as far as I
can tell from Vaughan's map -- jcm).
_________________________________________________| |__________________________
__________________________________Foothill______Light______Expressway_________
|P|
/.a|.....one way -
//|g| no auto entry to
// |e| Old Page Mill Road
// | |
Old || |M|
DIRECTIONS Page|| |i|
Take Page Mill to near 280 Mill|| |l|
Follow sign to Old Page Mill Road Road|| |l|
Gerth Lane has row of 12 mailboxes........ || Light_________________________
Cross small wooden bridge.............. : || |E --------------------------
2215 is second on left....... : : || |x| Deer Creek Road
Parking at %'s ........... : : : || |p|
: : : : ||% |w|
Gerth Lane %% : : : ||=== y|.....only entry to
=================:=========#==:== \ | ! Old Page Mill Road
Vaughan & Margot Pratt %2215 % || M %\\ | | and Gerth Lane
2215 Old Page Mill Road 2209 % \\| |
Palo Alto, CA 94304, 415-494-2545 \ |
_________________________________________________| |_________________________
_<--SF______________________Route 280____________ _____________________SJ-->
------- End of Forwarded Message
∂03-Nov-89 1610 stager@Polya.Stanford.EDU CS323 Winter Quarter
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 3 Nov 89 16:10:44 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA17263; Fri, 3 Nov 89 16:10:30 -0800
Date: Fri, 3 Nov 1989 16:10:29 PST
From: "Claire E. Stager" <stager@Polya.Stanford.EDU>
To: jmc@sail.Stanford.EDU
Subject: CS323 Winter Quarter
Message-Id: <CMM.0.88.626141429.stager@Polya.Stanford.EDU>
We've lost access to 260-264 for the rest of the year because of the quake, so
have been scrambling to find another room for your course. We've been able
to reserve TTh 1:15-2:30 in Terman 156. Will this work for you?
(If I can have a confirmation from you today--Friday--we can get the change
listed in the Winter Qtr. schedule.)
As to textbooks:
I'm assuming you still need no texts ordered. Please let me know if I'm
mistaken.
Thanks.
Claire
∂03-Nov-89 1632 MPS Room reservation
I received the room confirmation for your noon meeting at Tresidder
on the 9th of Nov.
Also, your mail slot downstairs has been moved to opposite
from where it use to be.
∂03-Nov-89 1735 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
AUTOEPISTEMIC MODAL LOGICS
Grigory Shvarts
Program Systems Institute
of the USSR Academy of Sciences
Pereslavl-Zalessky, USSR
Monday, November 6, 3:15pm
MJH 252
A modal approach to nonmonotonic reasoning was proposed by Drew
McDermott and Jon Doyle in 1980-82. Almost immediately some
disadvantages of that approach were pointed out. Robert Moore
(1983) proposed his autoepistemic logic, which overcomes these
difficulties. Later, some authors (Kurt Konolige, Paul Morris
and others) found peculiarities of different kinds in Moore's
logic and proposed rather complicated solutions to these problems.
A careful mathematical analysis of Moore's and McDermott's
approaches shows that Moore's logic is merely a special case
of McDermott's logic, at least formally. The problems that arose
in Moore's logic may find a simple and uniform solution by
coming back to McDermott's original concept.
∂04-Nov-89 0121 B.BSK@Macbeth.Stanford.EDU re: One Week until Egg-Drop I
Received: from Macbeth.Stanford.EDU by SAIL.Stanford.EDU with TCP; 4 Nov 89 01:21:25 PST
Date: Sat 4 Nov 89 01:20:11-PST
From: Brian Keller <B.BSK@Macbeth.Stanford.EDU>
Subject: re: One Week until Egg-Drop I
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <2C6sPl@SAIL.Stanford.EDU>
Message-ID: <12539498075.12.B.BSK@Macbeth.Stanford.EDU>
Sure can! It's open to anyone who's interested.
bsk
-------
∂04-Nov-89 1037 beeson@ucscd.UCSC.EDU Re: Shrinking Lattice Polyhedra
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 4 Nov 89 10:37:39 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA29558; Sat, 4 Nov 89 10:37:38 -0800
Date: Sat, 4 Nov 89 10:37:38 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8911041837.AA29558@ucscd.UCSC.EDU>
To: JMC@SAIL.Stanford.EDU, beeson@ucscd.UCSC.EDU
Subject: Re: Shrinking Lattice Polyhedra
I have that report in hand, and have had it in hand for some time.
Susan Landau sent it to me in response to the letter I wrote her at
your suggestion. In fact, last week I finally found the day
needed to revise the triangle paper and submit it for publication
(I sent it to Wilf, the editor of the Math Monthly, on the grounds
that the paper is of general interest and doesn't require specialized
knowledge). It contains a reference to the report in question.
Incidentally, you might wonder (I did) whether the smallest triangle
similar to a given lattice triangle in 4-space can be embedded in
3-space, as we know SOME triangle similar to it can. Well, Susan
Landau says not always, and I presume that is a simple consequence
of the proof-method of their result, i.e. gcds in the quaternions,
though I did not have time to calculate it myself. By the way,
I thought you told me at that original Mexican dinner that you had
proved that result which is now in Landau and Cremona. Did you
give the idea away or was it independent work? My revised version
also presents your idea of looking at lattice angles instead of
triangles. It seems to simplify the first section a little but
made less impact on the paper than I originally thought. (On the
actual proofs that is.) I should have gotten around to these
revisions long ago.
As for your question about the earthquake, 17 is still closed. I have
tried every possible alternate route including the 6am special train from
Watsonville. The best route is 152 which takes me 75 minutes. I am
more or less cut off from the possibility of attending Stanford seminars
and had to miss Mateyisevich (sp?) last week. We suffered various
damage to the house, most of which was not immediately obvious (except
the chimney of course) so there has been a succession of pieces of
bad news, conversations with contractors, insurance people, phone calls,
not to mention a wife with a bad case of "earthquake nerves", all of
which has taken far too much time from the project of finishing MATHPERT,
and promises to continue doing so. From my point of view the main
impact of the earthquake is as a drain on research time, already a scarce
resource.
∂04-Nov-89 1610 ACT Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 4 November 1989
Current Charges 0.30 (vending machine)
-------
TOTAL AMOUNT DUE 0.30
PAYMENT DELIVERY LOCATION: CSD Receptionist.
Make checks payable to: STANFORD UNIVERSITY.
Please deliver payments to the Computer Science Dept receptionist, Jacks Hall.
To ensure proper crediting, please include your PONY ACCOUNT NAME on your check.
Note: The recording of a payment takes up to three weeks after the payment is
made, but never beyond the next billing date. Please allow for this delay.
Bills are payable upon presentation. Interest of 1.0% per month will be
charged on balances remaining unpaid 25 days after bill date above.
An account with a credit balance earns interest of .33% per month,
based on the average daily balance.
∂05-Nov-89 0029 ME NS
∂04-Nov-89 1856 JMC
What is the problem with NS?
ME - I don't know. I spent time on Friday trying to fix it, but I'm
not sure what's going on.
∂05-Nov-89 0939 ME AP
The AP wire is working again.
∂05-Nov-89 2259 alex@jessica.Stanford.EDU Thesis news AND questions
Received: from jessica.Stanford.EDU by SAIL.Stanford.EDU with TCP; 5 Nov 89 22:59:19 PST
Received: by jessica.Stanford.EDU (5.59/25-eef) id AA04841; Sun, 5 Nov 89 23:00:06 PDT
Date: Sun, 5 Nov 1989 23:00:05 PST
From: Alex Bronstein <alex@jessica.Stanford.EDU>
To: clt@sail.Stanford.EDU, jmc@sail.Stanford.EDU, rwf@sail.Stanford.EDU,
horning@src.dec.com
Cc: alex@jessica.Stanford.EDU
Subject: Thesis news AND questions
Message-Id: <CMM.0.88.626338805.alex@jessica.Stanford.EDU>
Dear Reading Committee Members,
Well, epsilon (the distance between me and the end of my thesis)
is painfully but finally becoming really small. I intend to have a final
draft ready this Thursday, for review by those who requested it (so far:
Carolyn Talcott and Jim Horning). I will then go for a week to an IFIP
conference to present a 20 page summary of my thesis which was accepted
there. When I come back on Nov. 20, I intend to take a week to make
the final final final corrections. In the last couple of days of
November, or at the latest, on December 1st, I intend to ask everybody
their signature.
This ends the news and leads me to the following questions:
1) Prof. McCarthy and Prof. Floyd: do you want a final draft this
coming Thursday?
2) Everybody: do you have any plans to be OUT-OF-TOWN or otherwise
unreachable during the period: Nov. 28 to Dec. 2nd?
Thanks,
Alex
∂06-Nov-89 0515 cross@vax.darpa.mil Re: When
Received: from vax.darpa.mil (darpa.mil) by SAIL.Stanford.EDU with TCP; 6 Nov 89 05:15:15 PST
Received: from sun46.darpa.mil by vax.darpa.mil (5.61/5.61+local)
id <AA02808>; Mon, 6 Nov 89 08:14:57 -0500
Posted-Date: Mon 6 Nov 89 08:14:15-EST
Received: by sun46.darpa.mil (4.1/5.51)
id AA06476; Mon, 6 Nov 89 08:14:17 EST
Date: Mon 6 Nov 89 08:14:15-EST
From: Steve Cross <CROSS@DARPA.MIL>
Subject: Re: When
To: JMC@sail.stanford.edu
Cc: CROSS@vax.darpa.mil, cross@vax.darpa.mil
Message-Id: <626361255.0.CROSS@SUN46.DARPA.MIL>
In-Reply-To: <S7Cr5@SAIL.Stanford.EDU>
Mail-System-Version: <SUN-MM(217)+TOPSLIB(128)@SUN46.DARPA.MIL>
John: Still planning on being there this Friday, Nov 10 around
2:00 PM. Steve
-------
∂06-Nov-89 0617 @Score.Stanford.EDU,@MCC.COM:greene@mcc.com Elephant 2000
Received: from Score.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Nov 89 06:17:45 PST
Received: from MCC.COM by SCORE.STANFORD.EDU with TCP; Mon 6 Nov 89 06:18:00-PST
Received: from luke.aca.mcc.com by MCC.COM with TCP/SMTP; Mon 6 Nov 89 08:17:33-CST
Date: Mon, 6 Nov 89 08:16:59 CST
From: greene@mcc.com (Kevin Greene)
Posted-Date: Mon, 6 Nov 89 08:16:59 CST
Message-Id: <8911061416.AA02250@luke.aca.mcc.com>
Received: by luke.aca.mcc.com (3.2/ACTv4.1i)
id AA02250; Mon, 6 Nov 89 08:16:59 CST
To: JMC@SAIL.Stanford.EDU
In-Reply-To: John McCarthy's message of 17 Oct 89 0756 PDT <ZmVnc@SAIL.Stanford.EDU>
Subject: Elephant 2000
Reply-To: greene@mcc.com
Postal-Address: MCC, 3500 West Balcones Center Drive, Austin, TX, 78759
>>Thanks for the interest. I'll send you the next draft. I will probably
>>be at MCC November 16 and 17 and will be happy to talk about Elephant
>>if anyone is interested.
Thanks in advance for the "next draft". Thanks also for offerring to speak to
us--unfortunately we will not be able accept your offer as we will be
presenting our annual "in depth" review to our shareholders on 16&17 Nov.
Thanks again, Kevin
∂06-Nov-89 0825 stager@Polya.Stanford.EDU re: CS323 Winter Quarter
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Nov 89 08:25:47 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA26703; Mon, 6 Nov 89 08:25:36 -0800
Date: Mon, 6 Nov 1989 8:25:33 PST
From: "Claire E. Stager" <stager@Polya.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: CS323 Winter Quarter
In-Reply-To: Your message of 03 Nov 89 1748 PST
Message-Id: <CMM.0.88.626372733.stager@Polya.Stanford.EDU>
Thanks for the reply.
I'll check on the time of VTSS160.
Claire
∂06-Nov-89 1047 perrault@ai.sri.com illocutionary vs perlocutionary
Received: from Warbucks.AI.SRI.COM by SAIL.Stanford.EDU with TCP; 6 Nov 89 10:47:16 PST
Received: from agate.ai.sri.com by Warbucks.AI.SRI.COM with INTERNET ;
Mon, 6 Nov 89 10:47:00 PST
Received: by agate.ai.sri.com (4.1/4.16)
id AA03273 for jmc@sail.stanford.edu; Mon, 6 Nov 89 10:46:32 PST
Date: Mon 6 Nov 89 10:46:27-PST
From: PERRAULT@Warbucks.AI.SRI.COM (Ray Perrault)
Subject: illocutionary vs perlocutionary
To: jmc@sail.stanford.edu
cc: perrault@Warbucks.AI.SRI.COM
Message-Id: <626381187.0.PERRAULT@ai.sri.com>
Mail-System-Version: <SUN-MM(229)+TOPSLIB(128)@ai.sri.com>
John,
A belated comment following your presentation at CSLI. *I* think the
main distinction Searle and Austin try to make between the two terms
concerns *how* the agent's act affects the state of the world.
Although perlocutionary effects (e.g. coming to have a certain belief,
or coming to have a certain intention to do something (or even doing
it)) can be the result of acts of communication, they can also come
about without communication playing any role. In Austin's old
example, if you know I'm a greedy person and you want me to leave your
office, you might throw a coin out the window, expecting that I will
go fetch it. No act of communication (i.e. no illocutionary act, in
the way they use the term) has been performed, although perlocutionary
consequences have certainly been achieved. (You haven't *said*
anything, and not just because you haven't used language, because they
want to allow communicative acts to be performed non-linguistically,
which is one reason why *speech* act is such a bad term). Attempts to
characterize The illocutionary act have hinged on the necessary
intended (by the speaker) recognition (by the hearer) of some
intention of the speaker. Thus if I know that you know that I'm
greedy, and I see you tossing a coin out of the window, then I might
leave, in part in virtue of my recognizing that you intend me to
recognize your intention that I do so. The other intuition they try
to capture is that the illocutionary act is completely in the control
of the speaker, whereas the perlocutionary act is not. This is
supposed to account for the fact that it is only possible to insert
illocutionary verbs in utterances of the form "I hereby ___ that..."
The conditions defining illocutionary acts need to be completely
statable in terms of the speaker's mental state (including his beliefs
about the environment, and the mental state of the hearer). The
perlocutionary acts generally depend directly on the state of the
hearer, so that I haven't convinced you of p unless you actually come
to believe p, but I can have asserted p without your coming to believe
p, or anything else for that matter. If I'm not mistaken, that is the
aspect of the distinction you'r trying to embed in Elephant.
By the way, I know of no satisfactory precise definition of the
illocutionary-perlocutionary distinction that captures these
intuitions, although many have tried.
Ray
-------
∂06-Nov-89 1219 MPS
Who do I send the IBM invoice to and where?
Also, Does Dennis Bahler get the NC State one?
Thanks
∂06-Nov-89 1332 stager@Polya.Stanford.EDU re: CS323 Winter Quarter
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Nov 89 13:32:16 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA09942; Mon, 6 Nov 89 13:32:04 -0800
Date: Mon, 6 Nov 1989 13:32:02 PST
From: "Claire E. Stager" <stager@Polya.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Cc: jones@Polya.Stanford.EDU, stager@Polya.Stanford.EDU
Subject: re: CS323 Winter Quarter
In-Reply-To: Your message of 03 Nov 89 1748 PST
Message-Id: <CMM.0.88.626391122.stager@Polya.Stanford.EDU>
Looks like the plan won't work....
VTSS160 is scheduled for TTh 1:15-2:30.
How about teaching CS323 TTh 4:15-5:30 in Terman 156? (This slot is the
only other TTh one available.) That would mean two afternoon classes for
you, instead of an 11:00 and a 1:15. What do you think?
Thanks.
Claire
∂06-Nov-89 1459 stager@Polya.Stanford.EDU re: CS323 Winter Quarter
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Nov 89 14:59:16 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA13660; Mon, 6 Nov 89 14:59:00 -0800
Date: Mon, 6 Nov 1989 14:58:58 PST
From: "Claire E. Stager" <stager@Polya.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: CS323 Winter Quarter
In-Reply-To: Your message of 06 Nov 89 1410 PST
Message-Id: <CMM.0.88.626396338.stager@Polya.Stanford.EDU>
Great.
Thanks.
Claire
∂06-Nov-89 1515 MPS
138.41
∂06-Nov-89 1636 COLLEEN@SUWATSON.stanford.edu Message from Arcady Blinov
Received: from SUWATSON.stanford.edu by SAIL.Stanford.EDU with TCP; 6 Nov 89 16:36:29 PST
Received: from suWatson.Stanford.EDU by SUWATSON.stanford.edu (IBM VM SMTP R1.2.1) with BSMTP id 7509; Mon, 06 Nov 89 16:37:10 PDT
Received: by SUWATSON (Mailer X1.25) id 7508; Mon, 06 Nov 89 16:37:09 PDT
Date: Mon, 06 Nov 89 16:29:44 PDT
From: COLLEEN@SUWATSON.stanford.edu
Subject: Message from Arcady Blinov
To: jmc@sail.stanford.edu
John,
This is a message from Arcady Blinov who is now visiting from Moscow.
Please reply to me (Colleen Crangle) and I will pass on the message to
Arcady. Arcady arrived November 4th and will be leaving November 17.
He remembers that you asked him to give a talk on elections from the
Soviet Academy. He is happy to do this any time (but not Nov. 16 4pm
when he will be giving a lecture here at Ventura Hall on the logical
form of action sentences).
His hotel telephone number is 322-7666, room number 11, if you'd like to
phone him early morning or in the evening.
Colleen
∂06-Nov-89 1748 alex@jessica.Stanford.EDU re: Thesis news AND questions
Received: from jessica.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Nov 89 17:47:56 PST
Received: by jessica.Stanford.EDU (5.59/25-eef) id AA18950; Mon, 6 Nov 89 17:48:44 PDT
Date: Mon, 6 Nov 1989 17:48:43 PST
From: Alex Bronstein <alex@jessica.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: Thesis news AND questions
In-Reply-To: Your message of 06 Nov 89 0939 PST
Message-Id: <CMM.0.88.626406523.alex@jessica.Stanford.EDU>
OK, will do.
Alex
∂06-Nov-89 1802 alex@jessica.Stanford.EDU re: Thesis news AND questions
Received: from jessica.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Nov 89 18:02:08 PST
Received: by jessica.Stanford.EDU (5.59/25-eef) id AA19153; Mon, 6 Nov 89 18:02:55 PDT
Date: Mon, 6 Nov 1989 18:02:54 PST
From: Alex Bronstein <alex@jessica.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: Thesis news AND questions
In-Reply-To: Your message of 06 Nov 89 0939 PST
Message-Id: <CMM.0.88.626407374.alex@jessica.Stanford.EDU>
Prof. McCarthy,
Would it be OK if the thesis copy you get is velo-bound? (as opposed
to spiral bound which would cut 12 hours of extra work I could do on the
draft because I would have to give it to a shop)
Alex
∂06-Nov-89 2258 eaf@sumex-aim.stanford.edu [MAILER-DAEMON@sumex-aim.stanford.edu (Mail Delivery Subsystem) :
Received: from sumex-aim.stanford.edu by SAIL.Stanford.EDU with TCP; 6 Nov 89 22:57:54 PST
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AA00332; Mon, 6 Nov 89 22:59:45 PST
Date: Mon, 6 Nov 1989 22:59:44 PST
From: Edward A. Feigenbaum <eaf@sumex-aim.stanford.edu>
To: jmc@sail.stanford.edu
Subject: [MAILER-DAEMON@sumex-aim.stanford.edu (Mail Delivery Subsystem) :
Returned mail: User unknown ]
Message-Id: <CMM.0.88.626425184.eaf@sumex-aim.stanford.edu>
Return-Path: <MAILER-DAEMON@sumex-aim.stanford.edu>
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AB29821; Mon, 6 Nov 89 22:50:12 PST
Date: Mon, 6 Nov 89 22:50:12 PST
From: MAILER-DAEMON@sumex-aim.stanford.edu (Mail Delivery Subsystem)
Full-Name: Mail Delivery Subsystem
Subject: Returned mail: User unknown
Message-Id: <8911070650.AB29821@sumex-aim.stanford.edu>
To: eaf@sumex-aim.stanford.edu
----- Transcript of session follows -----
Connected to polya.Stanford.EDU:
>>> RCPT To:<jmc@polya.stanford.edu>
<<< 550 <jmc@polya.stanford.edu>... User unknown
550 jmc@cs... User unknown
----- Unsent message follows -----
Return-Path: <eaf@sumex-aim.stanford.edu>
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AA29808; Mon, 6 Nov 89 22:50:12 PST
Date: Mon, 6 Nov 1989 22:50:12 PST
From: Edward A. Feigenbaum <eaf@sumex-aim.stanford.edu>
To: ksl@sumex-aim.stanford.edu
Cc: jmc@Polya.Stanford.EDU, nilsson@Polya.Stanford.EDU
Subject: National Medal of Science/ History
Message-Id: <CMM.0.88.626424612.eaf@sumex-aim.stanford.edu>
Our friend and close colleague, Joshua Lederberg (now President of
Rockefeller
University) was awarded on Oct. 18 the National Medal of Science by President
Bush. Lederberg, of course, is also a winner of the Nobel Prize for Medicine.
Josh Lederberg was co-founder of the Heuristic Programming Project, (which
grew into the Knowledge Systems Laboratory), and co-inventor of the expert
systems concept. He was also a leader of our effort to offer excellent
computer
resources for ourselves and the national AI-in-Medicine group (SUMEX-AIM).
His citation for the National Medal of Science cited three things:
a. <words relating to his biological discoveries that led to the Nobel Prize>
b. "his seminal research in artificial intelligence in biochemistry and
medicine"
c. <words relating to his extensive public service>
Additional historical notes:
The co-investigator with Lederberg, Buchanan, and me on the DENDRAL
project, and for many years the Principal Investigator of the DENDRAL
Project,
Carl Djerassi, is also a winner of the National Medal of Science. Djerassi
has many other important scientific contributions, the most important of
which is
the first birth control pill.
A winner of both the National Medal of Science and the National Medal of
Technology (which was also awarded on Oct. 18) is Stanley Cohen, co-inventor
of
the recombinant DNA method that launched modern genetic engineering. Stan
Cohen was co-investigator with Buchanan on the MYCIN project and one of Ted
Shortliffe's mentors on Ted's prize winning thesis work.
Finally, another National Medal of Science winner, Herbert Simon of Carnegie-
Mellon, is one of the "founding fathers" of AI (also a Nobel Prize winner,
in
Economics).
(p.s. please excuse the poor formatting of this message)
∂07-Nov-89 0717 beeson@ucscd.UCSC.EDU re: Shrinking Lattice Polyhedra
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 7 Nov 89 07:17:31 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA26878; Tue, 7 Nov 89 07:17:39 -0800
Date: Tue, 7 Nov 89 07:17:39 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8911071517.AA26878@ucscd.UCSC.EDU>
To: JMC@SAIL.Stanford.EDU, beeson@ucscd.UCSC.EDU
Subject: re: Shrinking Lattice Polyhedra
OK, I'll send you a copy of the triangle paper.
In addition to missint Matijasevich's talks, I missed YOURS, on
Elephant 2000, which sounded quite interesting from the description.
If there is anything written about it, please send me a copy.
And what is the story of the rather strange name "Elephant 2000"?
Regarding Hilbert's 10th you probably already know the work of
Chaitin, presented in his nice little book 'Algorithmic Information
Theory'. If not, have a look: you would be interested (if for
no other reason) because he makes a THEORETICAL use of LISP.
Regarding the earthquake: I was jolted out of self-pity by
the task of writing letters of recommendation for Kushner, who is
living in Philadelphia with the help of Jewish refugee organizations,
and has no job to commute to and no house with a foundation that can
crack. (Kushner is the successor to Markov and led the school of
constructive mathematics at Steklov Institute before emigrating.)
∂07-Nov-89 0827 MPS
Chris Anderson, Scientist Newspaper, called again about
the AI question.
202 857 0355
∂07-Nov-89 1021 VAL Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU, su-events@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
REASONING IN KNOWLEDGE SYSTEMS
Michael Gelfond
University of Texas at El Paso
Monday, November 13, 3:15pm
MJH 252
We will present an approach to building a theory of a reasoning
agent trying to behave intelligently in the world. The world view
of such an agent will be represented by a knowledge system K - a
quadruple consisting of:
1. A theory T containing current information about the world.
2. A set O of formulae representing new observations.
3. A set H of formulae representing possible explanations of new
observations from O.
4. A relation < defining the relative plausibility of different
sets of possible hypotheses from H.
In this setting the knowledge from K can be used for answering
queries about the world and beliefs of the agent as well as for
finding explanations to the new observations. We will present
formalisations of both of these types of reasoning for knowledge
systems with T represented by logic programs expanded by classical
negation and disjunction, and demonstrate how this formalisation
can be used to solve problems in diagnostic reasoning, logic
programming and disjunctive databases which seem to cause
difficulties for some other approaches.
∂07-Nov-89 1434 COLLEEN@SUWATSON.stanford.edu re: Message from Arcady Blinov
Received: from SUWATSON.stanford.edu by SAIL.Stanford.EDU with TCP; 7 Nov 89 14:34:08 PST
Received: from suWatson.Stanford.EDU by SUWATSON.stanford.edu (IBM VM SMTP R1.2.1) with BSMTP id 0198; Tue, 07 Nov 89 14:34:50 PDT
Received: by SUWATSON (Mailer X1.25) id 0197; Tue, 07 Nov 89 14:34:48 PDT
Date: Tue, 07 Nov 89 14:32:03 PDT
From: COLLEEN@SUWATSON.stanford.edu
Subject: re: Message from Arcady Blinov
To: John McCarthy <JMC@SAIL.Stanford.EDU>
In-Reply-To: Your message of 07 Nov 89 1151 PST
Unfortunately, Arkady's office (Ventura 29) has no phone.
But a message could be left for him through Pat's secretary, Laura
Tickle --- 3-3111.
Colleen
∂07-Nov-89 1435 pasquale%cs@ucsd.edu ucsd visit
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 7 Nov 89 14:35:02 PST
Received: from cassano.ucsd.edu by ucsd.edu; id AA26518
sendmail 5.61/UCSD-2.0-sun via SMTP
Tue, 7 Nov 89 14:34:51 -0800 for JMC@sail.stanford.edu
Received: by cassano.UCSD.EDU (3.2/UCSDPSEUDO.2)
id AA07646 for delivery to JMC@sail.stanford.edu; Tue, 7 Nov 89 14:34:39 PST
Date: Tue, 7 Nov 89 14:34:39 PST
From: pasquale%cs@ucsd.edu (Joseph Pasquale)
Message-Id: <8911072234.AA07646@cassano.UCSD.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: ucsd visit
Cc: christos@cs.UCSD.EDU
Dear Professor McCarthy,
I am helping Christos organize your visit to UCSD this coming
Monday. I will give you a call tomorrow (Wednesday) to check with you
if there is anything I can do to help in making arrangements. The main
thing I need right now is the title and abstract of your talk. If you
could send this to me electronically, I would appreciate it very much.
Thanks.
-Joseph Pasquale
∂08-Nov-89 0851 MPS Bureaucratic Meetings
Hi,
I will be at a staff meeting at 9:00 today and following that
a safety committee meeting at 10:00.
Pat
∂08-Nov-89 1029 VAL Re: visit
∂08-Nov-89 1020 bledsoe@cs.utexas.edu Re: visit
Received: from cs.utexas.edu by SAIL.Stanford.EDU with TCP; 8 Nov 89 10:18:31 PST
Received: from alfalfa.cs.utexas.edu by cs.utexas.edu (5.59/1.44)
id AA20711; Wed, 8 Nov 89 12:15:03 CST
Posted-Date: Wed, 8 Nov 1989 12:11:56 CST
Received: by alfalfa.cs.utexas.edu (5.59/1.4-Client)
id AA15530; Wed, 8 Nov 89 12:11:57 CST
Date: Wed, 8 Nov 1989 12:11:56 CST
From: Woody Bledsoe <bledsoe@cs.utexas.edu>
To: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
Cc: boyer@cli.com, apt@cs.utexas.edu
Subject: Re: visit
In-Reply-To: Your message of 26 Oct 89 1512 PDT
Message-Id: <CMM.0.86.626551916.bledsoe@alfalfa.cs.utexas.edu>
Vladimir,
There was a lot of interest in you after your visit last week, and apparantly
also in Philosophy.
As I said before you left, I will be in touch with you when we have reviewed
the situation in the Budget council, and have decided what we want to do.
PLEASE check with me before you seriously consider something else, and I
will definitely keep you informed of our possibilities.
Best regards, Woody
∂08-Nov-89 1043 CLT qlisp
To: weening@Gang-of-Four.Stanford.EDU, JMC@SAIL.Stanford.EDU
rpg will be here sometime between 1 and 1:30 to talk about qlisp
∂08-Nov-89 1055 Mailer re: The education president
Received: from rml2.sri.com by SAIL.Stanford.EDU with TCP; 8 Nov 89 10:55:26 PST
Date: Wed 8 Nov 89 10:53:29-PST
From: Ric Steinberger <RIC@RML2.SRI.COM>
Subject: re: The education president
To: JMC@SAIL.STANFORD.EDU
Cc: RIC@RML2.SRI.COM, comments@KL.SRI.COM, su-etc@SAIL.STANFORD.EDU
Message-ID: <626554409.350000.RIC@RML2.SRI.COM>
In-Reply-To: <o8Xtk@SAIL.Stanford.EDU>
Mail-System-Version: <VAX-MM(246)+TOPSLIB(136)@RML2.SRI.COM>
> From: John McCarthy <JMC@SAIL.Stanford.EDU>
> Steinberger says:
>> And thus we perpetuate a crisis where the unborn
>> citizens of the future will have all the more reason to
>> damn their ancestors of the twentieth century for their
>> self-centered short-sightedness.
> I don't think our descendants with their robot servants and
> 20 room houses and 30 weeks of work per year will have any
> more reason to damn us than we have to damn our ancestors
> who worked 12 hours a day and saw half of their children
> die.
Robot servants? 20 room houses? 30 weeks of work per year? Pardonez-moi,
but what does any of this speculation have to do with global warming and
the failure of the US government to act? Is there any evidence outside
of science fiction novels that presents the above as a highly-probable,
much less desirable, future of human society?
> Before that he [Ric S.] says,
>> And so we postpone once again the day where significant
>> environmental laws are passed.
> The environmentalists world-wide are on a ``power trip''.
> Their leaders are on a real power trip and the followers
> are on a vicarious power trip. What's the evidence?
> They are more interested in getting ``significant environmental
> laws'' than in deciding what to do. On CO2 they propose various
> tokens, but can't bring themselves to agree that nuclear energy
> is the solution to the electricity problem, because it would
> spoil their "no enemy on the left" attitude toward their
> extremists.
What does ``power trip'' seem to mean here? It means that environmentalists
fail to embrace nuclear power production as an ecological holy grail.
> The CO2 problem really does require more study before expensive (many
> trillion dollar) action is taken. That's right, trillion not billion.
> There are at least two reasons. First the effect of increased CO2 is
> unknown because the current models don't take into account several
> relevant effects, e.g. they are only starting to take into account the
> effect on cloudiness. Second the largest sources and sinks of CO2
> seem to be geological (see a Scientific American article of the last
> year), namely volcanoes and deposition of carbonates on ocean bottoms
> followed by subduction.
We do know that 25% of the man-made CO2 that enters the atmosphere originates
in the US. We know that if there were the political initiative to reduce
that amount in this country, and that other countries would be cooperating
to reduce their emissions, that the net result could be considerably less
CO2 being emitted into the atmosphere. We know that this would be desirable
in terms of global warming effects.
> However, we really ought to be taking a quite different long term
> attitude towards the atmosphere and the global temperature
> balance. If it is really true that small amounts of various
> substances in the upper atmosphere affect the temperature
> equilibria, then this means that we can control the
> temperature by adding substances to the upper atmosphere.
> In the long run we will need to do this in order to prevent
> the next ice age. Our rich descendants will figure it out.
This seems to me to be a somewhat arrogant attitude. [I use the word
attitude because I have nothing against JMC personally or professionally;
it is only the opinion expressed that I wish to take issue with]. To
assume that we can somehow play chemist-in-the-sky, adding a little of
this, removing a bit of that, stirring it all around, while continuing
to produce greenhouse gases on the earth's surface is to assume that
we know far more than we can possibly know. It is also a highly
reckless attitude where the stakes are very large.
To claim that "our rich descendents will figure it out" is equivalent
to stating that future residents of Love Canal will figure out how to
live there.
> The environmentalist anti-earthman attitude is that Nature
> is benign and that people are intruders in a wilderness
> planet. Both propositions are false. The old image of
> conquering nature is better, although regarding the
> world as a garden to be rearranged to suit humanity
> is still a better analogy.
I don't know any environmentalist that would agree with the above
representation of environmental philosophy. I like to think that one of
the guiding principles of the environmental movement is balance: establishing
a balance of man in nature. Neither "back to the land" nor "holding
dominion over the earth to suit humanity" are the wisest attitudes.
Where the balance is to be is the subject of considerable debate, for
sure. At present I believe the scales need tipping in the direction
of restraint and a re-evaluation of various industrial policies.
-ric steinberger
-------
∂08-Nov-89 1238 pasquale%cs@ucsd.edu re: ucsd visit
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 8 Nov 89 12:37:58 PST
Received: from cassano.ucsd.edu by ucsd.edu; id AA14076
sendmail 5.61/UCSD-2.0-sun via SMTP
Wed, 8 Nov 89 12:37:41 -0800 for JMC@sail.stanford.edu
Received: by cassano.UCSD.EDU (3.2/UCSDPSEUDO.2)
id AA16843 for delivery to JMC@sail.stanford.edu; Wed, 8 Nov 89 12:37:32 PST
Date: Wed, 8 Nov 89 12:37:32 PST
From: pasquale%cs@ucsd.edu (Joseph Pasquale)
Message-Id: <8911082037.AA16843@cassano.UCSD.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: re: ucsd visit
Cc: pasquale@cs.UCSD.EDU
Professor McCarthy,
Thanks for the information. Here are some details. Please let
me know if any of them are unsuitable, and I will be happy to make any
changes.
Plane: I see you've taken care of this already. Either Christos or myself
(most likely Christos as this is what he told me) will come and get
you at the airport.
Hotel: You'll be staying at the Residence Inn in La Jolla. You really
don't need to worry about transportation to and from the hotel
as we will take care of this.
Sched: The main event on Monday is your talk, which will be held at the
San Diego Supercomputer Center Auditorium, and the time is 4:00.
There are other events which we are currently fixing times for, such
as a round table discussion with yourself and some of the faculty
members here, and various meetings that Christos asked to be set up
for you. Christos will come and pick you up on Monday morning, say
between 9 and 10.
The abstract of your talk sounds most interesting, and we are all excited
about your visit. I personally am looking forward to meeting you. Please
rest assured that we will do our best to make your visit enjoyable.
-Joseph Pasquale
∂08-Nov-89 1347 POSTMASTER%McMaster.CA@Forsythe.Stanford.EDU Policies
Received: from Forsythe.Stanford.EDU by SAIL.Stanford.EDU with TCP; 8 Nov 89 13:46:58 PST
Received: by Forsythe.Stanford.EDU; Wed, 8 Nov 89 13:45:56 PST
Date: Wed, 8 Nov 89 16:34 EDT
From: "Fred Whiteside, PostMaster" <POSTMASTER%McMaster.CA@Forsythe.Stanford.EDU>
Subject: Policies
To: jmc@sail.stanford.edu
X-VMS-To: IN%"jmc@sail.stanford.edu"
Hi,
I have been trying to get a copy of the text of a report from
the Stanford Faculty Senate Committee on Libraries regarding the
prohibition of the rec.humor.funny newsgroup at Stanford. Your name
was suggested to me by June Genis:
> You might try John McCarthy,
> jmc@sail.stanford.edu, who has the faculty member responsible for
> getting the Senate to review the issue.
> /June
The reason for this request is that we currently have no
position on news here at McMaster, and a good reference can go a long
way.
Sorry if this inconveniences you ...
Thanks in advance,
Fred Whiteside POSTMAST@MCMASTER.BitNet
McMaster NetNorth Postmaster POSTMASTER@McMaster.CA
Development Analyst WHITESID@SSCvax.McMaster.CA
McMaster University ...!uunet!utai!utgpu!maccs!fred
Hamilton, Canada
∂08-Nov-89 1624 rick@hanauma.stanford.edu global warming seminar tonight 7:30 in Cubberly
Received: from hanauma.stanford.edu by SAIL.Stanford.EDU with TCP; 8 Nov 89 16:24:06 PST
Received: by hanauma.stanford.edu (5.51/7.0)
id AA17875; Wed, 8 Nov 89 16:23:40 PST
Date: Wed, 8 Nov 89 16:23:40 PST
From: rick@hanauma.STANFORD.EDU (Richard Ottolini)
Message-Id: <8911090023.AA17875@hanauma.stanford.edu>
To: JMC@sail
Subject: global warming seminar tonight 7:30 in Cubberly
According to an ad in today's Daily.
I may go to make sure that the panel doesn't only consider a subset of the facts.
∂08-Nov-89 1722 POSTMASTER%McMaster.CA@Forsythe.Stanford.EDU RE: re: Policies
Received: from Forsythe.Stanford.EDU by SAIL.Stanford.EDU with TCP; 8 Nov 89 17:22:34 PST
Received: by Forsythe.Stanford.EDU; Wed, 8 Nov 89 17:21:48 PST
Date: Wed, 8 Nov 89 20:25 EDT
From: "Fred Whiteside, PostMaster" <POSTMASTER%McMaster.CA@Forsythe.Stanford.EDU>
Subject: RE: re: Policies
To: JMC@SAIL.Stanford.EDU
X-VMS-To: IN%"JMC@SAIL.Stanford.EDU"
Hi,
Thanks for the quick reply. I had pieced together a rough
approximation of that sequence of events from news.groups. What I was
wondering was if there is an electronic copy of:
> 11. Committee on Libraries says that electronic media shouldn't be
> treated differently from print media, and the policy is universality
> tempered only by cost.
or something similar. Our news feed has just yesterday moved
due to machine tourbles on our old feed, and the new feed has
generated newgroups for alt.sex.bestiality, alt.drugs, et al.
Previously we only received alt.flame from the alt heirarchy, and of
course, the Vice-President (Computing) has just discovered news and
people are asking if this is an appropriate thing for the university
to be doing, etc.
I'm hoping to have some well-respected University opinions
collected indicating that news is similar to print media and that
censorship of news is a bad thing, generally.
Hopefully, such arguments are self-evident and I worry
needlessly. On the other hand, I *have* been here some 15 years and
wouldn't mind having prepared arguments ....
Anyway, if such a document is available, I'd appreciate a
pointer to it. And thanks for your time in any event.
Cheers,
Fred Whiteside POSTMAST@MCMASTER.BitNet
McMaster NetNorth Postmaster POSTMASTER@McMaster.CA
Development Analyst WHITESID@SSCvax.McMaster.CA
McMaster University ...!uunet!utai!utgpu!maccs!fred
Hamilton, Canada
∂09-Nov-89 0812 Mailer re: The education president
Received: from rml2.sri.com by SAIL.Stanford.EDU with TCP; 9 Nov 89 08:12:21 PST
Date: Thu 9 Nov 89 08:10:22-PST
From: Ric Steinberger <RIC@RML2.SRI.COM>
Subject: re: The education president
To: JMC@SAIL.STANFORD.EDU
Cc: RIC@RML2.SRI.COM, comments@KL.SRI.COM, su-etc@SAIL.STANFORD.EDU
Message-ID: <626631022.350000.RIC@RML2.SRI.COM>
In-Reply-To: <1S8#q4@SAIL.Stanford.EDU>
Mail-System-Version: <VAX-MM(246)+TOPSLIB(136)@RML2.SRI.COM>
> Ric addresses some of the issues but not all. Unaddressed are the
> specific reasons for delaying actions - uncertainty about the CO2
> cycle and its effects on temperature.
Time limitations prevent me from responding to all of JMC's issues, and
thus I will attempt to address those that seem most important to me.
> Ric doubts that our descendants will be richer. Note the date
> on the following quotation.
> [quotation removed, present in JMC's posting].
My contention is that if present population trends continue, and if the
source of much of our energy remains fossil fuels, *and* if we maintain the
same types of industrial processes and economic structures that have
resulted in the depletion of vast reserves of fossil fuels, that our
descendents will, in fact, have less material wealth than we now have. I am
not attempting to place any particular value judgement of the desirability
of such a condition.
I suspect that if there had been only 20% of the petroleum, natural
gas and coal within the earth's surface than have been discovered, Western
societies would have devoted massive efforts to the development of
solar thermal and photovaltaic energies. In fairness to JMC, it is likely
that significant efforts to develop nuclear energy would also have been
made (unless there had been no Uranium within the earth's crust).
> Ric has misunderstood me and probably also misunderstood the U.S.
> and Japanese positions at the conference. It isn't proposed
> to leave the CO2 problem to our descendants, but to do the research
> needed to understand it.
This is the same old chestnut that the Reagan administration kept using
when attempting to justify reasons for failing to support legislation
that might have curbed acid rains. No one would doubt that further
researches would improve our understanding of the interaction of
CO2 and the global warming phenomena. However, it is very clear that
continued increasing production of this gas does have comprensible,
though not deterministically-predictable, effects on the earth's climate.
It is because we know (based on existing research and hydrogeologic
models) that failure to reduce CO2 emissions will have potentially severe
effects on most life on earth that the time to act is now.
> Since Ric's discussion is purely qualitative, he doesn't address
> my estimate of trillion dollar costs.
I didn't address JMC's estimate. I don't have a better number and JMC may
be right. If so, this number is roughly the value that was assigned as
a cost for deploying SDI by various budgeting agencies a few years ago.
A large number, to be sure.
> Ric refers to my ideas as science fiction. They aren't intended
> as fiction but as proposals for study. There inclusion in these
> discussions has a lesser purpose. It is merely to show that
> taking a different attitude than that of the environmentalists
> is possible and necessary.
"Science fiction" was my reaction to JMC's remarks about our descendents living
in 20-room houses, served by robots, and working 30-hour weeks. I have tried
to treat his other statements with seriousness and respect.
> Humanity is likely to suffer greatly if we don't eventually think
> ambitiously about how to rearrange our planet. However, the
> consequences are unlikely to emerge in my grandson's lifetime, and the
> understanding to decide what to do and the resources to do it will be
> much greater then.
Sounds like Buckminster Fuller to me. %-) I would maintain, however,
that much of the ecological destruction we have encountered and continue
to face are a direct result of our "ambitious rearrangement of our planet."
The fault is not that we have been ambitious, but rather that we have
equated "the good life" with *maximization* of material acquisition
and consumption. And thus, we have all too often, assumed that none
of our technological endeavors, regardless of scale, would ever have any
adverse effects.
What I believe we should ambitiously think about is, "What is the
good life?" and, "How shall the diverse peoples of earth cooperatively
work to live it while peacefully resolving their differences?" And
these questions will surely lead to, "What kinds of technologies and
industrial policies are conducive to such an effort?" [Not a bad
choice of topic for a college course? Any takers?]
> Ric says that no environmentalist takes the positions I
> ascribe to them. Here are some quotes from my bad guys file.
JMC is correct. Many of the quotations he provided were examples of
extremism. I don't believe those isolated positions respresent mainstream
environmental thinking today. I have read material from all of the
writers, and while I have agreed with much of it, cannot support every
position. I suspect I would have little difficulty finding examples
of extreme right-wing positions on environmentalism that JMC would not
support.
And so what seems to remain of the discussion between JMC and me is a
call for balance, and a dispute over where balance lies.
JMC: Is this a fair statement?
regards,
ric steinberger
-------
∂09-Nov-89 0913 beeson@ucscd.UCSC.EDU Elephant
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 9 Nov 89 09:13:27 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA27021; Thu, 9 Nov 89 09:13:37 -0800
Date: Thu, 9 Nov 89 09:13:37 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8911091713.AA27021@ucscd.UCSC.EDU>
To: JMC@SAIL.Stanford.EDU, beeson@ucscd.UCSC.EDU
Subject: Elephant
I have seen the abstract. Can it be so terribly difficult to implement
that we have to wait 16 years? Please send me the full paper.
What are you plans for implementing the language?
(oops, "your plans", not "you plans").
∂09-Nov-89 0940 beeson@ucscd.UCSC.EDU cute little proof
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 9 Nov 89 09:40:35 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA27330; Thu, 9 Nov 89 09:40:27 -0800
Date: Thu, 9 Nov 89 09:40:27 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8911091740.AA27330@ucscd.UCSC.EDU>
To: blufox@ucscd.UCSC.EDU, jmc@sail.stanford.edu,
joel@pericles.tcp.cs.umass.edu
Subject: cute little proof
The following allegedly was found on the restroom wall in the Princeton
Math Department. Proof of the existence of infinitely many primes by
topology: Topologize the integers by taking the open sets to be
the unions of arithmetic progressions, i.e. the arithmetic progressions
are a basis for the topology. (You need gcd's to see this is ok.)
The set of numbers not divisible by a particular prime p is open, being
the union of p-1 arithmetic progressions (the numbers congruent to
1, 2,..., p-1 mod p). The set of numbers not divisible by any prime
is just {1}, which is not open. But if there were finitely many primes,
it would be the intersection of open sets, hence open. QED.
∂09-Nov-89 1250 JJW SAIL DD monitors
To: ARK, CLT, JMC
Tom D. told me he is trying to clean up the machine room, and there
are a lot of SAIL DD monitors that are no longer being used. What
should we do with them?
∂09-Nov-89 1342 weening@Gang-of-Four.Stanford.EDU Re: SAIL DD monitors
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 9 Nov 89 13:41:50 PST
Received: from LOCALHOST by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA09387; Thu, 9 Nov 89 13:43:25 -0800
Message-Id: <8911092143.AA09387@Gang-of-Four.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: Re: SAIL DD monitors
In-Reply-To: Your message of 09 Nov 89 13:37:00 -0800.
<1W9qnG@SAIL.Stanford.EDU>
Date: Thu, 09 Nov 89 13:43:23 PST
From: Joe Weening <weening@Gang-of-Four.Stanford.EDU>
Marty should be able to arrange this. (Or say if it requires extra
wiring, which might be the case.)
∂09-Nov-89 1413 ME extra DD
∂09-Nov-89 1343 JMC dds
I understand there are extras. If there is an available desk stand and
if no new building wiring is required, I'd like another right next to the
one I have. I want to experiment with its usefulness before SAIL goes
away.
ME - We can probably do that. There are certainly spare keyboards and
monitors. You can even control two monitors (screens) from one keyboard.
∂09-Nov-89 1532 stager@Polya.Stanford.EDU re: CS323 Winter Quarter
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 9 Nov 89 15:31:53 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA08965; Thu, 9 Nov 89 15:31:39 -0800
Date: Thu, 9 Nov 1989 15:31:39 PST
From: "Claire E. Stager" <stager@Polya.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: CS323 Winter Quarter
In-Reply-To: Your message of 06 Nov 89 1410 PST
Cc: stager@Polya.Stanford.EDU
Message-Id: <CMM.0.88.626657499.stager@Polya.Stanford.EDU>
And yet again....
We seem to be endlessly in quest of the optimal solution.
There have been complaints about my last plan (i.e.--schedule CS323 for
TTh 4:15) VTSS has agreed to move your VTSS 160 to TTh 11-12:15 so that
we can put CS323 TTh 1:15-2:30 in Terman 156. This would put you back at
your original schedule, but with the times for VTSS and CS switched.
What about it? Willing to try one more time? (I do believe this will
be the last pass.)
Thanks again.
Claire
∂10-Nov-89 0153 cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK E-mail address (cliff jones)
Received: from NSFnet-Relay.AC.UK by SAIL.Stanford.EDU with TCP; 10 Nov 89 01:53:41 PST
Received: from sun.nsfnet-relay.ac.uk by vax.NSFnet-Relay.AC.UK
via Janet with NIFTP id aa02798; 10 Nov 89 9:36 GMT
Date: Thu, 9 Nov 89 18:30:42 GMT
From: cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK
Message-Id: <8911091830.AA28642@ipse2pt5.cs.man.ac.uk>
To: williams@almaden.ibm.com, lamport@src.dec.com, rushby@csl.sri.com,
dan%ora.on.ca@NSFnet-Relay.AC.UK, liskov@xx.lcs.mit.edu,
liskov@xx.lcs.mit.edu, pamela%allegra.uucp@NSFnet-Relay.AC.UK,
kh%ddc.uucp@NSFnet-Relay.AC.UK, cbn%ddc.uucp@NSFnet-Relay.AC.UK,
jsp%ddc.uucp@NSFnet-Relay.AC.UK, sup%ddc.uucp@NSFnet-Relay.AC.UK,
sestoft%diku.dk@NSFnet-Relay.AC.UK, ph%ddc.uucp@NSFnet-Relay.AC.UK,
sp%ddc.uucp@NSFnet-Relay.AC.UK, weber%unipas.uucp@NSFnet-Relay.AC.UK,
stengel%unipas.uucp@NSFnet-Relay.AC.UK,
wilhelm%stzdo.uucp@NSFnet-Relay.AC.UK,
amir <@cunyvm.cuny.edu:amir@wisdom.bitnet>,
rww%ibm-b.rutherford.ac.uk@NSFnet-Relay.AC.UK,
ianh%prg.oxford.ac.uk@NSFnet-Relay.AC.UK, ttp@kestrel.edu,
rco%ipse2pt5.cs.man.ac.uk@NSFnet-Relay.AC.UK,
pdm%daimi.uucp@NSFnet-Relay.AC.UK,
adafd%icnucevm.bitnet@NSFnet-Relay.AC.UK, Luca@src.dec.com,
MBJ@almaden.ibm.com, waite@boulder.colorado.edu,
lloyd@boulder.colorado.edu, meersman%htikub5.bitnet@NSFnet-Relay.AC.UK,
dbeech%oracle.com@NSFnet-Relay.AC.UK, scherlis@vax.darpa.mil,
pamela%allegra.uucp@NSFnet-Relay.AC.UK,
HERZOG%ds0lilog.earn@NSFnet-Relay.AC.UK,
neuhold%darmstadt.gmd.dbp.de@NSFnet-Relay.AC.UK,
eike%infhil.uucp@NSFnet-Relay.AC.UK,
BACKRJ%finabo.bitnet@NSFnet-Relay.AC.UK, jmc@sail.stanford.edu
Subject: E-mail address (cliff jones)
My official e-mail address is now:
cliff@uk.ac.man.cs
I also suspect that I have lost some messages - so, if you're awaiting
a reply to something you sent a week or more ago, please resend.
cliff jones
∂10-Nov-89 0907 bhr@sumex-aim.stanford.edu CS300 Winter Quarter
Received: from sumex-aim.stanford.edu by SAIL.Stanford.EDU with TCP; 10 Nov 89 09:07:17 PST
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AA02823; Fri, 10 Nov 89 09:08:24 PST
Date: Fri, 10 Nov 1989 9:08:22 PST
From: Barbara Hayes-Roth <bhr@sumex-aim.stanford.edu>
To: CAB@sail.stanford.edu, binford@coyote.stanford.edu,
cheriton@pescadero.stanford.edu, Gail@sol-margret.stanford.edu,
dill@amadeus.stanford.edu, feigenbaum@SUMEX-AIM.Stanford.EDU,
rwf@sail.stanford.edu, RPG@sail.stanford.edu,
Genesereth@score.stanford.edu, Ginsberg@polya.stanford.edu,
goldberg@polya.stanford.edu, golub@patience.stanford.edu,
ag@pepper.stanford.edu, halpern@ibm.com,
hayes-roth@SUMEX-AIM.Stanford.EDU, jlh@amadeus.stanford.edu,
Herriot@score.stanford.edu, Iwasaki@SUMEX-AIM.Stanford.EDU,
ARK@sail.stanford.edu, OK@coyote.stanford.edu, DEK@sail.stanford.edu,
lam@mojave.stanford.edu, latombe@coyote.stanford.edu,
val@sail.stanford.edu, linton@amadeus.stanford.edu,
zm@sail.stanford.edu, mayr@polya.stanford.edu, jmc@sail.stanford.edu,
ejm@sierra.stanford.edu, miller@kl.sri.com, jcm@polya.stanford.edu,
bmoore@ai.sri.com, nii@SUMEX-AIM.Stanford.EDU,
nilsson@tenaya.stanford.edu, oliger@pride.stanford.edu,
pratt@jeeves.stanford.edu, ALS@sail.stanford.edu,
shoham@score.stanford.edu, singh@score.stanford.edu,
shortliffe@SUMEX-AIM.Stanford.EDU, tobagi@sierra.stanford.edu,
ullman@score.stanford.edu, waldinger@ai.sri.com,
Wiederhold@SUMEX-AIM.Stanford.EDU, wilf@score.stanford.edu,
winograd@csli.stanford.edu
Cc: bhr@sumex-aim.stanford.edu
Subject: CS300 Winter Quarter
Message-Id: <CMM.0.88.626720902.bhr@sumex-aim.stanford.edu>
We have decided to continue CS300, the Faculty Lecture Series, during
Winter Quarter. It will be held on Thursdays at 4:15-5:30pm in
room 380-380W. If you would like to give a lecture, please suggest
a couple of alternative Thursdays that fit your schedule.
Thanks,
Barbara Hayes-Roth
∂10-Nov-89 1119 paek@Neon.Stanford.EDU Nonmonotonic Reasoning class
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 10 Nov 89 11:19:35 PST
Received: by Neon.Stanford.EDU (5.61/25-eef) id AA17009; Fri, 10 Nov 89 11:19:25 -0800
Date: Fri, 10 Nov 1989 11:19:24 PST
From: Eunok Paek <paek@Neon.Stanford.EDU>
To: jmc@sail
Subject: Nonmonotonic Reasoning class
Message-Id: <CMM.0.88.626728764.paek@Neon.Stanford.EDU>
Hello.
I would like to know a little bit about the class you're going to
teach next term. I am thinking of doing TA work for your class (of
course, if you need one). I will be very grateful if you let me know
what your course plan is like (such as content of lectures, homeworks,
exams and/or papers).
-Eunok Paek
∂10-Nov-89 1152 winograd@loire.stanford.edu AI QUAL SCHEDULE - ABOUT TO BE MADE OFFICIAL
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 10 Nov 89 11:52:05 PST
Received: by loire.stanford.edu (5.59/25-eef) id AA18792; Fri, 10 Nov 89 11:51:13 PDT
Date: Fri, 10 Nov 89 11:51:13 PDT
Message-Id: <8911101951.AA18792@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: ai-qual-faculty@loire.stanford.edu
Subject: AI QUAL SCHEDULE - ABOUT TO BE MADE OFFICIAL
The following is my attempt to schedule committees meeting everyone's stated
constraints. There are two times, each with two students. Some of you
are on two committees, some only one. Let me know by next Tuesday if
there is a problem with this schedule, and then I will send it to the students.
On each committee there is a (C)hair, a (D)epth examiner and one (O)ther.
AI Qualifying Exam schedule, Wednesday morning, December 6, 1989
PRELIMINARY VERSION FOR APPROVAL
8:45 - 10:15
Becky Thomas: bthomas@neon - representation
C: Winograd D: Hayes O: Binford
David Ash: ash@sumex-aim - planning
C: LaTombe D: Nilsson O: McCarthy
10:30 - 12:00
Fangzhen Lin: lin@neon - nonmonotonic
C: LaTombe D: McCarthy O: Hayes
Ramanathan Guha: guha@sumex-aim - representation
C: Nilsson D: Ginsberg O: Hayes-Roth
--t
∂10-Nov-89 1405 MPS phone call
A call from Ron Paslucci came before I went to lunch. He
wants you to call. work 213-390-8600 - home is okay 213-378-7105
∂10-Nov-89 1424 ME failed mail returned
Usual problem:
The following message was undeliverable to the address(es) below
because the destination Host Name(s) are Unknown:
singh@sierra.UUCP
------- Begin undelivered message: -------
10-Nov-89 1330 Mailer re: Reunification of Germany [was Re: Thought for today]
To: singh@sierra.UUCP, su-etc@SAIL.Stanford.EDU
From: John McCarthy <JMC@SAIL.Stanford.EDU>
[In reply to message from singh@sierra.UUCP sent 10 Nov 89 17:29:09 GMT.]
That one's easy. I don't think we should oppose German
reunification if that's what the Germans on both sides of the
(former) Iron Curtain want. It certainly makes some people
nervous, but if the Germans are for it, a threat of war would be
required to prevent it. Anything less would merely make the
parties in Germany that might develop revanchist ideas more
inclined to get votes by peddling such ideas. It has become a moot
point anyway. Bush has said he doesn't care, Mitterand has said
he doesn't care and so has Shevardnadze or maybe Gorbachev himself.
Perhaps Galbraith is indeed a skeleton by now.
------- End undelivered message -------
∂10-Nov-89 1431 @loire.stanford.edu:hayes@kanga.parc.xerox.com AI QUAL SCHEDULE - ABOUT TO BE MADE OFFICIAL
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 10 Nov 89 14:31:21 PST
Received: from arisia.Xerox.COM by loire.stanford.edu (5.59/25-eef) id AA19466; Fri, 10 Nov 89 14:30:11 PDT
Received: from kanga.parc.Xerox.COM by arisia.Xerox.COM with SMTP
(5.61+/IDA-1.2.8/gandalf) id AA05252; Fri, 10 Nov 89 14:28:01 -0800
Received: by kanga.parc.xerox.com
(5.61+/IDA-1.2.8/gandalf) id AA00559; Fri, 10 Nov 89 14:30:23 PST
Message-Id: <8911102230.AA00559@kanga.parc.xerox.com>
Date: Fri, 10 Nov 89 14:30:23 PST
From: Pat Hayes <hayes@parc.xerox.com>
To: Winograd@csli.stanford.edu
Cc: ai-qual-faculty@loire.stanford.edu
In-Reply-To: Terry Winograd's message of Fri, 10 Nov 89 11:51:13 PDT <8911101951.AA18792@loire.stanford.edu>
Subject: AI QUAL SCHEDULE - ABOUT TO BE MADE OFFICIAL
Fine by me.
Pat
∂10-Nov-89 1525 MPS
Please read your mail for next Tuesday. There are some changes.
I think you will have to change your reservation.
Pat
∂10-Nov-89 1548 VAL Notes on Cross's visit
He needs simple examples explaining the essence of our work that will
help him design more impressive examples, illustrating potential applications.
Boehm seems to suspect that AI methods can be replaced by much simpler
software technology. It's important to explain the complexity of the AI
problem.
∂11-Nov-89 2311 larson@unix.sri.com Re: The education president
Received: from unix.sri.com by SAIL.Stanford.EDU with TCP; 11 Nov 89 23:11:06 PST
Received: by unix.sri.com (4.0/SMI-4.0)
id AA06413; Sat, 11 Nov 89 23:11:02 PST
Date: Sat, 11 Nov 89 23:11:02 PST
From: larson@unix.sri.com (Alan Larson)
Message-Id: <8911120711.AA06413@unix.sri.com>
To: JMC@SAIL.STANFORD.EDU
Subject: Re: The education president
Newsgroups: su.etc
In-Reply-To: <1S8#q4@SAIL.Stanford.EDU>
Reply-To: larson@unix.sri.com
Organization: SRI International, Menlo Park, CA
Cc:
In article <1S8#q4@SAIL.Stanford.EDU> you write:
>Finally, my position as to the direction of proper balance is the
>reverse of Ric's. I think environmentalism has gone to far and
>bears a nontrivial part of the responsibility for the fact that
>my children can't afford to buy the kind of house I could afford
>when I was their age. Bush has made too many concessions to the
>environmentalists already.
This is a problem I have been looking at as I have developed the urge
to buy a house. Of course, I can afford what I want, but at prices 3
or 4 years ago.
One of the problems I see is that the shortage of available places to
live near work - this forces people to face long, wasteful, commutes.
There is the energy cost (even with mass transit), the resulting
environmental costs, the waste of time from personal or productive use,
and probably others.
Even so, the prices of a decent home in thearea are so far out of line
with the rest of the world, that when I tell the real estate agent that
I want to be able to bicycle (or perhaps walk, if patient) to work,
but that prices over $400K are out, they look at me as if I just fell
off of the turnip truck.
3 bedroom houses in my area are going for over $700K, and the last 2
bedroom / 1 bath house I saw was $475K. So much for living in Menlo
Park.
The liberal approach seems to be to attempt to raise the price of
commuting in dollars or in inconvenience. Will nobody address the
real problem?
Actually, I don't think it is fair to blame environmental concerns
alone.
Alan
∂12-Nov-89 1341 perlis@cs.UMD.EDU Putnam's Theorem
Received: from yoohoo.cs.UMD.EDU by SAIL.Stanford.EDU with TCP; 12 Nov 89 13:40:57 PST
Received: by yoohoo.cs.UMD.EDU (5.61/UMIACS-0.9/04-05-88)
id AA21994; Sun, 12 Nov 89 16:40:50 -0500
Date: Sun, 12 Nov 89 16:40:50 -0500
From: perlis@cs.UMD.EDU (Don Perlis)
Message-Id: <8911122140.AA21994@yoohoo.cs.UMD.EDU>
To: jmc@sail.stanford.edu
Subject: Putnam's Theorem
John, I recently recaled our lengthy discussion when you
were last here, on the issue of intentionality/meaning
in an agnet's linguistic behavior. You had some rather
powerful argument that countered Putnam's Theorem.
Putnam (and others) have claimed that there can be
no unique tie between language and reality; you claimed
the opposite. I think that you might find it interesting to
take a look at Lakoff's discussion of Putnam, in Lakoff's
book "Women, FIre, and Dangerous Things". He presents
in almost axiomatic form the assumptions (requirements,
as I think he puts it) on which a theory of meaning is
to be based, and then pro$ proceeds to derive P's result.
Lakoff also discussed a varieyt of attempts to circumvent
P's result. It woul dbe interesting to know whether your
argument (based on information theory, as I recall) fits
one of those.
Best,
Don
∂12-Nov-89 1414 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
REASONING IN KNOWLEDGE SYSTEMS
Michael Gelfond
University of Texas at El Paso
Monday, November 13, 3:15pm
MJH 252
We will present an approach to building a theory of a reasoning
agent trying to behave intelligently in the world. The world view
of such an agent will be represented by a knowledge system K - a
quadruple consisting of:
1. A theory T containing current information about the world.
2. A set O of formulae representing new observations.
3. A set H of formulae representing possible explanations of new
observations from O.
4. A relation < defining the relative plausibility of different
sets of possible hypotheses from H.
In this setting the knowledge from K can be used for answering
queries about the world and beliefs of the agent as well as for
finding explanations to the new observations. We will present
formalisations of both of these types of reasoning for knowledge
systems with T represented by logic programs expanded by classical
negation and disjunction, and demonstrate how this formalisation
can be used to solve problems in diagnostic reasoning, logic
programming and disjunctive databases which seem to cause
difficulties for some other approaches.
∂12-Nov-89 1649 gstuck@note.nsf.gov Re: electronic reviewing
Received: from note.nsf.gov by SAIL.Stanford.EDU with TCP; 12 Nov 89 16:49:20 PST
To: John McCarthy <JMC@SAIL.STANFORD.EDU>
Subject: Re: electronic reviewing
In-reply-to: Your message of 24 Jul 89 16:19:00 -0700.
<pBzEd@SAIL.Stanford.EDU>
Date: Sun, 12 Nov 89 19:24:40 EST
From: Jerry Stuck <gstuck@note.nsf.gov>
Message-ID: <8911121924.aa06713@Note.NSF.GOV>
This note acknowledges the receipt of your response to NSF's
request for reviewer e-mail addresses (dated 7/14/89). We are in
the process of loading the addresses to our reviewer data base
for use by NSF staff during the proposal review process. We
appreciate the response you have sent and will be communicating
with you further in the near future. If you have any questions
or comments, please address them to gstuck@note.nsf.gov
(Internet) or gstuck@nsf (Bitnet). Thanks again.
Jerry Stuck, NSF
∂13-Nov-89 0553 qphysics-owner@neat.cs.toronto.edu QP-Workshop 1990 - Call for papers
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 13 Nov 89 05:53:27 PST
Received: by neat.cs.toronto.edu id 2949; Mon, 13 Nov 89 08:50:41 EST
Received: from ugw.utcs.utoronto.ca by neat.cs.toronto.edu with SMTP id 2582; Mon, 13 Nov 89 08:47:12 EST
Received: from CLSEPF51.BITNET (stdin) by ugw.utcs.utoronto.ca with SMTP id 57371; Mon, 13 Nov 89 08:45:30 EST
Received: from elma.epfl.ch by SIC.Epfl.CH with VMSmail ;
Mon, 13 Nov 89 11:00:23 N
Date: Mon, 13 Nov 89 08:30:23 EST
From: faltings@elma.epfl.ch
Message-Id: <891113110023.20602914@SIC.Epfl.CH>
Subject: QP-Workshop 1990 - Call for papers
To: qphysics@ai.utoronto.ca
X-ST-Vmsmail-To: ELCC::"qphysics@ai.utoronto.ca"
Resent-From: qphysics-owner@cs.toronto.edu
Resent-To: qphysics@cs.toronto.edu
Resent-Reply-To: faltings@elma.epfl.ch
Resent-Message-Id: <89Nov13.085041est.2949@neat.cs.toronto.edu>
Resent-Date: Mon, 13 Nov 89 08:48:52 EST
============ 1st Call for Papers; Deadline: Feb. 27th, 1990 ==============
4th International Workshop on Qualitative Physics
Lugano, Switzerland
July 9th - 12th, 1990
Following the success of the previous workshops in Urbana, Paris and
Palo Alto, the fourth international workshop on Qualitative Physics will
be held at the Istituto Dalle Molle di Studi sull'Intelligenza Artificiale
(IDSIA), Lugano, Switzerland. It will be sponsored jointly by IDSIA
and the Swiss Group for Artificial Intelligence and Cognitive Science
(SGAICO).
Papers are invited on all topics of Qualitative Physics, including
(but not limited to):
- Qualitative Representations of Space and Time
- Naive Physics
- Techniques for Automatic Model Management
- Applications of Qualitative Physics
- Implementation Strategies and Performance Studies
- Causal Reasoning
- Qualitative Theories of Physical Domains
Attendance of the workshop is limited in order to permit free discussion
among participants. Invitations will be made by the program committee
on the basis of submitted papers. The program committee consists of:
Johan de Kleer (Xerox PARC), Boi Faltings (EPF Lausanne, co-chairman)
Ken Forbus (University of Illinois), Peter Struss (Siemens Munchen, co-chairman)
,
Olivier Raiman (IBM Paris and Yorktown Heights)
Papers must not exceed 11 pages of single-spaced text using 12pt. type.
Submit 6 copies by February 27th, 1990, to:
Boi Faltings
Att: QP-Workshop
EPFL, LIA-DI
MA-Ecublens
1015 Lausanne, Switzerland
Each paper will be reviewed by at least two reviewers. Papers will be
selected according to their originality and potential to generate
interesting discussions.
Invitations to the workshop will be mailed by mid-April. If this is too late
for you, please send a letter explaining you situation along with your paper
submission. If you would like to be notified by FAX, please state your FAX
number with the paper submission.
The local organization will be handled by IDSIA and the Department of Computer S
cience of
the Universit\`{a} degli Studi di Milano. For questions, please contact Francesc
o Gardin,
Dipartimento di Scienze dell'Informazione, Universit\`{a} degli Studi di Milano,
Via
Moretto da Brescia 9, 20133 Milano, Italy.
For further information, please contact Boi Faltings at the above adress or
Tel. +41-21-693-2735, FAX +41-21-693-5225, E-mail: faltings@elma.epfl.ch or
faltings@CLSEPF51.bitnet, or Peter Struss, SIEMENS AG, Otto-Hahn Ring 6,
D-8000 Munchen 83, West Germany, Tel.+49-89-636-2414, FAX +49-89-636-44150,
E-mail: struss@ztivax.siemens.com.
∂13-Nov-89 1036 korf@CS.UCLA.EDU Stanford Visit
Received: from shemp.cs.ucla.edu by SAIL.Stanford.EDU with TCP; 13 Nov 89 10:36:21 PST
Return-Path: <korf@CS.UCLA.EDU>
Received: from Denali.CS.UCLA.EDU by shemp.cs.ucla.edu
(Sendmail 5.61a/2.23) id AA15680;
Mon, 13 Nov 89 10:36:08 -0800
Message-Id: <8911131836.AA15680@shemp.cs.ucla.edu>
Received: by denali; Mon, 13 Nov 89 10:46:32 pst
Date: Mon, 13 Nov 89 10:46:32 pst
From: Richard E Korf <korf@CS.UCLA.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: Stanford Visit
John,
Anoop Gupta has suggested Jan. 23 as a good day for me to give a colloquium at
Stanford. Before I confirm any date with him, I want to make sure it's a good
day for you. Will you be around that day?
-rich
∂13-Nov-89 1701 VAL Commonsense and nonmonotonic reasoning seminar - no meeting
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
There will be no meeting of the nonmonotonic seminar on Monday, November 20.
--Vladimir Lifschitz
∂14-Nov-89 0111 root@Gang-of-Four.Stanford.EDU
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Nov 89 01:11:50 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA27000; Tue, 14 Nov 89 01:13:11 -0800
Date: Tue, 14 Nov 89 01:13:11 -0800
From: root@Gang-of-Four.Stanford.EDU (Charlie Root)
Message-Id: <8911140913.AA27000@Gang-of-Four.Stanford.EDU>
To: jmc@sail.Stanford.EDU
>From rdz Mon Nov 13 14:43:59 1989
Return-Path: <rdz@Gang-of-Four.Stanford.EDU>
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA24759; Mon, 13 Nov 89 14:43:55 -0800
Date: Mon, 13 Nov 89 14:43:55 -0800
>From: rdz@Gang-of-Four.Stanford.EDU (Ramin Zabih)
Full-Name: Ramin Zabih
Message-Id: <8911132243.AA24759@Gang-of-Four.Stanford.EDU>
To: jmc
Subject: This might interest you
Return-Path: <@Score.Stanford.EDU:dgleason.PA@Xerox.COM>
Sender: dgleason.PA@Xerox.COM
Date: 10 Nov 89 16:48:11 PST (Friday)
Subject: PARC Forum November 16, "Computers, Perestroika, and Glasnost"
>From: "PARC_Forum_Host.PA"@Xerox.COM
To: ForumsAtParc↑.X@Xerox.COM, All-Colloquiums@SCORE.Stanford.EDU
Cc: Mulhern.PA@Xerox.COM
Reply-To: dgleason:PA.PA@Xerox.COM;
Date: Thursday, 16 November 1989
Time: 4:00 pm
Place: PARC Auditorium
"Computers, Perestroika, and Glasnost"
Gary Chapman
Executive Director of
Computer Professionals for Social Responsibility
The Soviet Union is almost uniquely backward in computer technology for an
otherwise industrialized nation. The Soviets have great difficulty producing
reliable computers, or computer hardware in sufficient numbers to meet demand.
This talk will describe the unusual features of the computing world in the Soviet
Union, such as the astronomical prices for computing equipment, the new
computer cooperatives, the work of Soviet programmers, Soviet viruses, and
government policies relating to computers in the country. It will address the
role of computers in promoting democracy and human rights in the USSR, and the
issue of technology transfer between the U.S. and the Soviet Union.
Gary Chapman is executive director of Computer Professionals for Social
Responsibility, a national public interest organization of computer
professionals concerned about the impact of computers in society. At the
invitation of the International Computer Club in Moscow, he travelled to
the Soviet Union to talk to people first hand about the impact of information
technologies in the Soviet Union. An account of their trip was the cover story
of the June 12 issue of Forbes magazine, and was also featured in reports in
The CPSR Newsletter and The Communications of the ACM.
---
This Forum is OPEN. All are invited, but seating is limited and doors will
be closed when available seats are taken. Refreshments will be served at
3:45.
Upcoming Forums:
Nov 20 (Monday): Jaron Lanier "Virtual Realities"
Nov 30 (Thursday): Paul Doherty on the S.F Exploratorium
The PARC Auditorum is located at 3333 Coyote Hill Rd. in Palo Alto. We are
between Page Mill Road (west of Foothill Expressway) and Hillview Avenue,
in the Stanford Research Park. The easiest way here is to get onto Page
Mill Road, and turn onto Coyote Hill Road. As you drive up Coyote Hill
past the horse pastures, PARC is the only building on the left after you
crest the hill. Park in the large parking lot, and enter the auditorium
at the upper level of the building. (The auditorum entrance is located
down the stairs and to the left of the main doors.)
Hosted by Dana Gleason (Systems Software Unit 408-737-4946, DGleason.pa@Xerox.COM)
CARPOOLING FROM SSU IN SUNNYVALE: Meet at 3:20 Thursday in the Information
Center, bldg 5 room H1104.
Requests for videotapes for should be sent to Susie Mulhern (415-494-4068).
∂14-Nov-89 0800 JMC
notebook
∂14-Nov-89 1054 nilsson@Tenaya.Stanford.EDU reminder
Received: from Tenaya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Nov 89 10:54:38 PST
Received: by Tenaya.Stanford.EDU (NeXT-1.0 (From Sendmail 5.52)/25-eef) id AA13686; Tue, 14 Nov 89 10:48:51 PST
Date: Tue, 14 Nov 89 10:48:51 PST
From: Nils Nilsson <nilsson@tenaya.stanford.edu>
Message-Id: <8911141848.AA13686@Tenaya.Stanford.EDU>
To: abbas@isl.stanford.edu, losleben@cis.stanford.edu,
plummer@sierra.stanford.edu, nanni@galileo.stanford.edu,
jlh@vsop.stanford.edu, horowitz@mojave.stanford.edu, toole@darpa.mil,
cross@darpa.mil, gibbons@sierra.stanford.edu,
levinthal@sierra.stanford.edu, feigenbaum@sumex aim.stanford.edu,
engelmore@sumex-aim.stanford.edu, goodman@sierra.stanford.edu,
linvill@cis.stanford.edu, jmc@sail.stanford.edu, jct@sail.stanford.edu,
val@sail.stanford.edu, wiederhold@sail.stanford.edu,
ullman@cs.stanford.edu, latombe@cs.stanford.edu,
binford@cs.stanford.edu, cannon@sierra.stanford.edu,
wiederhold@cs.stanford.edu, luckham@sierra.stanford.edu,
cheriton@cs.stanford.edu, gupta@cs.stanford.edu,
jutta@coyote.stanford.edu, bhr@sumex-aim.stanford.edu,
khatib@cs.stanford.edu
Subject: reminder
Cc: nilsson@Tenaya.Stanford.EDU
This is a reminder about the schedule for the visit
by Barry Boehm next Monday and Tuesday. Please
let me know if you have questions. (I am assuming
that Hennessy/Plummer have organized the Tuesday
sessions and have arranged for CIS 101.) -Nils
---
Visit by Barry Boehm, Director, DARPA/ISTO
Monday and Tuesday, November 20 and 21, 1989
Session 1; leader Bob Engelmore
Monday, November 20, 1989
Knowledge Systems Lab
701 Welch Road, Building A and C, Palo Alto
8:00 - 9:30 Welcome and Overview of Stanford EE and
CS Departments
Dean James Gibbons, Joe Goodman,
Nils Nilsson
9:30- noon Review of KSL projects
Bob Engelmore and KSL staff
noon - 1:15 Discussion and working lunch at KSL
followed by travel to Cedar Hall
Session 2; leader Jean-Claude Latombe
Monday, November 20, 1989
Robotics Laboratory
Cedar Hall
1:15 - 2:30 Foundations of AI and Related Projects
John McCarthy, Vladimir Lifschitz, and
Carolyn Talcott
2:30 - 5:00 Stanford Robotics Projects
Jean-Claude Latombe, Tom Binford, and
Bob Cannon
Session 3; leader Bob Cannon
Monday, November 20, 1989
Durand Building
5:00 - 5:30 Tour of Aero/Astro Robotics Lab
5:30 - 6:30 Roundtable Discussion (Durand Conference
Room 450)
[Steve suggests that this roundtable
focus on AI-related topics;
so all of the robotics and AI people are
invited.]
7:00 - Possible small group for dinner? (Nils to
follow up and arrange.)
Sessions 4 & 5; leaders John Hennessy and Jim
Plummer
Tuesday, November 21, 1989
CIS Building
Conference Room 101
8:30 - 9:15 Database Research
Gio Wiederhold
9:15 - 10:00 Parallel Computation
Jeff Ullman
10:00 - 5:00
John Hennessy and Jim Plummer will be arranging the
rest of the day to
include, possibly, the following or similar topics (as
suggested by Steve Cross):
Software, David Luckham
Microsupercomputers, John Hennessy and Mark
Horowitz
Structured Process Flow, Jim Plummer
CAD for EEPROM, Abbas El Gammal
Logic Synthesis, Giovanni DiMicheli
Distributed Systems, David Cheriton
the agenda for these topics can spill over into
Wednesday morning if needed.
12:15 - 1:15 on Tuesday: Lunch with CS and CSL
faculty in MJH 146
Sandwiches/drinks provided.
∂14-Nov-89 1600 tajnai@Polya.Stanford.EDU AI Day on Campus, 6/7/90
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Nov 89 16:00:15 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA26123; Tue, 14 Nov 89 15:59:23 -0800
Date: Tue, 14 Nov 1989 15:59:22 PST
From: "Carolyn E. Tajnai" <tajnai@Polya.Stanford.EDU>
To: binford@Coyote.Stanford.EDU, eaf@SUMEX-AIM.Stanford.EDU,
Genesereth@polya.Stanford.EDU, latombe@Coyote.Stanford.EDU,
lenat@SUMEX-AIM.Stanford.EDU, ZM@Sail.Stanford.EDU,
jmc@sail.Stanford.EDU, nilsson@TENAYA.STANFORD.EDU,
shoham@Polya.Stanford.EDU, Marty@CIS.STANFORD.EDU,
waldinger@ai.sri.com, Winograd@csli.stanford.edu,
Shortliffe@SUMEX-AIM.Stanford.EDU, musen@SUMEX-AIM.Stanford.EDU,
barraqua@Coyote.Stanford.EDU, rse@sumex-aim.Stanford.EDU,
Fagan@SUMEX-AIM.Stanford.EDU, ginsberg@polya.stanford.edu,
gruber@sumex-aim.stanford.edu, Bhayes-roth@sumex-aim.Stanford.EDU,
Iwasaki@sumex-aim.stanford.edu, KELLER@SUMEX-AIM.Stanford.EDU,
OK@coyote.stanford.edu, VAL@Sail.Stanford.EDU, IAM@Sail.Stanford.EDU,
Nii@SUMEX-AIM.Stanford.EDU, Ponce@Coyote.Stanford.EDU,
Rindfleisch@SUMEX-AIM.Stanford.EDU, rit@coyote.stanford.edu,
singh@hudson.stanford.edu, CLT@Sail.Stanford.EDU,
weening@gang-of-four.stanford.edu, em.der@psych.Stanford.EDU,
cooper@SUMEX-AIM.Stanford.EDU, tajnai@Polya.Stanford.EDU
Cc: hiller@Polya.Stanford.EDU
Subject: AI Day on Campus, 6/7/90
Message-Id: <CMM.0.88.627091162.tajnai@Polya.Stanford.EDU>
The Computer Forum gave a KSL Day on Campus in 88, Systems Day on
Campus in 89, and we are planning AI Day on Campus for Thursday,
June 7, 1990.
This is to alert you so you have advance notice and can mark your
calendar. I have prepared a special AIfaculty distribution list
(including research associates). If you do not wish to participate
or hear any more about our plans, let me know, and I'll delete you
from the list.
We will block the program the first of March, prepare a brochure and
mail it by the first of April. We will have all day parallel sessions,
probably demos, and a barbeque lunch at noon.
I'll send you more information after the first of the year.
Carolyn
∂14-Nov-89 1638 sanu@cs.utexas.edu elephant 2000
Received: from cs.utexas.edu by SAIL.Stanford.EDU with TCP; 14 Nov 89 16:38:51 PST
Received: from megalon.cs.utexas.edu by cs.utexas.edu (5.59/1.44)
id AA08140; Tue, 14 Nov 89 18:38:36 CST
Posted-Date: Tue, 14 Nov 89 18:38:27 CST
Message-Id: <8911150038.AA07794@megalon.cs.utexas.edu>
Received: by megalon.cs.utexas.edu (5.59/1.4-Client)
id AA07794; Tue, 14 Nov 89 18:38:29 CST
From: sanu@cs.utexas.edu (Sankrant Sanu)
Date: Tue, 14 Nov 89 18:38:27 CST
X-Mailer: Mail User's Shell (6.5.6 6/30/89)
To: jmc@sail.stanford.edu
Subject: elephant 2000
Cc: sanu@cs.utexas.edu
Prof. McCarthy,
This refers to your talk at the University of Texas on October 12, '89,
which I missed. I am an incoming Ph.D. Student, and, from the abstract, this
seemed like an interesting topic to me. Could you provide some reference to
material I could read on what was covered in your talk?
Thanks,
Sankrant Sanu.
∂14-Nov-89 1649 sanu@cs.utexas.edu re: elephant 2000
Received: from cs.utexas.edu by SAIL.Stanford.EDU with TCP; 14 Nov 89 16:49:30 PST
Received: from ghidrah.cs.utexas.edu by cs.utexas.edu (5.59/1.44)
id AA08566; Tue, 14 Nov 89 18:49:27 CST
Posted-Date: Tue, 14 Nov 89 18:49:16 CST
Message-Id: <8911150049.AA01294@ghidrah.cs.utexas.edu>
Received: by ghidrah.cs.utexas.edu (5.59/1.4-Client)
id AA01294; Tue, 14 Nov 89 18:49:19 CST
From: sanu@cs.utexas.edu (Sankrant Sanu)
Date: Tue, 14 Nov 89 18:49:16 CST
X-Mailer: Mail User's Shell (6.5.6 6/30/89)
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: elephant 2000
My U.S. mail address is
Sankrant Sanu
Graduate Student
Department of Computer Science
The University of Texas at Austin
Austin, Texas.
Thank you.
Sankrant.
∂14-Nov-89 1754 rdz@Gang-of-Four.Stanford.EDU MIT talk of possible interest
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Nov 89 17:54:15 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA00206; Tue, 14 Nov 89 17:55:41 -0800
Date: Tue, 14 Nov 89 17:55:41 -0800
From: rdz@Gang-of-Four.Stanford.EDU (Ramin Zabih)
Message-Id: <8911150155.AA00206@Gang-of-Four.Stanford.EDU>
To: jmc@sail.Stanford.EDU
Subject: MIT talk of possible interest
Return-Path: <JCMA@reagan.ai.mit.edu>
Date: Tue, 14 Nov 89 02:42 EST
>From: John C. Mallery <JCMA@reagan.ai.mit.edu>
Subject: Victor Sergeev: Contemporary Soviet AI Research
To: ai-seminar@ai.mit.edu, ai-seminar@ai.mit.edu
Start-Date: Tue, 14 Nov 89 17:00 EST
Expiration-Date: Tue, 14 Nov 89 17:45 EST
Tuesday, Nov 14, 5:00-5:45 p.m. NE43 8th floor playroom.
Contemporary Soviet Research in
Artificial Intelligence and Computational Linguistics
Victor Sergeev
Head of Laboratory for Decision Analysis
Institute of The United States and Canada
Academy of Sciences of The U.S.S.R
The talk overviews cognitive modeling at the Soviet Academy of Sciences,
situating it within more general Soviet research in artificial intelligence
and computer science. At the Laboratory for Decision Analysis, cognitive
modeling spans knowledge representation, text analysis, syntactic analysis,
symbolic decision processes, as well as rationality and learning. The
conclusion contrasts disciplinary origins of Soviet research to U.S. research
and reviews particularly strong Soviet traditions, which include very well
developed linguistic theory (semantics and pragmatics), mathematical logic,
and pattern recognition.
∂15-Nov-89 1003 chandler@Polya.Stanford.EDU Faculty Reports
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 15 Nov 89 10:03:18 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA23040; Wed, 15 Nov 89 10:03:06 -0800
Date: Wed, 15 Nov 1989 10:03:05 PST
From: "Joyce R. Chandler" <chandler@Polya.Stanford.EDU>
To: cab@sail.Stanford.EDU, binford@coyote.Stanford.EDU,
cheriton@pescadero.Stanford.EDU, dill@amadeus.Stanford.EDU,
feigenbaum@sumex-aim.Stanford.EDU, rwf@sail.Stanford.EDU,
genesereth@score.Stanford.EDU, goldberg@Polya.Stanford.EDU,
golub@patience.Stanford.EDU, guibas@dec.com, ag@amadeus.Stanford.EDU,
dek@sail.Stanford.EDU, lam@mojave.Stanford.EDU,
latombe@coyote.Stanford.EDU, zm@sail.Stanford.EDU,
jmc@sail.Stanford.EDU, ejm@sierra.Stanford.EDU, jcm@Polya.Stanford.EDU,
oliger@pride.Stanford.EDU, plotkin@hudson.Stanford.EDU,
pratt@Polya.Stanford.EDU, shoham@Polya.Stanford.EDU,
wiederhold@sumex-aim.Stanford.EDU, winograd@csli.Stanford.EDU
Cc: chandler@Polya.Stanford.EDU
Subject: Faculty Reports
Message-Id: <CMM.0.88.627156185.chandler@Polya.Stanford.EDU>
Have mercy, dear faculty, on your Department Secretary, who has a deadline of
12/1 to have these reports into the School of Engineering.
Please send me your faculty report at your earliest opportunity.
Thanks.
∂15-Nov-89 1308 chandler@Polya.Stanford.EDU DARPA Visit Agenda
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 15 Nov 89 13:08:14 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA29856; Wed, 15 Nov 89 13:05:14 -0800
Date: Wed, 15 Nov 1989 13:05:11 PST
From: "Joyce R. Chandler" <chandler@Polya.Stanford.EDU>
To: abbas@isl.Stanford.EDU, losleben@cis.Stanford.EDU,
plummer@sierra.Stanford.EDU, nanni@galileo.Stanford.EDU,
jlh@vsop.Stanford.EDU, horowitz@mojave.Stanford.EDU, toole@darpa.mil,
cross@darpa.mil, gibbons@sierra.Stanford.EDU,
levinthal@sierra.Stanford.EDU, feigenbaum@sumex-aim.Stanford.EDU,
engelmore@sumex-aim.Stanford.EDU, goodman@sierra.Stanford.EDU,
linvill@cis.Stanford.EDU, jmc@sail.Stanford.EDU, jct@sail.Stanford.EDU,
val@sail.Stanford.EDU, wiederhold@sail.Stanford.EDU,
ullman@cs.Stanford.EDU, latombe@cs.Stanford.EDU,
binford@cs.Stanford.EDU, cannon@sierra.Stanford.EDU,
wiederhold@cs.Stanford.EDU, luckham@sierra.Stanford.EDU,
cheriton@cs.Stanford.EDU, gupta@cs.Stanford.EDU,
jutta@coyote.Stanford.EDU, bhr@sumex-aim.Stanford.EDU,
khatib@cs.Stanford.EDU
Cc: chandler@Polya.Stanford.EDU, fincher@darpa.mil
Subject: DARPA Visit Agenda
Message-Id: <CMM.0.88.627167111.chandler@Polya.Stanford.EDU>
Visit by Barry Boehm, Director, DARPA/ISTO
Monday and Tuesday, November 20 and 21, 1989
Session 1; leader Bob Engelmore
Monday, November 20, 1989
Knowledge Systems Lab
701 Welch Road, Building A and C, Palo Alto
8:00 - 9:30 Welcome and Overview of Stanford EE and
CS Departments
Dean James Gibbons, Joe Goodman,
Nils Nilsson
9:30- noon Review of KSL projects
Bob Engelmore and KSL staff
noon - 1:15 Discussion and working lunch at KSL
followed by travel to Cedar Hall
Session 2; leader Jean-Claude Latombe
Monday, November 20, 1989
Robotics Laboratory
Cedar Hall
1:15 - 2:30 Foundations of AI and Related Projects
John McCarthy, Vladimir Lifschitz, and
Carolyn Talcott
2:30 - 4:00 Stanford Robotics Projects
Jean-Claude Latombe, Tom Binford, and
Bob Cannon
Session 3; leader Bob Cannon
Monday, November 20, 1989
Durand Building
4:00 - 5:30 Tour of Aero/Astro Robotics Lab
5:30 - 6:30 Roundtable Discussion (Durand Conference
Room 450)
[Steve suggests that this roundtable
focus on AI-related topics;
so all of the robotics and AI people are
invited.]
7:00 - Possible small group for dinner? (Nils to
follow up and arrange.)
Sessions 4 & 5; leaders John Hennessy and Jim
Plummer
Tuesday, November 21, 1989
CIS Building
Conference Room 101
8:30 - 9:15 Database Research
Gio Wiederhold
9:15 - 10:00 Parallel Computation
Jeff Ullman
10:00 - 5:00
John Hennessy and Jim Plummer will be arranging the
rest of the day to
include, possibly, the following or similar topics (as
suggested by Steve Cross):
Software, David Luckham
Microsupercomputers, John Hennessy and Mark
Horowitz
Structured Process Flow, Jim Plummer
CAD for EEPROM, Abbas El Gammal
Logic Synthesis, Giovanni DiMicheli
Distributed Systems, David Cheriton
the agenda for these topics can spill over into
Wednesday morning if needed.
12:15 - 1:15 on Tuesday: Lunch with CS and CSL
faculty in MJH 146
Sandwiches/drinks provided.
∂15-Nov-89 1359 kirsh%cogsci@ucsd.edu hewitt
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 15 Nov 89 13:59:44 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA04157
sendmail 5.61/UCSD-2.0-sun via SMTP
Wed, 15 Nov 89 13:59:33 -0800 for JMC@sail.stanford.edu
Received: from cogsci-sun.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA21469 for JMC@sail.stanford.edu; Wed, 15 Nov 89 13:57:58 PST
Date: Wed, 15 Nov 89 14:01 PST
From: David Kirsh <kirsh%cogsci-sun@ucsd.edu>
Subject: hewitt
To: JMC@Sail.Stanford.EDU
Message-Id: <19891115220102.4.KIRSH@cogsci-sun.ucsd.edu>
John,
Your mailer will not accept this message, so I'm cutting it up in 2 bits.
The first has a bout half of Carl's paper, the remainder of the paper is in
the second message.
Message failed for the following:
JMC@Sail.Stanford.EDU.#Internet: 552 Message text too long! (Try sending it in smaller pieces.)
------------
Received: from ucsd.edu by SCORE.STANFORD.EDU with TCP; Tue 14 Nov 89 14:59:02-PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA08948
sendmail 5.61/UCSD-2.0-sun via SMTP
Tue, 14 Nov 89 14:58:48 -0800 for jmc@score.stanford.edu
Received: from cogsci-sun.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA07264 for jmc@score.stanford.edu; Tue, 14 Nov 89 14:57:16 PST
Date: Tue, 14 Nov 89 15:00 PST
From: David Kirsh <kirsh%cogsci-sun@ucsd.edu>
Subject: essay hewitt
To: jmc@score.stanford.edu
Cc: kirsh@cogsci.ucsd.edu
Message-Id: <19891114230032.4.KIRSH@cogsci-sun.ucsd.edu>
I had trouble sending yesterday.
John,
It was fun chatting with you tonight. Carl just sent me an updated version
of his paper. So now we are reasonably close to a final version.
Please send both Carl and me your revised manuscript. You probably want to
enlarge it too. Everyone is interested in hearing what you have to say, so
why not take the opportunity to say some of the major things you believe.
-- David
\documentstyle[11pt]{article}
\def\startline{\par\nobreak\noindent}
{\obeylines\obeyspaces\gdef\procedure{\bigbreak\begingroup
\parindent=0pt \parskip=0pt plus0pt minus0pt
\obeylines \obeyspaces \eg\let↑↑M=\startline \eg}
\gdef\endproc{\par\endgroup\bigbreak}}
\def\wyn#1{$\spadesuit${\tt #1}}
\newcommand\code[1]{\mbox{\eg{#1}}}
\newcommand\eol{\hfill\break}
\newcommand\bline{\vskip12pt plus0pt minus0pt}
\newcommand\eg{\tt}
\title{Artificial Intelligence\\and/or\\Open Systems Technoscience}
\author{\copyright\ 1989 Carl Hewitt}
\date{Draft of \today}
\begin{document}
% \bibliographystyle{plain}
% \bibliography{/home/tx/wsnow/biblio}
\maketitle
\section{Abstract}
Open Systems Technoscience (OST) is the technology and science of large
scale human/telecomputer organizational work. Examples of this kind of
work include flexible semiconductor manufacturing, constructing a
permanent space station, and the software engineering of electronic
fund transfer systems. Open Systems Technoscience addresses issues of
participant values, robustness, and scalability in large scale
organizational work. In contrast, Artificial Intelligence (AI) is the science
and technology of intelligent agents and robots. The relationship
between OST and AI has much of the same complementary and
supplementary character as the relationship between sociology and
psychology.
Artificial Intelligence has created new technologies of taxonomies,
inference-based systems, and problem spaces. All of these technologies
can be useful within organizations that engage in large-scale work.
However, each of these technologies is useless without the extensive
organizational support that is necessary to make it work.
\section{Introduction}
All large-scale Open Systems are concurrent, asynchronous, decentralized,
and indeterminate. They are composed of numerous participants which
operate {\em concurrently\/} in order to accomplish the multitude of
tasks that are performed. They are {\em distributed\/} in order to deal
with the influx of information from many sources and to convey
information to the places where it is needed.
In any short span of time, each participant of a large-scale Open System
operates {\em autonomously\/} and {\em asynchronously\/} in accordance
with its own local needs and procedures. No truly simultaneous change of
all the participants in a large-scale Open System is ever possible, and new
information may arrive from any source at any time. Thus in general, one
participant of the system will start using new information before it
reaches the others.
Furthermore, asynchronous operation means that any large-scale Open
System is {\em indeterminate\/} in a physical sense: it does not have a
determinable current state which (together with new information that
arrives) determines its future operation. In fact, attempts to pin down
an instantaneous state of a large system by gathering more information
about the finest details of its internal operation makes the system {\em
more\/} indeterminate, because gathering the information affects its
operation. Furthermore, another large-scale Open System is needed just
to gather, interpret, and store the information about the smallest-scale
activities of another large system.
\subsection{Advantages of Open Systems}
Open Systems Technoscience addresses issues of gaining advantages from the
Open Systems environment in which large-scale organizations operate:
\begin{itemize}
\item
{\bf Asynchrony:} enables each participant of an organization to operate as
quickly as possible, given local circumstances. Otherwise, the pace of
the activity of each participant would be locked to a single organizational
scheduler down to the lowest level activities.
\item
{\bf Local autonomy:} enables each participant of an organization to react
immediately to changing circumstances. Otherwise, each participant would
have to consult a single organizational decision maker for each decision.
\item
{\bf Late-arriving information:} enables the organization to increase the
effectiveness of its decision making by incorporating information as it
arrives.
\item
{\bf Multiple authorities:} increases an organization's pluralism,
diversity, and robustness. For example, the engineering department of a
utility wants to build a new kind of nuclear reactor which is inherently
safer than existing reactors, while the finance department maintains that
the financial risk is too great.
\item
{\bf Arm's length relationships:} enables actors to conceal their internal
activities from other other actors. For example, the Nuclear Regulatory
Commission operates at arm's length from nuclear utilities that it
regulates in order to be more effective in detecting and prosecuting
violations of its regulations.
\item
{\bf Division of labor:} specialization of functionality can increase
organizational effectiveness by enabling each participant to concentrate
and focus its efforts on a narrow range of goals. For example, an
electric utility has separate finance and engineering suborganizations,
each specializing in different functions: finance takes care of raising
the money, and engineering directs the construction of power plants.
\end{itemize}
\section{Trials of Strength}
Any situation in which forces are pulling in different directions
constitutes a {\em trial of strength}. All trials of strength are local in
the sense that they occur at a particular time and place among local
participants. For example negotiation provides a mechanism for
representatives of different participants to come together in a trial of
strength.
\subsection{Commitments}
A {\em commitment\/} is a course of action, usually involving several
actors, which the actors are {\em on course\/} to carry through. An
actor's {\em responsibility\/} is its part in the commitment.
Commitments are used in this way in order to avoid becoming embroiled
in issues of {\em intentionality\/}. Instead, an actor is said to be {\em
on course\/} or {\em off course\/} with respect to an activity. This
terminology allows inanimate objects to participate in commitments.
The course of action deliberately chosen by an organization in order to
accomplish its goals can also be a commitment.
Conflict is a trial of strength in which participants have incompatible
commitments. As we have seen, conflict is a fundamental aspect of any
Open System for a variety of reasons. Resources are finite and limited;
choices must be made about how to use them. In addition, most
organizations have built-in checks and balances which more or less
deliberately generate conflicting commitments. The commitment of the
safety department to safe conservative procedures conflicts with the
need for the engineering department to lower costs of construction.
This often places them in conflict with one another.
Sometimes the commitments of the affected participants are compatible
and there is no conflict. One participant says, ``Let's do it this way,''
and everybody agrees, so the negotiation is a trivial one, but there is
always the {\em potential\/} for conflict. No one can be certain in
advance about the outcome, so any negotiation is to that extent a trial
of strength and therefore indeterminate.
Our new discipline of Open Systems Technoscience needs to have an
intimate understanding of the nature of commitment. In particular, Open
Systems Technoscience is especially concerned with the nature of
overall organizational commitments and how they relate to the
commitments of participants inside and outside the organization,
because these broader commitments are what make large-scale projects
possible.
An organizational commitment is one which the organization is on course
to carry out. For example, a utility can have an organizational
commitment to build a new power plant. The utility's Finance department
has a commitment to fund the cost of the plant, and its Engineering
department has a commitment to design and build it. Both Finance and
Engineering must work together in a fairly detailed way, and the
organizational commitment must address the specialized needs and
commitments of both participants. The organizational commitment of the
utility goes beyond the responsibilities of Financing, Engineering, and
other departments. The utility has organizational commitments and
authority that go beyond just the individual ones of its departments and
members.
We said before that a participant's responsibility for a commitment is
its part in a commitment. Alternative courses of action can affect
participant use of resources which is often a source of conflict. There
are several kinds of resources: money, space/time/material,
mechanism/technology, and sentiment.
Money and space/time/material are self-explanatory. Mechanism/\break
technology is anything that transforms the world: for example, a nuclear
power plant transforms radioactive material into electricity. Sentiment
deals with how various actors feel about each other: goodwill,
reputation, obligation, etc. For example, when a public utility says
that its reactor is safe, it is staking part of its reputation on that
statement. Also, when a utility asks for leniency on an issue from a
regulatory board, it is putting to use some of its goodwill for the
utility, which means that it is inhibited in its freedom later on.
However, the past is really gone and the future is never here. Both
planning for the future and reflecting upon past experience are
activities that take place in the present. In addition to the daily,
familiar use of resources, some activities which look to the future and
the past are also commitments. For example, planning can create a
commitment; it chooses a particular course of action that allocates
resources in one way instead of another and sets the planner on course to
keep the commitment. Creating a financial record also generates a
commitment. When a utility submits its annual report, it makes a
statement which says how much money it earned that year. Once it reports
this statement to the regulators, that's a commitment. If the regulators
come and investigate, the utility has to substantiate its claim that it
really did earn that much during the year.
Two commitments are said to be {\em conflicting\/} if they give rise to a
trial of strength which results in at least one of the commitments not
being kept. For example, the commitment of a utility to operate a nuclear
power plant is in conflict with the commitment of an environmental group
to shut it down. These two commitments are incompatible, and cannot both
be kept.
The notion of commitment used in this paper is closely related the
terminology used by those sociologists who emphasize a commitment as a
choice among incompatible alternatives
\cite{concept-of-commitment,quality-of-life}. In this terminology, {\em
choice\/} is analyzed in terms of trials of strength. Commitments have
been discussed by Winograd and Flores in their discussion of Heiddiger's
notion of ``throwness'' \cite{winograd-flores}, and by Richard Fikes
\cite{commitment-based-framework} in the context of contract nets as
developed by Reid Smith and Randy Davis \cite{contract-nets}.
% also \cite{improvised-news}.
% CH NOTE: Get {\em Improvised News\/} by Tom Shibutani, study of
% rumor in organizations, citation goes in structure of commitments.
\subsection{Open Systems Semantics}
Open Systems Semantics is the study of the meaning of Open Systems
Action. Taking any action entails changes in commitments, and that
change is the meaning of the action. This is an open-world
characterization of meaning---as opposed to previous, closed-world
attempts based on possible worlds in which the meaning of a set of
sentences is defined to be the set of all possible worlds that
satisfy the meaning conditions. Building on previous actor theory, Gul Agha
\cite{agha-phd} has developed an
open-world, mathematical semantics for concurrent systems in which the
meaning of an action is characterized by its effect on the evolution of
the system.
In general, Open Systems Action involves conflict---and therefore
indeterminacy. Current situations and events influence, but do not
determine, the future because the outcomes of numerous trials of
strength cannot be known and some trials of strength cannot even be
anticipated.
In logical semantics, representation is the mapping between a sentence
proposition and specified meaning conditions. Meaning is built on and
grows out of representation. Two participants agree about the meaning of a
sentence when they agree about the meaning conditions.
In Open System semantics, however, representation is taken to be how
effects commitments. The important difference is conveyed in the
following Open Systems Semantics slogan:
\begin{quote}
{\Large {\bf No representation without communication!}}
\end{quote}
The test of the degree to which one participant has adequately
communicated its meaning to another participant is: whether the recipient's
commitments change in the way which constitutes the meaning.
Large organizations have extensive policies, procedures, and regulations
which govern the meaning of organizational actions.
Open Systems Semantics is a research programme
\cite{methodology-of-research-programmes} for studying the meaning of
Open Systems Actions. Just as there is no global synchrony or cause
and effect, the meaning of an Open Systems Action is also localized: it
begins in the participants at a particular time and place and then
can causally propagates to wider contexts.
\subsection{Requirements of Open Systems}
Analysis of commitment relationships can be used to understand how the
characteristics of Open Systems engender requirements for Open
Systems implementations:
\begin{itemize}
\item
{\bf Asynchrony:} produces indeterminacy because arriving information
must be integrated with local information. The new information can
generate new commitments that conflict with pre-existing commitments.
\item
{\bf Local autonomy:} produces conflict because as organizational
commitments change, some of the new commitments will be incompatible with
pre-existing local commitments.
\item
{\bf Late-arriving information:} can produce conflict when it arrives at
an advanced stage of processing (e.g., an engineering group reports new
concerns about the capabilities to survive an earthquake just as the
utility is about to apply for a license to operate the reactor).
\item
{\bf Multiple authorities and division of labor:} can produce conflict
because the specialized commitments of multiple authorities may be
incompatible and come into conflict.
\item
{\bf Arm's length relationships:} can produce conflict because the
internal commitments of other organizations are not visible. This can
increase the severity of conflict because other organizations will
develop entrenched commitments before the conflict is discovered.
\end{itemize}
Another underlying constraint is {\em continuous operation}. In many
cases an organization cannot ``take a vacation'' in order to get itself
into better shape. It must continue operating though perhaps
at some reduced level of performance.
\subsection{Dealing with Unanticipated Conflict}
Robustness is keeping commitments in the face of unanticipated conflict.
Keeping a commitment often entails keeping subcommitments. For
example, the commitment to operate the Diablo Canyon reactor requires
keeping two subcommitments: constructing the plant and licensing it.
The licensing subcommitment might go smoothly at first and then run
into unanticipated conflict when seismologists discover new geological
faults near the location of the reactor. The robustness of the
commitment of the utility in part depends upon its ability to deal with
whatever unexpected trial of strength arises. In this case
the utility engages in a negotiation (i.e. a trial of strength)
with the Nuclear Regulatory Commission.
\section{Negotiation}
Negotiation is an important way in which an organization achieves
robustness (i.e. keeping commitments in the face of conflict). A
negotiation is a discourse in which the participants make
representations held in common which can engender joint actions that
are new joint commitments. The new commitments will often in turn
lead to more conflict with pre-existing commitments of individual
participants as they make adjustments for the new commitments.
The organizational commitment of the utility to operate a nuclear power
plant lead to its giving its Finance and Engineering departments more
specialized subcommitments. Engineering has the responsibility to
construct the reactor and Finance has the responsibility to raise the
money for construction. Negotiation gives both participants an
opportunity to negotiate how the utility can keep its commitment to
operate the plant.
Thus, we can see the value of bringing participants together to work out an
organizational response to conflict. Each participant has ongoing
commitments to attend to, and needs a way of figuring out how to allocate
its resources. Instead of being required to turn its full attention to
the new commitment, each participant can separate out a subpart---namely the
representative to the negotiation---which will be devoted to the new
commitment.
Negotiation can help an organization meet its existing commitments while
developing new ones---which is important for attaining robustness.
Negotiation also creates overall, organizational commitments---which are
essential to scaling. Thus, Open Systems Technoscience needs to support
mechanisms that can move a negotiation forward, and determine what
progress is being made.
Insights gleaned from the social sciences (law, sociology, anthropology,
organization theory, and philosophy of science) can help us conduct
human/telecomputer organizational work with increased participant
value. Human organizations have evolved methods for dealing
with conflict---and for turning them into strengths instead of
weaknesses. These methods can be adapted as a source of inspiration
for robustness in human/telecomputer organizations.
\subsection{What Happens During a Negotiation}
Each participant brings its own commitments to a negotiation. These
include decision-making criteria, such as preferences among predicted
outcomes. For example: ``It is preferable to have nuclear power plants
because they lessen our dependence on unstable foreign fuel supplies''
and ``It is preferable not to have nuclear power plants because they
create a threat of the release of significant amounts of radioactivity''
Conflicts among these preferences can be negotiated
\cite{negotiations}.
During a negotiation, the participants can make moves---where each move is the
description of a commitment. In this way, the various participants can arrive
at a joint commitment about the issues being addressed and the options
available for addressing those issues.
The various participants {\em clarify the issue\/} by discussing and
commenting upon each others' statements about the issue. This process
may expand or change the views of the various participants about the nature of
the issue. The representatives also discuss what the various {\em
options\/} are with respect to the commitments around this issue. An
option is a proposal for the rearrangement of the organization's
commitments. Here, discussion and commentary about options can lead to
the generation of new options, further clarification of the nature of the
issue, and further commentary on the commitments that are affected by the
issue.
\subsection{Contradictions}
{\em Contradictions\/} arise because each negotiating participant attempts to
keep its own commitments. When conflict leads to a negotiation, each
participant deliberately resists some of the statements that other participants make
to support their commitments. Each participant uses language as a tool to
further its own commitments, and that often produces resistance and even
contradictory statements.
The expression of conflict can be a very positive aspect of negotiation.
The diversity that produced the conflict and contradictions can also
produce new ideas, suggestions, and options. Thus, negotiation of
conflict can lead to more effective organizational information
processing.
\subsection{Cooperation}
If an organization is continually involved in negotiations, how does it
ever get anything accomplished? Won't bringing bureaucracy to the
machine create even worse messes than in human bureaucracies? While it's
true that negotiations can break down and result in deadlocks, they can
also come up with creative solutions. Two factors that help ``grease the
wheels'' of negotiation are cooperation and allies.
{\em Cooperation\/} is the process by which participants become committed to
each others' commitments. For example, Finance and Engineering might
both be committed to building a new power plant. Finance is committed to
raising the money to build the plant, and Engineering is committed to a
schedule for constructing the plant. Finance relies on Engineering's
commitment to build the plant on schedule in order to have credibility in
the financial marketplace---so Finance is committed to Engineering's
commitment. In a similar fashion, Engineering relies on Finance's
ability to provide the money to pay the construction cost as it comes due
so that construction can continue. So Engineering is committed to
Finance's commitment. Because we have these cross-commitments, we say
that Finance and Engineering are cooperating.
Another example deals with a utility and vendor. The utility generates a
purchase order for product $Q$. The vendor commits itself to shipping
product $Q$ at a certain time. The utility's commitment to the vendor's
shipping product $Q$ is represented by the purchase order. The vendor's
commitment to ship the product is dependent on the utility's commitment
to pay---as represented by the purchase order. Thus the purchase order
formalizes the cooperation between customer and vendor, and represents
the commitment of both participants to the others' commitment.
This process of mutual commitment---cooperation---is of fundamental
importance in moving organizations forward.
\subsection{Allies}
Another aspect of negotiation that helps make the negotiation process go
smoothly is the notion of an {\em ally}. One of the participants may claim an
ally: i.e., predict that in some future trial of strength, its ally would
behave in a certain way. In the payroll example, someone might claim
that if management doesn't allocate sufficient funds to meet the payroll,
the union will call for a strike!
The ally may prove to be faithful or unfaithful: faithful in the sense
that in the future trial of strength, it actually does behave as claimed.
\begin{quote}
{\em
\vbox{Glendower: I can call spirits from the vast deep.
Hotspur: Why, so can I, or so can any man; but will they come when you
do call for them?
\hfill---Shakespeare: {\em Henry IV}, {\em Act III}, {\em Scene 1}.}}
\end{quote}
Thus, when a negotiating participant claims an ally, the likely outcomes of
future trials of strength must be considered in deciding on what action
to take. In this way, claiming the support of allies can sway a
negotiation in a certain direction, but it is not necessarily decisive.
For example, a customer might invoke the Public Utility Board as an ally
when presenting a complaint to a utility that refuses to remove excess
charges from its bill. The utility needs to consider the implications of
an Utility Board investigation in deciding how to respond to the
customer.
Claiming an ally can be a very one-sided relationship. If the utility
decides to honor the customer's claim about the probable outcome of a
Utility Board investigation, then the Utility Board would never hear
about the incident, even though it was successfully invoked as an ally.
By contrast, an alliance is a {\em mutual commitment}. The participants to an
alliance make mutual commitments of their time, money, staff, and other
resources. Alliances are important outcomes of negotiation. Almost
every negotiation will affect alliances, either by creating new ones or
by strengthening, adjusting, or weakening old ones.
\section{Outcomes of Negotiation}
Many kinds of outcomes are possible, but the following three often
occur:
\begin{itemize}
\item
A {\bf resolution} to which the participants commit themselves.
\item
A {\bf deadlock} in which the participants at this particular negotiation
cannot reach an agreement. Quite often as a result of deadlock, another
negotiation is held with different representatives, and on a different
issue: namely, the fact that the other negotiation deadlocked. ``Those
guys didn't work it out, what are we going to do about it?''
\item
An {\bf appeal}. Some of the representatives might be unhappy about the
outcome and appeal to other parts of the organization---which might set
up another negotiation to deal with the issue of what to do about the
outcome of the previous negotiation.
\end{itemize}
Participants can be stalemated in conflict for a long period of time. For
example, an environmental group can work for decades attempting to revoke
the operating license of a nuclear power plant. Maintaining a
negotiation does consume resources, however. One has to keep track of
the other participant's position, plan strategy of how to continue carrying out
the negotiation, respond to the other participant's moves all the time, and so
on. Actions like these consume time, communications, storage space, and
other resources that would otherwise be put to different use. Thus,
maintaining a negotiation is a commitment.
In some cases, a negotiation ends when one participant runs out of the resources
needed to continue. In other cases, the process explicitly provides a back-up
procedure in the sense that it leaves the conflict potentially resumable, but
ends the current negotiation. An example of this would be a state that is
determined to oppose a public utility in its attempt to operate a nuclear
powerplant. The state can oppose the utility at every stage of its attempt to
get an operating license. Suppose that at each stage the negotiation is
broken off when the Nuclear Regulatory Commission decides in favor of the
utility. After each hearing the resources committed to the negotiation are
allowed to go their own way, with the intention that a whole new negotiation
might be convened at another time. Finally just before the nuclear power
plant goes into operation, the state might ``win'' the negotiation by offering
to compensate the utility for what it would earn by operating the nuclear plant
{\em provided} that it agrees to sell the plant to the state for \$1!
\subsection{Negotiations Create Commitments}
Negotiations are important, even if no conflict emerges, because they
create organizational commitments that go beyond the individual
commitments of the participants involved. Negotiations always have multiple
possible outcomes. Choices are made during a negotiation, which may
result in the creation of an organizational commitment: the participants agree
on a particular course of joint action. The negotiation might prove to
be a trivial one in which agreement is easily reached, but the outcome
still represents an organizational commitment. Late-arriving information
could have caused one of the participants to strive for a different outcome.
The significance of negotiations lies in their outcomes and the way those
outcomes affect other organizational actions. The country will
incrementally develop an electric power industry---and that industry will
influence energy costs, pollution levels, generating capacity, etc. For
example, Engineering and Finance can have a dispute as to whether to build
a plant with two reactors at once, as opposed to finishing one before
starting the next. Engineering prefers building both at once because it
can overlap similar activities to bring down the cost. Finance prefers
building them sequentially because the financial burden and risk is less.
The dispute between Finance and Engineering will have an outcome in terms
of the utility's profitability.
\subsection{Negotiation is Creative}
Negotiation is intrinsically creative. Often, the outcome is not as
predicted, or is unintended by participants, or may even be unwanted by
some participants. On the other hand, an outcome may turn out to be
better than expected. Even when a negotiation does not break new ground
and the outcome is one of those initially sought by one or more participants,
the process used to reach that outcome is fundamentally creative in the
sense that it creates an organizational commitment.
As we have seen, trials of strength embody conflict (because of
incompatible outcomes), and therefore indeterminacy (because no participant can
be certain what the outcome will be). Trials of strength are the
fundamental unit of activity that we want to understand and explore. The
actual unfolding of a trial of strength is a unique performance, so
strictly speaking, a trial of strength can never be repeated. A similar
one could be staged at a different place and time, but each performance
is unique.
This cycle of commitments leading to negotiations which lead to
commitments, some of which conflict with other commitments and thus lead
to further negotiations---this cycle is the way the world works.
\subsection{The Rationale}
The {\em rationale\/} for the outcome of a negotiation is stated at the
end of the negotiation. The rationale(s) given for the outcome are
partly generated during the negotiation process as the participants discuss
the proposed options. As each participant challenges each other's positions,
new beliefs and preferences are created. As the negotiation continues, a
rationale is often created in support of a particular outcome. For
example, in a conflict between Finance and Engineering about which of two
types of plant to build, the rationales supporting the outcome may
describe:
\begin{itemize}
\item
{\bf Predicted beneficial results:} A utility justifies the development
of a new plant: ``Nuclear power will cost less than burning fossil
fuel.''
\item
{\bf Policies guiding conduct:} The management of a utility makes a
policy: ``We must follow the rules and regulations of the Nuclear
Regulatory Commission in building our plant.''
\item
{\bf Reasons tied to specific institutional roles or processes:} A
utility sells a completed, ready to go, nuclear power plant to the state
government (which plans to demolish it) with the justification that the
state has agreed to compensate the utility in other ways.
\item
{\bf Precedent:} It is traditional to run diagnostics for the nuclear
reactor on Monday morning.
\end{itemize}
Precedent may seem like a weak rationale. However, deciding according to
precedent in the absence of strong alternatives has the consequences of
predictability, stability, and increased participant value in organizational work.
In the absence of strong alternatives, using precedent is usually less costly
than constantly redoing a decision process.
The rationale becomes part of the organizational history, and may become
a precedent.
This taxonomy not only describes characteristics of outcome rationales,
it also provides criteria for identifying problems and pointing out ways
in which the process can be made more effective. So the rationale is
much more than the big cheese standing up and saying, ``This course of
action will lead to wonderful results.'' Any rationale can claim
beneficial results. However, the rationale will be judged on its own
merits. Good decisions can have bad rationales, and vice versa.
\subsection{Assessment of Participant Values}
No one can stand outside the system and assess participant values (i.e.
benefits relative to costs) in organizational work. Anyone who wants to
increase participant value must become a player and participate in
organizational processes. All such assessments are made within a
framework of conflict: allies, commitments, incompleteness of
information, limitations of resources, etc. The {\em only way\/} to
increase participant value in organizational work is to become part of the
organizational processes.
For example, the Nuclear Regulatory Commission abolished the requirement
that localities and states must approve emergency evacuation plans before
a nuclear reactor will be granted an operating license, in large part
because of local officials who were refusing to approve the plans. The
commission felt that communities were using the evacuation plant process
to prevent nuclear plants from receiving operating licenses. After the
NRC announced that it would abolish the requirement, many participants
challenged its authority in court. They criticized the commission for
changing the rules in the middle of the licensing process.
Meta-commitments in this instance set new policies and procedures (i.e.,
new commitments) about how commitments get changed. These new
commitments address concerns about how the previous licensing procedure
was carried out (i.e, that the communities had veto power over the
emergency evacuation plans). As a result of this trial of strength, the
commission created new procedures and policies for changing its
regulations so that in the future, various participants will participate in a
better-defined process. The new procedures and policies arose from a
meta-commitment negotiation, and formed a commitment about how to change
other commitments.
All meta-commitment negotiations are concerned with how organizational
processes are changed. Their outcomes affect how subsequent negotiations
are conducted, and thus affect the outcomes (commitments) of subsequent
negotiations.
\subsection{Authority and Responsibility}
The meta-commitment described above changed the Nuclear Regulatory
Commission's authority and responsibilities. The utilities gained power
to participate in decisions about emergency evacuation plans, and the
commission took responsibility for spelling out its processes for
changing the licensing process.
Authority is power, and power is the ability to take action (i.e., use
resources). More precisely, authority is power legitimized by the
commitments of other authorities. For example, a utility has to register
with the Secretary of State in whichever state it operates, so its
organizational power is legitimized by the power of another authority,
namely the Secretary of State. If that authority withdraws this
legitimization, the utility's authority becomes problematical.
An organization's power is its control over resources, and its
responsibilities are its part in its commitments. (Accountability is whether or not it
actually takes those actions and meets those commitments.) So authority
and responsibility are both intimately tied to an organization's
commitments.
Authority can be delegated, but responsibility cannot. Responsibility is
established by the organization's undertaking a certain set of
commitments. An actor might get some help from other participants in meeting
those commitments, but they are still the actor's own responsibility. So
in the narrow sense, the actor cannot delegate responsibility. What it
can do instead is create other organizational arrangements that also
carry the same commitment---and hope that will be sufficient. But if
it's not, that commitment is still that actor's responsibility.
\section{Relationship to Artificial Intelligence}
The question now arises as to the relationship between Open Systems
and Artificial Intelligence. Various technologies have been developed in
Artificial Intelligence for providing a foundation for and structuring of
computation, including:
\begin{itemize}
\item microtheories
\item problem spaces
\item taxonomies
\end{itemize}
\subsection{Microtheories}
One of the most powerful ideas of science is the idea of a {\em
microtheory}. Microtheories are based on a closed-world
assumption---from a set of rules specified in advance, results can be
algorithmically checked for correctness. A spread-sheet is a good
example. The rules are the calculation procedures used in the body of
the spread-sheet. Given the previous values and formulas, an
automaton can algorithmically check whether the new values are correct
in real time. Our notion of a microtheory is very general in that
encompasses all known forms of deduction including first order logic,
nth order logics, modal logics, intuitionistic logic, relevance logic,
lambda calculi, circumscription, default logics, etc [de Kleer, Kripke,
McDermott, McCarthy, Nilsson, Reiter, Sandewall, etc.].
A microtheory has important strengths:
\begin{itemize}
\item
It is portable. A microtheories can be expressed as a stable inscription
that does not spontaneously change and can easily be moved and copied.
\item
The correctness of a derivation is algorithmically decidable solely from
the text of the derivation. Something as simple as an automaton can
decide in real time whether or not a derivation is correct.
\end{itemize}
\noindent
Within a microtheory, there are well-defined methods for dealing with any
conflict that might arise. Thus, negotiations are not very important
within a microtheory because the correctness of derivation can be algorithmically
decidable in a closed world.
Microtheories play an important role in negotiations because they can be
brought to bear on issues and provide support for commitments with
respect to those issues. For example, the utility's Finance and
Engineering departments might each have a different spread-sheet
model of the utility's financial condition with respect to the costs of a
new plant, and each representative can then bring that microtheory to
the negotiation. Their respective recommendations of how the utility
should spend its money might be contradictory. Comparing their
microtheories can help to determine what some of the underlying
conflicts are. They might discover that Finance's Comptroller does not
believe that Engineering can meet its construction schedule. Derivations
of a microtheory can be brought to bear as supporting arguments in the
negotiation. In general, however, there will be a lot of these
microtheories, and they will often have derivations that
formally contradict derivations.
Each microtheory compiles certain methods in a rigorous way. The
Comptroller tries to protect the utility against financial
difficulties---and has successfully negotiated a commitment from
Engineering that the utility will not borrow money if payments will
exceed 25 percent of its income. On the other hand, the Engineering
department maintains that more generating capacity is needed and that
with their construction schedule, the debt payment will never exceed 25
percent.
Many of the microtheories embody various commitments and allies. They
may have been endorsed by the organization itself, or by an outside
organization that has expertise in the field. A spread-sheet microtheory
derived from the tax code can be used to deduce the tax consequences of
differing proposals, and the participant holding this microtheory can claim the
IRS as an ally (i.e., claiming that the IRS will support the conclusions
drawn).
Thus, each side of any conflict (such as whether to pursue the
construction schedule developed by Engineering) will be able to marshall
its own body of microtheories, principles, precedents, and conclusions.
Having a deductive proof based on a microtheory usually does not thereby
carry the day and win the negotiation. The other participant will usually have
a competing microtheory. So microtheories are a strength, but they're
just one tool of negotiation. A powerful tool, but just one tool.
Having a microtheory facilitates negotiation, but in general does not
determine the outcome.
∂15-Nov-89 1401 kirsh%cogsci@ucsd.edu part 2 hewitt
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 15 Nov 89 14:01:06 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA04293
sendmail 5.61/UCSD-2.0-sun via SMTP
Wed, 15 Nov 89 14:01:05 -0800 for JMC@sail.stanford.edu
Received: from cogsci-sun.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA21480 for JMC@sail.stanford.edu; Wed, 15 Nov 89 13:59:40 PST
Date: Wed, 15 Nov 89 14:02 PST
From: David Kirsh <kirsh%cogsci-sun@ucsd.edu>
Subject: part 2 hewitt
To: JMC@SAIL.STANFORD.EDU
Message-Id: <19891115220249.5.KIRSH@cogsci-sun.ucsd.edu>
\subsection{The Role of Logical Deduction}
Several different negotiating strategies can be used with microtheories.
One is to include all the {\em if}s, {\em and}s, {\em but}s and {em
wherefore}s that one can imagine. This creates a cumbersome
microtheory that attempts to cover all possible special circumstances.
For example, ``If the utility uses more than 25 percent of its income for
debt payment, {\em and if furthermore\/} it does not have lots of liquid
assets, {\em and if furthermore\/} \ldots\ then the utility should not
take on more debt.''
A different strategy is to state a very simple rule and let it unfold in
the ongoing negotiation whether any exceptions apply. So the Comptroller
says: ``If construction is delayed, then the utility will spend much more
than 25 percent of its income for debt, so we shouldn't adopt the
construction plan.'' And the other participant replies, ``Yes,
but---Engineering has a good record for completing construction projects
on schedule at close to its estimated cost. Even though Engineering is
building a new kind of plant which can burn either coal or gas, it is not
very different from what it has built before.'' Having simple
microtheories that are parsimonious, easily understood, and clear in
their causality is often a better negotiating strategy than one which
tries to stipulate in advance all of the conditions which govern the
applicability of every rule.
The participants to a negotiation do not know for certain what sorts of
rationale for action and microtheories the others will present. The
applicability of one participant's microtheories depends on what happens during
the negotiation, not on the ability to assemble a large collection of
microtheories ahead of time. Either participant might come up with a
microtheory that the other has not thought of.
The Open Systems model of representations can be used to analyze a
deductive approach that has been explored in [McCarthy, etc.]
By inserting caveats into the axioms of conflicting microtheories,
interactions among the conditions of applicability of the axioms of the
microtheories can be expressed.
For example, various factors bear on the safety of the Diablo Canyon
nuclear reactor. Suppose that we attempt to conduct the negotiation by
writing rules with explicit caveats. Consider the following two rules:
\procedure %
{\rm Rule 1:}
if trained-operators and not(caveat-1), then safe-reactor
\endproc
\noindent
Having trained operators makes for a safe reactor unless it can be shown
that caveat-1 is true. Continuing to axiomatize,
\procedure %
{\rm Rule 2:}
if earthquake-zone and not(caveat-2), then not(safe-reactor)
\endproc
\noindent
Being in an earthquake zone means the reactor is not safe unless it can
be shown that caveat-2 is true.
In this way axioms for caveats can be developed over time
and gradually improved. For example, consider the following rule:
\procedure %
{\rm Interaction Rule 1:}
if trained-operators, then caveat-2
\endproc
\noindent
Having trained operators implies that being in an earthquake zone does
not necessarily imply that the reactor is unsafe.
\procedure %
{\rm Interaction Rule 2:}
if earthquake-zone, then caveat-1
\endproc
\noindent
Also, being in an earthquake zone implies that having trained operators
does not necessarily imply that the reactor is safe.
Unfortunately, the addition of Interaction Rules 1 and 2 blocks the
applicability of Rules 1 and 2. If we have both trained operators and
an earthquake zone, Rule 1 cannot be used since earthquake-zone implies
caveat-1, and Rule 2 cannot be used since trained-operators implies
caveat-2. The following axioms are needed instead:
\procedure %
{\rm Interaction Rule $1'$:}
if trained-operators and not(caveat-Interaction-Rule-1),
then caveat-2
\bline
{\rm{Interaction Rule $2'$:}}
if earthquake-zone and not(caveat-Interaction-Rule-2),
then caveat-1
\endproc
These interaction rules are needed to prevent contradiction and they can
be highly non-modular. For example, the use of interaction rules raises
the following question: Does being in an earthquake zone imply that the
presence of trained operators implies that being in an earthquake zone
implies that the reactor is safe? Questions like this can also be
expressed as rules:
\procedure %
{\rm{Second-Order Interaction Rule 1:}}
if trained-operators, then caveat-Interaction-Rule-2
\bline
{\rm{Second-Order Interaction Rule 2:}}
if earthquake-zone, then caveat-Interaction-Rule-1
\endproc
\noindent
Again, the Second-Order Interaction Rules 1 and 2 cannot be allowed to
stand. Instead, the following rules must be used:
\procedure %
{\rm{Second-Order Interaction Rule $1'$:}}
if trained-operator
and not(caveat-Second-Order-Interaction-Rule-1),
then caveat-Interaction-Rule-2
\bline
{\rm{Second-Order Interaction Rule $2'$:}}
if earthquake-zone
and not(caveat-Second-Order-Interaction-Rule-2),
then caveat-Interaction-Rule-1
\endproc
\noindent
At this point, it becomes very difficult to understand what we are talking
about.
Combining microtheories and adding caveats to rules does not resolve
all conflicts. In the power plant construction example, joining two
microtheories together with caveats leads to the derivation of: ``Yes,
the utility will get an operating license or no, the utility will not get an
operating license.'' But the utility cannot be told ``yes or no.'' Either
they must be told ``yes,'' or they must be told ``no.'' Adding caveats to
rules makes for a cumbersome negotiating strategy that does not
respond easily to changing circumstances.
Logical deduction can model the reasoning within a given microtheory, but
it cannot settle the dispute. Attempts to combine microtheories into a
larger theory by introducing caveats ultimately leads to ``yes or no''
derivations. So logical deduction is an appropriate and valuable tool
within the microtheories held by the respective participants. Quite often,
new microtheories will need to be created in order to better understand
the issues under negotiation. In general, these new microtheories will
not be logical consequences of the microtheories that were already
familiar to the participants before the negotiation began.
\subsection{Problem Spaces}
Problem spaces \cite{ProblemSpaces} can be used as a process modeling
technique. A problem space is a microtheory that provides for:
\begin{itemize}
\item
an {\bf initial state}. For example, the initial state may be the
financial state of the utility before it begins constructing a new plant
\item
one or more {\bf operators} that are applicable to each state. For
example, selling bonds is one of the operations that a utility can
usually take to change its financial condition.
\item
one or more {\bf success states} for one or more of the participants. For
example, the utility can specify its financial goals in terms of revenue,
investment, and earnings.
\end{itemize}
Problem spaces can also be used to model an ongoing negotiation.
Negotiation usually begins with discussion about issues. This process
can be viewed as negotiation about the initial state. Then there is
discussion about how the negotiation should proceed: who will speak, what
the agenda is, and so on---which is analogous to the possible operations
applicable to each state. Also, there is discussion about what
represents a successful outcome. A problem space attempts to model the initial
situation, trajectory, and criteria for when negotiation has ended.
Thus, problem spaces can be a useful way of characterizing an ongoing
process. However, we can rapidly encounter the same difficulty we had
with microtheories: each participant to the negotiation will have its own
problem space of how the negotiation should proceed which will in general
conflict with those of the other participants.
Consider for example the negotiating process to determine whether or
not to give a nuclear plant an operating license. The initial state of the
negotiations is quite problematical. Typically there are thousands of
pages of documentation and claims that have been submitted ahead of
time by the plant owners, the nearby local communities and states, by
environmental groups, and public utility commissions. One of the
participants might say, ``Okay, we are now in this particular initial
state; these are the ways that the negotiation can move forward from
the various states it might get into; and this is what we'll count as an
outcome.'' However, in this case the participants find that they agree
on very little about the starting state of the negotiations. In many
cases the problem spaces are not worked out in such an overt form.
Participants often come to a negotiation with {\em criteria for the
outcome\/} that they initially believe would represent success for
themselves (and possibly for others as well). Also, they would come
with their own understanding of the initial situation and what
negotiation moves would be legitimate. Furthermore in this case the
Nuclear Regulatory Commission decided to change the rules for dealing
with emergency evacuation plans in the middle of the negotiation
process; which neither it nor the other participants initially conceived
as a possible operation. The participants in this case differ greatly on
their characterization of the starting state, the allowable operations, or
which states constitute successful outcomes.
In order to use a problem space for conducting a negotiation, the participants
must first reach an agreement about the initial state of the negotiation. In
general it is difficult to come to an agreement as to exactly what are the
issues at stake. The representatives find it difficult to specify how the
organization they represent would characterize an issue. After some
representatives have stated their understanding of an issue, others may decide
to change their characterization. Thus attempting to precisely characterize
the initial state of a negotiation can be a very problematical undertaking
that has no termination and may even deadlock. To make the situation even
more problematical, a prolonged attempt to precisely specify the initial
state has the effect of itself changing the initial state! Furthermore some
of the representatives may not want to reveal all of their current plans and
understanding concerning of the issues under negotiation for a variety of
reasons--such as being in too preliminary a state of development to share
with others. Specifying in advance the operators that are applicable at each
stage of a negotiation as well as the success states are equally problematical.
In many cases, the problem spaces are not worked out in such an overt
form. Participants often come to a negotiation with {\em criteria for
the outcome\/} that they initially believe would represent success for
themselves (and possibly for others as well). Also, they would come with
their own understanding of the initial situation and what negotiation
moves would be legitimate. In the case of the emergency evacuation
plans, however, the Nuclear Regulatory Commission decided to change the
rules for dealing with such plans in the middle of the negotiation
process; neither the NRC nor the other participants initially conceived of
this as a possible operation.
Again, having a problem space does not guarantee what's going to happen
because each participant brings its own problem space. Participants with
conflicting commitments about a negotiation will have conflicting
problem spaces. Basically, these are conflicts over commitments on how
the negotiation should proceed. The conflicts between these problem
spaces---much as in the cases of microtheories---need to be deal with.
Given the prospects for conflict among problem spaces, it would be
difficult (and perhaps not even desirable) to try to create a problem
space in advance that governs represent the entire, ongoing negotiation.
Instead, problem spaces are probably best used by each participant to
communicate its own analysis of the current negotiating situation,
available options, and desired outcomes.
\section{Conclusions}
The new discipline of Open Systems Technoscience differs from Artificial
Intelligence in several crucial aspects:
\begin{itemize}
\item
In Open Systems Technoscience, the primary indicators of success are {\em
participant values} (which is roughly the benefits to a participant
relative to its costs (including externalities)), {\em robustness} (which is
the ability to keep commitments in the face of conflict), and {\em
scalability} (which is participant value in activity
increase some participant values). Open Systems Technoscience
is grounded in large scale human/telecomputer organizational work. The
primary indicator of success in Artificial Intelligence is the ability to
impress humans with behavior that they will call intelligent. It is
grounded in intelligent agents and robots. In contrast to Artificial
Intelligence, work can proceed on the development of foundations for
Open Systems Technoscience without the need to provide a characterization of
``intelligence''.
\item
In Open Systems Technoscience, {\em representation} is the activity of
communicating with others. Without communication there is no
representation. Communication takes its {\em meaning} from how it affects
the behavior of recipients. In Artificial Intelligence, representation is
traditionally about the correspondence between a structure in an intelligent
agent and a state of affairs in the world.
\item
Open Systems Technoscience views {\em commitment} as a {\em joint course
of action} in which the participants are {\em on course}. The {\em
responsibility} of a participant is its part in the commitment. Artificial
Intelligence has traditionally viewed commitment as a state of mind in
which there is {\em intentionality}
\cite{Cohen-Levesque}\cite{Dennet}.
\end{itemize}
In summary social processes especially those of science, technology, and
engineering \cite{Latour} inform Open Systems Technoscience, whereas Artificial
Intelligence has traditionally turned to neurophysiology, psychology, and
cognitive science.
Open Systems Technoscience provides a framework for analyzing
Artificial Intelligence technologies such as deductive theories,
taxonomies, and dictionaries. Conflict is ubiquitous in Open Systems.
It allows participants to consider and explore their alternatives in a
way that takes other commitments into account. As the participants to
the conflict negotiate their differences, they usually generate
justifications to support their position. They often use microtheories to
bolster their cases. Since microtheories are decontextualized, they can
be carried from place to place and used to seek additional leverage in
many different negotiations. Thus, the use of inference in microtheories
can be seen as a natural kind of specialized activity that often occurs in
the negotiation of conflict. The crucial characteristics of a microtheory
are that the rules are given in advance and that the derivations can be
checked for correctness in real time. The nonmathematical
microtheories of the participants usually conflict. Negotiation of
the conflict that arises can be a source of creativity and robustness.
% \cite{open-systems}
% \cite{robustness-reliability-and-overdetermination}
% \cite{regions-of-the-mind}
\section{Acknowledgments}
First, I would like to express my gratitude and admiration to Wyn Snow
for editorial assistance above and beyond the call of ordinary duty.
Without her help, this paper would contain unboundedly many more
mindtraps.
Second, I wish to acknowledge the aid of Elihu Gerson in repeatedly
pulling me out of intellectual quicksand and setting me back on fruitful
paths.
Third, I wish to acknowledge the help of Les Gasser, David Kirsh, Bruno
Latour, John McCarthy, and Susan Leigh Star for pushing forward in new
directions as well as helping to reconceptualize old ones.
Fourth, I wish to thank members of the Message Passing Semantics
Group for helping to find obscurities and errors.
Fifth (and perhaps most important), I wish to thank Randy Fenstermacher,
Ron Flemming, Sue Gerson, Fanya Montalvo, John Stutz, and other close friends for
helping me to continue to grow.
\section{Related Work}
\vbox{\noindent
Hewitt, Carl,
``Viewing Control Structures as Patterns of Passing Messages,''
{\em A.I.\ Journal}, Vol.~8, No.~3, June 1977, pp.~323--364.}
This paper re-examined the issue of control structures in Artificial
Intelligence. Control structures were previously defined as looking for
the best choice in moving from the current global state to the next one.
The control structure was supposed to accomplish this either by guiding
the production system or by guiding a theorem-prover that was attempting
to search through the realm of possibilities. This paper pointed out that
traditional programming language control structures (such as iteration,
recursion and co-routine) could be analyzed in terms of patterns:
stereotypical or stylized patterns of communication among different
participants. Instead of looking at the behavior of an {\em individual\/}
intelligent agent as Newell and Simon did, this paper initiated the idea
that communities of people are a primary existence proof and analog for
how to extend these ideas.
\bigskip\vbox{\noindent
Kornfeld, William A.\ and Carl Hewitt, ``The Scientific Community
Metaphor,'' {\em IEEE Transactions on Systems, Man, and Cybernetics},
Vol.~SMC-11, No.~1, January 1981, pp.~24--33.}
This paper introduced several important concepts into the Artificial
Intelligence arena, and further develops the ideas Hewitt first discussed
in ``Viewing Control Structures.'' It uses the scientific community as a
model for the problem-solving process, and speaks generally about how
principles and mechanisms of scientific communities might be incorporated
into the problem-solving technology of Artificial Intelligence. Several
fundamental properties of scientific communities have nice analogs for
computing systems that aspire to intelligent behavior. Among these
properties are monotonicity, commutativity, parallelism and pluralism.
The paper also introduces the notion of having sceptics as well as
proponents of different kinds of ideas, and explicates how those kinds of
questions can be investigated concurrently.
This work was somewhat preliminary in nature and is still in need of
further development. Subsequent work in the Message-Passing Semantics
Group at MIT has focused on organizational processes and structures.
\bigskip\vbox{\noindent
Kornfeld, William Arthur,
``Concepts in Parallel Problem Solving,''
Ph.D.\ Thesis, Dept.\ of EECS, MIT, February 1982.}
This is a further development of the work in ``The Scientific Community
Metaphor.'' Kornfeld here shows that by developing a concurrent process
that has critics as well as proponents of ideas, the amount of resources
consumed can, in some cases, be vastly reduced. This results in a kind of
combinatorial implosion instead of the usual combinatorial explosion where
the number of alternatives proliferate indefinitely. Such exponential
proliferation of possibilities is typical of backward-chaining reasoning.
The negotiation described here is very primitive in form, and consists of
entering absolute objections---a very cut-and-dried situation. We would
like to apply this type of process in more relaxed situations where one
has less hard knowledge, and the objections aren't guaranteed to be always
fatal to what they're objecting to.
\bigskip\vbox{\noindent
Barber, Gerald Ram{\'o}n,
``Office Semantics,''
Ph.D.\ thesis, Dept.\ of EECS, MIT, February 1982.}
This paper shows how the viewpoint mechanism introduced in Kornfeld and
Hewitt's ``Scientific Community Metaphor'' can be used to model changing
situations in terms of multiple points of view. It also introduces some
of the kinds of mechanisms for dealing with contradictory microtheories.
\bigskip\vbox{\noindent
Huberman, B.~A. editor,
{\em The Ecology of Computation}
North Holland, 1988}
This book is an excellent collection of articles which deal with the nature,
design, description, implementation, and management of Open Systems. The
articles are grouped in three major sections. Papers in the first section
deal with general issues underlying Open Systems, studies of computational
ecologies, and their similarities with social organizations. Papers in the
second section deal with implementation issues of distributed computation, and
those in the third section discuss the issues of developing suitable languages
and information media for Open Systems.
\bigskip\vbox{\noindent
Stefik, Mark,
``The Next Knowledge Medium,''
{\em The AI Magazine}, Vol.~7, No.~1, Spring 1986, pp.~34--46.}
Stefik describes the growth and spread of cultural knowledge: the kinds of
things that communities of humans do---and shows how the existence of a
technical infrastructure (such as railroads) can greatly facilitate and
accelerate cultural change. Our current knowledge market is static and
pretty much confined to inscriptions: things that can be reduced to a
string of bits (such as a diagram or sentence or literary work) and thus
transported and copied at very small price. Stefik portrays a dynamic
knowledge market that would supplement our current product market. It
would move intelligent models around that have the capability of taking
action. An active knowledge medium could interact with both its human
users and with various kinds of expert systems. He also describes several
current projects, such as the Co-Lab at Xerox Park, that are beginning to
show rudimentary characteristics of an active knowledge medium.
\bigskip\vbox{\noindent
Alvarado, Sergio J., Michael G.~Dyer and Margot Flowers,
``Editorial Comprehension in OpEd Through Argument Units,''
Fifth National Conference on Artificial Intelligence,
Aug 11--15, 1986, Philadelphia, PA,
{\em AAAI-86}, Vol.~1, pp.\ 250--256.}
This paper shows how arguments can be diagrammed in much the same way that
debate contests are often diagrammed by their judges. Such diagramming
examines the beliefs, the tree-structure of the supporting beliefs, and
the way one side can attack the other side's beliefs. (There are really
two kinds of important relationships between the two sides: support
relationships and attack relationships.) The paper presents an analysis
that looks at both the achievement of plans and goals, and the development
of editorials that critique other sides, showing how other sides have
beliefs that are supporting to the opinion that's being reported. This is
quite interesting work in terms of starting to build technology that can
do argument analysis, because that's an important component of
negotiation. Of course, there are other kinds of representations---as John
McCarthy pointed out---in terms of making threats and other kinds of
speech acts, but argument analysis is certainly a very important component
of negotiation.
\bigskip\vbox{\noindent
Devereux, Erik August,
``Processing Political Debate: A Methodology for Data Production
with Special Application to the Lincoln-Douglas Debates,''
B.S.\ thesis, MIT Dept.\ of Political Science, June 1985.}
Devereux expands on something very similar to the argument units in
Alvarado {\em et. al.}. Devereux takes the whole of the Lincoln-Douglas
debates and attempts to identify both attacking statements between Douglas
and Lincoln and supporting links within the individual arguments
themselves. Interestingly enough, there are no supporting links between
the two debaters, so in that respect, the argument units of Alvarado {\em
et al}.\ represent an advance over the analysis that was done by Devereux.
\begin{thebibliography}{9999}
\bibitem[Agha 1986]{agha-phd}
Agha, G., {\em Actors: A Model of Concurrent Computation in Distributed
Systems}, Cambridge, MA: MIT Press, 1986.
\bibitem[Becker 1960]{concept-of-commitment}
Becker, Howard S., ``Notes on the Concept of Commitment,'' {\em American
Journal of Sociology}, Vol.~66, July 1960, pp.~32--40.
\bibitem[Fikes 1982]{commitment-based-framework}
Fikes, Richard E., ``A commitment-based framework for describing informal
cooperative work,'' {\em Cognitive Science}, Vol.~6, 1982, pp.~331--347.
\bibitem[Gerson 1976]{quality-of-life}
Gerson, Elihu M., ``On the Quality of Life,'' {\em American Sociological
Review}, Vol.~41, October 1976, pp.~793--806.
\bibitem[Latour 1987]{science-in-action}
Latour, Bruno, {\em Science In Action}, Cambridge, MA: Harvard University
Press, 1987.
\bibitem[Smith and Davis 1981]{contract-nets}
Smith, R. and Davis, R., ``Frameworks for cooperation in distributed
problem solving,'' {\em IEEE Transactions on Systems, Man, and
Cybernetics}, Vol.~SMC-11, 1981, pp.~61--70.
\bibitem[Star 1983]{simplification-in-scientific-work}
Star, S.L., ``Simplification in scientific work: An example from
neuroscience research,'' {\em Social Studies of Science}, Vol.~13, No.~2,
1983, pp.~205--228.
\bibitem[Strauss 1978]{negotiations}
Strauss, Anselm, {\em Negotiations}, Jossey-Bass, 1978.
\bibitem[Winograd and Flores 1987]{winograd-flores}
Winograd, Terry, and Flores, Fernando, {\em Understanding Computers and
Cognition}, Reading, MA: Addison-Wesley, 1987.
\end{thebibliography}
\end{document}
-------
∂15-Nov-89 1817 jlh@vsop.Stanford.EDU Tuesday schedule for DARPA visit
Received: from vsop.Stanford.EDU by SAIL.Stanford.EDU with TCP; 15 Nov 89 18:17:49 PST
Received: by vsop.Stanford.EDU (5.57/Ultrix2.4-C)
id AA01899; Wed, 15 Nov 89 18:15:07 PST
Message-Id: <8911160215.AA01899@vsop.Stanford.EDU>
To: abbas@isl.stanford.edu, losleben@cis.stanford.edu,
plummer@sierra.stanford.edu, nanni@galileo.stanford.edu,
jlh@vsop.stanford.edu, horowitz@mojave.stanford.edu,
linvill@cis.stanford.edu, jmc@sail.stanford.edu, jct@sail.stanford.edu,
ullman@cs.stanford.edu, gupta@amadeus.Stanford.EDU,
wiederhold@cs.stanford.edu, luckham@sierra.stanford.edu,
cheriton@cs.stanford.edu, gupta@cs.stanford.edu
Cc: margaret@vsop.Stanford.EDU
Subject: Tuesday schedule for DARPA visit
Date: Wed, 15 Nov 89 18:15:05 PST
From: jlh@vsop.Stanford.EDU
Here's a shot at the Tuesday schedule; let me know if something will
not work out and we can iterate. I have tried to balance the time
allocations. Let me know if you need more. This is time only;
time slots coming next.
Time Persons
:30 Weiderhold (need :45?)
:30 Ullman ( need :45?)
:90 Hennessy/Horowitz (Gupta, Lam)
:30 Luckham
1:30 Break for lunch
1:30 Plummer, et. al. (need 2:00?)
:30 Cheriton (need :45?)
:30 DeMicheli
:30 El Gamal
∂15-Nov-89 2117 fuller@sierra.STANFORD.EDU Myrhvold@microsoft
Received: from sierra.STANFORD.EDU by SAIL.Stanford.EDU with TCP; 15 Nov 89 21:17:49 PST
Received: by sierra.STANFORD.EDU (3.2/4.7); Wed, 15 Nov 89 21:16:47 PST
Date: Wed, 15 Nov 89 21:16:47 PST
From: fuller@sierra.STANFORD.EDU (Dwain N. Fullerton)
To: jmc@sail
Cc: nilsson@tenaya, gibbons@sierra.STANFORD.EDU, ct.pes@forsythe,
ct.jfk@forsythe, fuller@sierra.STANFORD.EDU
Subject: Myrhvold@microsoft
Message-Id: <CMM.0.88.627196606.fuller@>
John,
Today I met with Nathan Myrhvold at Microsoft--he is coming on the
Engineering School Advisory Council, and because I was in the area stopped by
to brief him on dates and duties. In the course of the conversation he
mentioned an essay you wrote--"brilliant essay" he called it--on the need to
make it easy to use electronic mail. Myrhvold says that stimulated him to
start a project, and they are very close to having it ready to go. As simple
to use as the telephone, delivers either screen or hard copy off the FAX, etc.
At least that's how I understood it. He said to let you know what you'd
started up there.
Best,
Dwain
∂16-Nov-89 0650 CLT show and tell
Dick, Vladimir, Joe and I met yesterday afternoon.
We agreed that there should be a 15 minute presentation
by each of us with you giving an intro or a summary in the
remaining 15 min. The proposed order of the project presentations
is Jjw, Val, Clt, Rpg.
Does this seem ok?
Dick is going to talk to Scherlis and see if there are any hints
from him as how to improve presentations.
∂16-Nov-89 1058 pehoushe@Gang-of-Four.Stanford.EDU New low-cost extensions to Qlisp
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Nov 89 10:58:19 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA08281; Thu, 16 Nov 89 10:59:37 -0800
Date: Thu, 16 Nov 89 10:59:37 -0800
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8911161859.AA08281@Gang-of-Four.Stanford.EDU>
To: qlisp@Gang-of-Four.Stanford.EDU
Subject: New low-cost extensions to Qlisp
As before, the extensions run on top of /u/pehoushe/bin/new-qlisp,
which is an old copy of new-qlisp.
To load the low cost extensions to Qlisp, (load "/qlisp/nqn-newload"). You
must recompile all task creating forms (old version is still obtained
from (load "/qlisp/nqn-load")).
The improvements include a slightly faster task creation time, about 20
microseconds per fork-join task, and a fast, self-contained predicate for
dynamic spawning (roughly 1.2 microseconds). For reference, a simple
function call takes about 9 microseconds, and a closure creation
takes about 50 microseconds.
Both theory and practice show that the major source of overhead when
using dynamic spawning is actually computing the predicate,
dynamic-spawn-p, for large tree-like computations.
The dynamic spawning predicates in the new vesion are:
Dynamic-Spawn-P1 : T iff local task stack has less than 1 element
Dynamic-Spawn-P2 : T iff local task stack has less than 2 elements
These are inlined functions with no arguments. The first takes 4
machine instructions the second uses 5 machine instructions. For
balanced computation trees, Dynamic-Spawn-P1 is best. For less
balanced trees, Dynamic-Spawn-P2 might be better.
The file /qlisp/nqn-newqlet.lisp contains some useful examples and
/qlisp/nqn-fib.lisp contains some hairy fibonacci examples.
-Dan Pehoushek
∂16-Nov-89 1158 RPG@SAIL.Stanford.EDU New Qlisp Stuff
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Nov 89 11:58:42 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA08455; Thu, 16 Nov 89 11:59:59 -0800
Message-Id: <1CrZQH@SAIL.Stanford.EDU>
Date: 16 Nov 89 1158 PST
From: Dick Gabriel <RPG@SAIL.Stanford.EDU>
Subject: New Qlisp Stuff
To: qlisp@GANG-OF-FOUR.STANFORD.EDU
I guess I don't know how to use this stuff. I loaded nqn-newload
as instructed. I wrote the following test program which does a stupid
thing, but is designed to show me that multiprocessing is happening
as well as some idea of the fairness of the processing:
(defun test (n)
(let ((a 0))
(qlet t ((x (dotimes (i n a) (when (evenp i) (incf a))))
(y (dotimes (i n a) (when (oddp i) (incf a)))))
(list x y))))
It works fine in the base Qlisp Dan suggested.
When I compiled it with Dan's stuff loaded, it told me that the variable
named I was special and then it complained about the syntax of DOTIMES. I
restructured my program as follows, and it never terminated:
(defun test (n)
(let ((a 0))
(qlet t ((x (progn (dotimes (i n) (when (evenp i) (incf a))) a))
(y (progn (dotimes (i n) (when (oddp i) (incf a))) a)))
(list x y))))
I ran it by saying:
(qeval (test 10000))
-rpg-
∂16-Nov-89 1206 iam@Gang-of-Four.Stanford.EDU New Qlisp Stuff
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Nov 89 12:06:19 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA08490; Thu, 16 Nov 89 12:07:37 -0800
Date: Thu, 16 Nov 89 12:07:37 -0800
From: iam@Gang-of-Four.Stanford.EDU (Ian Mason)
Message-Id: <8911162007.AA08490@Gang-of-Four.Stanford.EDU>
To: RPG@SAIL.Stanford.EDU
Cc: qlisp@GANG-OF-FOUR.STANFORD.EDU
In-Reply-To: Dick Gabriel's message of 16 Nov 89 1158 PST <1CrZQH@SAIL.Stanford.EDU>
Subject: New Qlisp Stuff
the forms (other than the last) that get spawned need to be function calls, for example
(defun test (n)
(let ((a 0))
(labels ((task-1 () (dotimes (i n a) (when (evenp i) (incf a)))))
(qlet t ((x (task-1))
(y (dotimes (i n a) (when (oddp i) (incf a)))))
(list x y)))))
also rather than qeval, try (cpu (test 10000))
∂16-Nov-89 1252 RPG@SAIL.Stanford.EDU New Qlisp Stuff
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Nov 89 12:52:19 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA08560; Thu, 16 Nov 89 12:53:42 -0800
Message-Id: <1urwCw@SAIL.Stanford.EDU>
Date: 16 Nov 89 1251 PST
From: Dick Gabriel <RPG@SAIL.Stanford.EDU>
Subject: New Qlisp Stuff
To: qlisp@GANG-OF-FOUR.STANFORD.EDU
I guess I don't know how to use this stuff. I loaded nqn-newload
as instructed. I wrote the following test program which does a stupid
thing, but is designed to show me that multiprocessing is happening
as well as some idea of the fairness of the processing:
(defun test (n)
(let ((a 0))
(qlet t ((x (dotimes (i n a) (when (evenp i) (incf a))))
(y (dotimes (i n a) (when (oddp i) (incf a)))))
(list x y))))
It works fine in the base Qlisp Dan suggested.
When I compiled it with Dan's stuff loaded, it told me that the variable
named `I' was special and then it complained about the syntax of DOTIMES. I
restructured my program as follows, and when I started running it, it never
terminated:
(defun test (n)
(let ((a 0))
(qlet t ((x (progn (dotimes (i n) (when (evenp i) (incf a))) a))
(y (progn (dotimes (i n) (when (oddp i) (incf a))) a)))
(list x y))))
I ran it by saying:
(qeval (test 10000))
-rpg-
∂16-Nov-89 1254 rpg@lucid.com New Qlisp Stuff
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Nov 89 12:54:23 PST
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA08579; Thu, 16 Nov 89 12:55:32 -0800
Received: from rose ([192.31.212.83]) by LUCID.COM id AA20546g; Thu, 16 Nov 89 12:50:02 PST
Received: by rose id AA01647g; Thu, 16 Nov 89 12:51:36 PST
Date: Thu, 16 Nov 89 12:51:36 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911162051.AA01647@rose>
To: iam@Gang-of-Four.Stanford.EDU
Cc: qlisp@GANG-OF-FOUR.STANFORD.EDU
In-Reply-To: Ian Mason's message of Thu, 16 Nov 89 12:07:37 -0800 <8911162007.AA08490@Gang-of-Four.Stanford.EDU>
Subject: New Qlisp Stuff
Ah, I guess these messages should point out that the new Qlisp stuff isn't
common lisp.
-rpg-
∂16-Nov-89 1336 RPG@SAIL.Stanford.EDU New Qlisp Stuff
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 16 Nov 89 13:36:20 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA08747; Thu, 16 Nov 89 13:37:31 -0800
Message-Id: <krwlb@SAIL.Stanford.EDU>
Date: 16 Nov 89 1335 PST
From: Dick Gabriel <RPG@SAIL.Stanford.EDU>
Subject: New Qlisp Stuff
To: qlisp@GANG-OF-FOUR.STANFORD.EDU
Hm, now I guess I don't know what CPU does. I modified my program
as follows:
(defun test (n)
(multiple-value-bind (f g h)
(f)
(qlet t ((x (funcall h (funcall f n)))
(y (funcall h (funcall g n))))
(list x y))))
(defun f ()
(let ((a 0))
(flet ((f (n) (dotimes (i n) (when (evenp i) (incf a))))
(g (n) (dotimes (i n) (when (oddp i) (incf a))))
(h (x) (declare (ignore x)) a))
(values #'f #'g #'h))))
I thought this satisfied all the constraints Dan's stuff imposes,
but when I ran (CPU (TEST 100000)) it printed out a bunch of statistics
and then the list (50003 100000), which means virtually no parallel processing
took place.
-rpg-
∂16-Nov-89 1402 SJG (on TTY73, at TV-112 1402)
I never left SAIL! I log on from time to time to print things that I've
ftp'ed from my micro; then I (apparently at least) disappear again. Sort
of like my periodic but brief visits to the third floor ...
Hope you're well --
Matt
∂17-Nov-89 0817 Mailer Re: Fitzwater speaks
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 17 Nov 89 08:16:59 PST
Received: from LOCALHOST by Neon.Stanford.EDU with SMTP (5.61/25-eef) id AA23081; Fri, 17 Nov 89 08:16:57 -0800
Message-Id: <8911171616.AA23081@Neon.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Cc: su-etc@SAIL
Subject: Re: Fitzwater speaks
In-Reply-To: Your message of 17 Nov 89 08:11:00 -0800.
<qs7r#@SAIL.Stanford.EDU>
Date: Fri, 17 Nov 89 08:16:55 -0800
From: holstege@Neon.Stanford.EDU
Well, as a person who holds non-violence as a high principle, I interpret
'fight' rather differently. Delete it, or substitute 'work', if you like.
The point is that someone who tailors his principles to suit the convenience
of the moment has no such principles, protestations to the contrary
notwithstanding. -- mary
∂17-Nov-89 0859 MPS Faculty Report
Good morning;
Joyce is asking for the Fac. Rpt. Have you had the
chance to do yours? Thanks.
Pat
∂17-Nov-89 0959 VAL Seminar at Berkeley
SPECIAL SEMINAR; Vladimir Lifschitz; Stanford University;
Applications Of Formal Nonmonotonic Reasoning; Wednesday,
November 29, 1989; 4:00 p.m. - 5:30 p.m.; 120C Bechtel
∂17-Nov-89 1141 ceb@might.Stanford.EDU possibility of a referee?
Received: from might.Stanford.EDU by SAIL.Stanford.EDU with TCP; 17 Nov 89 11:40:57 PST
Received: by might.Stanford.EDU (5.57/Ultrix3.0-C)
id AA02331; Fri, 17 Nov 89 11:38:25 PST
Date: Fri, 17 Nov 89 11:38:25 PST
From: ceb@might.stanford.edu (Charles Buckley)
Message-Id: <8911171938.AA02331@might.Stanford.EDU>
To: jmc@sail.Stanford.EDU
Subject: possibility of a referee?
I was a student of yours (in your course) several years ago, and I was
wondering if you might agree to act as a referee for a job I am
considering.
The situation is as follows: This would be a faculty position, with
research specialty concentrating in modelling and simulation, and
including programming and architecture pardigms for that purpose. It
would be a joint appointment, mechanical engineering and CS (my
doctorate was in ME).
As a general reference you hardly know me, but there are other
referees for that purpose - the open question is whether or not
what I do is "of interest to computer science".
You seem particularly suited for this, since during the past several
years I have been working on parallelizable algorithms for geometric
modelling in Lisp. I had in fact contacted Joe Weening some time ago
from far away about the possibility of porting my implemented and
working code to your Qlisp system (there was also an Alliant there),
both as a means of studying the algorithm and of the efficacy of Qlisp
(answering questions concerning how it performs on a large system
using lots of bignum arithmetic, requiring lots of garbage collection,
etc.) Naive divide-and-conquer parallelism will be trivial to
implement; but there are also questions of intra merge-step
parallelism, and I am now considering the application of things like
futuring and eager evaluation to this problem.
Even before this job possibility came up, I had taken advantage of my
being once again in the area to resume contacts with Joe about doing
this, and he has set me up with an account on gang-of-four. Dan
Pehoushek has filled me in on the ins and outs of using the
experimental new Lisp system with processor-local queues. I am
currently porting my code to common Lisp, after which time it will
appear on your machine (it takes up disk space, and I wanted to
minimize this impact).
So, while I ask you to evaluate me, I also give you a means to do it
which is beyond reproach - the kinds of questions expected to be
answered in such an evaluation would have their answers become evident
in the course of my carrying out this experiment, which I had wanted to
do anyway - you could ask your staff, look at the code and results,
ask me, etc.
What do you think?
∂17-Nov-89 1533 VAL job hunting news
Howard Barringer from Manchester told me that the Chair in AI had been given
to somebody else. (He couldn't tell me to whom, because it's not official yet.
He said I probably wouldn't recognize the name.) They will probably offer me
(according to Barringer) a visiting professorship in "computational logic",
for 4 or 5 years. This implies that I would be expected to be involved in
teaching and research, rather than administration. There is a possibility
that that position will become permanent.
Alan Mackworth from UBC called and said they'd like to receive an application
from me.
Yale is looking for an AI person. I wrote to McDermott and asked him whether I
should apply.
Wilensky invited me to give a seminar at Berkeley on Nov 29.
The places that Gibbons suggested to me for the purpose of playing games with
Stanford, besides Berkeley, are MIT and CMU. I understand that MIT is impossible;
is there a possibility that CMU may be interested?
∂17-Nov-89 1545 MPS
I will be in around 9:30-10:00 on Monday. Have an
appointment.
Pat
∂17-Nov-89 1555 pehoushe@Gang-of-Four.Stanford.EDU The time of basic operations
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 17 Nov 89 15:54:52 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA14047; Fri, 17 Nov 89 15:56:05 -0800
Date: Fri, 17 Nov 89 15:56:05 -0800
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8911172356.AA14047@Gang-of-Four.Stanford.EDU>
To: qlisp@Gang-of-Four.Stanford.EDU
Subject: The time of basic operations
For our papers, it will be useful to have the times for
some basic operations. The following numbers are moderately
accurate. I am less certain about the numbers without
decimal points.
Estimated time of a few basic lisp operations:
car & cdr: 1.2 Microseconds
(null X): 1.2 microseconds
(setf s i): .9 microseconds
(incf s): 1.9 microseconds
funcall: 9.0 microseconds
cons: 12.5 microseconds
closure: 55 microseconds
The funcall is the time for a function of 1 argument
that returns nil; the closure closes over 1 variable.
The (null x) is branched on.
And in the low-cost extensions:
Via qlet, qvalues, #!, or #?:
Creating a task: 20.0 microseconds
Dynamic-spawn-P: 1.2 microseconds
Swapping: 30 microseconds
(SPAWN T form): 100 microseconds
If anyone wants to know how I got these numbers, I can provide
the programs.
-dan
P.S.
The high cost version of Qlet tasks (in the latest new-qlisp) take
roughly 2080 microseconds per task; but, you can't mix both high and
low cost extensions in the same program; in my papers, I won't mention
the high-cost version of spawning.
∂17-Nov-89 1643 Mailer re: A Sense of Direction
To: JMC@SAIL.Stanford.EDU, su-etc@SAIL.Stanford.EDU
From: Robert W Floyd <RWF@SAIL.Stanford.EDU>
[In reply to message from JMC rcvd 17-Nov-89 08:17-PT.]
Fine, let intelligent people have lots of children. But
then what happens if ALL the intelligent people want to
be lawyers and stockbrokers and send their kids to day care?
The kids get raised by some not-so-intelligent people.
A quotation:
While I - and most of my friends - were saying our minds
were "too good" to stay at home and raise our children, none
of us ever asked the question, "Then what sorts of minds
_should_ be raising our children - minds that were _not_
very good?"
My carefully worded advertisements for childcare literally
came back to haunt me...I wanted someone who would encourage
my children's creativity, take them on interesting outings,
answer all their little questions, and rock them to sleep.
I wanted someone who would be a "part of the family."
Slowly, painfully, after really thinking about what I wanted
for my children and rewriting advertisement after advertisement,
I came to the stunning realization that the person I was looking
for was right under my nose. I had been desperately trying to
hire me.
∂17-Nov-89 1650 rpg@lucid.com The time of basic operations
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 17 Nov 89 16:50:00 PST
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA14150; Fri, 17 Nov 89 16:51:18 -0800
Received: from rose ([192.31.212.83]) by LUCID.COM id AA01076g; Fri, 17 Nov 89 16:46:24 PST
Received: by rose id AA02609g; Fri, 17 Nov 89 16:48:02 PST
Date: Fri, 17 Nov 89 16:48:02 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911180048.AA02609@rose>
To: pehoushe@Gang-of-Four.Stanford.EDU
Cc: qlisp@Gang-of-Four.Stanford.EDU
In-Reply-To: Dan Pehoushek's message of Fri, 17 Nov 89 15:56:05 -0800 <8911172356.AA14047@Gang-of-Four.Stanford.EDU>
Subject: The time of basic operations
I've been thinking about Dan's thing - what it is and what to call it.
First, it seems like the target of a compiler transformation. If a compiler
sees this:
(qlet t ((x (f <1form1>...<1formn>))
(y (g <2form1>...<2formm>))) ...)
It might replace it with Dan's thing if computing <iformj> does not
seem to take up too much of the parent process's time (otherwise, do
the usual Qlet thing). I think this is the ultimate best use for Dan's
idea, unless the functions f and g have behavioral constraints on them
too.
I think it is a very bad idea to continue calling Dan's thing Qlet,
because it just isn't the same. Internally we can call a cat a dog,
but not externally. I propose the following name, though the syntax is
pretty lousy:
(pfuncall f1 (1arg1...1argn1)
f2 (2arg1...2argn2)
...
fm (marg1...margnm)) => multiple values
Each iargji is evaluated, in left to right order, and then processes
for each of the functions fi are spawned with the corresponding
arguments passed. It returns the values the parallel call of these
functions returns. The bad syntax is enclosing the argument forms in
parentheses.
Here is a way to get rid of the weird parentheses, but at the expense
of keywords - though this is a special form or macro:
(pfuncall :function f1 :args 1arg1...1argn1
:function f2 :args 2arg1...2argn2
...
:function fm :args marg1...margnm) => multiple values
We might as well have the apply version:
(papply f1 (1arg1...1argn1*)
f2 (2arg1...2argn2*)
...
fm (marg1...margnm*)) => multiple values
This is the apply version of the above. There maybe should be two
others, one for each of these with multiple values.
To get a binding form you can write:
(multiple-value-call #'(lambda (...)...) (pfuncall ...) ...)
Or, you can think of some other name and syntax, but don't use Qlet.
The key thing is to separate the evaluation of arguments from the
spawning of function/processes. Here is another try:
(process-let ((x1 (make-function-process <f1> <1arg1>...<1argn>))
...
(xn (make-function-process <fn> <1arg1>...<1argn>)))
...)
-rpg-
∂17-Nov-89 1717 VAL Boehm's visit
I'm planning to make a 10-15 minute presentation using the two slides I'm leaving
on your terminal. It would be a good idea to let me speak after you explain why
formalizing common sense is important.
∂17-Nov-89 1957 Mailer re: Berlin@Deutschland
Received: from akbar.cac.washington.edu by SAIL.Stanford.EDU with TCP; 17 Nov 89 19:50:06 PST
Received: from tomobiki-cho.cac.washington.edu by akbar.cac.washington.edu
(5.61/UW-NDC Revision: 2.6 ) id AA01799; Fri, 17 Nov 89 19:50:04 -0800
Date: Fri, 17 Nov 1989 19:41:40 PST
From: Mark Crispin <MRC@CAC.Washington.EDU>
Sender: mrc@Tomobiki-Cho.CAC.Washington.EDU
Subject: re: Berlin@Deutschland
To: John McCarthy <JMC@sail.stanford.edu>
Cc: su-etc@sail.stanford.edu
In-Reply-To: <1Gs7GW@SAIL.Stanford.EDU>
Message-Id: <MailManager.627363700.5474.mrc@Tomobiki-Cho.CAC.Washington.EDU>
JMC, I assume that reporting what the people in question themselves say they
want isn't accurate enough for you?
-------
∂17-Nov-89 2257 GLB
At the conference Logic from Computer Science
the following paper was presented:
S.Buss, Department of Mathematics, UC San Diego
THE UNDECIDABILITY OF k-PROVABILITY FOR THE SEQUENT CALCULUS.
ABSTRACT: The k-provability problem is: given a first-order formula
and an integer k, decide if the formula has a proof of k or fewer
lines of length. This is undecidable for the sequent calculus.
The proof is based on two techniques: (a) careful analysis of the
`logical flow graph' of a proof which traces the influence of
a formula through a proof and (b) reducing to the problem of
second-order unification with partial substitution.
-----
The remarkable thing for us is that (a) is essentially the technique
of chains introduced by Jussi, on which I have been working so far.
Buss' paper provides evidence that this notion is basic:
it allows to minimize the bureaucracy of logical calculi and attack
minimality problems.
∂18-Nov-89 0600 CLT@SAIL.Stanford.EDU qlet or not qlet
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 18 Nov 89 06:00:13 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA16212; Sat, 18 Nov 89 06:01:30 -0800
Message-Id: <GsUAr@SAIL.Stanford.EDU>
Date: 18 Nov 89 0559 PST
From: Carolyn Talcott <CLT@SAIL.Stanford.EDU>
Subject: qlet or not qlet
To: RPG@SAIL.Stanford.EDU
Cc: qlisp@SAIL.Stanford.EDU
I agree that calling a cat a dog externally is a bad idea. Even
internally. As to syntax -- maybe I have missed something, but
why not
(pfuncall (f1 1arg1...1argn1)
(f2 2arg1...2argn2)
...
(fm marg1...margnm)) => multiple values
if you really want multiple values. Otherwise process-let seems
pretty good.
∂18-Nov-89 0938 RPG@SAIL.Stanford.EDU re: qlet or not qlet
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 18 Nov 89 09:38:30 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA16420; Sat, 18 Nov 89 09:39:52 -0800
Message-Id: <ksXMo@SAIL.Stanford.EDU>
Date: 18 Nov 89 0938 PST
From: Dick Gabriel <RPG@SAIL.Stanford.EDU>
Subject: re: qlet or not qlet
To: CLT@SAIL.Stanford.EDU
Cc: qlisp@SAIL.Stanford.EDU
[In reply to message from CLT sent 18 Nov 89 0559 PST.]
There are two reasons to not make the forms looks like function applications,
one of them easily fixable. The fixable one is that funcall and apply imply
evaluated function positions (evaluated in the ordinay sense). The other
one is that the expression (f a1...an) always looks like it expresses a
series of operations starting with evaluating the arguments, and it is
easy to think, then, that those evaluations are part of the operations
that happen in parallel. What you really want to write is something like:
(pfuncall f1 ... fn)
where each fi is the function along with its arguments.
Probably this objection is not damning, but worth thinking about.
This leads to what I think might be a really interesting idea, which is
there ought to be no explicitly parallel constructs and there ought to be
subtypes of functions, one of which is such that calling one spawns a
process. So, you don't write:
(defun f (x) ...)
(defun g (x) ...)
(qlet t ((x (f a))
(y (g b)))
...)
but
(pdefun f (x) ...)
(pdefun g (x) ...)
(let ((x (f a))
(y (g b)))
...)
When you reason about whether to write the qlet in the first one, you
reason about the natures of f and g - can it be run orthogonally. So the
latter let's you express this in the natural places - the functions f and
g have the property that they can be run independently, and therefore in
parallel.
These functions can, implementationally, return futures or the
implementation of let can cause a join just before binding. There is no
difference in semantics, and it is up to the implementor to decide which
to use. If one runs faster than the other, this is the same sort of
criticism as one's floating point arithmetic runs faster than the other.
In any event, it seems that the significant thing Dan is doing is identifying
processes and functions, which needs to be explicit.
-rpg-
∂18-Nov-89 1000 JMC
Acme security systems
∂18-Nov-89 1044 @SAIL.Stanford.EDU:jlm@lucid.com qlet or not qlet
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 18 Nov 89 10:44:21 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA16494; Sat, 18 Nov 89 10:45:42 -0800
Received: from lucid.com by SAIL.Stanford.EDU with TCP; 18 Nov 89 10:43:59 PST
Received: from edsel ([192.31.212.1]) by LUCID.COM id AA04427g; Sat, 18 Nov 89 10:40:33 PST
Received: from chappaquiddick. by edsel id AA11290g; Sat, 18 Nov 89 10:34:34 PST
Received: by chappaquiddick. (4.0/SMI-4.0)
id AA02663; Sat, 18 Nov 89 10:46:20 PST
Date: Sat, 18 Nov 89 10:46:20 PST
From: jlm@lucid.com (Jim McDonald)
Message-Id: <8911181846.AA02663@chappaquiddick.>
To: RPG@SAIL.Stanford.EDU
Cc: CLT@SAIL.Stanford.EDU, qlisp@SAIL.Stanford.EDU
In-Reply-To: Dick Gabriel's message of 18 Nov 89 0938 PST <ksXMo@SAIL.Stanford.EDU>
Subject: qlet or not qlet
(pdefun f (x) ...)
(pdefun g (x) ...)
(let ((x (f a))
(y (g b)))
...)
When you reason about whether to write the qlet in the first one, you
reason about the natures of f and g - can it be run orthogonally. So the
latter let's you express this in the natural places - the functions f and
g have the property that they can be run independently, and therefore in
parallel.
Is it generally the case that all such functions can be run
independently, or do you sometimes get sets of functions such that
functions within a set can be run orthogonally but functions from
distinct sets cannot? (This seems to be the reverse of the
restriction I'd normally expect--that functions within a set would be
non-orthogonal while functions from distinct sets would be
orthogonal.) When the orthogonality is not global, you might need to
be able to express finer constraints than pdefun allows.
jlm
P.S. The syntax that used :ARGS and :FUNCTIONS has the (minor?) problem
that it becomes hard to pass the keywords :ARGS and :FUNCTIONS as
arguments to the functions.
P.P.S. In considering alternatives, I thought about something like
(peval :eval-args-first (f x) (g y)) where a set of keywords
would tell the special form what the rules are. I don't
particularly like this approach, but it seems there's an idea
lurking there--that one might somehow reify the evaluation
scheme and make it an argument to be passed along as part of
the environment. Then arbitary evaluation templates become
possible.
∂18-Nov-89 1114 CLT@SAIL.Stanford.EDU qlet or not qlet
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 18 Nov 89 11:14:07 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA16539; Sat, 18 Nov 89 11:15:29 -0800
Message-Id: <osYgK@SAIL.Stanford.EDU>
Date: 18 Nov 89 1113 PST
From: Carolyn Talcott <CLT@SAIL.Stanford.EDU>
Subject: qlet or not qlet
To: RPG@SAIL.Stanford.EDU
Cc: qlisp@SAIL.Stanford.EDU
% This leads to what I think might be a really interesting idea, which is
% there ought to be no explicitly parallel constructs and there ought to be
% subtypes of functions, one of which is such that calling one spawns a
% process.
This sounds to me very much like actors -- which for our purposes can
be thought of as closure like objects which when invoked (receive a message)
start up a new process (task) to do the computation. That is the only
means of specifying parallelism in actor languages. Other constructs
are defined in terms of actors. To reason about whether two actors
can meaningfully be run in parallel one needs to know what other
actors they share access to, if any.
∂18-Nov-89 1406 @SAIL.Stanford.EDU:rpg@lucid.com qlet or not qlet
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 18 Nov 89 14:06:48 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA16898; Sat, 18 Nov 89 14:08:02 -0800
Received: from lucid.com by SAIL.Stanford.EDU with TCP; 18 Nov 89 14:06:10 PST
Received: from rose ([192.31.212.83]) by LUCID.COM id AA04801g; Sat, 18 Nov 89 14:02:48 PST
Received: by rose id AA03036g; Sat, 18 Nov 89 14:04:27 PST
Date: Sat, 18 Nov 89 14:04:27 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911182204.AA03036@rose>
To: jlm@lucid.com
Cc: CLT@SAIL.Stanford.EDU, qlisp@SAIL.Stanford.EDU
In-Reply-To: Jim McDonald's message of Sat, 18 Nov 89 10:46:20 PST <8911181846.AA02663@chappaquiddick.>
Subject: qlet or not qlet
I was assuming there were categories of functions:
a. must be run alone
b. can be run with any others in the same class
c. must not be run with any others in the same class, but other
classes are ok.
There may be others, plus complicated combinations - can be
run with any in its class but with no others in some other classes.
I guess there is an algebra of acceptable class structures during
execution.
-rpg-
∂18-Nov-89 1938 pehoushe@Gang-of-Four.Stanford.EDU processes and functions
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 18 Nov 89 19:37:55 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA17984; Sat, 18 Nov 89 19:39:01 -0800
Date: Sat, 18 Nov 89 19:39:01 -0800
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8911190339.AA17984@Gang-of-Four.Stanford.EDU>
To: rpg@lucid.com
Cc: qlisp@Gang-of-Four.Stanford.EDU
In-Reply-To: Richard P. Gabriel's message of Fri, 17 Nov 89 16:48:02 PST <8911180048.AA02609@rose>
Subject: processes and functions
rpg wrote:
>In any event, it seems that the significant thing Dan is doing is identifying
>processes and functions, which needs to be explicit.
There was some motivation, other than efficiency, behind the
pre-evaluation of arguments when spawning tasks. I took the view that
a process should be:
Program + Inputs (and special variable environment)
Why use this functional form, instead permitting arbitrary
expressions? There are two answers. The first, uninteresting, answer
is efficiency. Spawning arbitrary forms requires creating a closure,
while the program+inputs approach is similar in complexity to function
calling.
The second, most interesting answer, from a language designer's
perspective is effectiveness, or, in the case of closures,
ineffectiveness. Closures are just not highly transparent for most
programmers. Allowing arbitrary forms (necessitating closures) to be
processes yields unexpected results in many common situations.
The following example illustrates the ineffectiveness of closures as
processes:
(defun foo (n)
(let ((l nil))
(dotimes (i n)
(push
(spawn (t) (fib i))
l))
l))
At first glance, you would expect (foo 20) to spawn 20 calls to fib on
the integers 0..19. Serially, I got the expected result:
(4181 2584 1597 987 610 377 233 144 89 55 34 21 13 8 5 3 2 1 1 0)
In parallel, I tried several times, getting a different undesirable
result each time:
(6765 6765 2584 1597 987 610 377 233 144 89 55 34 21 13 8 5 3 2 1 1)
(6765 6765 4181 1597 987 610 377 233 144 89 55 21 21 13 8 5 3 2 1 1)
(6765 6765 4181 1597 987 610 377 233 144 89 55 34 13 13 8 5 3 2 1 0)
(6765 6765 2584 1597 987 610 233 233 144 89 55 21 21 13 8 3 3 2 1 0)
It gets even worse for larger values of n.
Obviously, I did not get the expected result from this program. The
problem is that each spawned task references the lexical variable i,
which changes with each iteration of the loop.
However, using the Program + Inputs formulation, albeit simple, gives
the expected result. In this, more functional paradigm, sharing
lexical variables among processes requires explicit creation of
closures. It seems to me that the purpose of lexical scope is to
make program behavior fairly transparent; allowing arbitrary forms to
be processes clouds the transparency (due to lexical variable
sharing), while the functional style retains the "transparency".
-dan
P.S. In consideration of pfuncall, papply, pdefun, how about adding
some DECLARE and THE annotations, as a non-intrusive mechanism for
locating and exploiting parallelism? Something like this:
;; side effect free definition of fib
(defun fib (n)
(declare (annotation pure-function))
(if (< n 2)
n
;; THE is common lisp, meaning hint-for-compiler
(the fork-join
(+ (the branch (fib (- n 2)))
(the branch (fib (1- n)))))))
∂19-Nov-89 0847 RPG reply to message
[In reply to message rcvd 19-Nov-89 02:05-PT.]
Not yet. I tried calling Scherlis a few times. I'll try again today.
-rpg-
∂19-Nov-89 1141 rpg@lucid.com processes and functions
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 19 Nov 89 11:41:45 PST
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA19351; Sun, 19 Nov 89 11:43:01 -0800
Received: from rose ([192.31.212.83]) by LUCID.COM id AA00225g; Sun, 19 Nov 89 11:01:47 PST
Site:
Received: by rose id AA03455g; Sun, 19 Nov 89 11:03:28 PST
Date: Sun, 19 Nov 89 11:03:28 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911191903.AA03455@rose>
To: pehoushe@Gang-of-Four.Stanford.EDU
Cc: qlisp@Gang-of-Four.Stanford.EDU
In-Reply-To: Dan Pehoushek's message of Sat, 18 Nov 89 19:39:01 -0800 <8911190339.AA17984@Gang-of-Four.Stanford.EDU>
Subject: processes and functions
The analysis that shows that closures are ineffective because of its
interaction with DOTIMES is not compelling. First, I couldn't
understand what ``effective'' means in this situation. Certainly it is
not intended to mean ``computable'', which would be its classic
meaning. Second, the analysis does not seem to address the macro
DOTIMES. Its description in CLtL indicates that side effects like SETQ
if the iteration variable might have implementation dependent behavior.
This indicates that various expansions are allowable.
To simplify, the expansion of (DOTIMES (I 2) (F I)) could be any of these:
(LET ((I 0))
(F I)
(INCF I)
(F I)
NIL)
(PROGN (LET ((I 0)) (F I))
(LET ((I 1)) (F I))
NIL)
(PROGN (F 0)
(F 1)
NIL)
(LET ((I 0))
(TAGBODY
(GO TEST)
LOOP
(F I)
(INCF I)
TEST
(WHEN (< I 2) (GO LOOP))))
(LET ((GENSYM 0)) ;where GENSYM is a gensym
(TAGBODY
(GO TEST)
LOOP
(LET ((I GENSYM))
(F I))
(INCF GENSYM)
TEST
(WHEN (< GENSYM 2) (GO LOOP))))
The first and fourth have closure ``problems'' when I is captured.
The second, third, and fifth do not.
In sequential Common Lisp programs, the capture of the variable I
must be explicit, in Qlisp it is implicit. In general the introduction
of closures by the implementation should be transparent, and here, because
the expansion of DOTIMES (and DOLIST) is incorrect in the context of
Qlisp, the introduction of closures is not transparent. Therefore, the
solution seems to me to repair DOTIMES and DOLIST, not to redefine
the concept of process.
This does not mean that there should not be a notion of process/function,
and, in fact, in Qlisp there is - the process closure. If FIB were made a
process closure in Dan's example, there would not have been any problem.
By the way, the output of the following program is *unspecified* in
Common Lisp:
(dolist (x (let ((a nil))
(dotimes (i 10 a)
(push #'(lambda () (print i)) a))))
(funcall x))
Implementations may specify it, and Qlisp should specify it as indicated
by Dan's analysis.
-rpg-
∂19-Nov-89 1548 RPG@SAIL.Stanford.EDU Restrictions on Process Functions
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 19 Nov 89 15:48:36 PST
Received: from Sail.Stanford.EDU by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA19612; Sun, 19 Nov 89 15:49:43 -0800
Message-Id: <gtsha@SAIL.Stanford.EDU>
Date: 19 Nov 89 1548 PST
From: Dick Gabriel <RPG@SAIL.Stanford.EDU>
Subject: Restrictions on Process Functions
To: qlisp@Gang-of-Four.Stanford.EDU
Dan says that the basic unit of process is
Program + Inputs (and special variable environment)
The actual restriction seems to be
Program (with null lexical environment and trivial dynamic environment)
+ Inputs (and special variable environment)
Neither of the following seems to work, though they appear to satisfy the
constraints of the QLET forms having a simple function invocation pattern.
I think part of the reason these are low cost extensions is that they are
also low functionality extensions as well.
(defun test (n m)
(flet ((f (x) (dotimes (i x) (when (< (/ n 2) i) (return-from test 'f))))
(g (x) (dotimes (i x) (when (< (/ n 2) i) (return-from test 'g)))))
(qlet t ((x (f n))
(y (g m)))
(values x y))))
(defun test (n m)
(catch 'test
(flet ((f (x) (dotimes (i x) (when (< (/ n 2) i) (throw 'test 'f))))
(g (x) (dotimes (i x) (when (< (/ n 2) i) (throw 'test 'g)))))
(qlet t ((x (f n))
(y (g m)))
(values x y)))))
∂19-Nov-89 1846 RPG Boehm
To: JMC@SAIL.Stanford.EDU, CLT@SAIL.Stanford.EDU, VAL@SAIL.Stanford.EDU,
JSW@SAIL.Stanford.EDU
I finally got a hold of Scherlis. Here is what he suggests:
Boehm is an engineering type of guy who is interested in the impact of
DARPA programs on computing and software engineering. He is not as well
grounded in computer science, but he is more thoughtful and refeclective
than Schwartz.
Here is what Scherlis says about each presentation:
Qlisp - He suggests I start that topic by commenting on the relationship
between stanford and lucid. He thinks we have a good leg up with Boehm
because he respects people who have had commercial success with very large
systems. I will then remark that the collaboration is aimed at producing
a research prototype based on a commercial product. Then I should turn
to Joe for the ideas in Qlisp and what has been accomplished. I think
my little spiel will take 5 minutes.
CLT - If there is anything you can say about the impact of your work on
computing/scalability in the forseeable future, you should mention it.
If you have any interesting relationships with other groups you should mention
it.
VAL - You're on your own.
CPL - Explain that the work proceeded in two phases, with completing
the working report as the first phase. Mention that the second phase is
to work with Lucid on building a prototype of an architecture for
an environment (and that Lucid is supplying this prototype) and to
build on the foundation for prototyping. I will handle all of this.
I will also invite him to Lucid for a demo.
JMC - You're on your own, but make sure we all have time to make our
presentations.
The upshot is that we can proceed as we planned, but with my introducing
Qlisp as well.
-rpg-
∂19-Nov-89 2148 rz@cs.cornell.edu Parallel Computer Algebra Workshop
Received: from cu-arpa.cs.cornell.edu by SAIL.Stanford.EDU with TCP; 19 Nov 89 21:48:00 PST
Received: from LOKI.CS.CORNELL.EDU by cu-arpa.cs.cornell.edu (5.61+2/1.91d)
id AA02105; Mon, 20 Nov 89 00:48:26 -0500
Received: from algron.cs.cornell.edu by loki.cs.cornell.edu (4.0/I-1.91f)
id AA18016; Mon, 20 Nov 89 00:48:23 EST
Date: Mon, 20 Nov 89 00:48:21 EST
From: rz@cs.cornell.edu (Richard Zippel)
Message-Id: <8911200548.AA00694@algron.cs.cornell.edu>
Received: by algron.cs.cornell.edu (4.0/N-0.08)
id AA00694; Mon, 20 Nov 89 00:48:21 EST
To: jmc@sail.stanford.edu
Subject: Parallel Computer Algebra Workshop
Dear Prof. McCarthy:
I don't think we've ever met, but you might know of me through
Ramin Zabih (I was his thesis advisor at MIT) or from my work on Macsyma.
From Sunday, May 6, to Saturday, May 12, 1990, a group of persons
with a common interest in computer algebra are organizing two
contiguous workshops entitled ``Computer Algebra and Differential
Equations (CADE-90)'' (Sunday morning--Wednesday noon) and ``Computer
Algebra and Parallelism (CAP-90)'' (Wednesday afternoon--Saturday
afternoon) at the Mathematical Sciences Institute (MSI) at Cornell
University. The CADE-90 organizers are Micheal Singer at North
Carolina State University and Evelyne Tournier at the University of
Grenoble. The CAP-90 organizers are John Fitch at the University of
Bath in England, Jean Della Dora at the University of Grenoble, and I.
Both committees also include Richard Zippel at Cornell University. We
are combining efforts such that persons with overlapping interest can
stay for the entire week.
We would be delighted if you would agree to come to our CAP-90
workshop (or both) and present a 45-minute to one hour lecture on
parallelism in algebraic computation. The nature of the CAP-90
workshop is that we will ask approxi- mately 10 more persons to
present single lectures, and that we will strongly encourage students
to participate and give presentations on their work. Therefore, we
would like to encourage you to ask your students to also participate
in this workshop. Following a suggestion of John Fitch, we will
organize several discussion sessions, and we might try to publish the
script of these discussions as the workshop record.
The conference budget allows us to subsidize your expenses,
including all local hotel accommodation and meal costs. The remaining
funds we will distribute fairly among the invitees to cover travel
expenses.
I am looking forward hearing from you.
Yours, Richard Zippel
∂20-Nov-89 0900 JMC
verify usair number with Dina Bolla
∂20-Nov-89 0938 pehoushe@Gang-of-Four.Stanford.EDU Implicit Closures or Process Functions
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 20 Nov 89 09:37:56 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA22122; Mon, 20 Nov 89 09:39:18 -0800
Date: Mon, 20 Nov 89 09:39:18 -0800
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8911201739.AA22122@Gang-of-Four.Stanford.EDU>
To: RPG@SAIL.Stanford.EDU
Cc: qlisp@Gang-of-Four.Stanford.EDU
In-Reply-To: Dick Gabriel's message of 19 Nov 89 1548 PST <gtsha@SAIL.Stanford.EDU>
Subject: Implicit Closures or Process Functions
The main point of my message was that closures are obscure for many
average lisp programmers; and spawning arbitrary forms as processes
implicitly requires closures. Therefore, qlisp will be even more
obscure for many qlisp programmers.
By obscure, I mean that the programmer might not get the expected
behavior from his program. Take any simple looping structure and try
spawning tasks in it, and you'll get the idea. I think implicit
closures will cause no end of trouble, and that is what I meant by
Ineffective, in the software development sense.
You might call the issue "User Friendliness in Language Design", and
implicit closures bring along implicit unfriendliness.
Functions + arguments (and spec var environment) are moderately clear,
to the average lisp programmer, and that's what I mean by effective.
If you want a shared lexical environment (closure) in my system you
have to be explicit about it, via an flet or labels or #'(lambda ...);
it doesn't automatically share lexical variables for the programmer.
rpg wrote:
>I think part of the reason these are low cost extensions is that they are
>also low functionality extensions as well.
I think this is a low cost blow (are :)'s allowed?). The low cost
extensions should be nothing less than fully functional.
rpg's example uses throw. In the low-cost system, once upon a time I
had a qcatch and qthrow which were built on top of ordinary common
lisp catch and throw. The scheme really worked quite well. But then
catch and throw got redefined (they still don't work very well),
trashing my implementation of qcatch and qthrow.
So the example would have worked, a long time ago, but doesn't
anymore.
-Dan
∂20-Nov-89 1008 rpg@lucid.com Implicit Closures or Process Functions
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 20 Nov 89 10:08:19 PST
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA22334; Mon, 20 Nov 89 10:09:29 -0800
Received: from rose ([192.31.212.83]) by LUCID.COM id AA05450g; Mon, 20 Nov 89 10:04:32 PST
Received: by rose id AA03928g; Mon, 20 Nov 89 10:06:17 PST
Date: Mon, 20 Nov 89 10:06:17 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911201806.AA03928@rose>
To: pehoushe@Gang-of-Four.Stanford.EDU
Cc: qlisp@Gang-of-Four.Stanford.EDU
In-Reply-To: Dan Pehoushek's message of Mon, 20 Nov 89 09:39:18 -0800 <8911201739.AA22122@Gang-of-Four.Stanford.EDU>
Subject: Implicit Closures or Process Functions
Dan writes:
``Take any simple looping structure and try spawning tasks in it, and
you'll get the idea.''
I guess he didn't see my note that said that DOTIMES and DOLIST have bugs
in them. I'll ask Goldman to fix them.
Dan writes:
``rpg wrote:
>I think part of the reason these are low cost extensions is that they are
>also low functionality extensions as well.
I think this is a low cost blow (are :)'s allowed?). The low cost
extensions should be nothing less than fully functional.''
He missed my point. There is nothing wrong with low functionality
constructs if they are efficient or have other virtues. In the case of
Dan's constructs, they should be
a. given another name so that people don't confuse them with the
high-cost, high-functionality ones
b. made the target of compiler optimizations that transform simple cases
of the high-cost constructs.
The only criticism I have of this work is that Dan used the wrong
names for the constructs - this is minor.
By the way, I misspoke about the restrictions on Dan's constructs.
Functions should have a null lexical control environment and trivial
dynamic control requirements. I'm not sure how easy it is to detect the
latter in separately compiled modules, but local functions can be easily
tested.
-rpg-
∂20-Nov-89 1151 pehoushe@Gang-of-Four.Stanford.EDU Implicit Closures or Process Functions
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 20 Nov 89 11:51:41 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA22822; Mon, 20 Nov 89 11:52:42 -0800
Date: Mon, 20 Nov 89 11:52:42 -0800
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8911201952.AA22822@Gang-of-Four.Stanford.EDU>
To: rpg@lucid.com
Cc: qlisp@Gang-of-Four.Stanford.EDU
In-Reply-To: Richard P. Gabriel's message of Mon, 20 Nov 89 10:06:17 PST <8911201806.AA03928@rose>
Subject: Implicit Closures or Process Functions
> ``Take any simple looping structure and try spawning tasks in it, and
> you'll get the idea.''
>
>I guess he didn't see my note that said that DOTIMES and DOLIST have bugs
>in them. I'll ask Goldman to fix them.
While he's at it, maybe he could fix DO, LOOP, and LET, too. Possibly
this is nitpicking, but I don't think so.
Personally, I don't think the problem is in DOTIMES. The problem is
implicit use of closures, implicitly sharing lexical variables among
processes. In my opinion, the programmer should have to make such
sharing explicit, because that's where alot of bugs get introduced.
>From a previous rpg message,
>The first and fourth have closure ``problems'' when I is captured.
>The second, third, and fifth do not.
Sarcastic Moral, when using implicit closures:
Before adding parallelism to a serial program, be sure to notice and
rewrite all parts of code written in the first and fourth styles!!
Come on Dick, give in. You don't really think implicit closures
are a good idea, do you? Anyone who likes lexical scope shouldn't
like implicit closures.
rpg wrote:
> ``rpg wrote:
> >I think part of the reason these are low cost extensions is that they are
> >also low functionality extensions as well.
>
> I think this is a low cost blow (are :)'s allowed?). The low cost
> extensions should be nothing less than fully functional.''
>
>He missed my point. There is nothing wrong with low functionality
>constructs if they are efficient or have other virtues. In the case of
Low cost should not imply or otherwise indicate low functionality. It
may just indicate inefficiency in the high cost version.
>The only criticism I have of this work is that Dan used the wrong
>names for the constructs - this is minor.
Thanks!
However, you keep using the term "low functionality", and in some way,
that seems like a criticism; for instance, catch/throw and return-from
definitely should work (but they don't), and at low cost, too.
>By the way, I misspoke about the restrictions on Dan's constructs.
>Functions should have a null lexical control environment and trivial
>dynamic control requirements.
For Spawns and Futures, I completely agree with you; such constructs
should not be allowed to throw or "return" control to the parent.
But fork-join constructs such as qlet and qdotimes should be allowed
to throw control back to the parent. Fork-join constructs imply some
sort of control connection with the parent, while Spawn implies a
disconnection from the parent.
-Dan
∂20-Nov-89 1543 rpg@lucid.com Implicit Closures or Process Functions
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 20 Nov 89 15:42:56 PST
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA24029; Mon, 20 Nov 89 15:44:03 -0800
Received: from rose ([192.31.212.83]) by LUCID.COM id AA07822g; Mon, 20 Nov 89 15:39:03 PST
Received: by rose id AA04216g; Mon, 20 Nov 89 15:40:45 PST
Date: Mon, 20 Nov 89 15:40:45 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911202340.AA04216@rose>
To: pehoushe@Gang-of-Four.Stanford.EDU
Cc: qlisp@Gang-of-Four.Stanford.EDU
In-Reply-To: Dan Pehoushek's message of Mon, 20 Nov 89 11:52:42 -0800 <8911201952.AA22822@Gang-of-Four.Stanford.EDU>
Subject: Implicit Closures or Process Functions
Dan writes:
``Personally, I don't think the problem is in DOTIMES. The problem is
implicit use of closures.''
If you really agree with your own argument, then you simply do not
like closures, since you can achieve exactly the same unclarities in
sequential code using closures.
``While he's at it, maybe he could fix DO, LOOP, and LET, too.''
If you ``fix LET'' you can never achieve information sharing. You must not
have meant this.
``Anyone who likes lexical scope shouldn't like implicit closures.''
I couldn't figure out what you meant by this. I would have guessed
anyone who liked and understood lexical scope wouldn't notice implicit
closures. Note, I never had any problems writing Qlisp code until I
used your constructs.
I think the simple issue is that Common Lisp is not a language
designed with parallel processing in mind. Therefore, parallelization
must involve careful thought about how things interact. I believe it's
foolish at this stage of understanding to be able to write anything
but the most trivial parallel programs (like AND-parallelism) without
a lot of deep thought. Fixing the problem by implicitly requiring a
small bottleneck - which is what process functions does - is not the
smart way out of this problem.
Well, but this silly argument is boring. Let's resolve it this way;
Please change the names of your constructs from QLET to something
else and try to make it possible for your constructs and the Qlisp
ones to co-exist.
-rpg-
∂20-Nov-89 1543 rpg@lucid.com Round Table
Received: from lucid.com by SAIL.Stanford.EDU with TCP; 20 Nov 89 15:43:32 PST
Received: from rose ([192.31.212.83]) by LUCID.COM id AA07838g; Mon, 20 Nov 89 15:40:00 PST
Received: by rose id AA04219g; Mon, 20 Nov 89 15:41:44 PST
Date: Mon, 20 Nov 89 15:41:44 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911202341.AA04219@rose>
To: clt@sail.stanford.edu, jmc@sail.stanford.edu
Subject: Round Table
HBS and I were planning to go to it so as to try to press CPL on them
more. Are you going?
-rpg-
∂20-Nov-89 1754 ME mail address
∂20-Nov-89 1555 JMC
SAIL refuses the address cliff@uk.ac.man.cs. What's the tinker?
ME - Turn it around: cs.man.ac.uk. For some reason, somewhere they
put the parts in the other order.
∂20-Nov-89 1843 VAL Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU, su-events@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
INCREMENTAL, APPROXIMATE PLANNING
(Preliminary Report)
Charles Elkan
Department of Computer Science
University of Toronto
Monday, November 27, 3:15pm
MJH 252
I shall present evidence for three claims.
(1) Locally stratified definite clauses with the perfect model
semantics are adequate for specifying the Yale shooting problem
and other traditional planning domains, using the following
frame axioms:
causes(A,S,P) => holds(P,do(A,S))
holds(P,S) & ~ cancels(A,S,P) => holds(P,do(A,S))
(2) With the right control strategy, Green's method is a practical
way of inventing plans given a locally stratified domain theory.
(3) Allowing default conditions to be assumed is an abductive
planning strategy that produces plausible plans quickly, and those
plans can be incrementally refined towards guaranteed plans.
∂21-Nov-89 0155 weening@Gang-of-Four.Stanford.EDU
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 21 Nov 89 01:54:56 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA25843; Tue, 21 Nov 89 01:56:12 -0800
Date: Tue, 21 Nov 89 01:56:12 -0800
From: weening@Gang-of-Four.Stanford.EDU (Joe Weening)
Message-Id: <8911210956.AA25843@Gang-of-Four.Stanford.EDU>
To: jmc@sail.Stanford.EDU, val@sail.Stanford.EDU
>From: djb@wjh12.harvard.edu (David J. Birnbaum)
Newsgroups: comp.std.internat
Subject: ISO standards for non-Latin alphabets
Date: 20 Nov 89 21:29:01 GMT
Reply-To: djb@wjh12.UUCP (David J. Birnbaum)
Organization: Harvard University, Cambridge MA
Note: The following is a slightly revised version of a paper pre-
sented earlier this year at the Fourth International Conference
on Symbolic and Logical Computing, held at Dakota State Univer-
sity, Madison, South Dakota. Endnote numbers within the text are
enclosed in parentheses. Readers may wish to consult a character
map for ISO 8859/5 (= ECMA 113). While some computational issues
mentioned here will be obvious to USENET readers, I hope the
philological and linguistic perspective will prove interesting.
=================================================================
Issues in Developing International Standards
for Encoding non-Latin Alphabets(1)
David J. Birnbaum
Department of Slavic Languages, University of Pittsburgh
Russian Research Center, Harvard University
djb@wjh12.harvard.edu [Internet]
djb@harvunxw.bitnet [Bitnet]
Copyright (c) 1989 by David J. Birnbaum
All rights reserved
Introduction
Defining an appropriate character set is the most impor-
tant preliminary to any text processing. The generally accepted
system for encoding English language texts is the American Stan-
dard Code for Information Interchange (ASCII),(2) but the devel-
opment of appropriate standards for other languages and alphabets
has been less successful. As a result of this lack of agreement,
idiosyncratic systems have proliferated, producing predictable
obstacles to the efficient exchange of data.
Recently the International Standards Organization (ISO)
promulgated the 8859 series of standards for a variety of writing
systems. One of these standards, 8859/5,(3) is designed to serve
all six modern Slavic languages that use the Cyrillic alphabet
(Russian, Ukrainian, Belorussian, Bulgarian, Macedonian, and Ser-
bocroatian). My discussion today focuses on general methodologi-
cal issues involved in determining appropriate international
standards, which I illustrate through a specific critique of
8859/5.
The theoretical issues involved include
1) the languages to be covered;
2) the alphabets for these languages;
3) the uses to be served by the standard;
4) the coding for the individual characters.
Additional problems that influence issue 4 include
1) 7-bit or 8-bit sets or other representations;
2) compatibility with other existing standards for the same
languages;
3) compatibility with character sets for other languages;
4) balancing priorities of different languages combined in a
single set;
5) upper and lower case relationships;
6) sorting and string comparison order;
7) differences in information processing and information in-
terchange requirements;
8) hardware and software limitations.
Information Processing and Information Interchange:
the Difference between Local and International Standards
One important preliminary consideration is that informa-
tion processing and information interchange may require different
standards. Information processing is a local concern, where com-
patibility with existing local standards may be important. Fur-
thermore, users may be constrained not only by local conventions,
but also by specific hardware or software configurations and any
peculiarities of their texts. For example, different operating
systems, printers, and application software may reserve particu-
lar control characters.(4) For my own research, which requires
fairly exact transcriptions of orthographically complex medieval
Cyrillic manuscripts, I have to provide for a variety of non-
standard characters, ligatures, and diacritics; other users will
customize their systems in different ways. It is difficult for a
single international Slavic Cyrillic standard to anticipate all
the needs of all users, but there is no real impediment to devel-
oping a sensible standard for modern languages.
Since the optimal solution to a specific limited problem
may be incompatible with a more general standard that must serve
a wider range of users, the most reasonable compromise is that
standards for information interchange should be designed to serve
as many uses as possible as efficiently as possible, with
whatever compromises that entails. Local standards for informa-
tion processing, on the other hand, should be designed separately
to deal effectively with specific local tasks. Filtering text
files is a trivial matter and users can easily convert local
formats to an accepted interchange standard if the material is to
be shared with users who may have different local information
processing standards.
This internationalist approach can be contrasted to the
philosophy behind 8859/5, where, as we shall see, general re-
quirements for dealing with multilingual Slavic Cyrillic texts
have been needlessly subordinated to local, strictly Russian con-
cerns.(5) It would have been more sensible to expect Russians to
use their own well-established national standard locally, but to
compromise on an international interchange standard that is truly
international.
The Independence of Binary Representations
from Keyboards, Monitors, and Printers
Humanists unfamiliar with computers often fail to realize
that the internal binary representation(6) of a character set is
completely independent of keyboard layouts and screen or printer
displays. Striking a specific key on a keyboard generates a
hardware scan code,(7) which is not the same as the binary repre-
sentation of a character. The operating system is then
responsible for interpreting the scan code, checking for shift or
control keys and other details, and generating a binary character
representation. Typing a lower case {a} on an IBM PC generates a
scan code of 1Eh (30) with no shift mask, which the BIOS will
translate into 61h (97). This translation can be modified by the
user, so that the physical location of a certain letter on a key-
board may determine the scan code generated, but this scan code
is irrelevant to the binary representation that will be assigned
to that character.
Similarly, the relationship between the internal repre-
sentation of a character and its screen or printed display can be
defined separately by the user to suit the application. To con-
tinue the preceding example, the binary representation 61h does
not have to put a lower case {a} on the screen. The user can
revise the relationship between binary representations and the
character display just as he can revise the relationship between
keyboard scan codes and binary representations.
Although technically more complicated than simple remap-
ping, there are situations where it is useful to allow a single
binary character representation to correspond to multiple screen
or printer representations. For example, most letters of the
Arabic alphabet have four separate shapes, depending on whether
they appear in isolation or at the beginning, middle, or end of a
word (or of a sequence of connected letters). In the dark days
of typewriters, it was necessary for the typist to use multiple
shift keys to enter the correct form of the character to be dis-
played. The most efficient scheme for encoding such contextually
dependent information today is to store each Arabic letter as a
separate binary code, and to make the display or printing soft-
ware responsible for selecting the appropriate graphic
variant.(8)
Efficient Use of 8-Bit Systems
One initially encouraging decision reflected in 8859/5 is
the use of an 8-bit representation, providing 256 characters per
set instead of the 128 available in a 7-bit standard. But this
sensible procedure is vitiated by the decision to retain Latin
characters (with standard ASCII assignments) in the lower half of
all 8859 sets, so that at most 128 positions could be available
for Cyrillic characters. An additional sixty-four positions of
every set are needlessly reserved for control characters, reduc-
ing the actual number of slots potentially available for Cyrillic
characters to ninety-six.(9) This is barely enough to encode up-
per and lower case variants of all letters used in the modern
Slavic languages.
Most documents are monolingual, do not combine Latin and
Cyrillic, and would be better served if a larger inventory was
available for the relevant alphabet. Multiple-alphabet documents
could be accommodated by a standard for switching between non-
overlapping sets of 256 characters. Such a standard will be
necessary in any case for documents that combine, for example,
Cyrillic (8859/5) with Greek (8859/7).(10) Serbocroatian is
unique among the Slavic languages in its official use of both the
Latin and Cyrillic alphabets, which means that documents includ-
ing both versions would require 8859/5 for the Cyrillic portion
and 8859/1(11) for the Latin. Another advantage of combining
character sets is that control characters could be defined for
only a single set, which would open additional positions in the
extended alphabet sets. Although 8859/5 is technically an 8-bit
system, the combination of Latin and Cyrillic in a single set and
the prodigal assignment of control characters results in the same
limitations that constitute the principal liabilities of 7-bit
systems.
Although all of the standard characters of the modern
languages are included in ISO 8859/5, a system providing more
positions could be put to wider use. For example, there is no
room in 8859/5 for European quotation marks (guillemets), which
are a regular feature of Cyrillic typography.(12) Additionally,
8859/5 was designed only for *modern* Slavic Cyrillic languages
and is inadequate even for basic work with historical sources
that use a slightly different character set than the modern lan-
guages. The Russian alphabet was reformed in 1918 and the Bul-
garian one as late as 1945; in both cases letters were deleted
that would be useful to people working with earlier sources. The
Ukrainian alphabet includes a separate 'g' character, the use of
which has at times been considered an act of sedition by Soviet
authorities and a mark of national pride by many Ukrainians, par-
ticularly in the west. Even if the matter were not politically
sensitive, there are no free positions in 8859/5 for this and
other obsolete letters that are important for work with histori-
cal sources.
The Internationality of International Standards
Although 8859/5 purports to be an international standard
for all modern Slavic languages that use the Cyrillic alphabet,
it is needlessly and offensively Russocentric. The ISO is under-
standably concerned with maintaining compatibility with accepted
national standards, but this concern should be paramount only for
monolingual standards. 8859/5 is not supposed to be a Russian
standard and it should have been established by a disinterested
evaluation of the requirements for dealing with six languages,
rather than by slavishly adopting a Russian national system at
the expense of the other languages.
Those who are familiar with Russian will note that
columns 11 through 14 of 8859/5 contain the letters of the Rus-
sian alphabet in order. The thirty-third Russian character, the
{e} with diaresis is tucked away on the side. In almost all Rus-
sian writing, the diaresis is omitted and this letter is treated
as identical to {e}, so that it is, in some respects, a marginal
part of the Russian alphabet and a good candidate for special
treatment.
Case Folding in Multialphabet Sets
Reducing the 33-character Russian alphabet to 32 is
desirable not only because one letter is orthographically
marginal, but because 32 is a convenient number for binary com-
puters and can facilitate case folding. Note, however, that the
Russian characters begin in an odd-numbered column, while the
Latin characters begin in an even-numbered one, which means that
Latin and Cyrillic case folding require different algorithms.(13)
If the Russian alphabet is to be reduced to 32 characters to fa-
cilitate case folding, it would seem sensible in a two-alphabet
character set to establish a mapping that would allow a single
procedure to accomplish case folding for both alphabets.
Of the remaining languages served by 8859/5, only the
Bulgarian alphabet is a perfect subset of the Russian. Ukraini-
an, Belorussian, Macedonian, and Serbocroatian all include addi-
tional characters not present in Russian. In 8859/5 these have
been tucked away in columns 10 and 15. This entails yet a third
relationship between upper and lower case and means that case
folding even for monolingual texts in languages other than Rus-
sian requires two separate procedures, one to fold 11 and 12 in
with 13 and 14 and another to fold 10 and 15 together.(14)
Character Order
One advantage to following alphabetic order in character
coding is that it enables alphabetic sorting by comparing strings
according to machine order. This type of unfiltered sorting in
8859/5 is impossible for Ukrainian, Belorussian, Serbocroatian,
or Macedonian, since the characters from columns 10 and 15 would
have to be inserted into their proper places. This is a com-
pletely unnecessary limitation, because with one minor excep-
tion(15) all modern Slavic languages that use the Cyrillic al-
phabet follow a single order. Not all characters will occur in
each language, but a single order for the entire character set
would have made it possible to sort all languages in machine or-
der.(16)
The Problem of 8859/5 as an International Standard
The upper half of 8859/5 is an excellent example of how
not to organize an international standard. It is an imperfect
Russian national standard that is poorly suited to the other
Slavic languages it is supposed to represent. As I mentioned
earlier, standards for local information processing may differ
from standards for international information interchange and a
Russian writing exclusively in Russian should use the resources
that best answer his requirements. A multilingual international
standard, on the other hand, should balance the requirements of
all the languages involved. Filtering text files to convert be-
tween local and international standards is not difficult and to
favor one national system over all others as a basis for an in-
ternational interchange standard is not justifiable technically,
intellectually, or diplomatically.(17)
Alternative Standards
If we abandon the idea of combining Latin and Cyrillic
into a single 8-bit set, it is possible to deal more effectively
with Cyrillic requirements. One possible approach is that imple-
mented in the ISO 6861 draft standard, which provides a system of
extended Cyrillic sets that incorporates most of the characters
required for work with modern and medieval Slavic sources and
Romanian Cyrillic.(18) A standard control sequence can be used
to select the appropriate set for an application, as well as to
switch sets within a single text.
Another desirable extension of the Cyrillic inventory
would be the addition of characters from non-Slavic languages of
the Soviet Union that use the Cyrillic alphabet. Either the
medieval letters of the 6861 draft standard or an extended modern
Cyrillic set would provide a more efficient use of character
positions than the combination of Latin and Cyrillic found in
8859/5. A single set, similar to ISO 8859/1, could serve for
most of the Latin alphabet languages of Europe, while other sets
could provide better support for languages using Cyrillic.
This type of approach, which overcomes the limitations
inherent in any 8-bit set, which can have room for no more than
256 characters, is exemplified by the recent multi-octet (or
multiple-byte) ISO draft proposal 10646. This three-dimensional
representation has room for over 16 million characters, each of
which could be fully specified by three bytes. Of course, a
three-byte representation would be wasteful for most applications
and the preliminary description of the standard includes modifi-
cations that would permit simpler representations when appropri-
ate. These include:
1) a two-octet form, restricted exclusively to a single
plane, which would suffice for most purely alphabetic ap-
plications;
2) a compacted form, permitting strings of related charac-
ters to be used as single-octets.
According to this latter modification, a string of Cyrillic
characters with two of the three octets in common could be
represented by a control sequence indicating that those two would
be in force until further notice, whereupon the specific individ-
ual characters could be identified merely by supplying the third
octet.
Conclusions
An ideal multilingual international standard would not
combine completely different alphabets, such as Latin and Cyril-
lic, into a single character set. It should not be designed
around the requirements of one language when an alternative is
available that serves all the languages with equal effectiveness.
If case folding is a priority, it should be implemented uniformly
throughout the set. If arranging characters in sorting order is
a priority, a mapping that supports all the languages equally
should be favored. Restrictions imposed by specific hardware and
software configurations, as well as conformity to existing na-
tional standards, which may be of primary importance for local
information processing, should not dictate international stan-
dards for information interchange. Continuity with existing na-
tional and international standards is desirable, but this desire
for compatibility should not allow obsolescent decisions to
retard the development of new standards that could better exploit
new resources.
Notes
1) I am grateful to Steven J. DeRose for help in obtaining in-
formation about ISO standards and especially to Harry Gaylord for
both help with materials and stimulating comments on many of the
issues mentioned here.
2) The most frequently encountered alternative is the Extended
Binary Coded Decimal Interchange Code (EBCDIC), which is used
primarily on IBM mainframes. Although the alphanumeric charac-
ters of ASCII and EBCDIC correspond, small differences between
EBCDIC variants (as well as variants in ASCII coding) make trans-
lation between ASCII and EBCDIC perilous and greatly complicate
the transfer of files between, e.g., Internet and Bitnet sites.
3) ISO 8859/5 has been adopted by the European Computer Manufac-
turers Association as their Standard ECMA-113 (2nd edition, July
1988, adopted by the General Assembly of the ECMA on 30 June
1988).
4) As an example of a hardware limitation, some display adapters
do not treat all characters identically. A number of MS-DOS
software packages use characters between B0h and DFh (176--223)
for lines and borders. In the traditional PC text display, all
characters are nine pixels wide, but only the eight leftmost
columns can be defined by the user. For characters between B0h
and DFh, the eighth pixel from the left is automatically dupli-
cated in the rightmost column, while for characters outside this
range, the rightmost column is blank. This enables the graphics
characters in the B0h--DFh range to connect, which is convenient
for continuous lines and borders. Unfortunately, this means that
any user-defined alphabetic characters assigned to this range
must be no more than seven pixels wide, since an 8-pixel wide
character would bleed into the rightmost column.
5) 8859/5 is based on the 1987 revision of the Soviet national
GOST Standard 19768.
6) I use the term "binary representation" to designate the ma-
chine coding for a character. Most standards implement 8-bit
representations, although an alternative is discussed later in
this paper.
7) For example, scan codes on IBM PCs essentially reflect the
physical order of the keys on the keyboard. New keyboard designs
have caused some keycaps to be moved, but old scan code assign-
ments were retained for compatibility. For example, the back-
slash key continues to generate a 2Bh even as it moves from one
location to another with each revision of the keyboard.
8) This simplifies data entry and editing, as well as sorting.
An escape code would be required to display a character outside
its usual context, but this extraordinarily rare situation cannot
justify commandeering four binary representations for every let-
ter of the alphabet. ISO 10646, which I discuss below, will use
a single character in the text file and allow the application to
transform it as appropriate for screen and printer output.
9) Proposals submitted to the ISO for 8-bit sets with a minimal
number of control characters (5 or less) have been resoundingly
rejected by most of the national delegations.
Many ISO standards continue to be influenced by
anachronistic concerns. Following the provisions of ISO 2022, 8-
bit standards are treated as two pages of 128 characters, rather
than one page of 256. The 256 characters of 8859 and other 8-bit
standards are divided into four sections: C0 (00/00--01/15), G0
(02/00--07/15), C1 (08/00--09/15), and G1 (10/0--15/15). C0 and
C1 are control sections and reserve thirty-two positions for con-
trol characters. G0 and G1 are available for two sets of
graphics characters, each containing up to 96 items.
Another striking anachronism that whittles 96 items down
to 95 is the designation of 07/15 as a control character. This
character, traditionally called delete (DEL), was previously used
to erase or obliterate erroneous or unwanted characters in
punched tape. There is no justification for reserving this posi-
tion today when the limited number of positions available for
characters in a multi-language standard is already so restricted.
Nonetheless, ISO DIS 6861 and DP 10646 reserve both 07/15 and
15/15 for control functions. ISO 8859/5, curiously, reserves
only 07/15, while assigning an alphabetic character to 15/15.
Although the number and coding of control characters can
be reduced with no loss of information, any such decision should
be taken in conjunction with a revision of International Telecom-
munications Union CCITT protocols, which use control characters
to regulate the transmission of digital information.
10) Equivalent to ECMA--118.
11) Equivalent to ECMA--94/1, second edition (June 1986). ISO
2022, Information processing --- ISO 7-bit and 8-bit coded
character sets --- Code extension techniques, establishes stan-
dards for switching among character sets within a document. ISO
Draft Proposal 4873 (currently being revised) also deals with
switching among C0, G0, C1, and G1.
12) Oddly enough, many Soviet Russian standards omit European
quotation marks. There is also no provision in 8859/5 for mark-
ing accented vowels, which might be required for textbooks, dic-
tionaries, or linguistic studies. The 8859 standards forbid
overstriking, so that any combination of character plus diacritic
must have a single binary representation. 8859/5 hardly has room
for accent marks, let alone fully formed accented vowel letters.
The 7-bit 646 standard, now under revision, allowed for
the use of a backspace combined with diacritics, which could be
entered after alphabetic characters. Other standards allowed for
nonspacing diacritics, similar to dead keys, which could be en-
tered before alphabetic characters.
13) A string of Latin alphabet text can be converted to lower
case by setting bit 6, which effectively adds 32 to the upper
case characters while leaving the lower case unchanged. The same
string can be converted to upper case by clearing bit 6. To con-
vert a string of Russian text to lower case requires setting bit
7 and toggling bit 6. Converting Russian text to upper case is
more complicated still. Note that ISO conventions call for num-
bering rows and columns in decimal from 00--15, rather than in
hexadecimal, and for numbering bits 1--8, instead of the more
common 0--7.
14) The last procedure involves setting (or clearing) bits 5 and
7.
15) The soft sign falls at the end of the alphabet in Ukrainian.
This letter never occurs in initial position in any Slavic lan-
guage and it is close to the end of the alphabet in the other
languages, so that this peculiarity of Ukrainian will have little
effect in real applications. On the other hand, the order of the
Cyrillic Old Church Slavonic alphabet, used for medieval texts,
differs in several places from the order in the modern languages,
so that even if the Old Church Slavonic characters were added to
the Cyrillic inventory, a different sorting algorithm would be
required.
16) According to John Clews, Language automation worldwide: the
development of character set standards, Harrogate: Sesame, 1988,
Section 5.1, transliteration standards exploiting this feature have been well known for over twenty years.
17) Lamentably, the fate of ISO standards depends on the voting
of national committees that may be more concerned with national
prestige than with enacting efficient international standards.
It is reported that an effort to establish a single character set
for the Far East with no duplication met with threats by China,
Japan, and Korea to withdraw if their entire national standards,
duplicate characters and all, were not included. It is unlikely
that sensible international standards will ever emerge from such
a chauvinist atmosphere; academic projects, such as the Text En-
coding Initiative, are more promising.
18) 6861 also includes a Glagolitic set, with characters as-
signed to the same positions as their alphabetic equivalents in
Cyrillic, a layout that facilitates transliterating between
Glagolitic and Cyrillic. There seems to be some uncertainty in
6861 about how to distinguish differences in character sets from
differences in typefaces, but the principle of not squandering
the limited inventory of binary representations available in a
Cyrillic set on Latin characters is sound.
∂21-Nov-89 0634 bajcsy@central.cis.upenn.edu Hello
Received: from central.cis.upenn.edu by SAIL.Stanford.EDU with TCP; 21 Nov 89 06:34:29 PST
Received: from LOCALHOST by central.cis.upenn.edu
id AA23934; Tue, 21 Nov 89 09:34:27 -0500
Posted-Date: Tue, 21 Nov 89 09:34:26 -0500
Message-Id: <8911211434.AA23934@central.cis.upenn.edu>
To: jmc@sail.stanford.edu
Subject: Hello
Date: Tue, 21 Nov 89 09:34:26 -0500
From: bajcsy@central.cis.upenn.edu
Dear John,
after 22 years I went back to my roots.
Bellow you will find my account of the trip.
I phoned prof. Gvozdjak, he is now retired (70 years old, we are all
getting older) and he rememebred you fundlly and sends his regards.
My daughter Klara and her son ,who live in E.Berlin were give an
official permission to emigrate to the USA. Their passports will be ready
on 28th of December, so with good luck on the American Visa,
they could be here around beginning of January. So finally, I am
going to be real grand mother instead of a virtual one. We are all
very excited about this event and also about all the changes that are
taking place in E.Europe. One must just pray for Gorbachov.
Letter from Bratislava
Dear Frends,
Some of you may know that last week, that is from November 6th til
November 12th, 1989 I was a guest of the Slovak Aacdemy of Sciences
in Bratislava, Czechoslovakia. I was invited to participate in the
International Conference on Artificial Intelligence and Control Systems
that took place in Strbske Pleso in the mountains of High Tatras.
First, the conference.
The conference had about 200 participants from all of the Europe.
There were two Americans, myself and Siklossy ( a former Hungerian).
The scientific level of the presentations was average, but not bellow
average in comparisson to other International meetings.
I was quite impressed by the paper from P.Hajek on formalisation
of the Prospector's uncertainity measure. He showed that the whole
thing can be nicely embeded in the traditional probability theory
without any need for new frameworks such as evidence reasoning
or their like. (I am sure Max will enjoye his paper).
The other paper that I liked was presented by Tibor Vamos from Hungary who
though gave a rather philosphical but deep evaluation of different
formalisms for knowledege representations.
The papers related to different technologies naturally lacked the
access to current technologies.
The conference was well organized , all presentations were in english
and by and large the spoken english was good. The E.Europeans are quite
well informed about what is going on but of course are handicapped
by lack of technology. Another handicap comes from the fact that for
last 40 years they were isolated and the governemnt requested from
them to develop all the technologies so that they are indipendent
from the West. Of course the consequence is that they were doing
everything but nothing with quality!
Now for my personal impressions:
Some of you may know that I went with great trepidation especially
since through my exhusband and my son the information I was getting was
that everybody will be scared to have any contacts with me, after all
I represented the evil empire!!
The contrary happend! People were very wellcomming and naturally curious
to learn how science is done in the USA, what programs are current,
what technologies we use,what consists of progress,etc.
It was indeed a very gratifying feeling.
I had no problems on the border, comming and going.
I visited my old grammer school which was replaced by an ugly department store,
but the bakery next to it is still there. I visited my old gymnazium
which was rather touching. And of course I visited the mass grave (of
about 2-3000 people) in Kremnicka (central part of Slovakia),
where my parents and my grand mother are burried
as a consequence of the Nazis. This visit was very important to me,
one must never forgett that these things did happend and could happen.
I met one of my schoolmates from gymnazium who updated me with the fate
of many of my contemporaries. At the expense of generalization,
I must say that my generation is the victim of the isolation and
of the 40 years of a failed experiment! In that view I did the right
thing to leave although at the expense of some personal losses,
i.e my childern. My son still did not want to meet me.
The governemnt repaired some historical buildings, castles but they
completely ignored the interest of people (contrary what they
promised in early 50s) when they build housing
which is a concrete forest without any parks, stores and cultural
entreprises. The other criminal thing they did is the polution.
One third of the Slovak forest is dead because of polution!!
Overall it was a good trip for me. And as my frend Tibor said
my presence first time after 22 years at this conference
signals a change. Let us hope so for everybody's sake.
LOve Ruzena
∂21-Nov-89 0733 beeson@ucscd.UCSC.EDU Re: Thanks for your paper.
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 21 Nov 89 07:33:16 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA02100; Tue, 21 Nov 89 07:33:31 -0800
Date: Tue, 21 Nov 89 07:33:31 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8911211533.AA02100@ucscd.UCSC.EDU>
To: JMC@SAIL.Stanford.EDU, beeson@ucscd.UCSC.EDU
Subject: Re: Thanks for your paper.
Thanks for the grammatical suggestion. You usually have a chance to
make revisions after the referee's report, and that's when I can
incorporate yours.
∂21-Nov-89 0738 beeson@ucscd.UCSC.EDU learning
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 21 Nov 89 07:38:44 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA02129; Tue, 21 Nov 89 07:39:03 -0800
Date: Tue, 21 Nov 89 07:39:03 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8911211539.AA02129@ucscd.UCSC.EDU>
To: jmc@sail.stanford.edu
Subject: learning
I am not very familiar with the research on AI and learning. Is there
any program (or design for one) that learns to play finite games by
somehow (perhaps AM-style) choosing logical concepts and inferring
theorems about them? For example, the way people actually play tic-tac-toe
is far different than the way programs play tic-tac-toe. I imagine a
program that would use unification to get for itself the concepts of
"row", "column", "diagonal", then generalize to "line", then deduce
" if you have two in a line and the third is blank, you can win by
moving in the other spot (if it's your move, I forgot to say)" and so on.
Such a program could (if done right) learn any simple board game--but
unfortunately its actual outward behavior (playing tic-tac-toe) wouldn't
be very impressive. The question is, has anybody done that already to
your knowledge?
∂21-Nov-89 0901 Mailer re: silver lining to the quake
Received: from akbar.cac.washington.edu by SAIL.Stanford.EDU with TCP; 21 Nov 89 09:01:05 PST
Received: from tomobiki-cho.cac.washington.edu by akbar.cac.washington.edu
(5.61/UW-NDC Revision: 2.6 ) id AA15321; Tue, 21 Nov 89 09:00:32 -0800
Date: Tue, 21 Nov 1989 8:47:44 PST
From: Mark Crispin <MRC@CAC.Washington.EDU>
Sender: mrc@Tomobiki-Cho.CAC.Washington.EDU
Subject: re: silver lining to the quake
To: John McCarthy <JMC@sail.stanford.edu>
Cc: su-etc@sail.stanford.edu
In-Reply-To: <GtzXz@SAIL.Stanford.EDU>
Message-Id: <MailManager.627670064.7590.mrc@Tomobiki-Cho.CAC.Washington.EDU>
Is it the environmentalists who've been screaming about the Embarcadero
Freeway? I always thought it was limp-wristed San Franciscans who objected to
the freeway blocking their view of their neighbor's houses (and bedrooms).
-------
∂21-Nov-89 1005 MPS Expenses
I need to have the receipt (or cost) of the hotel in
Texas.
Also, the same for San Diego.
Thanks
∂21-Nov-89 1045 pehoushe@Gang-of-Four.Stanford.EDU qlet& and qlet
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 21 Nov 89 10:45:38 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA26572; Tue, 21 Nov 89 10:46:58 -0800
Date: Tue, 21 Nov 89 10:46:58 -0800
From: pehoushe@Gang-of-Four.Stanford.EDU (Dan Pehoushek)
Message-Id: <8911211846.AA26572@Gang-of-Four.Stanford.EDU>
To: rpg@Gang-of-Four.Stanford.EDU
Subject: qlet& and qlet
Cc: qlisp@Gang-of-Four.Stanford.EDU
In the low-cost extensions to qlisp, QLET now takes accepts arbitrary
forms as potential processes. The other version, QLET&, requires the
forms to be function calls.
Note: QLET creates closures for the first n-1 forms. In parallel
mode, destructive operations to lexical variables in these forms could
cause erroneous and unpredictable bugs, which disappear serially.
The form QLET& is similar to QLET, except that the first N-1 forms
must be function calls, and destructive operations on lexical variables
are quite OK to use.
For your convenience, this version includes:
qdotimes, qmapcar, qmaplist, qmapc, qmapl, qmapcon, qmapcan
It is the programmer's responsibility to insure independence of the
individual iterations.
The list mapping functions are going to be the topic of a paper we are
sending to a conference.
Once again, to get this version of qlisp, run /u/pehoushe/bin/new-qlisp,
and (load "/qlisp/nqn-newload"). The parallel interface is qtime.
-Dan
P.S.
RPG's test program now works:
(defun test (n)
(let ((a 0))
(qlet t ((x (dotimes (i n a) (when (evenp i) (incf a))))
(y (dotimes (i n a) (when (oddp i) (incf a)))))
(list x y))))
(test 1000)
Ran in 11 milliseconds, and returned (1000 500).
In parallel, (Test 1000) ran in an average of 5.4 milliseconds,
over 10 trials, and (717 706) was a typical return value; this
program is interesting as the parallel answer seems quite
load dependent. The heavier the load, the fewer (incf a)'s
are "missed", and so the larger the return values.
∂21-Nov-89 1114 VAL re: book
[In reply to message rcvd 20-Nov-89 23:39-PT.]
I'll give Pat a copy of the manuscript used by the publisher.
I hope I'll be done with the index by the end of the month, and then we'll be
in a position to ask Ablex how things stand.
∂21-Nov-89 1901 VAL re: book
[In reply to message rcvd 20-Nov-89 23:39-PT.]
The manuscript is on Pat's desk.
∂21-Nov-89 1914 ME SAIL
∂20-Nov-89 2119 JMC
In the last two days SAIL has hiccupped in echoing to our home terminals.
ME - Is it working OK now? There was a pause yesterday morning.
∂21-Nov-89 1926 cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK Lifschitz
Received: from NSFnet-Relay.AC.UK by SAIL.Stanford.EDU with TCP; 21 Nov 89 19:26:10 PST
Received: from sun.nsfnet-relay.ac.uk by vax.NSFnet-Relay.AC.UK
via Janet with NIFTP id aa15765; 21 Nov 89 17:37 GMT
Date: Tue, 21 Nov 89 14:33:57 GMT
From: cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK
Message-Id: <8911211433.AA00465@ipse2pt5.cs.man.ac.uk>
To: JMC@sail.stanford.edu
Cc: jones%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK
In-Reply-To: John McCarthy's message of 20 Nov 89 2119 PST <Ktb#o@SAIL.Stanford.EDU>
Subject: Lifschitz
John,
Thank you for alerting me to this problem - I promise that I
will look into what was said. (Not by way of excuse, but
simply to explain that I cannot respond at once, I should
point out that I was not on the relevant appointments panel.)
If the information that there were other candidates were kept
from Lifschitz, I would share your view; even if it was a
mistake, there is clear need for an apology (at least). I am
away in Edinburgh for the next 2 days but will look into the
matter on Friday.
cliff
∂21-Nov-89 2020 rpg@lucid.com qlet& and qlet
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 21 Nov 89 20:20:52 PST
Received: from LUCID.COM by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA28766; Tue, 21 Nov 89 20:22:05 -0800
Received: from rose ([192.31.212.83]) by LUCID.COM id AA09606g; Tue, 21 Nov 89 20:17:08 PST
Received: by rose id AA05057g; Tue, 21 Nov 89 20:18:55 PST
Date: Tue, 21 Nov 89 20:18:55 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911220418.AA05057@rose>
To: pehoushe@Gang-of-Four.Stanford.EDU
Cc: qlisp@Gang-of-Four.Stanford.EDU
In-Reply-To: Dan Pehoushek's message of Tue, 21 Nov 89 10:46:58 -0800 <8911211846.AA26572@Gang-of-Four.Stanford.EDU>
Subject: qlet& and qlet
Qlet& looks like a good name. In your warning about shared variables,
do you mean random writing or even writing with proper locks? That
is, is it the fact of writing or the liklihood someone wrote a bug
that is the problem?
Note that I do not necessarily disagree with Dan's assertion that
implicit closures could be a bad idea for parallel languages. I simply
did not think his argument was compelling.
-rpg-
∂21-Nov-89 2048 ellis@src.dec.com Distribution of "su" newsgroups
Received: from decwrl.dec.com by SAIL.Stanford.EDU with TCP; 21 Nov 89 20:48:18 PST
Received: by decwrl.dec.com; id AA04890; Tue, 21 Nov 89 20:47:24 -0800
Received: by jumbo; id AA01295; Tue, 21 Nov 89 20:47:07 PST
Date: Tue, 21 Nov 89 20:47:07 PST
From: ellis@src.dec.com (John R. Ellis)
Message-Id: <8911220447.AA01295@jumbo>
To: jmc@sail.stanford.edu
Subject: Distribution of "su" newsgroups
I'm a big fan of your flames on su.etc, so I hope distribution isn't
restricted. To an outsider who's been following su.etc for 3 years
just for the fun of it (such as you versus Crispin), it sure looks like
the administrators don't like Crispin on su.etc now that he's gone.
∂22-Nov-89 0717 weening@Gang-of-Four.Stanford.EDU
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 22 Nov 89 07:17:24 PST
Received: from LOCALHOST by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA00275; Wed, 22 Nov 89 07:18:51 -0800
Message-Id: <8911221518.AA00275@Gang-of-Four.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
In-Reply-To: Your message of 21 Nov 89 22:35:00 -0800.
<KuCgt@SAIL.Stanford.EDU>
Date: Wed, 22 Nov 89 07:18:50 PST
From: Joe Weening <weening@Gang-of-Four.Stanford.EDU>
The recent concern has been from people posting to the su.gay
newsgroup. In the past, I've heard an su.etc contributor (Ramsey
Haddad) be quite surprised when an employer that he interviewed with
mentioned his postings -- he had no idea they went outside Stanford.
∂24-Nov-89 0937 cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK Lifschitz
Received: from NSFnet-Relay.AC.UK by SAIL.Stanford.EDU with TCP; 24 Nov 89 09:36:55 PST
Received: from sun.nsfnet-relay.ac.uk by vax.NSFnet-Relay.AC.UK
via Janet with NIFTP id aa13929; 24 Nov 89 17:27 GMT
Date: Fri, 24 Nov 89 17:18:59 GMT
From: cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK
Message-Id: <8911241718.AA08502@ipse2pt5.cs.man.ac.uk>
To: JMC@sail.stanford.edu
Cc: jrg%ipse2pt5.cs.man.ac.uk@NSFnet-Relay.AC.UK
In-Reply-To: John McCarthy's message of 20 Nov 89 2119 PST <Ktb#o@SAIL.Stanford.EDU>
Subject: Lifschitz
Dear John,
I have - as promised - spoken to (some of) the people who were
on the appointments committee for the "AI" chair. The matter
has been rather protracted because a strong attempt has been
made to find a way of making an offer to Vladimir that (whilst
not the original chair in question) we thought might attract
him. Brian Warboys in particular spent a long time
understanding Vladimir's situation with him and actually hoped
that a non-permanent position might be easier for Mrs
Lifschitz. With hindsight, and seeing that he must surely have
other offers to consider, we can see that our delay might well
have led to frustration. Knowing the people involved myself, I
can only assure you that none of them would have acted in bad
faith.
I, of course, conveyed your protest to the VC. A letter will
be sent to Vladimir in the near future which will hopefully
make the position entirely clear. (John Gurd - who is
Department Head - will write if the VC does not.) I hope that
this letter will satisfy Vladimir.
cliff
∂24-Nov-89 1415 harnad@clarity.Princeton.EDU Connectionist Learning/Representation: BBS Call for Commentators
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 24 Nov 89 14:15:30 PST
Received: from reason.Princeton.EDU by Princeton.EDU (5.58+++/2.25/mailrelay)
id AA09204; Fri, 24 Nov 89 17:00:58 EST
Received: by reason.Princeton.EDU (4.0/1.95)
id AA00826; Fri, 24 Nov 89 16:31:37 EST
Date: Fri, 24 Nov 89 16:31:37 EST
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8911242131.AA00826@reason.Princeton.EDU>
To: srh@flash.bellcore.com
Subject: Connectionist Learning/Representation: BBS Call for Commentators
Cc: AIList-request@AI.ai.mit.edu, Vision-List-Request@ADS.COM,
brain-database@athena.mit.edu, connectionists@cs.cmu.edu,
cvnet@yorkvm1.BITNET, jcha@csugreen.BITNET
(James Ha),
live-eye@yorkvm1.BITNET, neuron-request@hplabs.hp.com,
r2037@uqam.BITNET (Jacques Beaugrand),
epsynet%uhupvm1.bitnet@clarity.Princeton.EDU
Below is the abstract of a forthcoming target article to appear in
Behavioral and Brain Sciences (BBS), an international,
interdisciplinary journal that provides Open Peer Commentary on important
and controversial current research in the biobehavioral and cognitive
sciences. Commentators must be current BBS Associates or nominated by a
current BBS Associate. To be considered as a commentator on this article,
to suggest other appropriate commentators, or for information about how
to become a BBS Associate, please send email to:
harnad@confidence.princeton.edu harnad@pucc.bitnet or write to:
BBS, 20 Nassau Street, #240, Princeton NJ 08542 [tel: 609-921-7771]
____________________________________________________________________
WHAT CONNECTIONIST MODELS LEARN:
LEARNING AND REPRESENTATION IN CONNECTIONIST NETWORKS
Stephen J Hanson
Learning and Knowledge Acquisition Group
Siemens Research Center
Princeton NJ 08540
and
David J Burr
Artificial Intelligence and
Communications Research Group
Bellcore
Morristown NJ 07960
Connectionist models provide a promising alternative to the traditional
computational approach that has for several decades dominated cognitive
science and artificial intelligence, although the nature of
connectionist models and their relation to symbol processing remains
controversial. Connectionist models can be characterized by three
general computational features: distinct layers of interconnected
units, recursive rules for updating the strengths of the connections
during learning, and "simple" homogeneous computing elements. Using just
these three features one can construct surprisingly elegant and
powerful models of memory, perception, motor control, categorization
and reasoning. What makes the connectionist approach unique is not its
variety of representational possibilities (including "distributed
representations") or its departure from explicit rule-based models,
or even its preoccupation with the brain metaphor. Rather, it is that
connectionist models can be used to explore systematically the complex
interaction between learning and representation, as we try to
demonstrate through the analysis of several large networks.
∂24-Nov-89 1446 VAL
Spool el.dvi[1,val] to see new Elephant formulas.
∂24-Nov-89 2039 ME second DD monitor in your office
To: JMC@SAIL.Stanford.EDU
CC: ME@SAIL.Stanford.EDU, JJW@SAIL.Stanford.EDU
There is now a second DD monitor in your office and you can control both
monitors from your single keyboard.
At any given moment, the keyboard is associated with one or the other
of the two monitors. The commands to select the monitor are:
ESC , associates the keyboard with the primary line: old monitor, TV-140
BREAK , associates the keyboard with the second line: new monitor, TV-127
(That's ESCAPE or BREAK "comma".)
NOTHING resets the association to the primary line except a SAIL reload.
So it will stay associated where you leave it, even if you log out, type
BREAK CLEAR, or whatever.
Of course, on each monitor, you can do the usual DD line mapping stuff.
Since you have two "physical DD lines", the system thinks you have two
audio lines too, but I've hacked it to make any changes on your secondary
audio appear on your primary (only) audio. This can have some funny
effects. For instance, if you are listening to a radio station on one DD
and get a beep on the other, you'll find the radio gone when the beep
ends. You can fix this at that point by switching the second DD to the
same audio as the first. The alternative is to install a second speaker.
∂25-Nov-89 1338 rpg@lucid.com BAA
Received: from lucid.com by SAIL.Stanford.EDU with TCP; 25 Nov 89 13:38:12 PST
Received: from rose ([192.31.212.83]) by LUCID.COM id AA28900g; Sat, 25 Nov 89 13:34:33 PST
Received: by rose id AA01578g; Sat, 25 Nov 89 13:36:24 PST
Date: Sat, 25 Nov 89 13:36:24 PST
From: Richard P. Gabriel <rpg@lucid.com>
Message-Id: <8911252136.AA01578@rose>
To: clt@sail.stanford.edu, jmc@sail.stanford.edu
Subject: BAA
I've been encouraged to respond to it by Scherlis. I plan to do so
with a proposal from Stanford with a subcontract to Lucid to do the
prototype. The topic is
- persistent object management and typing;
- representation and management of software design information and
documentation, especially involving mixed formal/informal approaches;
- open-architecture environments and associated tools to support
risk-reduction process models;
- generic tools and integration technology for advanced open-architecture
environment prototypes.
Basically this will be the implementation of the CPL stuff for
environments.
First: John, are you willing to be the PI for this, assuming that I will
actually do the work?
Second: Carolyn, are you willing to act as administrator in your usual
fashion?
Third: If both answers are yes, then I need from Carolyn
A. A cover page including date, title, technical points of contact,
and administrative points of contact, including telephone number and
electronic mail address, if available. The cover page should be
plainly marked "PROPOSAL ABSTRACT" and should indicate the major
research area (Software Technology Research).
where I will supply the date and title. That is, I need the
``administrative points of contact, including telephone number and
electronic mail address.''
-rpg-
∂25-Nov-89 2210 ginsberg@Polya.Stanford.EDU Penrose
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 25 Nov 89 22:10:41 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA02460; Sat, 25 Nov 89 22:10:43 -0800
Date: Sat, 25 Nov 89 22:10:43 -0800
From: Matthew L. Ginsberg <ginsberg@Polya.Stanford.EDU>
Message-Id: <8911260610.AA02460@Polya.Stanford.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: Penrose
No. Does this mean you are reading his new book? I have it, but was
intending to save it for a vacation I'm taking in November.
Actually, all I can say for sure is that he wasn't a vegetarian in 1983.
Is the book any good?
Matt
∂26-Nov-89 0859 ginsberg@Polya.Stanford.EDU re: Penrose
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 26 Nov 89 08:59:36 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA12992; Sun, 26 Nov 89 08:59:36 -0800
Date: Sun, 26 Nov 89 08:59:36 -0800
From: Matthew L. Ginsberg <ginsberg@Polya.Stanford.EDU>
Message-Id: <8911261659.AA12992@Polya.Stanford.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: re: Penrose
Yes, I'd be interested in discussing the book; I should have it read by the
first of the year (that's my current plan, at least).
I don't remember about his politics (and I'm not sure that he really had
any, when I was there).
Matt
∂26-Nov-89 1358 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
INCREMENTAL, APPROXIMATE PLANNING
(Preliminary Report)
Charles Elkan
Department of Computer Science
University of Toronto
Monday, November 27, 3:15pm
MJH 252
I shall present evidence for three claims.
(1) Locally stratified definite clauses with the perfect model
semantics are adequate for specifying the Yale shooting problem
and other traditional planning domains, using the following
frame axioms:
causes(A,S,P) => holds(P,do(A,S))
holds(P,S) & ~ cancels(A,S,P) => holds(P,do(A,S))
(2) With the right control strategy, Green's method is a practical
way of inventing plans given a locally stratified domain theory.
(3) Allowing default conditions to be assumed is an abductive
planning strategy that produces plausible plans quickly, and those
plans can be incrementally refined towards guaranteed plans.
∂26-Nov-89 1556 VAL Elephant
File el1.tex[1,val] contains another formalization of the idea that a
passenger has a reservation if there is a time at which one was made
for him and which was not followed by a subsequent cancellation. It
differs from your formalization in two ways. First, it applies to
arbitary commitments, and not just to reservations. Second, in this
form it can be included in the semantics of Elephant.
∂26-Nov-89 1737 VAL reply to message
[In reply to message rcvd 26-Nov-89 17:34-PT.]
ok
∂27-Nov-89 0800 JMC
lock
∂27-Nov-89 1107 chandler@Polya.Stanford.EDU Faculty reports....again
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 27 Nov 89 11:07:34 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA12707; Mon, 27 Nov 89 11:07:12 -0800
Date: Mon, 27 Nov 1989 11:07:11 PST
From: "Joyce R. Chandler" <chandler@Polya.Stanford.EDU>
To: binford@coyote.Stanford.EDU, cheriton@pescadero.Stanford.EDU,
dill@amadeus.Stanford.EDU, feigenbaum@sumex-aim.Stanford.EDU,
rwf@sail.Stanford.EDU, genesereth@score.Stanford.EDU,
golub@patience.Stanford.EDU, lam@mojave.Stanford.EDU,
latombe@coyote.Stanford.EDU, zm@sail.Stanford.EDU,
jmc@sail.Stanford.EDU, ejm@sierra.Stanford.EDU, jcm@Polya.Stanford.EDU,
oliger@pride.Stanford.EDU, pratt@Polya.Stanford.EDU,
shoham@Polya.Stanford.EDU, winograd@csli.Stanford.EDU
Subject: Faculty reports....again
Message-Id: <CMM.0.88.628196831.chandler@Polya.Stanford.EDU>
I am persistent.....
This is yet another "gentle" reminder to please get me your faculty reports
as soon as possible. I have a 12/15 deadline to meet with the School of Engineering.
∂27-Nov-89 1216 ross%cs@ucsd.edu Distinguished Lecturer Expenses
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 27 Nov 89 12:16:25 PST
Received: from cseadmin.ucsd.edu by ucsd.edu; id AA02522
sendmail 5.61/UCSD-2.0-sun via SMTP
Mon, 27 Nov 89 12:16:29 -0800 for JMC@sail.stanford.edu
Received: by cseadmin.UCSD.EDU (3.2/UCSDPSEUDO.2)
id AA00197 for delivery to JMC@sail.stanford.edu; Mon, 27 Nov 89 12:16:00 PST
Date: Mon, 27 Nov 89 12:16:00 PST
From: ross%cs@ucsd.edu (Steve Ross)
Message-Id: <8911272016.AA00197@cseadmin.UCSD.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: Distinguished Lecturer Expenses
Cc: ross@cs.UCSD.EDU
Dr. McCarthy,
We received your letter of expenses today, but no receipts were inclosed.
Please forward any receipts you have to me at the same address. Ms. Shores
has taken another postion and I'll be handling your reimbursement. Thank you.
Steve Ross
CSE Dept.
UCSD, C-014
La Jolla, CA 92093
∂27-Nov-89 1722 RPG CV
I need a short version of your cv for a darpa proposal for more cpl work.
∂27-Nov-89 1744 winograd@loire.stanford.edu AI QUAL SCHEDULE
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 27 Nov 89 17:44:02 PST
Received: by loire.stanford.edu (5.59/25-eef) id AA11408; Mon, 27 Nov 89 17:38:46 PDT
Date: Mon, 27 Nov 89 17:38:46 PDT
Message-Id: <8911280138.AA11408@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: ai-qual-faculty@loire.stanford.edu, ash@sumex-aim, bthomas@neon,
guha@sumex-aim, lin@neon
Subject: AI QUAL SCHEDULE
This is the official schedule for the qualifying exams next
Wednesday morning, December 6, 1989. Let me know immediately
if there is a problem, or I will assume everyone can be there
as scheduled.
On each committee there is a (C)hair, a (D)epth examiner and one (O)ther.
8:45 - 10:15
Becky Thomas: bthomas@neon - representation
C: Winograd D: Hayes O: Binford
David Ash: ash@sumex-aim - planning
C: LaTombe D: Nilsson O: McCarthy
10:30 - 12:00
Fangzhen Lin: lin@neon - nonmonotonic
C: LaTombe D: McCarthy O: Hayes
Ramanathan Guha: guha@sumex-aim - representation
C: Nilsson D: Ginsberg O: Hayes-Roth
--t
∂27-Nov-89 1745 VAL Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU, su-events@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
A NEW APPROACH TO MODAL OPERATORS
Matt Ginsberg
Stanford University
Monday, December 4, 3:15pm
MJH 252
We describe a new formalization of modal operators that views them
not in terms of Kripke's possible worlds, but as functions on an
underlying set of truth values. Thus Moore's knowledge operator L,
where Lp means "I know that p," would correspond to a mapping taking
true into true (since we know p if we know it to be true) and taking
both false and unknown into false (since we do not know p if we
either know it to be false or know nothing about it at all). This
new approach has the following advantages over the conventional ones:
1. Intuitive simplicity
2. It provably generalizes both Kripke's construction and Moore's
autoepistemic logic, while making clear the distinctions between
them.
3. It allows for easy further generalization to modal operators
that are related to temporal reasoning and to causality.
4. The natural procedure for computing the truth value of a sentence
involving these modal operators is "incremental" in the sense that
it computes approximate answers that "converge" to the correct one
in the large runtime limit.
This talk will concentrate on the first two of these properties;
I will discuss the third and fourth as completely as time allows.
∂27-Nov-89 2312 golub@na-net.stanford.edu Re: Faculty reports....again
Received: from bravery.stanford.edu by SAIL.Stanford.EDU with TCP; 27 Nov 89 23:12:06 PST
Received: by bravery.stanford.edu (4.0/inc-1.5)
id AA06528; Mon, 27 Nov 89 23:16:07 PST
Date: Mon, 27 Nov 1989 23:15:53 PST
From: Gene Golub <golub@@na-net.stanford.edu>
To: "Joyce R. Chandler" <chandler@Polya.Stanford.EDU>
Cc: binford@coyote.Stanford.EDU, cheriton@pescadero.Stanford.EDU,
dill@amadeus.Stanford.EDU, feigenbaum@sumex-aim.Stanford.EDU,
rwf@sail.Stanford.EDU, genesereth@score.Stanford.EDU,
golub@patience.Stanford.EDU, lam@mojave.Stanford.EDU,
latombe@coyote.Stanford.EDU, zm@sail.Stanford.EDU,
jmc@sail.Stanford.EDU, ejm@sierra.Stanford.EDU, jcm@Polya.Stanford.EDU,
oliger@pride.Stanford.EDU, pratt@Polya.Stanford.EDU,
shoham@Polya.Stanford.EDU, winograd@csli.Stanford.EDU
Subject: Re: Faculty reports....again
In-Reply-To: Your message of Mon, 27 Nov 1989 11:07:11 PST
Message-Id: <CMM.0.88.628240553.golub@>
We are in good company!
Gene
∂28-Nov-89 1005 tajnai@Polya.Stanford.EDU Daniel Scales
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 28 Nov 89 10:05:15 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA18053; Tue, 28 Nov 89 10:05:22 -0800
Date: Tue, 28 Nov 1989 10:05:21 PST
From: "Carolyn E. Tajnai" <tajnai@Polya.Stanford.EDU>
To: jmc@sail.Stanford.EDU
Cc: clt@sail.Stanford.EDU
Subject: Daniel Scales
Message-Id: <CMM.0.88.628279521.tajnai@Polya.Stanford.EDU>
John, is there a reason why you did not nominate Daniel Scales for the
Bell Fellowship? I believe he is your advisee.
Carolyn
∂28-Nov-89 1044 tajnai@Polya.Stanford.EDU re: Daniel Scales
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 28 Nov 89 10:44:23 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA19153; Tue, 28 Nov 89 10:44:31 -0800
Date: Tue, 28 Nov 1989 10:44:31 PST
From: "Carolyn E. Tajnai" <tajnai@Polya.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: Daniel Scales
In-Reply-To: Your message of 28 Nov 89 1012 PST
Message-Id: <CMM.0.88.628281871.tajnai@Polya.Stanford.EDU>
I have to make a decision this week -- we can only nominate 3
students. There are 4 2nd year PHD students who have completed
their comps, quals and filed for candidacy. I'm trying to narrow
it down to 3, and thought that Dan should be considered.
Don't say anything to Dan at this point. Shall we consider him?
That commits you to writing a good letter of recommendation.
Carolyn
∂29-Nov-89 0915 MPS Vacation
Good morning,
I would like to take the 27, 28, and 29th off
after Xmas, if that is okay with you. Thanks.
Pat
∂29-Nov-89 1235 MPS
Mr. Todorovich called a couple of times. 408-252-6210
∂30-Nov-89 0824 MPS MCC
Good morning,
Would you like a separate folder of the MCC stuff to
be filed in your desk drawer rather than in the chron
file?
Pat
∂30-Nov-89 1059 qphysics-owner@neat.cs.toronto.edu Mailing List
Received: from neat.cs.toronto.edu by SAIL.Stanford.EDU with TCP; 30 Nov 89 10:59:24 PST
Received: by neat.cs.toronto.edu id 2899; Thu, 30 Nov 89 13:58:55 EST
Received: from crash.cs.umass.edu by neat.cs.toronto.edu with SMTP id 2591; Thu, 30 Nov 89 13:52:53 EST
Received: from jason.cs.umass.edu by crash.cs.umass.edu (5.61/Ultrix2.0-B)
id AA08644; Thu, 30 Nov 89 13:51:27 -0500
Message-Id: <2837444090-8577978@Jason>
Sender: LAASRI%Jason@cs.umass.edu
Date: Thu, 30 Nov 89 13:54:50 EST
Reply-To: Laasri%Jason@cs.umass.edu
From: Hassan Laasri <Laasri%Jason@cs.umass.edu>
To: qphysics@ai.utoronto.ca
Subject: Mailing List
Resent-From: qphysics-owner@cs.toronto.edu
Resent-To: qphysics@cs.toronto.edu
Resent-Reply-To: Hassan Laasri <Laasri%Jason@cs.umass.edu>
Resent-Message-Id: <89Nov30.135855est.2899@neat.cs.toronto.edu>
Resent-Date: Thu, 30 Nov 89 13:54:38 EST
Hi,
We're at UMass/Coins during a year of post-doc research. We work with
Prof. Victor Lesser on Distributed Artificial Intelligence. We're also
very interested by research on Qualitative Reasoning and Model-Based
Reasoning. Johan De Kleer suggested us to contact you in order to be
included in a mailing list concerning these topics. So, would you please
add our names in your mailing list? Here is our complete address:
Brigitte MAITRE and Hassan LAASRI
COINS - LGRC
University of Massachusetts
Amherst, MA 01003
USA
Email: maitre and laasri@cs.umass.edu
Thanks a lot,
Sincerely,
--- Brigitte and Hassan
∂30-Nov-89 1112 VAL re: Elephant programs as sentences
[In reply to message rcvd 30-Nov-89 00:57-PT.]
The new formulas look very nice. What happens to \xi now?
About the list of predicates that you propose to circumscribe:
$holds$, $arises$, $outputs$, $confirm$ and $revoke$. First, it
seems to me that $confirm$ is out of place here, because it is
not a predicate, but rather a function symbol of the i/o language.
Second, I'm not sure about $holds$. It seems that you can have a
complex reified formula as its second argument. For any A, we have
either holds(t,A) or holds(t,not A), so that it doesn't make mich
sense to minimize holds.
∂30-Nov-89 1125 GLB
To: JK@SAIL.Stanford.EDU, JMC@SAIL.Stanford.EDU
Idiocy and arrogance are at a historical low in Eastern Europe,
quite wonderfully! Dubcek was almost a hero also for us in Italy
20 years ago: it doesn't happen often that a hero survives
long enough to be vindicated!
The main question is whether we will ever have our European house
in order. About Western Europe, I don't know whether there is
anything that could be called a European political leadership.
Lady Thatcher is mainly engaged in a crusade against Socialism
and she seems to feel that she can win in Britain but not in Europe,
therefore she now resists European financial integration.
The W.German government -- quite understandably -- wants reunification,
but everyone else -- quite understandably too -- is weary
of the Germans. The Italian government does not inspire much respect.
The French government seems to have broader views, but it remains
to be seen how much European responsibility they may be willing
to take in practice.
As a consequence, everyone is asking the US to keep its political
and military protection for indefinite time. But this is precisely
what allows everyone to continue their parochial crusades and
to avoid the necessary `glasnost' and `peristroika', in the sense
of telling the public about common European responsibilities
and of restructuring the political system towards a real federal state.
I hope that budget constraints will force the US to withdraw
their troups from Europe (sure, gradually etc.) and European
politicians to face reality.
PS. Perhaps the statement that about Italian governments
does not require many explanations, but there are
more serious issues than the a government crisis every year.
E.g. the Christian Democrats, the main party, have not entirely
cleared their image w.r.t. some Mafia-connections, at a time when
Sicily is to Europe almost what Columbia it to the US.
The Italian government has still to provide an explanation
for a jetliner that vanished ten years ago in the see North of Sicily,
most likely hit by a stray missile in an area close to ongoing
NATO military manouvers, If I remember well, the military refused
to make their radar records public.
Foreign agents and holy warriors of all faiths seem to like vacationing
in Italy, so they don't mind doing their job there either. This includes
not only the well known Turkish lunatic Ali Agca and the much publicized
Arab attacks but also the less known actions of the Israelis,
who apparently kidnapped Mordechai Vanunu, -- the technician responsible
for providing evidence of Israeli nuclear buildup -- precisely in Rome.
Lack of respect for the government also creates periodic waves of
resentment for the foreign protectors of the government --- this
may or may not be justified --- e.g., the famous equation
Italy : US = Bulgaria : USSR.
∂30-Nov-89 1332 shoham@Hudson.Stanford.EDU agent-oriented programming
Received: from Hudson.Stanford.EDU by SAIL.Stanford.EDU with TCP; 30 Nov 89 13:32:44 PST
Received: by Hudson.Stanford.EDU (5.61/25-eef) id AA01027; Thu, 30 Nov 89 13:32:12 -0800
Date: Thu, 30 Nov 1989 13:32:10 PST
From: Yoav Shoham <shoham@Hudson.Stanford.EDU>
To: jmc@sail
Subject: agent-oriented programming
Message-Id: <CMM.0.88.628464730.shoham@Hudson.Stanford.EDU>
John,
Did you get a chance to look at that white paper?
Yoav
∂30-Nov-89 1455 winograd@loire.stanford.edu Locations and lunch for AI qual
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 30 Nov 89 14:55:30 PST
Received: by loire.stanford.edu (5.59/25-eef) id AA20746; Thu, 30 Nov 89 14:52:49 PDT
Date: Thu, 30 Nov 89 14:52:49 PDT
Message-Id: <8911302252.AA20746@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: ai-qual-faculty@loire.stanford.edu
Subject: Locations and lunch for AI qual
Thomas (Winograd, Hayes, Binford) and Lin (LaTombe, McCarthy, Hayes)
will be in 301. Ash (Latombe, Nilsson McCarthy) and Guha (Nilsson,
Ginsberg, Hayes-Roth) in Nils' conference room.
The evaluation meeting will be in Nils conferenc room at 12 sharp.
I am planning to get lunches from the Jordan Hall Cafe. Choose one
of the following if you plan to be there and want to eat. (Let me know
if you plan to come but not eat, too).
Thai food -vegetarian
Thai food -other
Sandwich [specify kind]
Feel free to give additional details, drinks, etc. if you know they are
available there.
--t
∂01-Dec-89 1224 @Score.Stanford.EDU:kirsh%cogsci@ucsd.edu aij
Received: from Score.Stanford.EDU by SAIL.Stanford.EDU with TCP; 1 Dec 89 12:24:36 PST
Received: from ucsd.edu by SCORE.STANFORD.EDU with TCP; Fri 1 Dec 89 12:24:13-PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA29503
sendmail 5.61/UCSD-2.0-sun via SMTP
Fri, 1 Dec 89 12:24:25 -0800 for jmc@score.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA21266 for jmc@score.stanford.edu; Fri, 1 Dec 89 12:23:05 PST
Date: Fri, 1 Dec 89 12:27 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: aij
To: jmc@score.stanford.edu
Message-Id: <19891201202701.7.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
John,
How are you proceeding with your reply to Carl. Other replies have been
clicking in at 10 -20 pages, and I am eager for you to feel that you can sya
as much as you want. I will be pout pof the country as of Dec 12th unti the
new year. SO please take the next week to prepare your text that I maight
have a good draft soon enough before I leave that I might review it and send
it back for revisions.
-- David
PS. Your mailer does not recognize jmc@sushi or jmc@polya. You probably
ought to alert it.
∂01-Dec-89 1807 kirsh%cogsci@ucsd.edu re: aij
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 1 Dec 89 18:07:16 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA16664
sendmail 5.61/UCSD-2.0-sun via SMTP
Fri, 1 Dec 89 18:07:23 -0800 for JMC@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA26064 for JMC@sail.stanford.edu; Fri, 1 Dec 89 18:06:04 PST
Date: Fri, 1 Dec 89 18:10 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: re: aij
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <KDx7A@SAIL.Stanford.EDU>
Message-Id: <19891202021009.1.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
What a snafu. I sent you a copy of Carl's paper by email, and Carl told me he
sent you a copy by post. I have had trouble with your mailer and my mailer so
I can understand if you never received my transmission of Carl's new version
of his aij essay. It has been substantially revised.
Your comments of two years ago are not beside the point but need a good deal
of revising to keep in line with the formality and breadth of the other
commentaries. When we spoke last I though you might use some of your new
ideas about the designer's knowledge of the environment to reply to Carl's
argument that Turing machines are inadequate because they can't explain
the effect of timed data arrivals on a computation. In any event I am
enclosing a copy of Carl's paper, and would greatly appreciate you're revising
your reply. I really think that you should say absolutely as much as you
like. The norm for replies is about 12 pages. We are now running past what
I had hoped was a real deadline, so tell me how much time you expect to need.
I am sending under separate email Carl's manuscript.
Please reply when you receive this.
-- David
∂02-Dec-89 1527 kirsh%cogsci@ucsd.edu hewitt part1
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 2 Dec 89 15:26:51 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA22865
sendmail 5.61/UCSD-2.0-sun via SMTP
Sat, 2 Dec 89 15:26:54 -0800 for jmc@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA02934 for jmc@sail.stanford.edu; Sat, 2 Dec 89 15:25:29 PST
Date: Sat, 2 Dec 89 15:29 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: hewitt part1
To: jmc@sail.stanford.edu
Message-Id: <19891202232928.8.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
Message bounced. I'm trying it in smaller bits.
OK John here's the paper, please signal when you receive it.
Date: Sun, 12 Nov 89 11:38 PST
From: Carl Hewitt <hewitt@ai.mit.edu>
Subject: my paper
To: David Kirsh <kirsh%cogsci-sun@ucsd.edu>, JMC@sail.edu
cc: hewitt@ai.mit.edu
John and David,
Here is an electronic copy my paper for the AI Journal. I have Federal
Expressed hardcopy to you which you should receive on Monday. I apologize for
taking so long to get it to you.
Sincerely,
Carl
\documentstyle[11pt]{article}
\def\startline{\par\nobreak\noindent}
{\obeylines\obeyspaces\gdef\procedure{\bigbreak\begingroup
\parindent=0pt \parskip=0pt plus0pt minus0pt
\obeylines \obeyspaces \eg\let↑↑M=\startline \eg}
\gdef\endproc{\par\endgroup\bigbreak}}
\def\wyn#1{$\spadesuit${\tt #1}}
\newcommand\code[1]{\mbox{\eg{#1}}}
\newcommand\eol{\hfill\break}
\newcommand\bline{\vskip12pt plus0pt minus0pt}
\newcommand\eg{\tt}
\title{Artificial Intelligence\\and/or\\Open Systems Technoscience}
\author{\copyright\ 1989 Carl Hewitt}
\date{Draft of \today}
\begin{document}
% \bibliographystyle{plain}
% \bibliography{/home/tx/wsnow/biblio}
\maketitle
\section{Abstract}
Open Systems Technoscience (OST) is the technology and science of large
scale human/telecomputer organizational work. Examples of this kind of
work include flexible semiconductor manufacturing, constructing a
permanent space station, and the software engineering of electronic
fund transfer systems. Open Systems Technoscience addresses issues of
participant values, robustness, and scalability in large scale
organizational work. In contrast, Artificial Intelligence (AI) is the science
and technology of intelligent agents and robots. The relationship
between OST and AI has much of the same complementary and
supplementary character as the relationship between sociology and
psychology.
Artificial Intelligence has created new technologies of taxonomies,
inference-based systems, and problem spaces. All of these technologies
can be useful within organizations that engage in large-scale work.
However, each of these technologies is useless without the extensive
organizational support that is necessary to make it work.
\section{Introduction}
All large-scale Open Systems are concurrent, asynchronous, decentralized,
and indeterminate. They are composed of numerous participants which
operate {\em concurrently\/} in order to accomplish the multitude of
tasks that are performed. They are {\em distributed\/} in order to deal
with the influx of information from many sources and to convey
information to the places where it is needed.
In any short span of time, each participant of a large-scale Open System
operates {\em autonomously\/} and {\em asynchronously\/} in accordance
with its own local needs and procedures. No truly simultaneous change of
all the participants in a large-scale Open System is ever possible, and new
information may arrive from any source at any time. Thus in general, one
participant of the system will start using new information before it
reaches the others.
Furthermore, asynchronous operation means that any large-scale Open
System is {\em indeterminate\/} in a physical sense: it does not have a
determinable current state which (together with new information that
arrives) determines its future operation. In fact, attempts to pin down
an instantaneous state of a large system by gathering more information
about the finest details of its internal operation makes the system {\em
more\/} indeterminate, because gathering the information affects its
operation. Furthermore, another large-scale Open System is needed just
to gather, interpret, and store the information about the smallest-scale
activities of another large system.
\subsection{Advantages of Open Systems}
Open Systems Technoscience addresses issues of gaining advantages from the
Open Systems environment in which large-scale organizations operate:
\begin{itemize}
\item
{\bf Asynchrony:} enables each participant of an organization to operate as
quickly as possible, given local circumstances. Otherwise, the pace of
the activity of each participant would be locked to a single organizational
scheduler down to the lowest level activities.
\item
{\bf Local autonomy:} enables each participant of an organization to react
immediately to changing circumstances. Otherwise, each participant would
have to consult a single organizational decision maker for each decision.
\item
{\bf Late-arriving information:} enables the organization to increase the
effectiveness of its decision making by incorporating information as it
arrives.
\item
{\bf Multiple authorities:} increases an organization's pluralism,
diversity, and robustness. For example, the engineering department of a
utility wants to build a new kind of nuclear reactor which is inherently
safer than existing reactors, while the finance department maintains that
the financial risk is too great.
\item
{\bf Arm's length relationships:} enables actors to conceal their internal
activities from other other actors. For example, the Nuclear Regulatory
Commission operates at arm's length from nuclear utilities that it
regulates in order to be more effective in detecting and prosecuting
violations of its regulations.
\item
{\bf Division of labor:} specialization of functionality can increase
organizational effectiveness by enabling each participant to concentrate
and focus its efforts on a narrow range of goals. For example, an
electric utility has separate finance and engineering suborganizations,
each specializing in different functions: finance takes care of raising
the money, and engineering directs the construction of power plants.
\end{itemize}
\section{Trials of Strength}
Any situation in which forces are pulling in different directions
constitutes a {\em trial of strength}. All trials of strength are local in
the sense that they occur at a particular time and place among local
participants. For example negotiation provides a mechanism for
representatives of different participants to come together in a trial of
strength.
\subsection{Commitments}
A {\em commitment\/} is a course of action, usually involving several
actors, which the actors are {\em on course\/} to carry through. An
actor's {\em responsibility\/} is its part in the commitment.
Commitments are used in this way in order to avoid becoming embroiled
in issues of {\em intentionality\/}. Instead, an actor is said to be {\em
on course\/} or {\em off course\/} with respect to an activity. This
terminology allows inanimate objects to participate in commitments.
The course of action deliberately chosen by an organization in order to
accomplish its goals can also be a commitment.
Conflict is a trial of strength in which participants have incompatible
commitments. As we have seen, conflict is a fundamental aspect of any
Open System for a variety of reasons. Resources are finite and limited;
choices must be made about how to use them. In addition, most
organizations have built-in checks and balances which more or less
deliberately generate conflicting commitments. The commitment of the
safety department to safe conservative procedures conflicts with the
need for the engineering department to lower costs of construction.
This often places them in conflict with one another.
Sometimes the commitments of the affected participants are compatible
and there is no conflict. One participant says, ``Let's do it this way,''
and everybody agrees, so the negotiation is a trivial one, but there is
always the {\em potential\/} for conflict. No one can be certain in
advance about the outcome, so any negotiation is to that extent a trial
of strength and therefore indeterminate.
Our new discipline of Open Systems Technoscience needs to have an
intimate understanding of the nature of commitment. In particular, Open
Systems Technoscience is especially concerned with the nature of
overall organizational commitments and how they relate to the
commitments of participants inside and outside the organization,
because these broader commitments are what make large-scale projects
possible.
An organizational commitment is one which the organization is on course
to carry out. For example, a utility can have an organizational
commitment to build a new power plant. The utility's Finance department
has a commitment to fund the cost of the plant, and its Engineering
department has a commitment to design and build it. Both Finance and
Engineering must work together in a fairly detailed way, and the
organizational commitment must address the specialized needs and
commitments of both participants. The organizational commitment of the
utility goes beyond the responsibilities of Financing, Engineering, and
other departments. The utility has organizational commitments and
authority that go beyond just the individual ones of its departments and
members.
We said before that a participant's responsibility for a commitment is
its part in a commitment. Alternative courses of action can affect
participant use of resources which is often a source of conflict. There
are several kinds of resources: money, space/time/material,
mechanism/technology, and sentiment.
Money and space/time/material are self-explanatory. Mechanism/\break
technology is anything that transforms the world: for example, a nuclear
power plant transforms radioactive material into electricity. Sentiment
deals with how various actors feel about each other: goodwill,
reputation, obligation, etc. For example, when a public utility says
that its reactor is safe, it is staking part of its reputation on that
statement. Also, when a utility asks for leniency on an issue from a
regulatory board, it is putting to use some of its goodwill for the
utility, which means that it is inhibited in its freedom later on.
However, the past is really gone and the future is never here. Both
planning for the future and reflecting upon past experience are
activities that take place in the present. In addition to the daily,
familiar use of resources, some activities which look to the future and
the past are also commitments. For example, planning can create a
commitment; it chooses a particular course of action that allocates
resources in one way instead of another and sets the planner on course to
keep the commitment. Creating a financial record also generates a
commitment. When a utility submits its annual report, it makes a
statement which says how much money it earned that year. Once it reports
this statement to the regulators, that's a commitment. If the regulators
come and investigate, the utility has to substantiate its claim that it
really did earn that much during the year.
Two commitments are said to be {\em conflicting\/} if they give rise to a
trial of strength which results in at least one of the commitments not
being kept. For example, the commitment of a utility to operate a nuclear
power plant is in conflict with the commitment of an environmental group
to shut it down. These two commitments are incompatible, and cannot both
be kept.
The notion of commitment used in this paper is closely related the
terminology used by those sociologists who emphasize a commitment as a
choice among incompatible alternatives
\cite{concept-of-commitment,quality-of-life}. In this terminology, {\em
choice\/} is analyzed in terms of trials of strength. Commitments have
been discussed by Winograd and Flores in their discussion of Heiddiger's
notion of ``throwness'' \cite{winograd-flores}, and by Richard Fikes
\cite{commitment-based-framework} in the context of contract nets as
developed by Reid Smith and Randy Davis \cite{contract-nets}.
% also \cite{improvised-news}.
% CH NOTE: Get {\em Improvised News\/} by Tom Shibutani, study of
% rumor in organizations, citation goes in structure of commitments.
\subsection{Open Systems Semantics}
Open Systems Semantics is the study of the meaning of Open Systems
Action. Taking any action entails changes in commitments, and that
change is the meaning of the action. This is an open-world
characterization of meaning---as opposed to previous, closed-world
attempts based on possible worlds in which the meaning of a set of
sentences is defined to be the set of all possible worlds that
satisfy the meaning conditions. Building on previous actor theory, Gul Agha
\cite{agha-phd} has developed an
open-world, mathematical semantics for concurrent systems in which the
meaning of an action is characterized by its effect on the evolution of
the system.
In general, Open Systems Action involves conflict---and therefore
indeterminacy. Current situations and events influence, but do not
determine, the future because the outcomes of numerous trials of
strength cannot be known and some trials of strength cannot even be
anticipated.
In logical semantics, representation is the mapping between a sentence
proposition and specified meaning conditions. Meaning is built on and
grows out of representation. Two participants agree about the meaning of a
sentence when they agree about the meaning conditions.
In Open System semantics, however, representation is taken to be how
effects commitments. The important difference is conveyed in the
following Open Systems Semantics slogan:
\begin{quote}
{\Large {\bf No representation without communication!}}
\end{quote}
The test of the degree to which one participant has adequately
communicated its meaning to another participant is: whether the recipient's
commitments change in the way which constitutes the meaning.
Large organizations have extensive policies, procedures, and regulations
which govern the meaning of organizational actions.
Open Systems Semantics is a research programme
\cite{methodology-of-research-programmes} for studying the meaning of
Open Systems Actions. Just as there is no global synchrony or cause
and effect, the meaning of an Open Systems Action is also localized: it
begins in the participants at a particular time and place and then
can causally propagates to wider contexts.
\subsection{Requirements of Open Systems}
Analysis of commitment relationships can be used to understand how the
characteristics of Open Systems engender requirements for Open
Systems implementations:
\begin{itemize}
\item
{\bf Asynchrony:} produces indeterminacy because arriving information
must be integrated with local information. The new information can
generate new commitments that conflict with pre-existing commitments.
\item
{\bf Local autonomy:} produces conflict because as organizational
commitments change, some of the new commitments will be incompatible with
pre-existing local commitments.
\item
{\bf Late-arriving information:} can produce conflict when it arrives at
an advanced stage of processing (e.g., an engineering group reports new
concerns about the capabilities to survive an earthquake just as the
utility is about to apply for a license to operate the reactor).
\item
{\bf Multiple authorities and division of labor:} can produce conflict
because the specialized commitments of multiple authorities may be
incompatible and come into conflict.
\item
{\bf Arm's length relationships:} can produce conflict because the
internal commitments of other organizations are not visible. This can
increase the severity of conflict because other organizations will
develop entrenched commitments before the conflict is discovered.
\end{itemize}
Another underlying constraint is {\em continuous operation}. In many
cases an organization cannot ``take a vacation'' in order to get itself
into better shape. It must continue operating though perhaps
at some reduced level of performance.
\subsection{Dealing with Unanticipated Conflict}
Robustness is keeping commitments in the face of unanticipated conflict.
Keeping a commitment often entails keeping subcommitments. For
example, the commitment to operate the Diablo Canyon reactor requires
keeping two subcommitments: constructing the plant and licensing it.
The licensing subcommitment might go smoothly at first and then run
into unanticipated conflict when seismologists discover new geological
faults near the location of the reactor. The robustness of the
commitment of the utility in part depends upon its ability to deal with
whatever unexpected trial of strength arises. In this case
the utility engages in a negotiation (i.e. a trial of strength)
with the Nuclear Regulatory Commission.
\section{Negotiation}
Negotiation is an important way in which an organization achieves
robustness (i.e. keeping commitments in the face of conflict). A
negotiation is a discourse in which the participants make
representations held in common which can engender joint actions that
are new joint commitments. The new commitments will often in turn
lead to more conflict with pre-existing commitments of individual
participants as they make adjustments for the new commitments.
The organizational commitment of the utility to operate a nuclear power
plant lead to its giving its Finance and Engineering departments more
specialized subcommitments. Engineering has the responsibility to
construct the reactor and Finance has the responsibility to raise the
money for construction. Negotiation gives both participants an
opportunity to negotiate how the utility can keep its commitment to
operate the plant.
Thus, we can see the value of bringing participants together to work out an
organizational response to conflict. Each participant has ongoing
commitments to attend to, and needs a way of figuring out how to allocate
its resources. Instead of being required to turn its full attention to
the new commitment, each participant can separate out a subpart---namely the
representative to the negotiation---which will be devoted to the new
commitment.
Negotiation can help an organization meet its existing commitments while
developing new ones---which is important for attaining robustness.
Negotiation also creates overall, organizational commitments---which are
essential to scaling. Thus, Open Systems Technoscience needs to support
mechanisms that can move a negotiation forward, and determine what
progress is being made.
Insights gleaned from the social sciences (law, sociology, anthropology,
organization theory, and philosophy of science) can help us conduct
human/telecomputer organizational work with increased participant
value. Human organizations have evolved methods for dealing
with conflict---and for turning them into strengths instead of
weaknesses. These methods can be adapted as a source of inspiration
for robustness in human/telecomputer organizations.
\subsection{What Happens During a Negotiation}
Each participant brings its own commitments to a negotiation. These
include decision-making criteria, such as preferences among predicted
outcomes. For example: ``It is preferable to have nuclear power plants
because they lessen our dependence on unstable foreign fuel supplies''
and ``It is preferable not to have nuclear power plants because they
create a threat of the release of significant amounts of radioactivity''
Conflicts among these preferences can be negotiated
\cite{negotiations}.
During a negotiation, the participants can make moves---where each move is the
description of a commitment. In this way, the various participants can arrive
at a joint commitment about the issues being addressed and the options
available for addressing those issues.
The various participants {\em clarify the issue\/} by discussing and
commenting upon each others' statements about the issue. This process
may expand or change the views of the various participants about the nature of
the issue. The representatives also discuss what the various {\em
options\/} are with respect to the commitments around this issue. An
option is a proposal for the rearrangement of the organization's
commitments. Here, discussion and commentary about options can lead to
the generation of new options, further clarification of the nature of the
issue, and further commentary on the commitments that are affected by the
issue.
\subsection{Contradictions}
{\em Contradictions\/} arise because each negotiating participant attempts to
keep its own commitments. When conflict leads to a negotiation, each
participant deliberately resists some of the statements that other participants make
to support their commitments. Each participant uses language as a tool to
further its own commitments, and that often produces resistance and even
contradictory statements.
The expression of conflict can be a very positive aspect of negotiation.
The diversity that produced the conflict and contradictions can also
produce new ideas, suggestions, and options. Thus, negotiation of
conflict can lead to more effective organizational information
processing.
\subsection{Cooperation}
If an organization is continually involved in negotiations, how does it
ever get anything accomplished? Won't bringing bureaucracy to the
machine create even worse messes than in human bureaucracies? While it's
true that negotiations can break down and result in deadlocks, they can
also come up with creative solutions. Two factors that help ``grease the
wheels'' of negotiation are cooperation and allies.
{\em Cooperation\/} is the process by which participants become committed to
each others' commitments. For example, Finance and Engineering might
both be committed to building a new power plant. Finance is committed to
raising the money to build the plant, and Engineering is committed to a
schedule for constructing the plant. Finance relies on Engineering's
commitment to build the plant on schedule in order to have credibility in
the financial marketplace---so Finance is committed to Engineering's
commitment. In a similar fashion, Engineering relies on Finance's
ability to provide the money to pay the construction cost as it comes due
so that construction can continue. So Engineering is committed to
Finance's commitment. Because we have these cross-commitments, we say
that Finance and Engineering are cooperating.
Another example deals with a utility and vendor. The utility generates a
purchase order for product $Q$. The vendor commits itself to shipping
product $Q$ at a certain time. The utility's commitment to the vendor's
shipping product $Q$ is represented by the purchase order. The vendor's
commitment to ship the product is dependent on the utility's commitment
to pay---as represented by the purchase order. Thus the purchase order
formalizes the cooperation between customer and vendor, and represents
the commitment of both participants to the others' commitment.
This process of mutual commitment---cooperation---is of fundamental
importance in moving organizations forward.
\subsection{Allies}
Another aspect of negotiation that helps make the negotiation process go
smoothly is the notion of an {\em ally}. One of the participants may claim an
ally: i.e., predict that in some future trial of strength, its ally would
behave in a certain way. In the payroll example, someone might claim
that if management doesn't allocate sufficient funds to meet the payroll,
the union will call for a strike!
The ally may prove to be faithful or unfaithful: faithful in the sense
that in the future trial of strength, it actually does behave as claimed.
\begin{quote}
{\em
\vbox{Glendower: I can call spirits from the vast deep.
Hotspur: Why, so can I, or so can any man; but will they come when you
do call for them?
\hfill---Shakespeare: {\em Henry IV}, {\em Act III}, {\em Scene 1}.}}
\end{quote}
Thus, when a negotiating participant claims an ally, the likely outcomes of
future trials of strength must be considered in deciding on what action
to take. In this way, claiming the support of allies can sway a
negotiation in a certain direction, but it is not necessarily decisive.
For example, a customer might invoke the Public Utility Board as an ally
when presenting a complaint to a utility that refuses to remove excess
charges from its bill. The utility needs to consider the implications of
an Utility Board investigation in deciding how to respond to the
customer.
Claiming an ally can be a very one-sided relationship. If the utility
decides to honor the customer's claim about the probable outcome of a
Utility Board investigation, then the Utility Board would never hear
about the incident, even though it was successfully invoked as an ally.
By contrast, an alliance is a {\em mutual commitment}. The participants to an
alliance make mutual commitments of their time, money, staff, and other
resources. Alliances are important outcomes of negotiation. Almost
every negotiation will affect alliances, either by creating new ones or
by strengthening, adjusting, or weakening old ones.
\section{Outcomes of Negotiation}
Many kinds of outcomes are possible, but the following three often
occur:
\begin{itemize}
\item
A {\bf resolution} to which the participants commit themselves.
\item
A {\bf deadlock} in which the participants at this particular negotiation
cannot reach an agreement. Quite often as a result of deadlock, another
negotiation is held with different representatives, and on a different
issue: namely, the fact that the other negotiation deadlocked. ``Those
guys didn't work it out, what are we going to do about it?''
\item
An {\bf appeal}. Some of the representatives might be unhappy about the
outcome and appeal to other parts of the organization---which might set
up another negotiation to deal with the issue of what to do about the
outcome of the previous negotiation.
\end{itemize}
Participants can be stalemated in conflict for a long period of time. For
example, an environmental group can work for decades attempting to revoke
the operating license of a nuclear power plant. Maintaining a
negotiation does consume resources, however. One has to keep track of
the other participant's position, plan strategy of how to continue carrying out
the negotiation, respond to the other participant's moves all the time, and so
on. Actions like these consume time, communications, storage space, and
other resources that would otherwise be put to different use. Thus,
maintaining a negotiation is a commitment.
In some cases, a negotiation ends when one participant runs out of the resources
needed to continue. In other cases, the process explicitly provides a back-up
procedure in the sense that it leaves the conflict potentially resumable, but
ends the current negotiation. An example of this would be a state that is
determined to oppose a public utility in its attempt to operate a nuclear
powerplant. The state can oppose the utility at every stage of its attempt to
get an operating license. Suppose that at each stage the negotiation is
broken off when the Nuclear Regulatory Commission decides in favor of the
utility. After each hearing the resources committed to the negotiation are
allowed to go their own way, with the intention that a whole new negotiation
might be convened at another time. Finally just before the nuclear power
plant goes into operation, the state might ``win'' the negotiation by offering
to compensate the utility for what it would earn by operating the nuclear plant
{\em provided} that it agrees to sell the plant to the state for \$1!
\subsection{Negotiations Create Commitments}
Negotiations are important, even if no conflict emerges, because they
create organizational commitments that go beyond the individual
commitments of the participants involved. Negotiations always have multiple
possible outcomes. Choices are made during a negotiation, which may
result in the creation of an organizational commitment: the participants agree
on a particular course of joint action. The negotiation might prove to
be a trivial one in which agreement is easily reached, but the outcome
still represents an organizational commitment. Late-arriving information
could have caused one of the participants to strive for a different outcome.
The significance of negotiations lies in their outcomes and the way those
outcomes affect other organizational actions. The country will
incrementally develop an electric power industry---and that industry will
influence energy costs, pollution levels, generating capacity, etc. For
example, Engineering and Finance can have a dispute as to whether to build
a plant with two reactors at once, as opposed to finishing one before
starting the next. Engineering prefers building both at once because it
can overlap similar activities to bring down the cost. Finance prefers
building them sequentially because the financial burden and risk is less.
The dispute between Finance and Engineering will have an outcome in terms
of the utility's profitability.
\subsection{Negotiation is Creative}
Negotiation is intrinsically creative. Often, the outcome is not as
predicted, or is unintended by participants, or may even be unwanted by
some participants. On the other hand, an outcome may turn out to be
better than expected. Even when a negotiation does not break new ground
and the outcome is one of those initially sought by one or more participants,
the process used to reach that outcome is fundamentally creative in the
sense that it creates an organizational commitment.
As we have seen, trials of strength embody conflict (because of
incompatible outcomes), and therefore indeterminacy (because no participant can
be certain what the outcome will be). Trials of strength are the
fundamental unit of activity that we want to understand and explore. The
actual unfolding of a trial of strength is a unique performance, so
strictly speaking, a trial of strength can never be repeated. A similar
one could be staged at a different place and time, but each performance
is unique.
This cycle of commitments leading to negotiations which lead to
commitments, some of which conflict with other commitments and thus lead
to further negotiations---this cycle is the way the world works.
\subsection{The Rationale}
The {\em rationale\/} for the outcome of a negotiation is stated at the
end of the negotiation. The rationale(s) given for the outcome are
partly generated during the negotiation process as the participants discuss
the proposed options. As each participant challenges each other's positions,
new beliefs and preferences are created. As the negotiation continues, a
rationale is often created in support of a particular outcome. For
example, in a conflict between Finance and Engineering about which of two
types of plant to build, the rationales supporting the outcome may
describe:
\begin{itemize}
\item
{\bf Predicted beneficial results:} A utility justifies the development
of a new plant: ``Nuclear power will cost less than burning fossil
fuel.''
\item
{\bf Policies guiding conduct:} The management of a utility makes a
policy: ``We must follow the rules and regulations of the Nuclear
Regulatory Commission in building our plant.''
\item
{\bf Reasons tied to specific institutional roles or processes:} A
utility sells a completed, ready to go, nuclear power plant to the state
government (which plans to demolish it) with the justification that the
state has agreed to compensate the utility in other ways.
\item
{\bf Precedent:} It is traditional to run diagnostics for the nuclear
reactor on Monday morning.
\end{itemize}
Precedent may seem like a weak rationale. However, deciding according to
precedent in the absence of strong alternatives has the consequences of
predictability, stability, and increased participant value in organizational work.
In the absence of strong alternatives, using precedent is usually less costly
than constantly redoing a decision process.
The rationale becomes part of the organizational history, and may become
a precedent.
This taxonomy not only describes characteristics of outcome rationales,
it also provides criteria for identifying problems and pointing out ways
in which the process can be made more effective. So the rationale is
much more than the big cheese standing up and saying, ``This course of
action will lead to wonderful results.'' Any rationale can claim
beneficial results. However, the rationale will be judged on its own
merits. Good decisions can have bad rationales, and vice versa.
∂02-Dec-89 1528 kirsh%cogsci@ucsd.edu hewitt part2
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 2 Dec 89 15:27:58 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA22929
sendmail 5.61/UCSD-2.0-sun via SMTP
Sat, 2 Dec 89 15:28:04 -0800 for jmc@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA02944 for jmc@sail.stanford.edu; Sat, 2 Dec 89 15:26:37 PST
Date: Sat, 2 Dec 89 15:30 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: hewitt part2
To: jmc@sail.stanford.edu
Message-Id: <19891202233036.9.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
\subsection{Assessment of Participant Values}
No one can stand outside the system and assess participant values (i.e.
benefits relative to costs) in organizational work. Anyone who wants to
increase participant value must become a player and participate in
organizational processes. All such assessments are made within a
framework of conflict: allies, commitments, incompleteness of
information, limitations of resources, etc. The {\em only way\/} to
increase participant value in organizational work is to become part of the
organizational processes.
For example, the Nuclear Regulatory Commission abolished the requirement
that localities and states must approve emergency evacuation plans before
a nuclear reactor will be granted an operating license, in large part
because of local officials who were refusing to approve the plans. The
commission felt that communities were using the evacuation plant process
to prevent nuclear plants from receiving operating licenses. After the
NRC announced that it would abolish the requirement, many participants
challenged its authority in court. They criticized the commission for
changing the rules in the middle of the licensing process.
Meta-commitments in this instance set new policies and procedures (i.e.,
new commitments) about how commitments get changed. These new
commitments address concerns about how the previous licensing procedure
was carried out (i.e, that the communities had veto power over the
emergency evacuation plans). As a result of this trial of strength, the
commission created new procedures and policies for changing its
regulations so that in the future, various participants will participate in a
better-defined process. The new procedures and policies arose from a
meta-commitment negotiation, and formed a commitment about how to change
other commitments.
All meta-commitment negotiations are concerned with how organizational
processes are changed. Their outcomes affect how subsequent negotiations
are conducted, and thus affect the outcomes (commitments) of subsequent
negotiations.
\subsection{Authority and Responsibility}
The meta-commitment described above changed the Nuclear Regulatory
Commission's authority and responsibilities. The utilities gained power
to participate in decisions about emergency evacuation plans, and the
commission took responsibility for spelling out its processes for
changing the licensing process.
Authority is power, and power is the ability to take action (i.e., use
resources). More precisely, authority is power legitimized by the
commitments of other authorities. For example, a utility has to register
with the Secretary of State in whichever state it operates, so its
organizational power is legitimized by the power of another authority,
namely the Secretary of State. If that authority withdraws this
legitimization, the utility's authority becomes problematical.
An organization's power is its control over resources, and its
responsibilities are its part in its commitments. (Accountability is whether or not it
actually takes those actions and meets those commitments.) So authority
and responsibility are both intimately tied to an organization's
commitments.
Authority can be delegated, but responsibility cannot. Responsibility is
established by the organization's undertaking a certain set of
commitments. An actor might get some help from other participants in meeting
those commitments, but they are still the actor's own responsibility. So
in the narrow sense, the actor cannot delegate responsibility. What it
can do instead is create other organizational arrangements that also
carry the same commitment---and hope that will be sufficient. But if
it's not, that commitment is still that actor's responsibility.
\section{Relationship to Artificial Intelligence}
The question now arises as to the relationship between Open Systems
and Artificial Intelligence. Various technologies have been developed in
Artificial Intelligence for providing a foundation for and structuring of
computation, including:
\begin{itemize}
\item microtheories
\item problem spaces
\item taxonomies
\end{itemize}
\subsection{Microtheories}
One of the most powerful ideas of science is the idea of a {\em
microtheory}. Microtheories are based on a closed-world
assumption---from a set of rules specified in advance, results can be
algorithmically checked for correctness. A spread-sheet is a good
example. The rules are the calculation procedures used in the body of
the spread-sheet. Given the previous values and formulas, an
automaton can algorithmically check whether the new values are correct
in real time. Our notion of a microtheory is very general in that
encompasses all known forms of deduction including first order logic,
nth order logics, modal logics, intuitionistic logic, relevance logic,
lambda calculi, circumscription, default logics, etc [de Kleer, Kripke,
McDermott, McCarthy, Nilsson, Reiter, Sandewall, etc.].
A microtheory has important strengths:
\begin{itemize}
\item
It is portable. A microtheories can be expressed as a stable inscription
that does not spontaneously change and can easily be moved and copied.
\item
The correctness of a derivation is algorithmically decidable solely from
the text of the derivation. Something as simple as an automaton can
decide in real time whether or not a derivation is correct.
\end{itemize}
\noindent
Within a microtheory, there are well-defined methods for dealing with any
conflict that might arise. Thus, negotiations are not very important
within a microtheory because the correctness of derivation can be algorithmically
decidable in a closed world.
Microtheories play an important role in negotiations because they can be
brought to bear on issues and provide support for commitments with
respect to those issues. For example, the utility's Finance and
Engineering departments might each have a different spread-sheet
model of the utility's financial condition with respect to the costs of a
new plant, and each representative can then bring that microtheory to
the negotiation. Their respective recommendations of how the utility
should spend its money might be contradictory. Comparing their
microtheories can help to determine what some of the underlying
conflicts are. They might discover that Finance's Comptroller does not
believe that Engineering can meet its construction schedule. Derivations
of a microtheory can be brought to bear as supporting arguments in the
negotiation. In general, however, there will be a lot of these
microtheories, and they will often have derivations that
formally contradict derivations.
Each microtheory compiles certain methods in a rigorous way. The
Comptroller tries to protect the utility against financial
difficulties---and has successfully negotiated a commitment from
Engineering that the utility will not borrow money if payments will
exceed 25 percent of its income. On the other hand, the Engineering
department maintains that more generating capacity is needed and that
with their construction schedule, the debt payment will never exceed 25
percent.
Many of the microtheories embody various commitments and allies. They
may have been endorsed by the organization itself, or by an outside
organization that has expertise in the field. A spread-sheet microtheory
derived from the tax code can be used to deduce the tax consequences of
differing proposals, and the participant holding this microtheory can claim the
IRS as an ally (i.e., claiming that the IRS will support the conclusions
drawn).
Thus, each side of any conflict (such as whether to pursue the
construction schedule developed by Engineering) will be able to marshall
its own body of microtheories, principles, precedents, and conclusions.
Having a deductive proof based on a microtheory usually does not thereby
carry the day and win the negotiation. The other participant will usually have
a competing microtheory. So microtheories are a strength, but they're
just one tool of negotiation. A powerful tool, but just one tool.
Having a microtheory facilitates negotiation, but in general does not
determine the outcome.
\subsection{The Role of Logical Deduction}
Several different negotiating strategies can be used with microtheories.
One is to include all the {\em if}s, {\em and}s, {\em but}s and {em
wherefore}s that one can imagine. This creates a cumbersome
microtheory that attempts to cover all possible special circumstances.
For example, ``If the utility uses more than 25 percent of its income for
debt payment, {\em and if furthermore\/} it does not have lots of liquid
assets, {\em and if furthermore\/} \ldots\ then the utility should not
take on more debt.''
A different strategy is to state a very simple rule and let it unfold in
the ongoing negotiation whether any exceptions apply. So the Comptroller
says: ``If construction is delayed, then the utility will spend much more
than 25 percent of its income for debt, so we shouldn't adopt the
construction plan.'' And the other participant replies, ``Yes,
but---Engineering has a good record for completing construction projects
on schedule at close to its estimated cost. Even though Engineering is
building a new kind of plant which can burn either coal or gas, it is not
very different from what it has built before.'' Having simple
microtheories that are parsimonious, easily understood, and clear in
their causality is often a better negotiating strategy than one which
tries to stipulate in advance all of the conditions which govern the
applicability of every rule.
The participants to a negotiation do not know for certain what sorts of
rationale for action and microtheories the others will present. The
applicability of one participant's microtheories depends on what happens during
the negotiation, not on the ability to assemble a large collection of
microtheories ahead of time. Either participant might come up with a
microtheory that the other has not thought of.
The Open Systems model of representations can be used to analyze a
deductive approach that has been explored in [McCarthy, etc.]
By inserting caveats into the axioms of conflicting microtheories,
interactions among the conditions of applicability of the axioms of the
microtheories can be expressed.
For example, various factors bear on the safety of the Diablo Canyon
nuclear reactor. Suppose that we attempt to conduct the negotiation by
writing rules with explicit caveats. Consider the following two rules:
\procedure %
{\rm Rule 1:}
if trained-operators and not(caveat-1), then safe-reactor
\endproc
\noindent
Having trained operators makes for a safe reactor unless it can be shown
that caveat-1 is true. Continuing to axiomatize,
\procedure %
{\rm Rule 2:}
if earthquake-zone and not(caveat-2), then not(safe-reactor)
\endproc
\noindent
Being in an earthquake zone means the reactor is not safe unless it can
be shown that caveat-2 is true.
In this way axioms for caveats can be developed over time
and gradually improved. For example, consider the following rule:
\procedure %
{\rm Interaction Rule 1:}
if trained-operators, then caveat-2
\endproc
\noindent
Having trained operators implies that being in an earthquake zone does
not necessarily imply that the reactor is unsafe.
\procedure %
{\rm Interaction Rule 2:}
if earthquake-zone, then caveat-1
\endproc
\noindent
Also, being in an earthquake zone implies that having trained operators
does not necessarily imply that the reactor is safe.
Unfortunately, the addition of Interaction Rules 1 and 2 blocks the
applicability of Rules 1 and 2. If we have both trained operators and
an earthquake zone, Rule 1 cannot be used since earthquake-zone implies
caveat-1, and Rule 2 cannot be used since trained-operators implies
caveat-2. The following axioms are needed instead:
\procedure %
{\rm Interaction Rule $1'$:}
if trained-operators and not(caveat-Interaction-Rule-1),
then caveat-2
\bline
{\rm{Interaction Rule $2'$:}}
if earthquake-zone and not(caveat-Interaction-Rule-2),
then caveat-1
\endproc
These interaction rules are needed to prevent contradiction and they can
be highly non-modular. For example, the use of interaction rules raises
the following question: Does being in an earthquake zone imply that the
presence of trained operators implies that being in an earthquake zone
implies that the reactor is safe? Questions like this can also be
expressed as rules:
\procedure %
{\rm{Second-Order Interaction Rule 1:}}
if trained-operators, then caveat-Interaction-Rule-2
\bline
{\rm{Second-Order Interaction Rule 2:}}
if earthquake-zone, then caveat-Interaction-Rule-1
\endproc
\noindent
Again, the Second-Order Interaction Rules 1 and 2 cannot be allowed to
stand. Instead, the following rules must be used:
\procedure %
{\rm{Second-Order Interaction Rule $1'$:}}
if trained-operator
and not(caveat-Second-Order-Interaction-Rule-1),
then caveat-Interaction-Rule-2
\bline
{\rm{Second-Order Interaction Rule $2'$:}}
if earthquake-zone
and not(caveat-Second-Order-Interaction-Rule-2),
then caveat-Interaction-Rule-1
\endproc
\noindent
At this point, it becomes very difficult to understand what we are talking
about.
Combining microtheories and adding caveats to rules does not resolve
all conflicts. In the power plant construction example, joining two
microtheories together with caveats leads to the derivation of: ``Yes,
the utility will get an operating license or no, the utility will not get an
operating license.'' But the utility cannot be told ``yes or no.'' Either
they must be told ``yes,'' or they must be told ``no.'' Adding caveats to
rules makes for a cumbersome negotiating strategy that does not
respond easily to changing circumstances.
Logical deduction can model the reasoning within a given microtheory, but
it cannot settle the dispute. Attempts to combine microtheories into a
larger theory by introducing caveats ultimately leads to ``yes or no''
derivations. So logical deduction is an appropriate and valuable tool
within the microtheories held by the respective participants. Quite often,
new microtheories will need to be created in order to better understand
the issues under negotiation. In general, these new microtheories will
not be logical consequences of the microtheories that were already
familiar to the participants before the negotiation began.
\subsection{Problem Spaces}
Problem spaces \cite{ProblemSpaces} can be used as a process modeling
technique. A problem space is a microtheory that provides for:
\begin{itemize}
\item
an {\bf initial state}. For example, the initial state may be the
financial state of the utility before it begins constructing a new plant
\item
one or more {\bf operators} that are applicable to each state. For
example, selling bonds is one of the operations that a utility can
usually take to change its financial condition.
\item
one or more {\bf success states} for one or more of the participants. For
example, the utility can specify its financial goals in terms of revenue,
investment, and earnings.
\end{itemize}
Problem spaces can also be used to model an ongoing negotiation.
Negotiation usually begins with discussion about issues. This process
can be viewed as negotiation about the initial state. Then there is
discussion about how the negotiation should proceed: who will speak, what
the agenda is, and so on---which is analogous to the possible operations
applicable to each state. Also, there is discussion about what
represents a successful outcome. A problem space attempts to model the initial
situation, trajectory, and criteria for when negotiation has ended.
Thus, problem spaces can be a useful way of characterizing an ongoing
process. However, we can rapidly encounter the same difficulty we had
with microtheories: each participant to the negotiation will have its own
problem space of how the negotiation should proceed which will in general
conflict with those of the other participants.
Consider for example the negotiating process to determine whether or
not to give a nuclear plant an operating license. The initial state of the
negotiations is quite problematical. Typically there are thousands of
pages of documentation and claims that have been submitted ahead of
time by the plant owners, the nearby local communities and states, by
environmental groups, and public utility commissions. One of the
participants might say, ``Okay, we are now in this particular initial
state; these are the ways that the negotiation can move forward from
the various states it might get into; and this is what we'll count as an
outcome.'' However, in this case the participants find that they agree
on very little about the starting state of the negotiations. In many
cases the problem spaces are not worked out in such an overt form.
Participants often come to a negotiation with {\em criteria for the
outcome\/} that they initially believe would represent success for
themselves (and possibly for others as well). Also, they would come
with their own understanding of the initial situation and what
negotiation moves would be legitimate. Furthermore in this case the
Nuclear Regulatory Commission decided to change the rules for dealing
with emergency evacuation plans in the middle of the negotiation
process; which neither it nor the other participants initially conceived
as a possible operation. The participants in this case differ greatly on
their characterization of the starting state, the allowable operations, or
which states constitute successful outcomes.
In order to use a problem space for conducting a negotiation, the participants
must first reach an agreement about the initial state of the negotiation. In
general it is difficult to come to an agreement as to exactly what are the
issues at stake. The representatives find it difficult to specify how the
organization they represent would characterize an issue. After some
representatives have stated their understanding of an issue, others may decide
to change their characterization. Thus attempting to precisely characterize
the initial state of a negotiation can be a very problematical undertaking
that has no termination and may even deadlock. To make the situation even
more problematical, a prolonged attempt to precisely specify the initial
state has the effect of itself changing the initial state! Furthermore some
of the representatives may not want to reveal all of their current plans and
understanding concerning of the issues under negotiation for a variety of
reasons--such as being in too preliminary a state of development to share
with others. Specifying in advance the operators that are applicable at each
stage of a negotiation as well as the success states are equally problematical.
In many cases, the problem spaces are not worked out in such an overt
form. Participants often come to a negotiation with {\em criteria for
the outcome\/} that they initially believe would represent success for
themselves (and possibly for others as well). Also, they would come with
their own understanding of the initial situation and what negotiation
moves would be legitimate. In the case of the emergency evacuation
plans, however, the Nuclear Regulatory Commission decided to change the
rules for dealing with such plans in the middle of the negotiation
process; neither the NRC nor the other participants initially conceived of
this as a possible operation.
Again, having a problem space does not guarantee what's going to happen
because each participant brings its own problem space. Participants with
conflicting commitments about a negotiation will have conflicting
problem spaces. Basically, these are conflicts over commitments on how
the negotiation should proceed. The conflicts between these problem
spaces---much as in the cases of microtheories---need to be deal with.
Given the prospects for conflict among problem spaces, it would be
difficult (and perhaps not even desirable) to try to create a problem
space in advance that governs represent the entire, ongoing negotiation.
Instead, problem spaces are probably best used by each participant to
communicate its own analysis of the current negotiating situation,
available options, and desired outcomes.
\section{Conclusions}
The new discipline of Open Systems Technoscience differs from Artificial
Intelligence in several crucial aspects:
\begin{itemize}
\item
In Open Systems Technoscience, the primary indicators of success are {\em
participant values} (which is roughly the benefits to a participant
relative to its costs (including externalities)), {\em robustness} (which is
the ability to keep commitments in the face of conflict), and {\em
scalability} (which is participant value in activity
increase some participant values). Open Systems Technoscience
is grounded in large scale human/telecomputer organizational work. The
primary indicator of success in Artificial Intelligence is the ability to
impress humans with behavior that they will call intelligent. It is
grounded in intelligent agents and robots. In contrast to Artificial
Intelligence, work can proceed on the development of foundations for
Open Systems Technoscience without the need to provide a characterization of
``intelligence''.
\item
In Open Systems Technoscience, {\em representation} is the activity of
communicating with others. Without communication there is no
representation. Communication takes its {\em meaning} from how it affects
the behavior of recipients. In Artificial Intelligence, representation is
traditionally about the correspondence between a structure in an intelligent
agent and a state of affairs in the world.
\item
Open Systems Technoscience views {\em commitment} as a {\em joint course
of action} in which the participants are {\em on course}. The {\em
responsibility} of a participant is its part in the commitment. Artificial
Intelligence has traditionally viewed commitment as a state of mind in
which there is {\em intentionality}
\cite{Cohen-Levesque}\cite{Dennet}.
\end{itemize}
In summary social processes especially those of science, technology, and
engineering \cite{Latour} inform Open Systems Technoscience, whereas Artificial
Intelligence has traditionally turned to neurophysiology, psychology, and
cognitive science.
Open Systems Technoscience provides a framework for analyzing
Artificial Intelligence technologies such as deductive theories,
taxonomies, and dictionaries. Conflict is ubiquitous in Open Systems.
It allows participants to consider and explore their alternatives in a
way that takes other commitments into account. As the participants to
the conflict negotiate their differences, they usually generate
justifications to support their position. They often use microtheories to
bolster their cases. Since microtheories are decontextualized, they can
be carried from place to place and used to seek additional leverage in
many different negotiations. Thus, the use of inference in microtheories
can be seen as a natural kind of specialized activity that often occurs in
the negotiation of conflict. The crucial characteristics of a microtheory
are that the rules are given in advance and that the derivations can be
checked for correctness in real time. The nonmathematical
microtheories of the participants usually conflict. Negotiation of
the conflict that arises can be a source of creativity and robustness.
% \cite{open-systems}
% \cite{robustness-reliability-and-overdetermination}
% \cite{regions-of-the-mind}
\section{Acknowledgments}
First, I would like to express my gratitude and admiration to Wyn Snow
for editorial assistance above and beyond the call of ordinary duty.
Without her help, this paper would contain unboundedly many more
mindtraps.
Second, I wish to acknowledge the aid of Elihu Gerson in repeatedly
pulling me out of intellectual quicksand and setting me back on fruitful
paths.
Third, I wish to acknowledge the help of Les Gasser, David Kirsh, Bruno
Latour, John McCarthy, and Susan Leigh Star for pushing forward in new
directions as well as helping to reconceptualize old ones.
Fourth, I wish to thank members of the Message Passing Semantics
Group for helping to find obscurities and errors.
Fifth (and perhaps most important), I wish to thank Randy Fenstermacher,
Ron Flemming, Sue Gerson, Fanya Montalvo, John Stutz, and other close friends for
helping me to continue to grow.
\section{Related Work}
\vbox{\noindent
Hewitt, Carl,
``Viewing Control Structures as Patterns of Passing Messages,''
{\em A.I.\ Journal}, Vol.~8, No.~3, June 1977, pp.~323--364.}
This paper re-examined the issue of control structures in Artificial
Intelligence. Control structures were previously defined as looking for
the best choice in moving from the current global state to the next one.
The control structure was supposed to accomplish this either by guiding
the production system or by guiding a theorem-prover that was attempting
to search through the realm of possibilities. This paper pointed out that
traditional programming language control structures (such as iteration,
recursion and co-routine) could be analyzed in terms of patterns:
stereotypical or stylized patterns of communication among different
participants. Instead of looking at the behavior of an {\em individual\/}
intelligent agent as Newell and Simon did, this paper initiated the idea
that communities of people are a primary existence proof and analog for
how to extend these ideas.
\bigskip\vbox{\noindent
Kornfeld, William A.\ and Carl Hewitt, ``The Scientific Community
Metaphor,'' {\em IEEE Transactions on Systems, Man, and Cybernetics},
Vol.~SMC-11, No.~1, January 1981, pp.~24--33.}
This paper introduced several important concepts into the Artificial
Intelligence arena, and further develops the ideas Hewitt first discussed
in ``Viewing Control Structures.'' It uses the scientific community as a
model for the problem-solving process, and speaks generally about how
principles and mechanisms of scientific communities might be incorporated
into the problem-solving technology of Artificial Intelligence. Several
fundamental properties of scientific communities have nice analogs for
computing systems that aspire to intelligent behavior. Among these
properties are monotonicity, commutativity, parallelism and pluralism.
The paper also introduces the notion of having sceptics as well as
proponents of different kinds of ideas, and explicates how those kinds of
questions can be investigated concurrently.
This work was somewhat preliminary in nature and is still in need of
further development. Subsequent work in the Message-Passing Semantics
Group at MIT has focused on organizational processes and structures.
\bigskip\vbox{\noindent
Kornfeld, William Arthur,
``Concepts in Parallel Problem Solving,''
Ph.D.\ Thesis, Dept.\ of EECS, MIT, February 1982.}
This is a further development of the work in ``The Scientific Community
Metaphor.'' Kornfeld here shows that by developing a concurrent process
that has critics as well as proponents of ideas, the amount of resources
consumed can, in some cases, be vastly reduced. This results in a kind of
combinatorial implosion instead of the usual combinatorial explosion where
the number of alternatives proliferate indefinitely. Such exponential
proliferation of possibilities is typical of backward-chaining reasoning.
The negotiation described here is very primitive in form, and consists of
entering absolute objections---a very cut-and-dried situation. We would
like to apply this type of process in more relaxed situations where one
has less hard knowledge, and the objections aren't guaranteed to be always
fatal to what they're objecting to.
\bigskip\vbox{\noindent
Barber, Gerald Ram{\'o}n,
``Office Semantics,''
Ph.D.\ thesis, Dept.\ of EECS, MIT, February 1982.}
This paper shows how the viewpoint mechanism introduced in Kornfeld and
Hewitt's ``Scientific Community Metaphor'' can be used to model changing
situations in terms of multiple points of view. It also introduces some
of the kinds of mechanisms for dealing with contradictory microtheories.
\bigskip\vbox{\noindent
Huberman, B.~A. editor,
{\em The Ecology of Computation}
North Holland, 1988}
This book is an excellent collection of articles which deal with the nature,
design, description, implementation, and management of Open Systems. The
articles are grouped in three major sections. Papers in the first section
deal with general issues underlying Open Systems, studies of computational
ecologies, and their similarities with social organizations. Papers in the
second section deal with implementation issues of distributed computation, and
those in the third section discuss the issues of developing suitable languages
and information media for Open Systems.
\bigskip\vbox{\noindent
Stefik, Mark,
``The Next Knowledge Medium,''
{\em The AI Magazine}, Vol.~7, No.~1, Spring 1986, pp.~34--46.}
Stefik describes the growth and spread of cultural knowledge: the kinds of
things that communities of humans do---and shows how the existence of a
technical infrastructure (such as railroads) can greatly facilitate and
accelerate cultural change. Our current knowledge market is static and
pretty much confined to inscriptions: things that can be reduced to a
string of bits (such as a diagram or sentence or literary work) and thus
transported and copied at very small price. Stefik portrays a dynamic
knowledge market that would supplement our current product market. It
would move intelligent models around that have the capability of taking
action. An active knowledge medium could interact with both its human
users and with various kinds of expert systems. He also describes several
current projects, such as the Co-Lab at Xerox Park, that are beginning to
show rudimentary characteristics of an active knowledge medium.
\bigskip\vbox{\noindent
Alvarado, Sergio J., Michael G.~Dyer and Margot Flowers,
``Editorial Comprehension in OpEd Through Argument Units,''
Fifth National Conference on Artificial Intelligence,
Aug 11--15, 1986, Philadelphia, PA,
{\em AAAI-86}, Vol.~1, pp.\ 250--256.}
This paper shows how arguments can be diagrammed in much the same way that
debate contests are often diagrammed by their judges. Such diagramming
examines the beliefs, the tree-structure of the supporting beliefs, and
the way one side can attack the other side's beliefs. (There are really
two kinds of important relationships between the two sides: support
relationships and attack relationships.) The paper presents an analysis
that looks at both the achievement of plans and goals, and the development
of editorials that critique other sides, showing how other sides have
beliefs that are supporting to the opinion that's being reported. This is
quite interesting work in terms of starting to build technology that can
do argument analysis, because that's an important component of
negotiation. Of course, there are other kinds of representations---as John
McCarthy pointed out---in terms of making threats and other kinds of
speech acts, but argument analysis is certainly a very important component
of negotiation.
\bigskip\vbox{\noindent
Devereux, Erik August,
``Processing Political Debate: A Methodology for Data Production
with Special Application to the Lincoln-Douglas Debates,''
B.S.\ thesis, MIT Dept.\ of Political Science, June 1985.}
Devereux expands on something very similar to the argument units in
Alvarado {\em et. al.}. Devereux takes the whole of the Lincoln-Douglas
debates and attempts to identify both attacking statements between Douglas
and Lincoln and supporting links within the individual arguments
themselves. Interestingly enough, there are no supporting links between
the two debaters, so in that respect, the argument units of Alvarado {\em
et al}.\ represent an advance over the analysis that was done by Devereux.
\begin{thebibliography}{9999}
\bibitem[Agha 1986]{agha-phd}
Agha, G., {\em Actors: A Model of Concurrent Computation in Distributed
Systems}, Cambridge, MA: MIT Press, 1986.
\bibitem[Becker 1960]{concept-of-commitment}
Becker, Howard S., ``Notes on the Concept of Commitment,'' {\em American
Journal of Sociology}, Vol.~66, July 1960, pp.~32--40.
\bibitem[Fikes 1982]{commitment-based-framework}
Fikes, Richard E., ``A commitment-based framework for describing informal
cooperative work,'' {\em Cognitive Science}, Vol.~6, 1982, pp.~331--347.
\bibitem[Gerson 1976]{quality-of-life}
Gerson, Elihu M., ``On the Quality of Life,'' {\em American Sociological
Review}, Vol.~41, October 1976, pp.~793--806.
\bibitem[Latour 1987]{science-in-action}
Latour, Bruno, {\em Science In Action}, Cambridge, MA: Harvard University
Press, 1987.
\bibitem[Smith and Davis 1981]{contract-nets}
Smith, R. and Davis, R., ``Frameworks for cooperation in distributed
problem solving,'' {\em IEEE Transactions on Systems, Man, and
Cybernetics}, Vol.~SMC-11, 1981, pp.~61--70.
\bibitem[Star 1983]{simplification-in-scientific-work}
Star, S.L., ``Simplification in scientific work: An example from
neuroscience research,'' {\em Social Studies of Science}, Vol.~13, No.~2,
1983, pp.~205--228.
\bibitem[Strauss 1978]{negotiations}
Strauss, Anselm, {\em Negotiations}, Jossey-Bass, 1978.
\bibitem[Winograd and Flores 1987]{winograd-flores}
Winograd, Terry, and Flores, Fernando, {\em Understanding Computers and
Cognition}, Reading, MA: Addison-Wesley, 1987.
\end{thebibliography}
\end{document}
∂03-Dec-89 1057 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
A NEW APPROACH TO MODAL OPERATORS
Matt Ginsberg
Stanford University
Monday, December 4, 3:15pm
MJH 252
We describe a new formalization of modal operators that views them
not in terms of Kripke's possible worlds, but as functions on an
underlying set of truth values. Thus Moore's knowledge operator L,
where Lp means "I know that p," would correspond to a mapping taking
true into true (since we know p if we know it to be true) and taking
both false and unknown into false (since we do not know p if we
either know it to be false or know nothing about it at all). This
new approach has the following advantages over the conventional ones:
1. Intuitive simplicity
2. It provably generalizes both Kripke's construction and Moore's
autoepistemic logic, while making clear the distinctions between
them.
3. It allows for easy further generalization to modal operators
that are related to temporal reasoning and to causality.
4. The natural procedure for computing the truth value of a sentence
involving these modal operators is "incremental" in the sense that
it computes approximate answers that "converge" to the correct one
in the large runtime limit.
This talk will concentrate on the first two of these properties;
I will discuss the third and fourth as completely as time allows.
∂03-Dec-89 2355 mdbomber@Portia.stanford.edu re: Bozo Bush vs. Chinese students [was Re: Recent su.etc Statistics]
Received: from Portia.stanford.edu by SAIL.Stanford.EDU with TCP; 3 Dec 89 23:55:21 PST
Received: by Portia.stanford.edu (5.59/25-eef) id AA26054; Sun, 3 Dec 89 23:52:08 PDT
Date: Sun, 3 Dec 89 23:52:08 PDT
From: Matt Bartley <mdbomber@Portia.stanford.edu>
Message-Id: <8912040752.AA26054@Portia.stanford.edu>
To: JMC@SAIL.Stanford.EDU
Subject: re: Bozo Bush vs. Chinese students [was Re: Recent su.etc Statistics]
Well, our side eventually won, so I guess we did something right!
MB
∂04-Dec-89 1000 JMC
Heffalump
∂04-Dec-89 1138 scherlis@vax.darpa.mil NSF-DARPA Formal Methods
Received: from vax.darpa.mil (darpa.mil) by SAIL.Stanford.EDU with TCP; 4 Dec 89 11:38:43 PST
Received: from sun45.darpa.mil by vax.darpa.mil (5.61/5.61+local)
id <AA01634>; Mon, 4 Dec 89 14:06:34 -0500
Posted-Date: Mon 4 Dec 89 14:05:34-EST
Received: by sun45.darpa.mil (4.1/5.51)
id AA03580; Mon, 4 Dec 89 14:05:35 EST
Date: Mon 4 Dec 89 14:05:34-EST
From: William L. Scherlis <SCHERLIS@DARPA.MIL>
Subject: NSF-DARPA Formal Methods
To: SW-PI@vax.darpa.mil
Cc: boesch@vax.darpa.mil, mettala@vax.darpa.mil, squires@vax.darpa.mil,
boehm@vax.darpa.mil, ktai@note.nsf.gov
Message-Id: <628801534.0.SCHERLIS@VAX.DARPA.MIL>
Mail-System-Version: <SUN-MM(217)+TOPSLIB(128)@VAX.DARPA.MIL>
To the Software PIs:
We expect a formal announcement of the Joint NSF-DARPA Initiative on
Formal Methods in Software Engineering to occur shortly. Below is a
statement of intent. Please note the following: (1) Proposals
submitted under the joint initiative will receive NSF-style scientific
peer review as part of the selection process. (2) Proposals can be
submitted both to our BAAs and to the NSF initiative, but they MUST be
so marked. The text below is not a formal announcement, so you should
delay sending proposals to NSF until you have received and read the
full announcment. (3) We will email the full announcement to this
list as soon as it is available. The full announcement will also be
distributed using the usual NSF channels. Thanks,
Bill
p.s. This is in addition to the currently active software BAA, which
was published in the Commerce Business Daily edition of 20 Oct 89.
================
December 4, 1989
Joint NSF-DARPA Initiative on
FORMAL METHODS IN SOFTWARE ENGINEERING
It is anticipated that beginning in fiscal year 1990, NSF and DARPA
will jointly support research projects related to formal methods
technology and software engineering. A brief description of this
joint initiative is given below. Copies of the formal announcement
will be mailed to all Computer Science Departments in the USA in late
December 1989.
The specific objectives of the joint initiative are:
- Strengthening the scientific basis for formal methods applied to
the specification, analysis, verification, testing, manipulation,
and adaptation of programs.
- Improving the engineering basis for experimentation with formal
methods techniques, with an emphasis on interface design and the
development of generic tools.
- Obtaining early empirical evaluation of formal methods
techniques.
- Promoting collaborative efforts between the research and
development communities.
NSF and DARPA expect to award a total of about $1.0 million during
this fiscal year. Proposals should be submitted to NSF following
regular NSF procedure. The following technical issues should be
explicitly addressed in the research narrative in proposals:
scalability, transition potential, and suitability for empirical
evaluation. Team efforts that include both academic and industrial
participants are encouraged. Awards to successful proposals will be
made through NSF with funding from both NSF and DARPA.
For technical information, prospective PI's may contact either NSF or
DARPA program offices:
Dr. William L. Scherlis Dr. K. C. Tai
Pgm Mgr, Software Technology Pgm Dir, Software Engineering
DARPA/ISTO National Science Foundation
(202) 694-5800 (202) 357-7345
scherlis@vax.darpa.mil ktai@note.nsf.gov
The deadline for submitting proposals is February 15, 1990 with awards
to be made by June 1, 1990.
================
-------
∂04-Dec-89 1820 VAL Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
IMPLEMENTING AUTOEPISTEMIC LOGIC ON A REASON MAINTENANCE SYSTEM
Kurt Konolige
SRI International
Monday, December 11, 3:15pm
MJH 252
Recent work shows that a Reason Maintenance System (RMS) can be
formalized as a type of autoepistemic theory. In this paper we
consider the inverse transformation: trying to implement an arbitrary
autoepistemic theory as an RMS. In so doing, we provide a
computationally attractive theorem-proving methodology for
autoepistemic logic.
∂04-Dec-89 1904 RWF re: World War II
[In reply to message rcvd 04-Dec-89 18:33-PT.]
If we take Lindbergh at his word, it isn't that he thought
Hitler was not so bad, but that he thought the European
civilization was doomed in Europe, and could be preserved
and defended in the western hemisphere. I recall an
estimate that the US was very easy to defend and hard to
attack by air. Browse a bit in the Lindbergh corner of
the library and tell me if you think he had a bad rap.
You might mention that the isolationist tradition is
much of it populist, with as many left as right wing
attitudes.
∂04-Dec-89 2211 ACT Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 4 December 1989
Previous Balance 0.30
Monthly Interest at 1.0% 0
Current Charges 0.30 (vending machine)
-------
TOTAL AMOUNT DUE 0.60
PAYMENT DELIVERY LOCATION: CSD Receptionist.
Make checks payable to: STANFORD UNIVERSITY.
Please deliver payments to the Computer Science Dept receptionist, Jacks Hall.
To ensure proper crediting, please include your PONY ACCOUNT NAME on your check.
Note: The recording of a payment takes up to three weeks after the payment is
made, but never beyond the next billing date. Please allow for this delay.
Bills are payable upon presentation. Interest of 1.0% per month will be
charged on balances remaining unpaid 25 days after bill date above.
An account with a credit balance earns interest of .33% per month,
based on the average daily balance.
You haven't paid your Pony bill since 10/89.
Accounts with balances remaining unpaid for more than 55 days are
considered delinquent and are subject to reduction of credit limit.
Please pay your bill and keep your account current.
∂04-Dec-89 2256 harnad@clarity.Princeton.EDU Emperor's New Mind: BBS Call for Commentators
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 4 Dec 89 22:56:39 PST
Received: from suspicion.Princeton.EDU by Princeton.EDU (5.58+++/2.25/mailrelay)
id AA23051; Tue, 5 Dec 89 01:38:51 EST
Received: by suspicion.Princeton.EDU (4.0/1.95)
id AA01784; Tue, 5 Dec 89 01:04:01 EST
Date: Tue, 5 Dec 89 01:04:01 EST
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8912050604.AA01784@suspicion.Princeton.EDU>
To: harnad@elbereth.rutgers.edu
Subject: Emperor's New Mind: BBS Call for Commentators
Below is the synopsis of a book that will be accorded a multiple book
review (20 - 30 multidisciplinary reviews, followed by the author's
response) in Behavioral and Brain Sciences (BBS), an international,
interdisciplinary journal that provides Open Peer Commentary on
important and controversial current research in the biobehavioral and
cognitive sciences. Reviewers must be current BBS Associates or
nominated by a current BBS Associate. To be considered as a reviewer
for this book, to suggest other appropriate reviewers, or for
information about how to become a BBS Associate, please send email to:
harnad@confidence.princeton.edu or write to:
BBS, 20 Nassau Street, #240, Princeton NJ 08542 [tel: 609-921-7771]
____________________________________________________________________
THE EMPEROR'S NEW MIND:
CONCERNING COMPUTERS, MINDS AND THE LAWS OF PHYSICS
Roger Penrose
Rouse Ball Professor of Mathematics
University of Oxford
The Emperor's New Mind is an attempt to put forward a scientific
alternative to the viewpoint of "Strong AI," according to which mental
activity is merely the acting out of some algorithmic procedure. John
Searle and other thinkers have likewise argued that mere calculation
does not, of itself, evoke conscious mental attributes, such as
understanding or intentionality, but they are still prepared to accept
that the action of the brain, like that of any other physical object,
could in principle be simulated by a computer. In my book I go further
than this and suggest that the outward manifestations of conscious
mental activity cannot even be properly simulated by calculation. To
support this view I use various arguments to show that the results of
mathematical insight, in particular, do not seem to be obtained
algorithmically. The main thrust of this work, however, is to present
an overview of the present state of physical understanding and to show
that an important gap exists at the point where quantum and classical
physics meet, and to speculate on how the conscious brain might be
taking advantage of whatever new physics is needed to fill this gap, in
order to achieve its non-algorithmic effects.
∂05-Dec-89 0630 CLT CLT itinerary
To: JMC, MPS
Tue Dec 5: sf -> Indianapolis USAIR 1556 12noon / 6:54pm
Sun Dec 10: Indianapolis -> sf USAIR 257 9:40am / 11:36am
Staying with
Alice ter Muelen
1217 a East Maxwell Lane, Bloomington, IN 47401
phone (812) 332-6225,
email: atm@GOLD.BACS.INDIANA.EDU
Contact in CS dept IU
Olivier Danvy
email: danvy@IUVAX.CS.INDIANA.EDU
office (812) 855-9756
home (812) 339-4565
∂05-Dec-89 0647 CLT cate
Call him -- sell Mesa, discuss Computer Assoc.
∂05-Dec-89 0914 phil@ub.d.umn.edu Re: Philosophy and Artificial Intelligence
Received: from ub.d.umn.edu ([131.212.32.6]) by SAIL.Stanford.EDU with TCP; 5 Dec 89 09:14:30 PST
Received: by ub.d.umn.edu (5.59/UMD-891202)
id AA18244; Tue, 5 Dec 89 11:14:42 CST
From: phil@ub.d.umn.edu (Philosophy Dept)
Message-Id: <8912051714.AA18244@ub.d.umn.edu>
Subject: Re: Philosophy and Artificial Intelligence
To: JMC@SAIL.Stanford.EDU (John McCarthy)
Date: Tue, 5 Dec 89 11:14:39 CDT
Cc: phil@ub.d.umn.edu
In-Reply-To: <oFAHJ@SAIL.Stanford.EDU>; from "John McCarthy" at Dec 04, 89 1946
X-Mailer: Elm [version 2.1 PL1]
John,
Thanks for agreeing to join the board. I am delighted. I also
welcome your remarks on the program verification controversy. I
wrote the piece as an exercise in the theory of knowledge, since
I perceived the position of Hoare to be tantamount to the claim
that there exists a special kind of synthetic a priori knowledge
in computer science that verification can provide. This seems to
me to be completely unfounded, for reasons such as those that were
outlined in my original paper and in the extensive discussion that
followed. The beat goes on: a response to Barwise will appear in
this month's Notices of the AMS and a debate will be held during
the forthcoming ACM CSC '90 in Washington during February. Never-
theless, I do not claim to be an expert on these things. The ACM
editor, Rob Kling, put my paper through four (4) drafts involving
four (4) prominent computer scientists. As Denning observed in
his comments in the March 1989 issue, the standards imposed by the
magazine for publication were surely met in this case. Since I am
still becoming acquainted with aspects of the literature, however,
I am glad to know of other material out there. In fact, I have put
together a collection of old and new papers on this subject for the
ACM Press. What interests me the most about your comments, though,
is that I would like to know if you might be willing to write some-
thing on this for an early issue of the journal (perhaps even the
first issue). I would be glad to send you copies of the entire de-
bate and would welcome whatever contribution your observations may
make toward its clarification. I would need something (perhaps a-
round 30 pages double-spaced/hardcopy) by 1 June 1990. Let me know
if this would be agreeable with you, because I would like to see it.
Jim
∂05-Dec-89 0915 MPS
Guten Tag
Where shall I send the letters concerning the concerned
scientist. To the Church Street address, or their individual
schools? Some are emeritus, so may not have an office at the
school. Thaks.
Pat
∂05-Dec-89 0925 MPS
Cancel that message about the concerned scientists. I
figured it out myself.
∂05-Dec-89 1000 JMC
Heffalump + suppes office
∂05-Dec-89 1225 MPS
Yoav does not know where Ma Xiwen is either, but also would
like to know. Do you want to answer that letter from Han
Guangyu, former student of Xiwen.
∂05-Dec-89 1300 JMC
invitation
∂05-Dec-89 1437 bscott@Polya.Stanford.EDU Your Questions
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 5 Dec 89 14:37:11 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA12091; Tue, 5 Dec 89 14:37:18 -0800
Date: Tue, 5 Dec 1989 14:37:17 PST
From: Betty Scott <bscott@Polya.Stanford.EDU>
To: JMC@Sail.Stanford.EDU
Cc: BScott@Polya.Stanford.EDU
Subject: Your Questions
Message-Id: <CMM.0.88.628900637.bscott@Polya.Stanford.EDU>
1. Sabbatical - at the end of the current year you will have 24 quarters
of eligibility (you used 6 qtrs. for a 50% sabbatical Spring 81-82).
So if you plan a 100% sabbatical for Autumn of 90-91 you will use up
18 quarters, and can request a banking of the remaining 6. If you are
sure about the 100% Autumn of next year, I suggest we go ahead and sub-
mit a leave form now. I would like very much to have this approved and
settled prior to my departure at the end of the month. Let me know.
2. Business class travel - if the flight is 10 hours or more, business
class fare will be approved. I am getting another copy of the memorandum
about this, and if there are other circumstances where it can be
approved, I will let you know.
Betty
∂05-Dec-89 1448 lamport@src.dec.com Re: Dec connections
Received: from decwrl.dec.com by SAIL.Stanford.EDU with TCP; 5 Dec 89 14:48:37 PST
Received: by decwrl.dec.com; id AA11393; Tue, 5 Dec 89 14:48:43 -0800
Received: by jumbo; id AA06698; Tue, 5 Dec 89 14:48:38 PST
Date: Tue, 5 Dec 89 14:48:38 PST
From: lamport@src.dec.com (Leslie Lamport)
Message-Id: <8912052248.AA06698@jumbo>
To: jmc@sail.stanford.edu
Cc: Carolyn Talcott <CLT@sail.stanford.edu>
Subject: Re: Dec connections
In-Reply-To: Your message of 03 Dec 89 0657 PST.
<CEV7F@SAIL.Stanford.EDU>
John,
It seems best to wait until Martin Abadi gets back from SOSP before
making any decisions. We'll contact you next week.
Leslie
∂05-Dec-89 1719 peters@russell.Stanford.EDU [Takayasu Ito: Your Proposal]
Received: from russell.Stanford.EDU by SAIL.Stanford.EDU with TCP; 5 Dec 89 17:19:22 PST
Received: from localhost by russell.Stanford.EDU (4.0/inc-1.0)
id AA10430; Tue, 5 Dec 89 17:20:51 PST
Message-Id: <8912060120.AA10430@russell.Stanford.EDU>
To: jmc@sail.stanford.edu
Subject: [Takayasu Ito: Your Proposal]
Date: Tue, 05 Dec 89 17:20:47 PST
From: peters@russell.Stanford.EDU
John,
I just got back from a trip to find this note from Prof. Ito. I'll
certainly put something in writing for him. If you have suggestions
about exactly what, I'd welcome them.
Do you think this means he would like to take a more active position
of advocacy and push our case in Japan? That he actually would like
to head whatever Japanese entity might exist in connection with
funding to Stanford? I don't want to read more into subtle
indications than is really there.
Stanley
------- Forwarded Message
Date: Wed, 22 Nov 89 20:36:55 JST
>From: Takayasu Ito <ito%ito.ecei.tohoku.ac.jp@RELAY.CS.NET>
Return-Path: <ito@ito.ecei.tohoku.ac.jp>
Message-Id: <8911221136.AA02517@ito.ito.ecei.tohoku.ac.jp>
To: peters%russell.stanford.edu%relay.cs.net%u-tokyo.junet@RELAY.CS.NET
Subject: Your Proposal
It was interesting what you said.I like to know about it in a more definite formto circulate (if possible) it to some of my friends.
If you are thinking seriously, would you write about your ideas and proposal in
a letter with your signature? I will talk with several people on the basis of
your letter,if I really become inrerested in it.
Takayasu Ito
------- End of Forwarded Message
∂05-Dec-89 1804 winograd@loire.stanford.edu REMINDER - AI QUAL TOMORROW (WED) MORNING AT 8:45
Received: from loire.stanford.edu by SAIL.Stanford.EDU with TCP; 5 Dec 89 18:03:55 PST
Received: by loire.stanford.edu (5.59/25-eef) id AA08260; Tue, 5 Dec 89 18:01:36 PDT
Date: Tue, 5 Dec 89 18:01:36 PDT
Message-Id: <8912060201.AA08260@loire.stanford.edu>
From: Terry Winograd <Winograd@csli.stanford.edu>
To: jmc@sail
Subject: REMINDER - AI QUAL TOMORROW (WED) MORNING AT 8:45
In MJH 301. --t
∂05-Dec-89 1951 kirsh%cogsci@ucsd.edu no comment
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 5 Dec 89 19:50:56 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA00203
sendmail 5.61/UCSD-2.0-sun via SMTP
Tue, 5 Dec 89 19:50:55 -0800 for JMC@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA28366 for JMC@sail.stanford.edu; Tue, 5 Dec 89 19:53:29 PST
Date: Tue, 5 Dec 89 19:53 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: no comment
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <#F#nE@SAIL.Stanford.EDU>
Message-Id: <19891206035335.1.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
Yes I received your email message. Carl has allegedly sent me a new paper.
If it seems any better would you be willing to read it? If we do not have a
commentator we will not include his paper in the special edition. A bit of a
blow I'd say. You could have 3 weeks to finish your reply if you thought it
were worthwhile.
-- David
∂06-Dec-89 0836 rabin@harvard.harvard.edu Re: sabbatical
Received: from harvard.harvard.edu (HARVARD-GW.HARVARD.EDU) by SAIL.Stanford.EDU with TCP; 6 Dec 89 08:36:36 PST
Received: by harvard.harvard.edu (5.54/a0.25)
(for JMC@SAIL.Stanford.EDU) id AA03030; Wed, 6 Dec 89 11:38:15 EST
Received: by endor.HARVARD.EDU; Wed, 6 Dec 89 11:38:14 EST
Date: Wed, 6 Dec 89 11:38:14 EST
From: rabin@harvard.harvard.edu (Michael Rabin)
To: JMC@SAIL.Stanford.EDU
Subject: Re: sabbatical
Dear John,
I spoke with the Dean and he was enthisiastic about
your planned stay at Harvard.
The title of visiting professor is strictly reserved
for people who come to formally teach, so that
he suggested to make you a visiting scholar. This
will give you access to all the Harvard facilities.
The office arrangement has also been made.
My recollection is that you are coming for the whole academic
year. When will you actually arrive?
Can we be of help in finding housing?
Your lecture here evoked a lot of interest, thank
you again.
Best regards. Michael
∂06-Dec-89 0900 JMC
street
∂06-Dec-89 1221 kirsh%cogsci@ucsd.edu re: no comment
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 6 Dec 89 12:20:56 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA02811
sendmail 5.61/UCSD-2.0-sun via SMTP
Wed, 6 Dec 89 12:20:59 -0800 for JMC@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA06192 for JMC@sail.stanford.edu; Wed, 6 Dec 89 12:23:30 PST
Date: Wed, 6 Dec 89 12:23 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: re: no comment
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <oFc13@SAIL.Stanford.EDU>
Message-Id: <19891206202335.5.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
I will forward a copy of the document as soon as I receive it. I am eager to
have you represented, John, so it is not just a matter of finding someone who
might comment on Carl.
-- David
∂06-Dec-89 1315 @Polya.Stanford.EDU:jcm@iswim.Stanford.EDU MTC qual reading list
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Dec 89 13:15:12 PST
Received: from iswim.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61+IDA/25-eef) id AA10326; Wed, 6 Dec 89 13:15:17 -0800
Received: by iswim.stanford.EDU (5.57/Ultrix3.0-C)
id AA00277; Wed, 6 Dec 89 13:12:30 PST
Date: Wed, 6 Dec 89 13:12:30 PST
From: John C. Mitchell <jcm@cs.Stanford.EDU>
Message-Id: <8912062112.AA00277@iswim.stanford.EDU>
To: hemenway@cs.Stanford.EDU
Subject: MTC qual reading list
Cc: jmc@sail.Stanford.EDU, lincoln@ghoti.Stanford.EDU
Could you send a copy to John McCarthy please.
As I recall, Pat Lincoln coordinated a revision
(and shortening) of this list last spring.
Thanks,
John
∂06-Dec-89 1322 hemenway@Sunburn.Stanford.EDU Re: MTC qual reading list
Received: from Sunburn.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Dec 89 13:22:23 PST
Received: by Sunburn.Stanford.EDU (5.61/25-eef) id AA06026; Wed, 6 Dec 89 13:23:14 -0800
Date: Wed, 6 Dec 1989 13:23:13 PST
From: "Sharon R. Hemenway" <hemenway@Sunburn.Stanford.EDU>
To: John C. Mitchell <jcm@cs.Stanford.EDU>
Cc: hemenway@cs.Stanford.EDU, jmc@sail.Stanford.EDU,
lincoln@ghoti.Stanford.EDU
Subject: Re: MTC qual reading list
In-Reply-To: Your message of Wed, 6 Dec 89 13:12:30 PST
Message-Id: <CMM.0.88.628982593.hemenway@Sunburn.Stanford.EDU>
The only MTC syllabus I have is one dated November, 1987. If there
was a revised list made last year, a copy never made its way to me.
I'll wait to pass on the list I have to McCarthy until I hear from
you.
Sharon
∂06-Dec-89 1505 @Polya.Stanford.EDU:lincoln@ghoti.Stanford.EDU RE: MTC qual reading list
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Dec 89 15:02:47 PST
Received: from ghoti.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61+IDA/25-eef) id AA14384; Wed, 6 Dec 89 15:02:49 -0800
Received: by ghoti.stanford.EDU (5.57/Ultrix3.0-C)
id AA16783; Wed, 6 Dec 89 15:02:54 EST
Date: Wed, 6 Dec 89 15:02:54 EST
From: Patrick Lincoln <lincoln@cs.Stanford.EDU>
Message-Id: <8912062002.AA16783@ghoti.stanford.EDU>
To: JMC@sail.stanford.edu
Cc: lincoln@ghoti.stanford.edu
Subject: RE: MTC qual reading list
Here is the on-line LaTeX version of the latest revision of
the MTC qual reading list. This list will be shortened (somehow)
before it is made 'official' for this spring's qual.
pdl.
P.S. If you want a hardcopy, let me know.
---------------------------------------------------
\documentstyle{article}
\setlength{\textwidth}{6in}
\setlength{\textheight}{9in}
\setlength{\oddsidemargin}{0in}
\setlength{\headheight}{0in}
\setlength{\headsep}{0in}
\begin{document}
\begin{center}
MATHEMATICAL THEORY OF COMPUTATION \\
Qualifying Exam Syllabus \\
*** DRAFT **************** Fall 1989 **************** DRAFT ***
\end{center}
%%% 350 assumed
%%% 500 background
%%% 630 subjects
%%% ------------
%%% 1480 total - still too much -
\begin{enumerate}
\item Assumed Knowledge %%% 350
\begin{itemize}
\item Davis, M.D., and Weyuker, E.J.:
{\em Computability, Complexity, and Languages}\\
(chapters 1-4; 70 p.)
Academic Press, 1982.
\item Enderton, H.B.:
{\em A Mathematical Introduction to Logic}\\
(except chapter 2.8; 280 p.)
Academic Press, 1972.
\end{itemize}
\item Background %%% 500
\begin{itemize}
\item Hindley, J.R., and Seldin, J.P:
{\em Introduction to Combinators and $\lambda$-calculus}\\
(chapters 1-12; 235 p.)
Cambridge University Press, 1986.
\item Manna, Z., and Waldinger, R.:
{\em The Logical Basis for Computer Programming Vol II}\\
("fast track", 200 p.)
In press 1989, see Manna or Waldinger.
\item Milner, R., {\em A Theory of Type Polymorphism in Programming}
(27 p.)\
in: Journal of Computer and System Sciences 17, 348-375 (1978).
\item Pierce, B.C.:
{\em A Taste of Category Theory for Computer Scientists}\\
(chapter 1-2, 40 p.)
CMU Technical Report CMU-CS-88-203.
\end{itemize}
\item Semantics %%% 250
\begin{itemize}
\item Allison, L.:
{\em A Practical Introduction to Denotational Semantics}
(chapters 1-3, 5-8, 80p.?)
Cambridge CS Texts 23, Cambridge University Press, 1986.
\item Ehrig, H., and Mahr, B.:
{\em Fundamentals of Algebraic Specification 1}\\
(chapters 1-6, 167 p.)
Springer, 1985.
\end{itemize}
\item Concurrency %%% 100
\begin{itemize}
\item Hoare, C.A.R.:
{\em Communicating Sequential Processes}
(10 p.)
in: CACM, 8/78.
\item Milner, R.:
{\em Lectures on a Calculus for Communicating Systems}\\
(20 p.)
in: {\em Seminar on Currency}, LNCS 197, Springer, 1985.
\item Pnueli, A.:
{\em Applications of Temporal Logic to the Specification and Verification
of Reactive Systems: A Survey of Current Trends}\\
(70 p.)
in: {\em Current Trends in in Concurrency}, LNCS 224, Springer, 1986.
\end{itemize}
\item Automated Deduction %%% 130
\begin{itemize}
\item Chang, C., and Lee, R.C.:
{\em Symbolic Logic and Mechanical Theorem Proving}\\
(chapters 1-5; 100 p.)
\item Huet, G., and Oppen, D.C.:
{\em Equations and Rewrite Rules: A Survey}
(30 p.),
in Book, R. (ed.):
{\em Formal Languages: Perspectives and Open Problems},
Academic Press, 1980.
\end{itemize}
\item Program Verification %%% 150
\begin{itemize}
\item Boyer, R., and Moore, R.:
{\em A Computational Logic}
(chapters 1-3; 55 p.)
\item Apt, K.:
{\em Ten Years of Hoare Logic, A Survey - Part I}\\
(52 p.)
ACM Transactions on Programming Languages and Systems 3,4
October, 1981, p 431-483.
\item Manna, Z., and Waldinger, R.:
{\em The Origin of a Binary Search Paradigm}\\
(46 p.)
in Science of Computer Programming 9, North Holland 1987.
\end{itemize}
\end{enumerate}
\end{document}
∂06-Dec-89 1633 @Polya.Stanford.EDU:lincoln@ghoti.Stanford.EDU RE: MTC qual reading list
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 6 Dec 89 16:33:07 PST
Received: from ghoti.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61+IDA/25-eef) id AA17823; Wed, 6 Dec 89 16:33:15 -0800
Received: by ghoti.stanford.EDU (5.57/Ultrix3.0-C)
id AA16845; Wed, 6 Dec 89 16:33:19 EST
Date: Wed, 6 Dec 89 16:33:19 EST
From: Patrick Lincoln <lincoln@cs.Stanford.EDU>
Message-Id: <8912062133.AA16845@ghoti.stanford.EDU>
To: JMC@sail.stanford.edu
Cc: lincoln@ghoti.Stanford.EDU
In-Reply-To: John McCarthy's message of 06 Dec 89 1507 PST <oGsBO@SAIL.Stanford.EDU>
Subject: RE: MTC qual reading list
I gave a hardcopy to your secretary. pdl.
∂06-Dec-89 1706 ortiz@itstd.sri.com Thanks
Received: from aai0.itstd.sri.com by SAIL.Stanford.EDU with TCP; 6 Dec 89 17:05:56 PST
Received: from localhost by aai0.itstd.sri.com (5.61/1.3davy)
id AA06560; Wed, 6 Dec 89 17:10:25 -0800
Message-Id: <8912070110.AA06560@aai0.itstd.sri.com>
To: jmc@sail.stanford.edu
Cc: ortiz@itstd.sri.com
Subject: Thanks
Date: Wed, 06 Dec 89 17:10:20 PST
From: Charles Ortiz <ortiz@itstd.sri.com>
Dear Professor McCarthy,
I want to thank you again for the recomendation you gave me a few
months ago for the University of Pennsylvania and let you know that I
will be attending Penn in January to pursue a PhD in AI. I hope to
combine my interests in knowledge representation, computational
linguistics, and logic there.
I also wanted to thank you for a most enjoyable, inspiring, and
stimulating course in nonmonotonic reasoning. I don't ever remember
being involved in a course that so influenced my thinking and
appreciation of a field.
Charlie Ortiz
∂06-Dec-89 1936 ash@sumex-aim.stanford.edu Qual
Received: from sumex-aim.stanford.edu by SAIL.Stanford.EDU with TCP; 6 Dec 89 19:36:18 PST
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AA18458; Wed, 6 Dec 89 19:38:26 PST
Date: Wed, 6 Dec 1989 19:38:25 PST
From: David Ash <ash@sumex-aim.stanford.edu>
To: jmc@sail.stanford.edu
Subject: Qual
Message-Id: <CMM.0.88.629005105.ash@sumex-aim.stanford.edu>
Is there a time when I could meet to talk with you about this qual? I'd like
to get an idea where you feel I went wrong this time and what I should do to
pass next time. I'm free most of the next 1.5 weeks.
Thanks.
-Dave
∂06-Dec-89 2100 kirsh%cogsci@ucsd.edu re: no comment
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 6 Dec 89 21:00:41 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA24807
sendmail 5.61/UCSD-2.0-sun via SMTP
Wed, 6 Dec 89 21:00:46 -0800 for JMC@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA12519 for JMC@sail.stanford.edu; Wed, 6 Dec 89 21:03:24 PST
Date: Wed, 6 Dec 89 21:03 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: re: no comment
To: JMC@SAIL.Stanford.EDU
In-Reply-To: <1qGq8B@SAIL.Stanford.EDU>
Message-Id: <19891207050329.3.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
John,
I just received the Nov 12th version and I think it is the same as the other
Nov 12th version. The title is not the title Carl gave me on the phone as his
new version. It was sent by his secretary. So until I confirm that this is
the version Carl thought was new PLEASE DON'T TAKE THE TIME OT READ IT.
I'll speak to you when I have news.
-- David
∂07-Dec-89 0942 MPS FAX
Goodday,
You received a fax yesterday from Jaap van den Ilerik, ICCA Journal -
also, he called. He will be calling today, here, around 1:00 our time
concerning your paper The Fruitfly on the Fly. Will you be in or should
I give him your home number. What he states in his letter is this.
We would like to have your permission for publication of this version as
it is. Of course, your improvements, corrections or amendments are welcome and
will be included. Specifically, we request you to include the initials
of Mr. Margan (p. 199) and to supply us with the exact details of your own 1983
publication (p.205). The figure from Huberman thesis and diagrams will be
produced by our department.
∂07-Dec-89 1012 MPS
Call the car shop 494-7676
∂07-Dec-89 1100 MPS DMV
I have to leave around 4 today as I have to register my car. Thanks
Pat
∂07-Dec-89 1356 VAL Manchester
Here is a message from Manchester and the reply I want to send.
Is the reply sufficiently polite, in your opinion?
-------------------
From: John Gurd <john%research7.computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK>
Date: Thu, 7 Dec 89 18:41:02 GMT
Message-Id: <13947.8912071841@r7f.r7.cs.man.ac.uk>
To: VAL@sail.stanford.edu
In-Reply-To: Vladimir Lifschitz's message of 31 Oct 89 1548 PST <w4yiO@SAIL.Stanford.EDU>
Subject: News from Manchester
Vladimir,
It appears that there might have been a particularly
unfortunate misunderstanding between us about the chair in artificial
intelligence, in that those members of this department who have spoken
to you over the past few months do not remember telling you explicitly
that there were other candidates under consideration. I certainly recall
that I did not mention this when I saw you in San Francisco, being under
the impression that you were already aware of it. Perhaps the same thing
happened to my colleagues. Whatever the circumstances, if you were not
told this, I apologise profusely for our impoliteness.
You may recall that I said there was uncertainty about the
nature of post that might be offered to you. This was caused principally
by the fact that you are seen as working in the "middle ground" between
AI and theoretical computer science, and that appointing you to what was
intended to be a "main-stream" AI chair did not make sense. Unfortunately,
the creation of a more suitable post has had financial implications which
are taking a long time to resolve.
The position here now is that the Department wants the Vice-
Chancellor to invite you to take up a visiting chair in computational logic
for a significant period of time, say 3 or 4 years, starting as soon as
possible. The Department has unconditionally offered to fund the first
year of such an appointment. The problem is where to find funds for the
subsequent period. We have made enquiries outside the University and
discovered that the SERC would be very likely to fund a further year.
However, this, in common with other possibilities, cannot be guaranteed
immediately. Hence, we have asked the University whether it would be
willing to fund an additional period. I still don't know the outcome
of this request.
I am, therefore, hoping that the Vice-Chancellor will
write to you very soon, offering perhaps a one or two year appointment.
Whatever period the VC offers, I would emphasise that the Department
will take steps after your arrival to increase the period by other
sources of funding. We are aware that it will be difficult for you to
respond to such an offer rapidly, and we have, therefore, requested
that the offer be open for you to take up any time at your convenience
within the next year. In case your family circumstances affect the length
of time you might consider staying in Manchester, we would appreciate some
feedback about the optimum length of appointment for you.
I'm sorry to have gone quiet for such a long time. I hope
that you were not unduly affected by the recent earthquake: It sounded
somewhat frightening from over here.
All the best, John Gurd.
-----------------
John,
Thank you for your message explaining the current situation in
Manchester. You are right: Until my recent meeting with Barringer,
I hadn't been told that other candidates were being considered
for the AI post. As a matter of fact, my last conversation with
Brian Warboys in Manchester left me under the impression that the
committee had decided to recommend me for that position. I don't
remember the exact words that Brian used; but, as I have mentioned
to you, he asked me to give an answer within two weeks. You will
agree that this was hardly consistent with the assumption that the
post had been offered, or would be offered, to someone else.
I appreciate that this incident was mostly due to the lack of
communication within the department, and I'll be happy to put this
matter behind us.
Best regards,
Vladimir
∂07-Dec-89 1610 VAL re: Elephant programs as sentences
[In reply to message rcvd 30-Nov-89 00:57-PT.]
Maybe you don't need $holds$ at all. You can use these abbreviations instead:
$$available(t,seat,flt) ≡ ¬∃psgr exists(t,commitment admit(psgr,flt,seat)),$$
$$full(t,flt) ≡ ¬∃seat available(t,seat,flt).$$
∂07-Dec-89 1719 bergman@Sunburn.Stanford.EDU proposals
Received: from Sunburn.Stanford.EDU by SAIL.Stanford.EDU with TCP; 7 Dec 89 17:19:30 PST
Received: by Sunburn.Stanford.EDU (5.61/25-eef) id AA29308; Thu, 7 Dec 89 17:19:04 -0800
Date: Thu, 7 Dec 1989 17:19:03 PST
From: Sharon Bergman <bergman@Sunburn.Stanford.EDU>
To: cheriton@pescadero, dill@amadeus, genesereth@score, ango@hudson,
golub@patience, guibas@src.dec.com, dek@sail, zm@sail, jmc@sail,
clt@sail, jcm@polya, rajeev@polya, nilsson@tenaya, oliger@pride,
plotkin@hudson, pratt@jeeves, ullman@nimbin, gio@sumex
Subject: proposals
Message-Id: <CMM.0.88.629083143.bergman@Sunburn.Stanford.EDU>
If you have a proposal deadline in the next few weeks, please note
that I will be gone for two weeks during the holidays. I will be
here through Dec. 22nd and then will be gone until Jan. 8th. Let me
know well in advance of Dec. 22nd if you need my assistance with a
proposal. Thanks.
-Sharon Bergman
∂08-Dec-89 0809 hewitt@ai.mit.edu aij
Received: from life.ai.mit.edu by SAIL.Stanford.EDU with TCP; 8 Dec 89 08:09:36 PST
Received: from rice-chex (rice-chex.ai.mit.edu) by life.ai.mit.edu (4.0/AI-4.10) id AA22960; Fri, 8 Dec 89 11:09:33 EST
From: hewitt@ai.mit.edu (Carl Hewitt)
Received: by rice-chex (4.0/AI-4.10) id AA04804; Fri, 8 Dec 89 11:09:01 EST
Date: Fri, 8 Dec 89 11:09:01 EST
Message-Id: <8912081609.AA04804@rice-chex>
To: JMC@sail.stanford.edu
Cc: kirsh%cogsci@ucsd.edu
In-Reply-To: John McCarthy's message of 01 Dec 89 2159 PST <WDcDl@SAIL.Stanford.EDU>
Subject: aij
Cc: Hewitt@ai.mit.edu
John,
Thanks for your comments. I apologize for removing the bank
withdrawal parallel program (I have published it previously in another
paper and it is included in another paper as well). Nevertheless I
agree that it belongs in this paper as well since it illustrates one
of the fundamental limitations of logic programming: indeterminate
outcomes can not be decided deductively.
Sincerely,
Carl
P.S. I am curious how you would use circumscription to address
the question of the safety of the Diablo Canyon nuclear reactor
example discussed in my paper. (Do you think that I am tilting
with windmills?)
∂08-Dec-89 0835 MPS Fruitfly paper
Mr. van der Herik called later yesterday. He is very happy
to have your paper to publish and said whatever changes you
want to make will be fine with him. They are going to submit
to the publisher next week so he would like your changes faxed
by Wednesday the 13th, or at the latest Friday the 15th. He
will be calling again today to have me tell him what day you
will have the paper ready to fax.
Pat
∂08-Dec-89 0846 MPS
∂08-Dec-89 0843 JMC re: Fruitfly paper
To: MPS
[In reply to message rcvd 08-Dec-89 08:35-PT.]
I'll send a fax to him later today with either the corrections
or a date when I'll have them.
the fax number is 31 43 252392
∂08-Dec-89 1043 shoham@Hudson.Stanford.EDU cheap pun
Received: from Hudson.Stanford.EDU by SAIL.Stanford.EDU with TCP; 8 Dec 89 10:43:39 PST
Received: by Hudson.Stanford.EDU (5.61/25-eef) id AA25297; Fri, 8 Dec 89 10:43:26 -0800
Date: Fri, 8 Dec 1989 10:43:24 PST
From: Yoav Shoham <shoham@Hudson.Stanford.EDU>
To: jmc@sail, val@sail
Subject: cheap pun
Message-Id: <CMM.0.88.629145804.shoham@Hudson.Stanford.EDU>
One of the typos I made recently resulted in the word "circuscription".
I decided to correct it anyway.
Yoav
∂08-Dec-89 1309 MPS
I have to go to the bank today to sign some papers.
I have to be there before 4:00, so I will leave here
around 3:15. Thanks. Pat
∂08-Dec-89 1332 cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK Lifschitz
Received: from NSFnet-Relay.AC.UK by SAIL.Stanford.EDU with TCP; 8 Dec 89 13:32:18 PST
Received: from sun.nsfnet-relay.ac.uk by vax.NSFnet-Relay.AC.UK
via Janet with NIFTP id aa17680; 8 Dec 89 21:17 GMT
Date: Fri, 8 Dec 89 19:21:17 GMT
From: cliff%computer-science.manchester.ac.uk@NSFnet-Relay.AC.UK
Message-Id: <8912081921.AA17941@ipse2pt5.cs.man.ac.uk>
To: jmc@sail.stanford.edu
Subject: Lifschitz
John,
The people involved in the committee have now been consulted
and the head-of-department has now been in touch with
Vladimir. I can only say how grateful I am to you for alerting
me to a problem that Vladimir was clearly not in a position to
bring up himself. I think that those involved see that a
serious mistake was made (there were all sorts of problems but
it is not my intention to make excuses). Vladimir has taken
John Gurd's apology very kindly and is even prepared to go on
discussing the alternative arrangement that we were working
on.
I'd just like to add that there was absolutely nothing
"dishonourable" uncovered in our investigation - but an
uncomfortable number of errors appear to have been made in a
position which required the utmost care.
thanks again,
cliff jones
∂08-Dec-89 1441 VAL re: Lifschitz
[In reply to message sent Fri, 8 Dec 89 19:21:17 GMT.]
> I can only say how grateful I am to you for alerting
> me to a problem that Vladimir was clearly not in a position to
> bring up himself.
I can only say how grateful I am to you for alerting Cliff Jones
to a problem that I was clearly not in a position to bring up myself.
∂08-Dec-89 1548 kirsh%cogsci@ucsd.edu hewitt
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 8 Dec 89 15:48:33 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA27696
sendmail 5.61/UCSD-2.0-sun via SMTP
Fri, 8 Dec 89 15:48:33 -0800 for jmc@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA07933 for jmc@sail.stanford.edu; Fri, 8 Dec 89 15:51:09 PST
Date: Fri, 8 Dec 89 15:51 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: hewitt
To: jmc@sail.stanford.edu
Supersedes: <19891208235046.3.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
Message-Id: <19891208235113.4.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
John,
I need some critical input on Carl's paper. I have been dissatisfied with
its abstract level and lack of technical detail. What is your opinion of the
quality of the paper? Is it worth publishing and finding another commentator?
According to Carl's secretary, we will receive a new paper Monday by Fedex.
I will be going away Tues early evening. If it is at all possible I would
appreciate your reading the new paper straight away and letting me know:
1) whether you wish to comment on it;
2) whether it is worthy of publishing.
-- David
∂08-Dec-89 1809 scales@Gang-of-Four.Stanford.EDU OPS5 in QLISP
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 8 Dec 89 18:08:55 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA05289; Fri, 8 Dec 89 18:09:47 -0800
Date: Fri, 8 Dec 89 18:09:47 -0800
From: scales@Gang-of-Four.Stanford.EDU (Dan Scales)
Message-Id: <8912090209.AA05289@Gang-of-Four.Stanford.EDU>
To: qlisp@Gang-of-Four.Stanford.EDU
Subject: OPS5 in QLISP
Here is a summary of some of the results I've gotten in trying to
parallelize OPS5 under QLISP, both the standard version and Dan's
extensions to it. (Hiroshi Okuno also worked on this problem, but never
got real speedup numbers, since he only used the QLISP simulator.) Send
questions if you want a clarification or more details on any of the
points.
OPS5 is a rule system which runs by repeatedly executing a two-phase
cycle: (1) determining which rules currently match the "working memory"
of the system and (2) executing ("firing") one of these matching rules.
OPS5 uses an algorithm known as the "Rete algorithm" to efficiently do
the match phase (which takes the majority of the time). It is this
algorithm that I've been working on parallelizing in QLISP.
As it reads a set of rules in, the Rete algorithm builds a discrimination
network corresponding to the set of rules. The network is like a
dataflow network, whose inputs are the changes to the working memory of
the system, and whose outputs are the rules that currently match working
memory. State is maintained in the network, in order to avoid having to
redo much of the matching work on every cycle. There is conceivably much
parallelism, since, with the proper locking, every activation of a node
in the network can be processed in parallel.
Here are some results of OPS5 runs on 8 processors. Times are in
milliseconds. I have labeled Dan's low-cost extensions as QNLISP, for
lack of a better name:
Serial QLISP QLISP QNLISP QNLISP
Task time time # procs time # procs
---- ------ ----- ------- ----- -------
Rubik 10061 5766 8643 2676 7583
Weaver 12710 28920 19594 8069 68872
Tourney 19449 13196 3464 8449 3700
Comments on these numbers:
1. For the rubik task, I get close to 8 out of 8 speedup on the matching
part of the OPS5 run, using Dan's low-cost extensions. (The speedup shown
by the above numbers is less, because these numbers also include the time
for the execution phase of OPS5, which is serial. For rubik, this time is
about 1600ms)
2. The parallel algorithm run on one processor is usually often a bit
slower than the serial algorithm, because much work that is done only
once in the serial algorithm must be duplicated to allow the algorithm to
run on multiple processors. In particular, this "loss of sharing" is why
the weaver task above shows such dismal speedup: for that task, the
parallel algorithm on one processor is three times slower than the serial
algorithm. The low speedup on the tourney task is for another reason:
contention for locks.
Differences between the Rete algorithm and some of the typical QLISP
benchmark programs:
1. The algorithm is completely non-functional; most functions do not
return values explicitly, but instead modify special variables and/or
information stored in the network.
2. The algorithm requires extensive use of locks at nodes in the network
that store state information.
3. The algorithm creates a large number of tasks, but the degree of
parallelism at different points in the run and the size of the tasks
created varies widely. Because of this variability, it is necessary to
spawn tasks at a fairly fine-grained level in order to get a good
speedup. Because of this, Dan's low-cost SPAWN and QLET constructs allow
for much better speedup than the standard QLISP versions.
4. Some limited tests indicate that using dynamic spawning does not much
affect the timings. This is probably because, as described above, there is
not an over-abundance of parallelism in the Rete algorithm. The amount of
parallelism in the algorithm is limited further by the necessity of
acquiring locks when accessing data in the network.
Some observations about QLISP:
1. I found it fairly easy to write the parallel Rete algorithm in QLISP.
For this non-functional type of algorithm, SPAWN and QWAIT were quite
useful. (That is, I created most tasks in the network via SPAWNs and
then just had one QWAIT around the entire matching phase that waited for
all the tasks to finish.) QLAMBDA mapped naturally to the locking needed
in several parts of the algorithm (though spin locks were applicable and
much less expensive in most cases).
2. I have definitely been tripped up by the fact that the form in a
SPAWN in QLISP is turned into an implicit closure. Most of my SPAWNs are
written in the form:
(let ((m memdp) (d dp))
(spawn (t :for-effect t) (sendto *flag-part* (cons (car m) d) 'left))
)
So, in fact, for most of my uses, a spawn that does not create an
implicit closure seems more convenient.
3. I definitely needed the SPAWN construct to achieve good speedup.
Dan's extensions originally did not provide a SPAWN construct, but I
found that I could not get a good speedup with just QLET. This is mainly
because a QLET has only a fixed degree of parallelism, whereas at various
points in the algorithm, I wanted to create a number of processes which
varied depending on the outcomes of matches. Taking advantage of this
variable parallelism via QLET required accumulating all the tasks to be
spawned and then creating them all at once, as opposed to spawning them
as they are discovered via SPAWN. Also, the synchronization point
enforced by each QLET (in that it waits for all child processes to
complete before returning) was unnecessary overhead, since most of the
tasks created are completely independent of each other and the parent.
4. I had a hard time determining where the bottlenecks were in the
parallel algorithm, because of the very limited timing and nonexistent
profiling facilities.
5. One tricky part of parallelizing OPS5 was dealing with global
variables, of which it uses a lot. Some of these variables had to be
rebound so they were private to each task, and the tricky part was
determining for which variables this was necessary and when the rebinding
should be done. I don't know that there is any aspect of QLISP that
might made this problem easier to deal with; I just mention it as a
difficulty in converting existing programs to QLISP.
∂10-Dec-89 0946 hewitt@ai.mit.edu aij
Received: from life.ai.mit.edu by SAIL.Stanford.EDU with TCP; 10 Dec 89 09:46:48 PST
Received: from wheat-chex (wheat-chex.ai.mit.edu) by life.ai.mit.edu (4.0/AI-4.10) id AA16056; Sun, 10 Dec 89 12:46:57 EST
From: hewitt@ai.mit.edu (Carl Hewitt)
Received: by wheat-chex (4.0/AI-4.10) id AA21459; Sun, 10 Dec 89 12:46:45 EST
Date: Sun, 10 Dec 89 12:46:45 EST
Message-Id: <8912101746.AA21459@wheat-chex>
To: JMC@sail.stanford.edu
Cc: kirsh%cogsci@ucsd.edu
Subject: aij
Cc: Hewitt@ai.mit.edu
John,
I was unable to mail you the current draft because the SAIL mail
server is unable to accept a message of this length. So I stored a
copy for you at SAIL in "DSK:[1,CDR]AIJ.TEX". Please have a SAIL
wizard get you access to the file.
Please dont't back out of your commitment to produce a commentary on
my paper at the 11th hour.
Sincerely,
Carl
P.S. Please ignore the Federal Express package with my paper that you
receive on Monday. The wrong version was (again!) sent by mistake!
You will get the correct version on Tuesday. The title of the correct
version is ``The Role of Artificial Intelligence in Open Systems
Science''.
P.P.S. I will be at Stanford on Dec. 21 and 22. Would it be possibel
for me to meet with you then?
∂10-Dec-89 1219 ME Here is the text of AIJ.TEX[1,CDR] (1417 lines)
∂10-Dec-89 1716 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.Stanford.EDU
From: Vladimir Lifschitz <VAL@SAIL.Stanford.EDU>
IMPLEMENTING AUTOEPISTEMIC LOGIC ON A REASON MAINTENANCE SYSTEM
Kurt Konolige
SRI International
Monday, December 11, 3:15pm
MJH 252
Recent work shows that a Reason Maintenance System (RMS) can be
formalized as a type of autoepistemic theory. In this paper we
consider the inverse transformation: trying to implement an arbitrary
autoepistemic theory as an RMS. In so doing, we provide a
computationally attractive theorem-proving methodology for
autoepistemic logic.
∂11-Dec-89 0009 kirsh%cogsci@ucsd.edu [hewitt@ai.mit.edu: My AIJ paper]
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 11 Dec 89 00:09:17 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA28047
sendmail 5.61/UCSD-2.0-sun via SMTP
Mon, 11 Dec 89 00:09:08 -0800 for jmc@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA26092 for jmc@sail.stanford.edu; Mon, 11 Dec 89 00:11:36 PST
Date: Mon, 11 Dec 89 00:11 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: [hewitt@ai.mit.edu: My AIJ paper]
To: jmc@sail.stanford.edu
Included-Msgs: <8912101759.AA21477@wheat-chex>,
The message of 10 Dec 89 09:59 PST from hewitt@ai.mit.edu,
The message of 10 Dec 89 09:59 PST from Carl Hewitt
Message-Id: <19891211081131.9.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
John,
Here is the revised paper which Carl tried to send you. I'll send it in two
pieces. Please confirm when you recieve both parts.
A quick glance and an appraisel that I might send Carl would be appreciated.
You needn't concern yourself with Carl's anger; had he sent you his long
promised revisions early enough, this problem would never have arisen.
- David
% -*- Mode: TeX; Default-character-style: (SAGE:SANS-SERIF-BODY :BOLD :NORMAL) -*-
\documentstyle[11pt]{article}
\def\startline{\par\nobreak\noindent}
{\obeylines\obeyspaces\gdef\procedure{\bigbreak\begingroup
\parindent=0pt \parskip=0pt plus0pt minus0pt
\obeylines \obeyspaces \eg\let↑↑M=\startline \eg}
\gdef\endproc{\par\endgroup\bigbreak}}
\def\wyn#1{$\spadesuit${\tt #1}}
\newcommand\code[1]{\mbox{\eg{#1}}}
\newcommand\eol{\hfill\break}
\newcommand\bline{\vskip12pt plus0pt minus0pt}
\newcommand\eg{\tt}
\title{The Role of Artificial Intelligence\\in\\Open Systems Science}
\author{\copyright\ 1989 Carl Hewitt}
\date{Draft of \today}
\begin{document}
% \bibliographystyle{plain}
% \bibliography{/home/tx/wsnow/biblio}
\maketitle
\section{Abstract}
Open Systems Science is the technology and science of large scale
information systems work. Examples of this kind of work
include flexible semiconductor manufacturing, constructing a permanent
space station, and the software engineering of an international
electronic fund transfer systems. Open Systems Science
addresses issues of systems commitments, robustness, and scalability in
large scale information systems work. In contrast, Artificial Intelligence
is the science and technology of intelligent agents and robots.
Open Systems science extends classical Artificial Intelligence by
introducing methods from the sociology of science, organizations theory,
and the theory of concurrent systems.
Artificial Intelligence has created new technologies of taxonomies,
inference-based systems, and problem spaces. All of these
technologies can be useful within systems that engage in large-scale
work. However, each of these technologies is useless without the
extensive systems support that is necessary to make it work.
\section{Introduction}
All large-scale Open Systems are concurrent, asynchronous, decentralized,
and indeterminate. They are composed of numerous participants which
operate {\em concurrently\/} in order to accomplish the multitude of
tasks that are performed. They are {\em distributed\/} in order to deal
with the influx of information from many sources and to convey
information to the places where it is needed.
In any short span of time, each participant of a large-scale Open System
operates {\em autonomously\/} and {\em asynchronously\/} in accordance
with its own local needs and procedures. No truly simultaneous change of
all the participants in a large-scale Open System is ever possible, and new
information may arrive from any source at any time. Thus in general, one
participant of the system will start using new information before it
reaches the others.
Furthermore, asynchronous operation means that any large-scale Open
System is {\em indeterminate\/} in a physical sense: it does not have a
determinable current state which (together with new information that
arrives) determines its future operation. In fact, attempts to pin down
an instantaneous state of a large system by gathering more information
about the finest details of its internal operation makes the system {\em
more\/} indeterminate, because gathering the information affects its
operation. Furthermore, another large-scale Open System is needed just
to gather, interpret, and store the information about the smallest-scale
activities of another large system.
\subsection{An Illustration of Indeterminacy}
Consider a shared financial account which is accessible from multiple
sites using electronic funds transfer. For concretness consier an
implementation of the shared account discussed above in our actor core
language. Each actor behavior script has message handlers (indicated
by \code{=>} below); one of these handlers will be applicable to the
incoming communication.
Changing behavior in actors is captured by the concept of {\em
replacement behavior}. In our actor core language a replacement behavior
can be specified by a change ind the parameters of the same behavior
by using a ``{\em ready}'' command.
The code for a communication handler to process withdrawals in an account
is as follows:\footnote{A communication handler is also called a
``method'' or ``virtual procedure'' in object-oriented languages.}
\procedure %
(DefName Account
(Behavior [balance owner]
(--> [(Any Withdrawal amount) \&Serializer]
(If (>= balance amount)
{\sl Withdraw amount requested}
(Then
(Let \{[newBalance = (- balance amount)]\}
{\sl let \code{newBalance} be balance less withdrawal}
(Ready Serializer [balance = newBalance])
{\sl Account is ready for the next message}
(Return (Create WithdrawalReceipt amount owner newBalance))))
(Else
(Ready Serializer)
{\sl Account is ready for the next message}
(Complain (Create OverDraftNotice amount owner balance)))))))
\endproc
A new account can be created with balance \code{100M} and owner \code{Clark}
can be creatd and bound to an identifier named \code{Acct1} as follows:
\procedure %
(DefName Acct1 (Create Account 100M Clark))
\endproc
\noindent
Suppose that Ueda and Shapiro need to share access to Clark's account. The
following commands give them the ability to communicate with \code{Acct1}:
\procedure %
(Send Ueda Acct1)
(Send Shapiro Acct1)
\endproc
\noindent
Now if Ueda attempts to withdraw 70M from \code{Acct1} using
\code{(Send Acct1 (Create Withdrawal 70M))}, while concurrently Shapiro
attempts to withdraw 80M, then the operation of the account will be
serialized so that one of them will get a withdrawal receipt and the
other an overdraft complaint.
A denotational semantics for actor languages based on system
configurations has been developed in Actor Theory which provides a meaning for
the scripts of actor programming languages, obtained recursively by
analyzing the script as a system of communicating actors
\cite{theriault-masters}. It is true that some
types of reasoning about systems implemented in actor languages can be
carried out using a declarative reading of the programs. Such
reasoning is a special case of actor {\em serializer
induction\/}\cite{serializers}.
However, it is physically indeterminate who will have theri witdrwaw
reequest honored. No amount of knowledge of the physical
circumstances in which the withdrawal requests are mad determines the
outcome. Therefore the outcome cannot be deductively determined from
knowledge of the circumstances of the withdrawal request.
\subsection{Advantages of Open Systems}
Open Systems provide environment in which participants can be more
self reliant in the following respects:
\begin{itemize}
\item
{\bf Asynchrony:} enables each participant to operate as
quickly as possible, given local circumstances. Otherwise, the pace of
the activity of each participant would be locked to a single
scheduler down to the lowest level activities.
\item
{\bf Local autonomy:} enables each participant to react
immediately to changing circumstances. Otherwise, each participant would
have to consult a single decision maker for each decision.
\item
{\bf Late-arriving information:} enables participants to increase the
effectiveness of their decision making by taking new information into
account as it arrives. For example new geological fault lines might
be discovered under a nuclear reactor after it has been constructed.
\item
{\bf Multiple authorities:} increases pluralism,
diversity, and robustness. For example, the engineering department of a
utility wants to build a new kind of nuclear reactor which is inherently
safer than existing reactors, while the finance department maintains that
the financial risk is too great. Creativity is needed to bridge
these differences.
\item
{\bf Arm's length relationships:} enables actors to conceal their internal
activities from other other actors. For example, the Nuclear Regulatory
Commission operates at arm's length from nuclear utilities that it
regulates in order to be more effective in detecting and prosecuting
violations of its regulations.
\item
{\bf Division of labor:} specialization of functionality can increase
system effectiveness by enabling each participant to concentrate
and focus its efforts on a narrow range of goals. For example, an
electric utility has separate finance and engineering suborganizations,
each specializing in different functions: finance takes care of raising
the money, and engineering directs the construction of power plants.
\end{itemize}
\section{Trials of Strength}
Any situation in which forces are pulling in different directions
constitutes a {\em trial of strength}. All trials of strength are local in
the sense that they occur at a particular time and place among local
participants. For example negotiation provides a mechanism for
representatives of different participants to come together in a trial of
strength.
\subsection{Commitments}
A {\em commitment\/} is a course of action, usually involving several
actors, which the actors are {\em on course\/} to carry through. An
actor's {\em role\/} is its part in the commitment.
Commitments are used in this way in order to avoid becoming embroiled
in issues of {\em intentionality\/}. Instead, an actor is said to be {\em
on course\/} or {\em off course\/} with respect to an activity. This
terminology allows inanimate objects to participate in commitments.
Conflict is a trial of strength in which participants have
incompatible commitments. As we have seen, conflict is a fundamental
aspect of any Open System for a variety of reasons. Resources are
finite and limited; choices must be made about how to use them. In
addition, most organizations have built-in checks and balances which
more or less deliberately generate potentially conflicting
commitments. The commitment of the safety department to safe
conservative procedures conflicts with the need for the engineering
department to lower costs of construction. This often places them in
conflict with one another.
Sometimes the commitments of the affected participants are compatible
and there is no conflict. One participant says, ``Let's do it this way,''
and everybody agrees, so the negotiation is a trivial one, but there is
always the {\em potential\/} for conflict. No one can be certain in
advance about the outcome, so any negotiation is to that extent a trial
of strength and therefore indeterminate.
Our new discipline of Open Systems Science needs to have an
intimate understanding of the nature of commitments. In particular,
Open Systems Science is especially concerned with the nature of
systems commitments and how they relate to the commitments of
participants inside and outside a system, because the broader
systems commitments are what make large-scale projects possible.
A systems is one in which a system is on course
to carry out. For example, a utility can have an organizational
commitment to build a new power plant. The utility's Finance department
has a commitment to fund the cost of the plant, and its Engineering
department has a commitment to design and build it. Both Finance and
Engineering must work together in a fairly detailed way, and the
organizational commitment must address the specialized needs and
commitments of both participants. The organizational commitment of the
utility goes beyond the responsibilities of Financing, Engineering, and
other departments. The utility has organizational commitments and
authority that go beyond just the individual ones of its departments and
members.
We said before that a participant's responsibility for a commitment is
its part in a commitment. Alternative courses of action can affect
participant use of resources which is often a source of conflict. There
are several kinds of resources: money, space/time/material,
mechanism/technology, and sentiment.
Money and space/time/material are self-explanatory. Mechanism/\break
technology is anything that transforms the world: for example, a nuclear
power plant transforms radioactive material into electricity. Sentiment
deals with how various actors feel about each other: goodwill,
reputation, obligation, etc. For example, when a public utility says
that its reactor is safe, it is staking part of its reputation on that
statement. Also, when a utility asks for leniency on an issue from a
regulatory board, it is putting to use some of its goodwill for the
utility, which means that it is inhibited in its freedom later on.
However, the past is really gone and the future is never here. Both
planning for the future and reflecting upon past experience are
activities that take place in the present. In addition to the daily,
familiar use of resources, some activities which look to the future and
the past are also commitments. For example, planning can create a
commitment; it chooses a particular course of action that allocates
resources in one way instead of another and sets the planner on course to
keep the commitment. Creating a financial record also generates a
commitment. When a utility submits its annual report, it makes a
statement which says how much money it earned that year. Once it reports
this statement to the regulators, that's a commitment. If the regulators
come and investigate, the utility has to substantiate its claim that it
really did earn that much during the year.
Two commitments are said to be {\em conflicting\/} if they give rise to a
trial of strength which results in at least one of the commitments not
being kept. For example, the commitment of a utility to operate a nuclear
power plant is in conflict with the commitment of an environmental group
to shut it down. These two commitments are incompatible, and cannot both
be kept.
The notion of commitment used in this paper is closely related the
terminology used by those sociologists who emphasize a commitment as a
choice among incompatible alternatives
\cite{concept-of-commitment,quality-of-life}. In this terminology, {\em
choice\/} is analyzed in terms of trials of strength. Commitments have
been discussed by Winograd and Flores in their discussion of Heiddiger's
notion of ``throwness'' \cite{winograd-flores}, and by Richard Fikes
\cite{commitment-based-framework} in the context of contract nets as
developed by Reid Smith and Randy Davis \cite{contract-nets}.
% also \cite{improvised-news}.
% CH NOTE: Get {\em Improvised News\/} by Tom Shibutani, study of
% rumor in organizations, citation goes in structure of commitments.
\subsection{Open Systems Semantics}
Open Systems Semantics is the study of the meaning of Open Systems
Action. Taking any action entails changes in commitments, and that
change is the meaning of the action. This is an open-world
characterization of meaning---as opposed to previous, closed-world
attempts based on possible worlds in which the meaning of a set of
sentences is defined to be the set of all possible worlds that
satisfy the meaning conditions. Building on previous actor theory, Gul Agha
\cite{agha-phd} has developed an
open-world, mathematical semantics for concurrent systems in which the
meaning of an action is characterized by its effect on the evolution of
the system.
In general, Open Systems Action involves conflict---and therefore
indeterminacy. Current situations and events influence, but do not
determine, the future because the outcomes of numerous trials of
strength cannot be known and some trials of strength cannot even be
anticipated.
In logical semantics, representation is the mapping between a sentence
proposition and specified meaning conditions. Meaning is built on and
grows out of representation. Two participants agree about the meaning of a
sentence when they agree about the meaning conditions.
In Open System semantics, however, representation is taken to be how
effects commitments. The important difference is conveyed in the
following Open Systems Semantics slogan:
\begin{quote}
{\Large {\bf No representation without communication!}}
\end{quote}
The test of the degree to which one participant has adequately
communicated its meaning to another participant is: whether the recipient's
commitments change in the way which constitutes the meaning.
Large organizations have extensive policies, procedures, and regulations
to control the meaning of organizational actions.
Open Systems Semantics is a research programme
\cite{methodology-of-research-programmes} for studying the meaning of
Open Systems Actions. Just as there is no global synchrony or cause
and effect, the meaning of an Open Systems Action is also localized: it
begins in the participants at a particular time and place and then
can causally propagates to wider contexts.
\subsection{Requirements of Open Systems}
Analysis of commitment relationships can be used to understand how the
characteristics of Open Systems engender requirements for Open
Systems implementations:
\begin{itemize}
\item
{\bf Asynchrony:} produces indeterminacy because arriving information
must be integrated with local information. The new information can
generate new commitments that conflict with pre-existing commitments.
\item
{\bf Local autonomy:} produces conflict because as systems
commitments change, some of the new commitments will be incompatible with
pre-existing local commitments.
\item
{\bf Late-arriving information:} can produce conflict when it arrives at
an advanced stage of processing (e.g., an engineering group reports new
concerns about the capabilities to survive an earthquake just as the
utility is about to apply for a license to operate the reactor).
\item
{\bf Multiple authorities and division of labor:} can produce conflict
because the specialized commitments of multiple authorities may be
incompatible and come into conflict.
\item
{\bf Arm's length relationships:} can produce conflict because the
internal commitments of other systems are not visible. This can
increase the severity of conflict because other systems may develop
entrenched incompatible commitments before the conflict is discovered.
\end{itemize}
Another underlying constraint is {\em continuous operation}. In many
cases a system cannot ``take a vacation'' in order to get itself
into better shape. It must continue operating though perhaps
at some reduced level of performance.
\subsection{Dealing with Unanticipated Conflict}
Robustness is keeping commitments in the face of unanticipated conflict.
Keeping a commitment often entails keeping subcommitments. For
example, the commitment to operate the Diablo Canyon reactor requires
keeping two subcommitments: constructing the plant and licensing it.
The licensing subcommitment might go smoothly at first and then run
into unanticipated conflict when seismologists discover new geological
faults near the location of the reactor. The robustness of the
commitment of the utility in part depends upon its ability to deal with
whatever unexpected trial of strength arises. In this case
the utility engages in a negotiation (i.e. a trial of strength)
with the Nuclear Regulatory Commission.
\section{Negotiation}
Negotiation is an important way in which a system achieves
robustness (i.e. keeping commitments in the face of conflict). A
negotiation is a discourse in which the participants make
representations held in common which can engender joint actions that
are new joint commitments. The new commitments will often in turn
lead to more conflict with pre-existing commitments of individual
participants as they make adjustments for the new commitments.
An organizational commitment of a utility to operate a nuclear power
plant leads to giving its Finance and Engineering departments more
specialized subcommitments. Engineering has the responsibility to
construct the reactor and Finance has the responsibility to raise the
money for construction. Negotiation gives both participants an
opportunity to negotiate how the utility can keep its commitment to
operate the plant. The commitment of Engineering is represented in its
consruction plans and the commitment of Finance in represented in its
financial plan. Each is committed to the others commitment.
Participants come together to work out a systems response to potential
conflict. Each participant has ongoing commitments to attend to, and
needs a way of figuring out how to allocate its resources. Instead of
being required to turn its full attention to the new commitment, each
participant can separate out a subpart which will be devoted to the
new commitment.
Negotiation can help a system to meet its existing commitments
while developing new ones---which is important for attaining
robustness. Negotiation also creates overall, systems
commitments---which are essential to scaling. Thus, systems to
support mechanisms that can move a negotiation forward, and determine
what progress is being made.
Insights gleaned from the social sciences (law, sociology,
anthropology, organizations theory, and philosophy of science) can help
us create systems that support systems commitments, robustness, and
scalability. Human organizations have evolved methods for dealing
with conflict---and for turning them into strengths instead of
weaknesses. These methods can be adapted as a source of inspiration
for robustness in human/telecomputer systems.
\subsection{What Happens During a Negotiation}
Each participant brings its own commitments to a negotiation. These
include decision-making criteria, such as preferences among predicted
outcomes. For example: ``It is preferable to have nuclear power plants
because they lessen our dependence on unstable foreign fuel supplies''
and ``It is preferable not to have nuclear power plants because they
create a threat of the release of significant amounts of radioactivity''
Conflicts among these preferences can be negotiated
\cite{negotiations}.
During a negotiation, the participants can make moves---where each move changes its commitments. In this way, the various participants can arrive
at a joint commitment about the issues being addressed and the options
available for addressing those issues.
The various participants {\em clarify the issue\/} by discussing and
commenting upon each others' statements about the issue. This process
may expand or change the views of the various participants about the nature of
the issue. The representatives also discuss what the various {\em
options\/} are with respect to the commitments around this issue. An
option is a proposal for the rearrangement of systems
commitments. Here, discussion and commentary about options can lead to
the generation of new options, further clarification of the nature of the
issue, and further commentary on the commitments that are affected by the
issue.
\subsection{Contradictions}
{\em Contradictions\/} arise because each negotiating participant
attempts to keep its own commitments. When conflict leads to a
negotiation, each participant deliberately resists some of the
statements that other participants make to support their commitments.
Each participant uses language as a tool to further its own
commitments, and that often produces resistance and even contradictory
statements.
The expression of conflict can be a very positive aspect of negotiation.
The diversity that produced the conflict and contradictions can also
produce new ideas, suggestions, and options.
\subsection{Against Bureaucracy}
Open Systems Science needs to provide methods to keep systems from
acting bureacratically in the following ways:
\begin{itemize}
\item The rigid application of rules.
\item The arbitrary application of authority.
\end{itemize}
If systems are continually potentially involved in negotiations, how
do they ever get anything accomplished? Won't bringing negotiation to
human/telcomputer systems create even worse messes than in human
systems? While it's true that negotiations can break down and
result in deadlocks, they can also come up with creative solutions.
Two factors that help ``grease the wheels'' of negotiation are
cooperation and allies.
\subsection{Cooperation}
{\em Cooperation\/} is the process by which participants become committed to
each others' commitments. For example, Finance and Engineering might
both be committed to building a new power plant. Finance is committed to
raising the money to build the plant, and Engineering is committed to a
schedule for constructing the plant. Finance relies on Engineering's
commitment to build the plant on schedule in order to have credibility in
the financial marketplace---so Finance is committed to Engineering's
commitment. In a similar fashion, Engineering relies on Finance's
ability to provide the money to pay the construction cost as it comes due
so that construction can continue. So Engineering is committed to
Finance's commitment. Because we have these cross-commitments, we say
that Finance and Engineering are cooperating.
Another example deals with a utility and vendor. The utility generates a
purchase order for product $Q$. The vendor commits itself to shipping
product $Q$ at a certain time. The utility's commitment to the vendor's
shipping product $Q$ is represented by the purchase order. The vendor's
commitment to ship the product is dependent on the utility's commitment
to pay---as represented by the purchase order. Thus the purchase order
formalizes the cooperation between customer and vendor, and represents
the commitment of both participants to the others' commitment.
This process of mutual commitment---cooperation---is of fundamental
importance.
\subsection{Allies}
Another aspect of negotiation that helps make the negotiation process go
smoothly is the notion of an {\em ally}. One of the participants may claim an
ally: i.e., predict that in some future trial of strength, its ally would
behave in a certain way. In the payroll example, someone might claim
that if management doesn't allocate sufficient funds to meet the payroll,
the union will call for a strike!
The ally may prove to be faithful or unfaithful: faithful in the sense
that in the future trial of strength, it actually does behave as claimed.
\begin{quote}
{\em
\vbox{Glendower: I can call spirits from the vast deep.
Hotspur: Why, so can I, or so can any man; but will they come when you
do call for them?
\hfill---Shakespeare: {\em Henry IV}, {\em Act III}, {\em Scene 1}.}}
\end{quote}
Thus, when a negotiating participant claims an ally, the likely outcomes of
future trials of strength must be considered in deciding on what action
to take. In this way, claiming the support of allies can sway a
negotiation in a certain direction, but it is not necessarily decisive.
For example, a customer might invoke the Public Utility Board as an ally
when presenting a complaint to a utility that refuses to remove excess
charges from its bill. The utility needs to consider the implications of
an Utility Board investigation in deciding how to respond to the
customer.
Claiming an ally can be a very one-sided relationship. If the utility
decides to honor the customer's claim about the probable outcome of a
Utility Board investigation, then the Utility Board would never hear
about the incident, even though it was successfully invoked as an ally.
By contrast, an alliance is a {\em mutual commitment}. The participants to an
alliance make mutual commitments of their time, money, staff, and other
resources. Alliances are important outcomes of negotiation. Almost
every negotiation will affect alliances, either by creating new ones or
by strengthening, adjusting, or weakening old ones.
\section{Outcomes of Negotiation}
Many kinds of outcomes are possible, but the following three often
occur:
\begin{itemize}
\item
A {\bf resolution} to which the participants commit themselves.
\item
A {\bf deadlock} in which the participants at this particular negotiation
cannot reach an agreement. Quite often as a result of deadlock, another
negotiation is held with different representatives, and on a different
issue: namely, the fact that the other negotiation deadlocked. ``Those
guys didn't work it out, what are we going to do about it?''
\item
An {\bf appeal}. Some of the representatives might be unhappy about
the outcome and appeal to other participants---which might set up
another negotiation to deal with the issue of what to do about the
outcome of the previous negotiation.
\end{itemize}
Participants can be stalemated in conflict for a long period of time. For
example, an environmental group can work for decades attempting to revoke
the operating license of a nuclear power plant. Maintaining a
negotiation does consume resources, however. One has to keep track of
the other participant's position, plan strategy of how to continue carrying out
the negotiation, respond to the other participant's moves all the time, and so
on. Actions like these consume time, communications, storage space, and
other resources that would otherwise be put to different use. Thus,
maintaining a negotiation is a commitment.
In some cases, a negotiation ends when one participant runs out of the resources
needed to continue. In other cases, the process explicitly provides a back-up
procedure in the sense that it leaves the conflict potentially resumable, but
ends the current negotiation. An example of this would be a state that is
determined to oppose a public utility in its attempt to operate a nuclear
powerplant. The state can oppose the utility at every stage of its attempt to
get an operating license. Suppose that at each stage the negotiation is
broken off when the Nuclear Regulatory Commission decides in favor of the
utility. After each hearing the resources committed to the negotiation are
allowed to go their own way, with the intention that a whole new negotiation
might be convened at another time. Finally just before the nuclear power
plant goes into operation, the state might ``win'' the negotiation by offering
to compensate the utility for what it would earn by operating the nuclear plant
{\em provided} that it agrees to sell the plant to the state for \$1!
\subsection{Negotiations Create Commitments}
Negotiations are important, even if no conflict emerges, because they
create systems commitments that go beyond the individual
commitments of the participants involved. Negotiations always have multiple
possible outcomes. Choices are made during a negotiation, which may
result in the creation of an systems commitment: the participants agree
on a particular course of joint action. The negotiation might prove to
be a trivial one in which agreement is easily reached, but the outcome
still represents an systems commitment. Late-arriving information
could have caused one of the participants to strive for a different outcome.
The significance of negotiations lies in their outcomes and the way
those outcomes affect other systems actions. For example a nation
will incrementally develop an electric power industry---and that
industry will influence energy costs, pollution levels, generating
capacity, etc. In the case of a utility constructing a nuclear power
plant with two reactors, Engineering and Finance can have a dispute as
to whether to construct both reactors concurrently, as opposed to
finishing one before starting the next. Engineering prefers building
both at once because it can overlap similar activities to bring down
the cost. Finance prefers building them sequentially because the
financial burden and risk is less. The dispute between Finance and
Engineering will have an outcome in terms of the utility's
profitability.
\subsection{Negotiation is Creative}
Negotiation is intrinsically creative. Often, the outcome is not as
predicted, or is unintended by participants, or may even be unwanted by
some participants. On the other hand, an outcome may turn out to be
better than expected. Even when a negotiation does not break new ground
and the outcome is one of those initially sought by one or more participants,
the process used to reach that outcome is fundamentally creative in the
sense that it creates an systems commitment.
As we have seen, trials of strength embody conflict (because of
incompatible outcomes), and therefore indeterminacy (because no participant can
be certain what the outcome will be). Trials of strength are the
fundamental unit of activity that we want to understand and explore. The
actual unfolding of a trial of strength is a unique performance, so
strictly speaking, a trial of strength can never be repeated. A similar
one could be staged at a different place and time, but each performance
is unique.
This cycle of commitments leading to negotiations which lead to
commitments, some of which conflict with other commitments and thus lead
to further negotiations---this cycle is the way the world works.
\subsection{The Rationale}
The {\em rationale\/} for the outcome of a negotiation is stated at the
end of the negotiation. The rationale(s) given for the outcome are
partly generated during the negotiation process as the participants discuss
the proposed options. As each participant challenges each other's positions,
new beliefs and preferences are created. As the negotiation continues, a
rationale is often created in support of a particular outcome. For
example, in a conflict between Finance and Engineering about which of two
types of plant to build, the rationales supporting the outcome may
describe:
\begin{itemize}
\item
{\bf Predicted beneficial results:} A utility justifies the development
of a new plant: ``Nuclear power will cost less than burning fossil
fuel.''
\item
{\bf Policies guiding conduct:} The management of a utility makes a
policy: ``We must follow the rules and regulations of the Nuclear
Regulatory Commission in building our plant.''
\item
{\bf Reasons tied to specific institutional roles or processes:} A
utility sells a completed, ready to go, nuclear power plant to the state
government (which plans to demolish it) with the justification that the
state has agreed to compensate the utility in other ways.
\item
{\bf Precedent:} It is traditional to run diagnostics for the nuclear
reactor on Monday morning.
\end{itemize}
Precedent may seem like a weak rationale. However, deciding according
to precedent in the absence of strong alternatives has the
consequences of predictability and stability. In the absence of
strong alternatives, using precedent is usually less costly than
constantly redoing a decision process.
The rationale becomes part of the systems history, and may become
a precedent.
This taxonomy not only describes characteristics of outcome rationales,
it also provides criteria for identifying problems and pointing out ways
in which the process can be made more effective. So the rationale is
much more than the big cheese standing up and saying, ``This course of
action will lead to wonderful results.'' Any rationale can claim
beneficial results. However, the rationale will be judged on its own
merits. Good decisions can have bad rationales, and vice versa.
\subsection{Assessment of Systems Commitments}
No one can stand outside the system and assess as system's
commitments. Anyone who wants to understand the commitments in place
must become a player and participate in system processes. All such
assessments are made within a framework of conflict: allies,
commitments, inconsitency, limitations of resources, etc. The {\em
only way\/} to assess systems commitments to become part of the
system processes.
For example, the Nuclear Regulatory Commission abolished the requirement
that localities and states must approve emergency evacuation plans before
a nuclear reactor will be granted an operating license, in large part
because of local officials who were refusing to approve the plans. The
commission felt that communities were using the evacuation plant process
to prevent nuclear plants from receiving operating licenses. After the
NRC announced that it would abolish the requirement, many participants
challenged its authority in court. They criticized the commission for
changing the rules in the middle of the licensing process.
Meta-commitments in this instance set new policies and procedures (i.e.,
new commitments) about how commitments get changed. These new
commitments address concerns about how the previous licensing procedure
was carried out (i.e, that the communities had veto power over the
emergency evacuation plans). As a result of this trial of strength, the
commission created new procedures and policies for changing its
regulations so that in the future, various participants will participate in a
better-defined process. The new procedures and policies arose from a
meta-commitment negotiation, and formed a commitment about how to change
other commitments.
∂11-Dec-89 0016 kirsh%cogsci@ucsd.edu 2 of Hewitt
Received: from ucsd.edu by SAIL.Stanford.EDU with TCP; 11 Dec 89 00:16:36 PST
Received: from cogsci.ucsd.edu by ucsd.edu; id AA28366
sendmail 5.61/UCSD-2.0-sun via SMTP
Mon, 11 Dec 89 00:16:40 -0800 for jmc@sail.stanford.edu
Received: from BULLWINKLE.cogsci.ucsd.edu (bullwinkle.UCSD.EDU) by cogsci.UCSD.EDU (4.0/UCSDPSEUDO.2)
id AA26186 for jmc@sail.stanford.edu; Mon, 11 Dec 89 00:19:11 PST
Date: Mon, 11 Dec 89 00:18 PST
From: David Kirsh <kirsh%cogsci@ucsd.edu>
Subject: 2 of Hewitt
To: jmc@sail.stanford.edu
Message-Id: <19891211081857.0.KIRSH@BULLWINKLE.cogsci.ucsd.edu>
\subsection{Authority and Responsibility}
The meta-commitment described above changed the Nuclear Regulatory
Commission's authority and responsibilities. The utilities gained power
to participate in decisions about emergency evacuation plans, and the
commission took responsibility for spelling out its processes for
changing the licensing process.
Authority is power, and power is the ability to take action (i.e., use
resources). More precisely, authority is power legitimized by the
commitments of other authorities. For example, a utility has to register
with the Secretary of State in whichever state it operates, so its
organizational power is legitimized by the power of another authority,
namely the Secretary of State. If that authority withdraws this
legitimization, the utility's authority becomes problematical.
An organization's power is its control over resources, and its
responsibilities are its part in its commitments. (Accountability is
whether or not it actually takes those actions and meets those
commitments.) So authority and responsibility are both intimately
tied to an organization's commitments.
Authority can be delegated, but responsibility cannot. Responsibility
is established by the organization's undertaking a certain set of
commitments. An actor might get some help from other
participants in meeting those commitments, but they are still the
actor's own responsibility. So in the narrow sense, the actor cannot
delegate responsibility. What it can do instead is create other
organizational arrangements that also carry the same commitment---and
hope that will be sufficient. But if it's not, that commitment is
still that actor's responsibility.
\section{Relationship to Artificial Intelligence}
The question now arises as to the relationship between Open Systems
and Artificial Intelligence. Various technologies have been developed in
Artificial Intelligence for providing a foundation for and structuring of
computation, including:
\begin{itemize}
\item microtheories
\item problem spaces
\item taxonomies
\end{itemize}
\subsection{Microtheories}
One of the most powerful ideas of science is the idea of a {\em
microtheory}. Microtheories are based on a closed-world
assumption---from a set of rules specified in advance, results can be
algorithmically checked for correctness. A spread-sheet is a good
example. The rules are the calculation procedures used in the body of
the spread-sheet. Given the previous values and formulas, an
automaton can algorithmically check whether the new values are correct
in real time. Our notion of a microtheory is very general in that
encompasses all known forms of deduction including first order logic,
nth order logics, modal logics, intuitionistic logic, relevance logic,
lambda calculi, circumscription, default logics, etc [de Kleer, Kripke,
McDermott, McCarthy, Nilsson, Reiter, Sandewall, etc.].
A microtheory has important strengths:
\begin{itemize}
\item
It is portable. A microtheories can be expressed as a stable inscription
that does not spontaneously change and can easily be moved and copied.
\item
The correctness of a derivation is algorithmically decidable solely from
the text of the derivation. Something as simple as an automaton can
decide in real time whether or not a derivation is correct.
\end{itemize}
\noindent
Within a microtheory, there are well-defined methods for dealing with any
conflict that might arise. Thus, negotiations are not very important
within a microtheory because the correctness of derivation can be algorithmically
decidable in a closed world.
Microtheories play an important role in negotiations because they can be
brought to bear on issues and provide support for commitments with
respect to those issues. For example, the utility's Finance and
Engineering departments might each have a different spread-sheet
model of the utility's financial condition with respect to the costs of a
new plant, and each representative can then bring that microtheory to
the negotiation. Their respective recommendations of how the utility
should spend its money might be contradictory. Comparing their
microtheories can help to determine what some of the underlying
conflicts are. They might discover that Finance's Comptroller does not
believe that Engineering can meet its construction schedule. Derivations
of a microtheory can be brought to bear as supporting arguments in the
negotiation. In general, however, there will be a lot of these
microtheories, and they will often have derivations that
formally contradict derivations.
Each microtheory compiles certain methods in a rigorous way. The
Comptroller tries to protect the utility against financial
difficulties---and has successfully negotiated a commitment from
Engineering that the utility will not borrow money if payments will
exceed 25 percent of its income. On the other hand, the Engineering
department maintains that more generating capacity is needed and that
with their construction schedule, the debt payment will never exceed 25
percent.
Many of the microtheories embody various commitments and allies. For
example a spread-sheet microtheory derived from the tax code can be
used to deduce the tax consequences of differing proposals, and the
participant holding this microtheory can claim the IRS as an ally
(i.e., claiming that the IRS will support the conclusions drawn).
Thus, each side of any conflict (such as whether to pursue the
construction schedule developed by Engineering) will be able to marshall
its own body of microtheories, principles, precedents, and conclusions.
Having a deductive proof based on a microtheory usually does not thereby
carry the day and win the negotiation. The other participant will usually have
a competing microtheory. So microtheories are a strength, but they're
just one tool of negotiation. A powerful tool, but just one tool.
Having a microtheory facilitates negotiation, but in general does not
determine the outcome.
\subsection{The Role of Logical Deduction}
Several different negotiating strategies can be used with microtheories.
One is to include all the {\em if}s, {\em and}s, {\em but}s and {em
wherefore}s that one can imagine. This creates a cumbersome
microtheory that attempts to cover all possible special circumstances.
For example, ``If the utility uses more than 25 percent of its income for
debt payment, {\em and if furthermore\/} it does not have lots of liquid
assets, {\em and if furthermore\/} \ldots\ then the utility should not
take on more debt.''
A different strategy is to state a very simple rule and let it unfold in
the ongoing negotiation whether any exceptions apply. So the Comptroller
says: ``If construction is delayed, then the utility will spend much more
than 25 percent of its income for debt, so we shouldn't adopt the
construction plan.'' And the other participant replies, ``Yes,
but---Engineering has a good record for completing construction projects
on schedule at close to its estimated cost. Even though Engineering is
building a new kind of plant which can burn either coal or gas, it is not
very different from what it has built before.'' Having simple
microtheories that are parsimonious, easily understood, and clear in
their causality is often a better negotiating strategy than one which
tries to stipulate in advance all of the conditions which govern the
applicability of every rule.
The participants to a negotiation do not know for certain what sorts of
rationale for action and microtheories the others will present. The
applicability of one participant's microtheories depends on what happens during
the negotiation, not on the ability to assemble a large collection of
microtheories ahead of time. Either participant might come up with a
microtheory that the other has not thought of.
The Open Systems model of representations can be used to analyze a
deductive approach that has been explored in [McCarthy, etc.]
By inserting caveats into the axioms of conflicting microtheories,
interactions among the conditions of applicability of the axioms of the
microtheories can be expressed.
For example, various factors bear on the safety of the Diablo Canyon
nuclear reactor. Suppose that we attempt to conduct the negotiation by
writing rules with explicit caveats. Consider the following two rules:
\procedure %
{\rm Rule 1:}
if trained-operators and not(caveat-1), then safe-reactor
\endproc
\noindent
Having trained operators makes for a safe reactor unless it can be shown
that caveat-1 is true. Continuing to axiomatize,
\procedure %
{\rm Rule 2:}
if earthquake-zone and not(caveat-2), then not(safe-reactor)
\endproc
\noindent
Being in an earthquake zone means the reactor is not safe unless it can
be shown that caveat-2 is true.
In this way axioms for caveats can be developed over time
and gradually improved. For example, consider the following rule:
\procedure %
{\rm Interaction Rule 1:}
if trained-operators, then caveat-2
\endproc
\noindent
Having trained operators implies that being in an earthquake zone does
not necessarily imply that the reactor is unsafe.
\procedure %
{\rm Interaction Rule 2:}
if earthquake-zone, then caveat-1
\endproc
\noindent
Also, being in an earthquake zone implies that having trained operators
does not necessarily imply that the reactor is safe.
Unfortunately, the addition of Interaction Rules 1 and 2 blocks the
applicability of Rules 1 and 2. If we have both trained operators and
an earthquake zone, Rule 1 cannot be used since earthquake-zone implies
caveat-1, and Rule 2 cannot be used since trained-operators implies
caveat-2. The following axioms are needed instead:
\procedure %
{\rm Interaction Rule $1'$:}
if trained-operators and not(caveat-Interaction-Rule-1),
then caveat-2
\bline
{\rm{Interaction Rule $2'$:}}
if earthquake-zone and not(caveat-Interaction-Rule-2),
then caveat-1
\endproc
These interaction rules are needed to prevent contradiction and they can
be highly non-modular. For example, the use of interaction rules raises
the following question: Does being in an earthquake zone imply that the
presence of trained operators implies that being in an earthquake zone
implies that the reactor is safe? Questions like this can also be
expressed as rules:
\procedure %
{\rm{Second-Order Interaction Rule 1:}}
if trained-operators, then caveat-Interaction-Rule-2
\bline
{\rm{Second-Order Interaction Rule 2:}}
if earthquake-zone, then caveat-Interaction-Rule-1
\endproc
\noindent
Again, the Second-Order Interaction Rules 1 and 2 cannot be allowed to
stand. Instead, the following rules must be used:
\procedure %
{\rm{Second-Order Interaction Rule $1'$:}}
if trained-operator
and not(caveat-Second-Order-Interaction-Rule-1),
then caveat-Interaction-Rule-2
\bline
{\rm{Second-Order Interaction Rule $2'$:}}
if earthquake-zone
and not(caveat-Second-Order-Interaction-Rule-2),
then caveat-Interaction-Rule-1
\endproc
\noindent
At this point, it becomes very difficult to understand what we are talking
about.
Combining microtheories and adding caveats to rules does not resolve
all conflicts. In the power plant construction example, joining two
microtheories together with caveats leads to the derivation of: ``Yes,
the utility will get an operating license or no, the utility will not get an
operating license.'' But the utility cannot be told ``yes or no.'' Either
they must be told ``yes,'' or they must be told ``no.'' Adding caveats to
rules makes for a cumbersome negotiating strategy that does not
respond easily to changing circumstances.
Logical deduction can model the reasoning within a given microtheory, but
it cannot settle the dispute. Attempts to combine microtheories into a
larger theory by introducing caveats ultimately leads to ``yes or no''
derivations. So logical deduction is an appropriate and valuable tool
within the microtheories held by the respective participants. Quite often,
new microtheories will need to be created in order to better understand
the issues under negotiation. In general, these new microtheories will
not be logical consequences of the microtheories that were already
familiar to the participants before the negotiation began.
\subsection{Problem Spaces}
Problem spaces \cite{ProblemSpaces} can be used as a process modeling
technique. A problem space is a microtheory that provides for:
\begin{itemize}
\item
an {\bf initial state}. For example, the initial state may be the
financial state of the utility before it begins constructing a new plant
\item
one or more {\bf operators} that are applicable to each state. For
example, selling bonds is one of the operations that a utility can
usually take to change its financial condition.
\item
one or more {\bf success states} for one or more of the participants. For
example, the utility can specify its financial goals in terms of revenue,
investment, and earnings.
\end{itemize}
Problem spaces can also be used to model an ongoing negotiation.
Negotiation usually begins with discussion about issues. This process
can be viewed as negotiation about the initial state. Then there is
discussion about how the negotiation should proceed: who will speak, what
the agenda is, and so on---which is analogous to the possible operations
applicable to each state. Also, there is discussion about what
represents a successful outcome. A problem space attempts to model the initial
situation, trajectory, and criteria for when negotiation has ended.
Thus, problem spaces can be a useful way of characterizing an ongoing
process. However, we can rapidly encounter the same difficulty we had
with microtheories: each participant to the negotiation will have its own
problem space of how the negotiation should proceed which will in general
conflict with those of the other participants.
Consider for example the negotiating process to determine whether or
not to give a nuclear plant an operating license. The initial state of the
negotiations is quite problematical. Typically there are thousands of
pages of documentation and claims that have been submitted ahead of
time by the plant owners, the nearby local communities and states, by
environmental groups, and public utility commissions. One of the
participants might say, ``Okay, we are now in this particular initial
state; these are the ways that the negotiation can move forward from
the various states it might get into; and this is what we'll count as an
outcome.'' However, in this case the participants find that they agree
on very little about the starting state of the negotiations. In many
cases the problem spaces are not worked out in such an overt form.
Participants often come to a negotiation with {\em criteria for the
outcome\/} that they initially believe would represent success for
themselves (and possibly for others as well). Also, they would come
with their own understanding of the initial situation and what
negotiation moves would be legitimate. Furthermore in this case the
Nuclear Regulatory Commission decided to change the rules for dealing
with emergency evacuation plans in the middle of the negotiation
process; which neither it nor the other participants initially conceived
as a possible operation. The participants in this case differ greatly on
their characterization of the starting state, the allowable operations, or
which states constitute successful outcomes.
In order to use a problem space for conducting a negotiation, the
participants must first reach an agreement about the initial state of
the negotiation. In general it is difficult to come to an agreement
as to exactly what are the issues at stake. The representatives often
find it difficult to specify how the organization they represent would
characterize an issue. After some representatives have stated their
understanding of an issue, others may decide to change their
characterization. Thus attempting to precisely characterize the
initial state of a negotiation can be a very problematical undertaking
that has no termination and may even deadlock. To make the situation
even more problematical, a prolonged attempt to precisely specify the
initial state has the effect of itself changing the initial state!
Furthermore some of the representatives may not want to reveal all of
their current plans and understanding concerning of the issues under
negotiation for a variety of reasons--such as being in too preliminary
a state of development to share with others. Specifying in advance
the operators that are applicable at each stage of a negotiation as
well as the success states are equally problematical.
In many cases, the problem spaces are not worked out in such an overt
form. Participants often come to a negotiation with {\em criteria for
the outcome\/} that they initially believe would represent success for
themselves (and possibly for others as well). Also, they would come with
their own understanding of the initial situation and what negotiation
moves would be legitimate. In the case of the emergency evacuation
plans, however, the Nuclear Regulatory Commission decided to change the
rules for dealing with such plans in the middle of the negotiation
process; neither the NRC nor the other participants initially conceived of
this as a possible operation.
Again, having a problem space does not guarantee what's going to happen
because each participant brings its own problem space. Participants with
conflicting commitments about a negotiation will have conflicting
problem spaces. Basically, these are conflicts over commitments on how
the negotiation should proceed. The conflicts between these problem
spaces---much as in the cases of microtheories---need to be deal with.
Given the prospects for conflict among problem spaces, it would be
difficult (and perhaps not even desirable) to try to create a problem
space in advance that governs represent the entire, ongoing negotiation.
Instead, problem spaces are probably best used by each participant to
communicate its own analysis of the current negotiating situation,
available options, and desired outcomes.
\section{Conclusions}
The new discipline of Open Systems Science extends classical Artificial
Intelligence in several aspects:
\begin{itemize}
\item
In Open Systems Science, the primary (non-numerical) indicators are
systems {\em commitments}, {\em robustness} (the ability to keep
commitments in the face of conflict), and {\em scalability}
(commitments to increase the scale of systms commitments). Open
Systems Science is grounded in large scale information systems work.
The primary indicator of success in Artificial Intelligence is the
ability to impress humans with behavior that they will call
intelligent. It is grounded in intelligent agents and robots. In
contrast to Artificial Intelligence, work can proceed on the
development of foundations for Open Systems Science without the
need to provide a characterization of ``intelligence''.
\item
In Open Systems Science, {\em representation} is the activity of
communicating with others. Without communication there is no
representation. Communication takes its {\em meaning} from how it affects
the behavior of recipients. In Artificial Intelligence, representation is
traditionally about the correspondence between a structure in an intelligent
agent and a state of affairs in the world.
\item
Open Systems Science views {\em commitment} as a {\em joint course
of action} in which the participants are {\em on course}. The {\em
responsibility} of a participant is its part in the commitment. Artificial
Intelligence has traditionally viewed commitment as a state of mind in
which there is {\em intentionality}
\cite{Cohen-Levesque}\cite{Dennet}.
\end{itemize}
In summary social processes especially those of science, technology, and
engineering \cite{Latour} inform Open Systems Science, whereas Artificial
Intelligence has traditionally turned to neurophysiology, psychology, and
cognitive science.
Open Systems Science provides a framework for analyzing
Artificial Intelligence technologies such as deductive theories,
taxonomies, and dictionaries. Conflict is ubiquitous in Open Systems.
It allows participants to consider and explore their alternatives in a
way that takes other commitments into account. As the participants to
the conflict negotiate their differences, they usually generate
justifications to support their position. They often use microtheories to
bolster their cases. Since microtheories are decontextualized, they can
be carried from place to place and used to seek additional leverage in
many different negotiations. Thus, the use of inference in microtheories
can be seen as a natural kind of specialized activity that often occurs in
the negotiation of conflict. The crucial characteristics of a microtheory
are that the rules are given in advance and that the derivations can be
checked for correctness in real time. The nonmathematical
microtheories of the participants usually conflict. Negotiation of
the conflict that arises can be a source of creativity and robustness.
% \cite{open-systems}
% \cite{robustness-reliability-and-overdetermination}
% \cite{regions-of-the-mind}
\section{Acknowledgments}
First, I would like to express my gratitude and admiration to Wyn Snow
for editorial assistance above and beyond the call of ordinary duty.
Without her help, this paper would contain unboundedly many more
mindtraps.
Second, I wish to acknowledge the aid of Elihu Gerson in repeatedly
pulling me out of intellectual quicksand and setting me back on fruitful
paths.
Third, I wish to acknowledge the help of Les Gasser, David Kirsh, Bruno
Latour, John McCarthy, and Susan Leigh Star for pushing forward in new
directions as well as helping to reconceptualize old ones.
Fourth, I wish to thank members of the Message Passing Semantics
Group for helping to find obscurities and errors.
Fifth (and perhaps most important), I wish to thank Randy Fenstermacher,
Ron Flemming, Sue Gerson, Fanya Montalvo, John Stutz, and other close friends for
helping me to continue to grow.
\section{Related Work}
\vbox{\noindent
Hewitt, Carl,
``Viewing Control Structures as Patterns of Passing Messages,''
{\em A.I.\ Journal}, Vol.~8, No.~3, June 1977, pp.~323--364.}
This paper re-examined the issue of control structures in Artificial
Intelligence. Control structures were previously defined as looking for
the best choice in moving from the current global state to the next one.
The control structure was supposed to accomplish this either by guiding
the production system or by guiding a theorem-prover that was attempting
to search through the realm of possibilities. This paper pointed out that
traditional programming language control structures (such as iteration,
recursion and co-routine) could be analyzed in terms of patterns:
stereotypical or stylized patterns of communication among different
participants. Instead of looking at the behavior of an {\em individual\/}
intelligent agent as Newell and Simon did, this paper initiated the idea
that communities of people are a primary existence proof and analog for
how to extend these ideas.
\bigskip\vbox{\noindent
Kornfeld, William A.\ and Carl Hewitt, ``The Scientific Community
Metaphor,'' {\em IEEE Transactions on Systems, Man, and Cybernetics},
Vol.~SMC-11, No.~1, January 1981, pp.~24--33.}
This paper introduced several important concepts into the Artificial
Intelligence arena, and further develops the ideas Hewitt first discussed
in ``Viewing Control Structures.'' It uses the scientific community as a
model for the problem-solving process, and speaks generally about how
principles and mechanisms of scientific communities might be incorporated
into the problem-solving technology of Artificial Intelligence. Several
fundamental properties of scientific communities have nice analogs for
computing systems that aspire to intelligent behavior. Among these
properties are monotonicity, commutativity, parallelism and pluralism.
The paper also introduces the notion of having sceptics as well as
proponents of different kinds of ideas, and explicates how those kinds of
questions can be investigated concurrently.
\bigskip\vbox{\noindent
Kornfeld, William Arthur,
``Concepts in Parallel Problem Solving,''
Ph.D.\ Thesis, Dept.\ of EECS, MIT, February 1982.}
This is a further development of the work in ``The Scientific Community
Metaphor.'' Kornfeld here shows that by developing a concurrent process
that has critics as well as proponents of ideas, the amount of resources
consumed can, in some cases, be vastly reduced. This results in a kind of
combinatorial implosion instead of the usual combinatorial explosion where
the number of alternatives proliferate indefinitely. Such exponential
proliferation of possibilities is typical of backward-chaining reasoning.
The negotiation described here is very primitive in form, and consists of
entering absolute objections---a very cut-and-dried situation. We would
like to apply this type of process in more relaxed situations where one
has less hard knowledge, and the objections aren't guaranteed to be always
fatal to what they're objecting to.
\bigskip\vbox{\noindent
Barber, Gerald Ram{\'o}n,
``Office Semantics,''
Ph.D.\ thesis, Dept.\ of EECS, MIT, February 1982.}
This paper shows how the viewpoint mechanism introduced in Kornfeld and
Hewitt's ``Scientific Community Metaphor'' can be used to model changing
situations in terms of multiple points of view. It also introduces some
of the kinds of mechanisms for dealing with contradictory microtheories.
\bigskip\vbox{\noindent
Huberman, B.~A. editor,
{\em The Ecology of Computation}
North Holland, 1988}
This book is an excellent collection of articles which deal with the nature,
design, description, implementation, and management of Open Systems. The
articles are grouped in three major sections. Papers in the first section
deal with general issues underlying Open Systems, studies of computational
ecologies, and their similarities with social organizations. Papers in the
second section deal with implementation issues of distributed computation, and
those in the third section discuss the issues of developing suitable languages
and information media for Open Systems.
\bigskip\vbox{\noindent
Stefik, Mark,
``The Next Knowledge Medium,''
{\em The AI Magazine}, Vol.~7, No.~1, Spring 1986, pp.~34--46.}
Stefik describes the growth and spread of cultural knowledge: the kinds of
things that communities of humans do---and shows how the existence of a
technical infrastructure (such as railroads) can greatly facilitate and
accelerate cultural change. Our current knowledge market is static and
pretty much confined to inscriptions: things that can be reduced to a
string of bits (such as a diagram or sentence or literary work) and thus
transported and copied at very small price. Stefik portrays a dynamic
knowledge market that would supplement our current product market. It
would move intelligent models around that have the capability of taking
action. An active knowledge medium could interact with both its human
users and with various kinds of expert systems. He also describes several
current projects, such as the Co-Lab at Xerox Park, that are beginning to
show rudimentary characteristics of an active knowledge medium.
\bigskip\vbox{\noindent
Alvarado, Sergio J., Michael G.~Dyer and Margot Flowers,
``Editorial Comprehension in OpEd Through Argument Units,''
Fifth National Conference on Artificial Intelligence,
Aug 11--15, 1986, Philadelphia, PA,
{\em AAAI-86}, Vol.~1, pp.\ 250--256.}
This paper shows how arguments can be diagrammed in much the same way that
debate contests are often diagrammed by their judges. Such diagramming
examines the beliefs, the tree-structure of the supporting beliefs, and
the way one side can attack the other side's beliefs. (There are really
two kinds of important relationships between the two sides: support
relationships and attack relationships.) The paper presents an analysis
that looks at both the achievement of plans and goals, and the development
of editorials that critique other sides, showing how other sides have
beliefs that are supporting to the opinion that's being reported. This is
quite interesting work in terms of starting to build technology that can
do argument analysis, because that's an important component of
negotiation. Of course, there are other kinds of representations---as John
McCarthy pointed out---in terms of making threats and other kinds of
speech acts, but argument analysis is certainly a very important component
of negotiation.
\bigskip\vbox{\noindent
Devereux, Erik August,
``Processing Political Debate: A Methodology for Data Production
with Special Application to the Lincoln-Douglas Debates,''
B.S.\ thesis, MIT Dept.\ of Political Science, June 1985.}
Devereux expands on something very similar to the argument units in
Alvarado {\em et. al.}. Devereux takes the whole of the Lincoln-Douglas
debates and attempts to identify both attacking statements between Douglas
and Lincoln and supporting links within the individual arguments
themselves. Interestingly enough, there are no supporting links between
the two debaters, so in that respect, the argument units of Alvarado {\em
et al}.\ represent an advance over the analysis that was done by Devereux.
\begin{thebibliography}{9999}
\bibitem[Agha 1986]{agha-phd}
Agha, G., {\em Actors: A Model of Concurrent Computation in Distributed
Systems}, Cambridge, MA: MIT Press, 1986.
\bibitem[Becker 1960]{concept-of-commitment}
Becker, Howard S., ``Notes on the Concept of Commitment,'' {\em American
Journal of Sociology}, Vol.~66, July 1960, pp.~32--40.
\bibitem[Clinger 1981]{clinger-phd}
W.~D.~Clinger. {\em Foundations of Actor Semantics}. AI-TR-~633, MIT
Artificial Intelligence Laboratory, May 1981.\filbreak
\bibitem[Fikes 1982]{commitment-based-framework}
Fikes, Richard E., ``A commitment-based framework for describing informal
cooperative work,'' {\em Cognitive Science}, Vol.~6, 1982, pp.~331--347.
\bibitem[Gerson 1976]{quality-of-life}
Gerson, Elihu M., ``On the Quality of Life,'' {\em American Sociological
Review}, Vol.~41, October 1976, pp.~793--806.
\bibitem[Hewitt and Atkinson 1979]{serializers}
Hewitt, Carl and Atkinson, Russell,
``Specification and Proof Techniques for Serializers,''
{\em IEEE Transactions on Software Engineering}
Vol.~SE-5, No.~1, January 1979, pp.~10--23.
\bibitem[Hewitt and Baker 1977]{laws}
C.~Hewitt and H.~Baker. ``Laws for communicating parallel processes.''
In {\em 1977 IFIP Congress Proceedings}, IFIP, August 1977, pp.~987--992.\filbreak
\bibitem[Latour 1987]{science-in-action}
Latour, Bruno, {\em Science In Action}, Cambridge, MA: Harvard University
Press, 1987.
\bibitem[Shapiro 1987]{shapiro-cp}
E.~Shapiro. ``A subset of concurrent prolog and its interpreter.'' In
{\em Concurrent Prolog: Collected Papers}, Cambridge, MA: MIT Press,
1987, pp.~27--83.\filbreak
\bibitem[Smith and Davis 1981]{contract-nets}
Smith, R. and Davis, R., ``Frameworks for cooperation in distributed
problem solving,'' {\em IEEE Transactions on Systems, Man, and
Cybernetics}, Vol.~SMC-11, 1981, pp.~61--70.
\bibitem[Star 1983]{simplification-in-scientific-work}
Star, S.L., ``Simplification in scientific work: An example from
neuroscience research,'' {\em Social Studies of Science}, Vol.~13, No.~2,
1983, pp.~205--228.
\bibitem[Strauss 1978]{negotiations}
Strauss, Anselm, {\em Negotiations}, Jossey-Bass, 1978.
\bibitem[Theriault 1983]{theriault-masters}
D.~Theriault. {\em Issues in the Design and Implementation of Act2}.
Technical Report~728, MIT Artificial Intelligence Laboratory, June 1983.\filbreak
\bibitem[Winograd and Flores 1987]{winograd-flores}
Winograd, Terry, and Flores, Fernando, {\em Understanding Computers and
Cognition}, Reading, MA: Addison-Wesley, 1987.
\end{thebibliography}
\end{document}
∂11-Dec-89 1100 JMC
Invitation
∂11-Dec-89 1250 siegman@sierra.STANFORD.EDU Re: Stanford costs and overhead
Received: from sierra.STANFORD.EDU by SAIL.Stanford.EDU with TCP; 11 Dec 89 12:49:39 PST
Received: by sierra.STANFORD.EDU (3.2/4.7); Mon, 11 Dec 89 12:49:14 PST
Date: Mon, 11 Dec 89 12:49:14 PST
From: siegman@sierra.STANFORD.EDU (Anthony E. Siegman)
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: Re: Stanford costs and overhead
In-Reply-To: Your message of 08 Dec 89 0830 PST
Message-Id: <CMM.0.88.629412554.siegman@>
1) I do myself agree with you that some action is probably needed --
at least picking up the two-by-four and waving it menacingly, if not
actually striking a blow.
2) I can't speak for others; but the evidence seems to be that there
is substantial hostility among some of the overhead paying faculty.
(I was unable to attend the meeting last week, but gather that it was
a fairly heated exchange.)
3) For myself, whether I'm held back by timidity or by loyalty to
Stanford is hard to say; but going directly to the government in some
fashion is probably not an option I'd like to take; has an aspect of
"fouling one's own nest". I'd rather work "within the system" for a
while longer.
4) My own protest actions at this point include:
a) No more academic year salary offsets (not an insignificant action
in the EE milieu).
b) No more service, ever again, on any "overhead committees" in any
way, shape or form (including APC).
c) No more service on any other Univ or Dept committees, or orals
committees for other departments, or any other form of voluntary
service to the university, on grounds that one simply can't at this
point maintain an outside-sponsored research program with the present
overhead burden; do an acceptable job of teaching; and do anything
else in addition for the university.
d) Make clear that I would no longer myself urge any promising
research star to come here as a new faculty member, with the present
overhead burden and working conditions.
5) I've considered a "faculty revolt" scenario (i.e., "PI revolt") at
times. I believe a first step would be formation of an organization
(labor union, if you wish) of sponsored-research PIs **completely
independent of the university**, including perhaps hiring a staff
member/lawyer/agent paid by the PIs (from unrestricted funds?) who was
NOT a university employee. (Analogy to agents who represent
professional athletes.) Then, negotiate with the university as a
group. Might even demand that the University pay (but NOT control)
our agent.
I'm not prepared to lead in forming such a group.
Other options:
a) Small group of faculty members go to Provost and say, here are our
renewal proposals for the programs we're presently running. Do you
want us to submit them with an overhead rate reduced by 15% from current
university rate -- or shall we not submit them at all? (thereby
terminating large amounts of student support, and cutting back on our
research activity within Stanford, in favor of taking our professional
efforts to consulting roles in local industry).
b) Running out of steam in this response . . .
∂11-Dec-89 2000 JMC
Etherington et al for Rathman
∂12-Dec-89 1110 harnad@clarity.Princeton.EDU BBS Penrose Book Review
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 12 Dec 89 11:10:36 PST
Received: from psycho.Princeton.EDU by Princeton.EDU (5.58+++/2.25/mailrelay)
id AA26704; Tue, 12 Dec 89 14:08:45 EST
Received: by psycho.Princeton.EDU (4.0/1.95)
id AA12381; Tue, 12 Dec 89 14:10:59 EST
Date: Tue, 12 Dec 89 14:10:59 EST
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8912121910.AA12381@psycho.Princeton.EDU>
To: jmc@sail.stanford.edu
Subject: BBS Penrose Book Review
Dear. Dr. McCarthy:
Invitation to prepare a BBS Multiple Book Review:
You are invited to write a review of Roger Penrose's, "The
Emperor's New Mind," which has been selected for multiple book
review treatment in Behavioral and Brain Sciences (BBS), an
international, interdisciplinary journal of open peer commentary.
(An abstract of the book appears below.)
The procedure is much the same as with commentaries on target articles
except that it is the book itself that is under review, not the
article-length precis that will co-appear with the reviews and author's
Response (to permit BBS readers who have not read the book to assess
the exchange).
Please let me know by email or regular mail whether you are willing to
submit a (1000-word) review of the book by: February 5, 1990.
If you do not have a copy of the book, please indicate this, and we
will have the volume sent to you by return first class or air mail.
(Please reply quickly as there are only a fixed number of copies
available.)
Note that as multiple reviews will be coappearing, you need only
comment on the aspects of the book relevant to your own specialty and
interests, not necessarily the book in its entirety. Any substantive
criticisms--including points calling for a detailed and substantive
response from the author--are appropriate. Hence, investigators who
have already reviewed or intend to review this book elsewhere are
still encouraged to submit a BBS review specifically written with this
specialized multilateral review-and-response feature in mind.
With many thanks for your collaboration in this book review treatment.
Stevan Harnad
Editor
____________________________________________________________________
THE EMPEROR'S NEW MIND:
CONCERNING COMPUTERS, MINDS AND THE LAWS OF PHYSICS
Roger Penrose
Rouse Ball Professor of Mathematics
University of Oxford
The Emperor's New Mind is an attempt to put forward a scientific
alternative to the viewpoint of "Strong AI," according to which mental
activity is merely the acting out of some algorithmic procedure. John
Searle and other thinkers have likewise argued that mere calculation
does not, of itself, evoke conscious mental attributes, such as
understanding or intentionality, but they are still prepared to accept
that the action of the brain, like that of any other physical object,
could in principle be simulated by a computer. In my book I go further
than this and suggest that the outward manifestations of conscious
mental activity cannot even be properly simulated by calculation. To
support this view I use various arguments to show that the results of
mathematical insight, in particular, do not seem to be obtained
algorithmically. The main thrust of this work, however, is to present
an overview of the present state of physical understanding and to show
that an important gap exists at the point where quantum and classical
physics meet, and to speculate on how the conscious brain might be
taking advantage of whatever new physics is needed to fill this gap, in
order to achieve its non-algorithmic effects.
∂12-Dec-89 1146 VAL index
I'm done with the index for your book. Would you like to take another look
at it before I mail it to Ablex?
∂12-Dec-89 1252 harnad@clarity.Princeton.EDU re: BBS Penrose Book Review
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 12 Dec 89 12:52:02 PST
Received: from suspicion.Princeton.EDU by Princeton.EDU (5.58+++/2.25/mailrelay)
id AA27691; Tue, 12 Dec 89 15:49:53 EST
Received: by suspicion.Princeton.EDU (4.0/1.95)
id AA04830; Tue, 12 Dec 89 15:04:07 EST
Date: Tue, 12 Dec 89 15:04:07 EST
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8912122004.AA04830@suspicion.Princeton.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: re: BBS Penrose Book Review
John, unfortunately, we cannot publish material that is
appearing or has appeared elsewhere. If you want to prepare
a 1000-word piece specifically for BB (and for Penrose to
respond to), that would be fine; but we couldn't copublish
all or part of the AMS Bulletin review. -- Stevan
∂12-Dec-89 1300 JMC
invitation
∂12-Dec-89 1307 scales@Neon.Stanford.EDU Re: cs499
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 12 Dec 89 13:07:51 PST
Received: by Neon.Stanford.EDU (5.61/25-eef) id AA20988; Tue, 12 Dec 89 13:08:00 -0800
Date: Tue, 12 Dec 89 13:08:00 -0800
From: Daniel J. Scales <scales@Neon.Stanford.EDU>
Message-Id: <8912122108.AA20988@Neon.Stanford.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: Re: cs499
Sorry I didn't reply right away to your message about my CS499 units. I
came by several times on Friday afternoon to talk with you, but you were
busy, and then I forgot to reply by mail over the weekend.
I only signed up for cs499 to fill out my 9 units of courses. I signed up
with you since I'm doing a QLISP RAship this quarter. I have talked a bit
with Anoop, since his thesis is related to the application (OPS5) that I'm
parallelizing in QLISP, but mostly I've been working on my own with advice
from Dan and Joe. I just sent out a summary of some of my results to the
QLISP mailing list. If you'd like to meet, I'm mostly free for the rest of
the week.
Let me know if there is any problem with you giving me a grade.
Dan
∂12-Dec-89 1309 harnad@clarity.Princeton.EDU re: BBS Penrose Book Review
Received: from Princeton.EDU by SAIL.Stanford.EDU with TCP; 12 Dec 89 13:09:53 PST
Received: from cognito.Princeton.EDU by Princeton.EDU (5.58+++/2.25/mailrelay)
id AA27920; Tue, 12 Dec 89 16:08:14 EST
Received: by cognito.Princeton.EDU (4.0/1.95)
id AA08780; Tue, 12 Dec 89 16:10:25 EST
Date: Tue, 12 Dec 89 16:10:25 EST
From: harnad@clarity.Princeton.EDU (Stevan Harnad)
Message-Id: <8912122110.AA08780@cognito.Princeton.EDU>
To: JMC@SAIL.Stanford.EDU
Subject: re: BBS Penrose Book Review
John, yes you can certainly cross-refer, as long as the BBS
piece is self-contained. We need the review by Feb 5, and
the usual BBS format applies (title, references, etc.). I
look forward to seeing your review! Cheers, Stevan
∂12-Dec-89 1613 weening@Gang-of-Four.Stanford.EDU Text needed for DARPA proposal
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 12 Dec 89 16:13:08 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA22004; Tue, 12 Dec 89 16:13:45 -0800
Date: Tue, 12 Dec 89 16:13:45 -0800
From: weening@Gang-of-Four.Stanford.EDU (Joe Weening)
Message-Id: <8912130013.AA22004@Gang-of-Four.Stanford.EDU>
To: jmc@sail.Stanford.EDU, rpg@sail.Stanford.EDU,
iam@Gang-of-Four.Stanford.EDU, pehoushek@Gang-of-Four.Stanford.EDU,
clt@sail.Stanford.EDU, stickel@warbucks.ai.sri.com,
waldinger@warbucks.ai.sri.com
Subject: Text needed for DARPA proposal
For the DARPA proposal abstract that I'm currently putting together
(parallel theorem proving and benchmarks), I currently have everyone's
background information and publications, but Carolyn and I feel that
it will be better to include some text describing our background in
more detail.
A paragraph or two from each person would be fine. Please describe
your interests and background in parallel processing, theorem proving,
Lisp, benchmarks, and anything else you feel is relevant to this
proposal. Thanks.
∂12-Dec-89 1754 PKR re: chapter 4
To: JMC
CC: PKR
[In reply to message rcvd 12-Dec-89 10:24-PT.]
All points are well taken. Category theory (colimits, etc.) is
indeed a very effective tool for burying useful results where
no one will read them. I am running the proofs together with
the text in the thesis draft, because that seems the easiest to
write. If I am not quite sure the theorem is true, I want the
proof right there. I am not opposed to eventually moving the
proofs to an appendix, even for the thesis.
A ghost is any universe element not denoted by a ground term.
They mess up the proofs, so I am wishing them away for the time being.
Thanks for the comments and encouragement.
-Peter
∂13-Dec-89 0939 MPS fruitfly
Hi,
Do I understand you correctly? You want me to key
in the entire paper on fruitfly? Thanks.
Pat
∂13-Dec-89 1012 @CUNYVM.CUNY.EDU:fraenkel@wisdom.weizmann.ac.il
Received: from CUNYVM.CUNY.EDU by SAIL.Stanford.EDU with TCP; 13 Dec 89 10:12:38 PST
Received: from WISDOM.WEIZMANN.AC.IL by CUNYVM.CUNY.EDU (IBM VM SMTP R1.2.2MX) with BSMTP id 7414; Wed, 13 Dec 89 13:11:48 EDT
Received: from wisdocs.wisdom.weizmann.ac.i by wisdom.wisdom.weizmann.ac.il (3.2
id AA28081; Wed, 13 Dec 89 20:11:32 +0200
Received: by wisdocs.wisdom.weizmann.ac.il (4.12/SMI-3.2)
id AA15950; Wed, 13 Dec 89 20:10:24 -0200
Date: Wed, 13 Dec 89 20:10:24 -0200
From: fraenkel%wisdom.weizmann.ac.il@CUNYVM.CUNY.EDU (Aviezri Fraenkel)
Message-Id: <8912131810.AA15950@wisdocs.wisdom.weizmann.ac.il>
To: jmc@sail.stanford.edu
Is this Dr John McCarthy?
Re recursive computer procedures, i.e., programs that call themselves,
I was wondering whether this was a contribution to CS from its AI community?
More specifically, my question is who invented or first used recursive
programs? Could you also provide suitable references?
With many thanks, Aviezri S. Fraenkel.
fraenkel at wisdom.weizmann.ac.il
∂13-Dec-89 1701 MPS Paper
I did a real rough job of corrections. I hope I got most of
them for you. See you tomorrow.
Pat
∂14-Dec-89 0801 HK.RLS@Forsythe.Stanford.EDU memorandum
Received: from Forsythe.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Dec 89 08:01:25 PST
Date: Wed, 13 Dec 89 20:38:33 PST
To: jmc@sail.stanford.edu
From: "Bob Street" <HK.RLS@Forsythe.Stanford.EDU>
Subject: memorandum
REPLY TO 12/12/89 16:40 FROM JMC@SAIL.STANFORD.EDU "John McCarthy": memorandum
John,
Thanks; we found them. They had gotten to someone other than the
secretaries.
Best,
Bob*
To: JMC@SAIL.STANFORD.EDU
∂14-Dec-89 1304 CLT bing
called -- timothy's things are in a plastic bag in the atrium
∂14-Dec-89 1424 irvine@sumex-aim.stanford.edu Betty Scott
Received: from sumex-aim.stanford.edu by SAIL.Stanford.EDU with TCP; 14 Dec 89 14:23:53 PST
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AA12860; Thu, 14 Dec 89 14:25:57 PST
Date: Thu, 14 Dec 1989 14:25:56 PST
From: Sue Irvine <irvine@sumex-aim.stanford.edu>
To: JMC@sail.stanford.edu
Cc: irvine@sumex-aim.stanford.edu
Subject: Betty Scott
Message-Id: <CMM.0.88.629677556.irvine@sumex-aim.stanford.edu>
As we all know (alas, alas, much-hand-wringing). Betty Scott
is retiring at the end of December. She has been a most
wonderful, competent, and helpful department administrator and
friend over the better part of two decades--our "chief of
staff", so to speak.
Last night, we presented her with gifts to remember us by, one
from the department, one from the Forum, and one from "faculty
and friends."
To you "faculty and friends", I would like to address a
request to help cover the costs of the gift ($230). We've
already received about $50, so we have $180 to go.
It's OK to use money from an "unrestricted funds" account for
this. All you need to do is send to me or to Carolyn Tajnai
(Tajnai@cs) an amount and an unrestricted funds account
number.
Or you can send a check to Carolyn (made out to Carolyn
Tajnai; she laid out the money) and mail it to her either in
the Forum mailbox in MJH or to ERL 450, Mailcode 4055.
Person gifts of ten or twenty dollars, or larger gifts from
unrestricted funds accounts, will quickly close the "gift
gap."
Many thanks to all, and to Carolyn for her role in this,
Ed Feigenbaum
∂14-Dec-89 1438 stager@Sunburn.Stanford.EDU Winter TA
Received: from Sunburn.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Dec 89 14:38:46 PST
Received: by Sunburn.Stanford.EDU (5.61/25-eef) id AA10226; Thu, 14 Dec 89 14:39:50 -0800
Date: Thu, 14 Dec 1989 14:39:49 PST
From: "Claire E. Stager" <stager@Sunburn.Stanford.EDU>
To: paek@cs
Cc: jmc@sail
Subject: Winter TA
Message-Id: <CMM.0.88.629678389.stager@Sunburn.Stanford.EDU>
Eunok:
You have been tentatively assigned as a 25% TA for CS323 Winter Quarter.
Please confirm this appointment with Claire Stager (stager@sunburn) as soon as
possible.
Happy Holidays!
Claire
∂14-Dec-89 1731 @Sunburn.Stanford.EDU,@Polya.Stanford.EDU:nilsson@Tenaya.Stanford.EDU reappointments
Received: from Sunburn.Stanford.EDU by SAIL.Stanford.EDU with TCP; 14 Dec 89 17:31:19 PST
Received: from Polya.Stanford.EDU by Sunburn.Stanford.EDU with SMTP (5.61/25-eef) id AA14909; Thu, 14 Dec 89 17:31:59 -0800
Received: from Tenaya.Stanford.EDU by Polya.Stanford.EDU with SMTP (5.61+IDA/25-eef) id AA01714; Thu, 14 Dec 89 17:30:38 -0800
Received: by Tenaya.Stanford.EDU (NeXT-1.0 (From Sendmail 5.52)/25-eef) id AA00584; Thu, 14 Dec 89 17:26:47 PST
Date: Thu, 14 Dec 89 17:26:47 PST
From: Nils Nilsson <nilsson@tenaya.Stanford.EDU>
Message-Id: <8912150126.AA00584@Tenaya.Stanford.EDU>
To: tenured@cs.stanford.edu
Subject: reappointments
There will be a meeting of the senior faculty immediately
following the regular faculty meeting, which latter
occurs on Tuesday, January 9, 1990 at 2:30 pm in
MJH 146. We will be considering the reappointment
of five of our assistant professors, namely:
Anoop Gupta
John Mitchell
David Dill
Yoav Shoham
Andrew Goldberg
The preliminary reappointment papers that have been
presented to the School of Engineering Excom (in order
to reserve a place on their agenda for possible
consideration of final papers) are available for your
inspection in my office (see Joyce). These preliminary
papers have lists of publications, vita, teaching
performance, etc. We are soliciting outside letters
on some of these candidates, and some of these
may be available before the meeting. I am presuming
that John Hennessy, Vaughan Pratt, Jean-Claude
Latombe, and Jeff Ullman will be at this senior
meeting to help with the presentation of these cases. I
hope to accomplish
the following at the Jan 9 meeting: For those
cases where a strong reappointment recommendation
is clearly and immediately indicated, we should confirm
the Department's
concurrence. For those where more information or
discussion is
needed, we should instruct the appropriate
Departmental
representatives and set a date for a subsequent
meeting in the very near future.
As mentioned in the immediately preceding message,
this is
advance notice because I will be away until Jan. 5.
Please do
mark your calendars and try to arrange your schedules
so that we can deal effectively and fairly with our
junior faculty.
Thanks,
-Nils
∂15-Dec-89 0821 VAL lunch
How about the Japanese place for today?
∂15-Dec-89 0916 VAL re: lunch
[In reply to message rcvd 15-Dec-89 09:15-PT.]
11:45 is fine.
∂15-Dec-89 1140 paek@Neon.Stanford.EDU re: Winter TA
Received: from Neon.Stanford.EDU by SAIL.Stanford.EDU with TCP; 15 Dec 89 11:40:46 PST
Received: by Neon.Stanford.EDU (5.61/25-eef) id AA16608; Fri, 15 Dec 89 11:40:33 -0800
Date: Fri, 15 Dec 1989 11:40:29 PST
From: Eunok Paek <paek@Neon.Stanford.EDU>
To: John McCarthy <JMC@SAIL.Stanford.EDU>
Subject: re: Winter TA
In-Reply-To: Your message of 14 Dec 89 1445 PST
Message-Id: <CMM.0.88.629754029.paek@Neon.Stanford.EDU>
When can I meet you?
I will be around today until 4 pm. If you prefer, I am available
any time next week.
-Eunok Paek
p.s. By the way, I am not "Mr.", but "Ms."
∂15-Dec-89 1655 jolle@cs.psu.edu
Received: from psuvax1.cs.psu.edu by SAIL.Stanford.EDU with TCP; 15 Dec 89 16:55:27 PST
Received: by psuvax1.cs.psu.edu (5.61/PSUCS-1.0)
id AA27528; Fri, 15 Dec 89 19:54:11 -0500
Date: Fri, 15 Dec 89 19:54:11 -0500
From: Reinhard von Hankleden <jolle@psuvax1.cs.psu.edu>
Message-Id: <8912160054.AA27528@psuvax1.cs.psu.edu>
To: jmc@sail.stanford.edu
Reinhard v. Hanxleden
712E West Beaver Avenue
State College, PA 16801
jolle@psuvax1.cs.psu.edu
(814) 867-2706
Prof. John McCarthy
Department of Computer Science
Stanford University, CA94305
Dear Prof. McCarthy,
I am currently applying to the Ph.D. program of the Department of Computer
Science, Stanford University. I visited your department this summer, and
I had the opportunity to talk to you about your current fields of research.
Later you introduced me to Vladimir Lifschitz and Dan Pehoushek, who
gave me then the thesis of Joseph Weening.
For 1988/89 I had a scholarship for graduate studies at Penn State, and
in spring I finished with Prof. Scott a thesis about the load balancing
of dynamic processes. Currently we are preparing two papers for submission.
The material is public readable via
ftp psuvax1.cs.psu.edu
(username) ftp
(password) jmc
cd pub/jolle
ls -l
get ...
quit
I read the articles recommended by Prof. Lifschitz and Weening's thesis,
and both non-monotonic reasoning and parallel Lisp are of special interest
of mine. In particular I noticed phenomena described in the thesis which
occurred similarly in my work. For example, dynamic partitioning is well
feasible and, even more, can be implemented algorithm independently.
A key difference between the two projects was the used architecture,
I used a message passing MIMD system (iPSC/2). The scalability was limited
only by the algorithms used, but the scheduling became more difficult in
terms of load balancing and communication.
Do you think about implementing Lisp on message passing architectures,
or do you have other projects integrating parallel computing and AI?
I would be very glad to hear from you again.
-- Reinhard v. Hanxleden
∂17-Dec-89 0731 jf@research.att.com
Received: from research.att.com (inet.att.com) by SAIL.Stanford.EDU with TCP; 17 Dec 89 07:31:42 PST
From: jf@research.att.com
Date: Sun, 17 Dec 89 10:28:28 EST
To: jmc@sail.stanford.edu
John:
I read your CACM contribution about email and agreed with it
completely. I've been moaning about FAX for about a year now. I've
also been wondering why AT&T wastes its time trying to build PC's
instead of figuring out how to sell email services to homes and businesses
that don't have them.
Thank you for expressing those ideas so well and so publicly. I hope that
your article causes the bureaucracies to think about things, if not to act
soon.
Joan Feigenbaum
(Stanford PhD program alumna, in case you don't remember me.)
∂17-Dec-89 1224 VAL
TIMES.TEX[1,VAL] is the letter I want to send to the New York Times.
What do you think about it?
∂18-Dec-89 1213 VAL re: reply to message
[In reply to message rcvd 18-Dec-89 01:31-PT.]
Thanks. Actually, Shipler's story is not about himself, but about another
correspondent, Christopher Wren. Vladik tells me that some another reporter
had a similar complaint about Solzhenitsin years ago.
∂18-Dec-89 1322 VAL Re: openings at Berkeley
∂18-Dec-89 1316 wilensky%larch.Berkeley.EDU@Berkeley.EDU Re: openings at Berkeley
Received: from larch.Berkeley.EDU by SAIL.Stanford.EDU with TCP; 18 Dec 89 13:16:27 PST
Received: by larch.Berkeley.EDU (5.57/1.25)
id AA00310; Mon, 18 Dec 89 13:16:56 PST
From: wilensky%larch.Berkeley.EDU@berkeley.edu (Robert Wilensky)
Message-Id: <8912182116.AA00310@larch.Berkeley.EDU>
To: Vladimir Lifschitz <VAL@sail.stanford.edu>
Subject: Re: openings at Berkeley
In-Reply-To: Your message of 15 Dec 89 15:40:00 PST.
<13Kyay@SAIL.Stanford.EDU>
Date: Mon, 18 Dec 89 13:16:52 PST
Dear Vladimir,
While we are recruiting in ``computer science'', the official
high-priority area is currently software. However, we are always in
principle looking for top people, and special cases can be made. Also,
in the past, we have made offers to folks who weren't in the area
identified as our top priority, when no one in the designated area
seemed too exciting, and an exceedingly interesting candidate in another
area was available.
A big factor in all this, however, is that this department is quite keen
on hiring junior people. The converse is also true, namely, that hiring
of more senior people is very rarely done. When it is, an extraordinary
case needs to be made.
In your own case, my suspicion is that it would be difficult to make the
case that your work ideally suits our programmatic needs. In English,
what this means is that folks around here would view you as a bit too
``theoretical'' for our current needs. That is, if and when I can make
the case for another AI position, it will probably be for someone for
whom system building is more of a direct concern.
Personally, I think it would be very interesting and useful to have
someone with your expertise and abilities around this department.
However, I can't say that I'm convinced your area of interest should be
our top AI hiring priority, especially since we may not have another AI
position for some time. Under these circumstances, I think that
something working out would be rather a long short.
Another possibility you might want to investigate further is to see if
something is available at the International Computer Science Institute
in Berkeley. This is this European-US cooperative center that Jerry
Feldman was brought in to head. It's also possible that some joint
appointment could be feasible. I don't want to falsely encourage you in
this respect, because I don't have any reason to believe that Feldman et
al. would want to hire someone with a strong logic orientation.
However, I thought for completeness I should mention it to you.
Anyway, these are my beliefs about the current situation here. I'd be
happy to discuss any of this further with you if you'd like. In any
case, best of luck to you!
Robert
∂18-Dec-89 1457 MPS Car
Kipp Skelton, Almaden Mazda 408-723-8800
MX6GY automatic (has cars)
its a go on the fleet deal.
Pat
∂18-Dec-89 1610 VAL Towers
The predicate tower(p,x), "p is a tower with the base x", can be defined
in second order logic as follows:
∀xp[∀y(p(y)≡y=x)⊃tower(p,x)],
∀xypq[tower(p,x)∧on(x,y)∧∀z(q(z)≡(p(z)∨z=y))⊃tower(q,y)].
Then we need to circumscribe "tower". This is an example of "higher order
circumscription", because one of the arguments of the circumscribed predicate
is a predicate variable.
∂18-Dec-89 2000 JMC
Searle bbs reference
∂19-Dec-89 1209 masahiko@sato.riec.tohoku.ac.jp NFS/JSPS proposal
Received: from argus.Stanford.EDU by SAIL.Stanford.EDU with TCP; 19 Dec 89 12:09:38 PST
Received: from mtfuji.gw.u-tokyo.ac.jp ([128.167.64.2]) by argus.Stanford.EDU with TCP; Mon, 18 Dec 89 23:15:46 PST
Received: from relay.cc.u-tokyo.ac.jp by mtfuji.gw.u-tokyo.ac.jp (5.61/2.7W)
id AA26760; Tue, 19 Dec 89 02:28:26 -0500
Received: from ccut.cc.u-tokyo.ac.jp by relay.cc.u-tokyo.ac.jp (5.61/2.7W)
id AA05061; Tue, 19 Dec 89 16:28:46 +0900
Received: by ccut.cc.u-tokyo.ac.jp (5.61/6.4J.6-ut2.2)
id AA28662; Tue, 19 Dec 89 16:28:14 +0900
Received: from clover.sato.riec.tohoku.ac.j by hirose.cc.tohoku.ac.jp (3.2/6.4J.6-Dec09)
id AA10678; Tue, 19 Dec 89 14:58:25 JST
Received: by clover.sato.riec.tohoku.ac.jp (4.0/6.4J.6-Sep01)
id AA05840; Tue, 19 Dec 89 14:56:57 JST
Received: by mars.sato.riec.tohoku.ac.jp (4.0/6.4J.6-Sep27)
id AA13517; Tue, 19 Dec 89 14:57:50 JST
Date: Tue, 19 Dec 89 14:57:50 JST
From: masahiko@sato.riec.tohoku.ac.jp (Masahiko Sato)
Return-Path: <masahiko@sato.riec.tohoku.ac.jp>
Message-Id: <8912190557.AA13517@mars.sato.riec.tohoku.ac.jp>
To: jmc@sail.stanford.edu, clt@sail.stanford.edu
Subject: NFS/JSPS proposal
I have just learned from JSPS that our proposal for the US/Japanese
collaboration on research in new foundations of computer science has
been accepted. It is a good news, isn't it?
I have to submit a plan for the first year, that is April 1990 - March
1991, by February 9, 1990. Perhaps we should start talking about
concrete plan soon.
masahiko
∂19-Dec-89 1255 CLT NFS/JSPS proposal
To: "@JAPAN.DIS[1,CLT]"@SAIL.Stanford.EDU
It appears that the proposal has been approved on both sides of the water
with a start date of 1-april-90.
For the US part the budget has been reduced somewhat (equivalent to
8 trips to Japan and no local travel -- but when its comes to spending
they are generally flexible).
Masahiko has to submit a plan for the first year (April 1990 - March 1991)
by February 9, 1990 (JSPS is bigger on paper work than NSF), so we should
think about when we might be able to travel to Japan and when we could
host the Japanese here.
One possible plan Ito and I discussed was to have the entire US group come
to Japan at the same time for a concentrated week of discussions with
the group of Japanese participants And dually (the alternate year) to have
the entire group of Japanese participants come to the US (presumably Stanford)
for a similar session.
Alternative suggestions would be welcome.
∂19-Dec-89 1640 @IU.AI.SRI.COM,@sunset.ai.sri.com:konolige@ai.sri.com nonmon90 workshop
Received: from IU.AI.SRI.COM by SAIL.Stanford.EDU with TCP; 19 Dec 89 16:40:06 PST
Received: from sunset.ai.sri.com by IU.AI.SRI.COM via SMTP with TCP;
Tue, 19 Dec 89 16:17:29-PST
Received: from kehoe.ai.sri.com by sunset.ai.sri.com (4.0/4.16) id
AA01814 for jmc@sail.stanford.edu; Tue, 19 Dec 89 16:17:05 PST
Received: by kehoe.ai.sri.com (4.0/4.16) id AA01640 for
val@sail.stanford.edu; Tue, 19 Dec 89 16:17:02 PST
Date: Tue, 19 Dec 89 16:17:02 PST
From: Kurt Konolige <konolige@ai.sri.com>
Message-Id: <8912200017.AA01640@kehoe.ai.sri.com>
To: jmc@sail.stanford.edu
Subject: nonmon90 workshop
Cc: val@sail.stanford.edu
John, I'd be delighted to have you at the nonmonotonic reasoning
workshop. I assume you've seen the call for participation and
know the place and time. I think it would be nice if you'ld present
some kind of retrospective on the field, or at least the part of it
that relates to your own work. We (the program committee) haven't
decided on the format of talks and panel discussions yet, but I'm
sure everyone would be interested in hearing from you.
--kk
∂19-Dec-89 1659 weening@Gang-of-Four.Stanford.EDU Alliant visit
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 19 Dec 89 16:59:09 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA03076; Tue, 19 Dec 89 16:59:56 -0800
Date: Tue, 19 Dec 89 16:59:56 -0800
From: weening@Gang-of-Four.Stanford.EDU (Joe Weening)
Message-Id: <8912200059.AA03076@Gang-of-Four.Stanford.EDU>
To: qlisp@Gang-of-Four.Stanford.EDU
Subject: Alliant visit
Alliant will visit us on Thursday, January 4 at 10:00 a.m. to describe
a new multiprocessor system that they will soon be announcing. You'll
have to sign a non-disclosure agreement if you want to attend. Please
let me know if you plan to come. It will be in MJH 352.
Joe
∂19-Dec-89 1915 @IU.AI.SRI.COM,@sunset.ai.sri.com:konolige@ai.sri.com re: nonmon90 workshop
Received: from IU.AI.SRI.COM by SAIL.Stanford.EDU with TCP; 19 Dec 89 19:15:40 PST
Received: from sunset.ai.sri.com by IU.AI.SRI.COM via SMTP with TCP;
Tue, 19 Dec 89 19:15:44-PST
Received: from kehoe.ai.sri.com by sunset.ai.sri.com (4.0/4.16) id
AA02741 for JMC@SAIL.Stanford.EDU; Tue, 19 Dec 89 19:14:25 PST
Received: by kehoe.ai.sri.com (4.0/4.16) id AA01729 for
JMC@SAIL.Stanford.EDU; Tue, 19 Dec 89 19:14:23 PST
Date: Tue, 19 Dec 89 19:14:23 PST
From: Kurt Konolige <konolige@ai.sri.com>
Message-Id: <8912200314.AA01729@kehoe.ai.sri.com>
To: JMC@SAIL.Stanford.EDU
Subject: re: nonmon90 workshop
Here's the info:
CALL FOR PARTICIPATION
THIRD INTERNATIONAL WORKSHOP ON NONMONOTONIC REASONING
The third international workshop on nonomontonic reasoning, sponsored by
AAAI and CSCSI, will be held May 31-June 3, 1990, at the Stanford Sierra
Camp in South Lake Tahoe, California. The aim of the workshop is to
bring together active researchers in the area of nonmonotonic reasoning
to discuss current research, results, and problems of both theoretical
and practical nature.
Topics (not necessarily exhaustive):
General theories of defeasible inference
Comparison of formal systems
Applications to planning, commonsense reasoning
Knowledge update and truth maintenance
Relation to probability models
Theories of inheritance with exceptions
Argument-based systems
Proof theory, complexity, and automation
Attendance will be limited to 30-40 people, by invitation only. Those
wishing to attend should submit five copies of a detailed abstracts of
current research to:
Kurt Konolige
SRI International EJ272
333 Ravenswood Avenue
Menlo Park, Ca. 94025
Phone: (415) 859-2788
E-mail: konolige@ai.sri.com
Electronic mail submissions are encouraged (one copy only!). Abstracts
should consist of no more than 10 double-spaced pages when printed (4000
words) and should include enough information to enable the program
committee to judge the contribution of the work. Abstracts will be
accepted on the basis of quality, originality, and significance. The
deadline for submission of papers is ***December 17, 1989*** (note that
this is later than the original announcement). Notification of
acceptance will be made by February 26, 1990. Accepted authors will be
asked to send a preprint for distribution at the workshop. Papers
presented at the workshop can be submitted for publication in a
collection to appear later.
Program Committee:
Johan DeKleer, Xerox Parc (dekleer.pa@xerox.com)
Jon Doyle, MIT (doyle@zermatt.lcs.mit.edu)
David Etherington, AT&T (ether@research.att.com)
Matt Ginsberg, Stanford (ginsberg@polya.stanford.edu)
Hector Geffner, UCLA (hector@cs.ucla.edu)
David Israel, SRI (israel@ai.sri.com)
Henry Kautz, AT&T (kautz@research.att.com)
Vladimir Lifschitz, Stanford (val@sail.stanford.edu)
David Poole, UBC (poole@vision.cs.ubc.ca)
Erik Sandewall, Linkoping (enea!lisbet.liu.se!e-sandewall@uunet.uu.net)
Richmond Thomason, Pittsburgh (rich.thomason@cad.cs.cmu.edu)
In addition to accepted authors, we have a limited number of slots for
students who have shown promise in the area. Sponsors for such students
should send a short justification to Kurt Konolige at the above address.
∂20-Dec-89 1551 GLB
To: JK@SAIL.Stanford.EDU, JMC@SAIL.Stanford.EDU
I have done in EKL a proof that represents the missionaries
and cannibals problem. It uses the fact that Petri Nets are
naturally represented in Direct (Linear) Logic.
In more detail, we have propositions
BL [BR] meaning ``the boat is on the left [right]''.
CL [CR] meaning ``a cannibal is on the left [right]''
and similarly ML, MR.
CONVERTED means that a conversion has taken place.
Using the properties of Direct Logic, we can represent the
presence of three distinct missionaries on the left as
ML∧ML∧ML and the passage of the boat across the river as
the implication BR ⊃ BL or BL ⊃ BR.
The fact that at each step no conversion happens is tested
as a failure to derive CONVERTED from the obvious axioms
from and the situation at that step.
Gianluigi
∂20-Dec-89 1553 GLB
To: JK@SAIL.Stanford.EDU, JMC@SAIL.Stanford.EDU
;the missionnaries and cannibals problem
(wipe-out)
(proof miscan)
(decl (ml mr cl cr bl br converted) (type: truthval)(syntype: constant))
;2
(axiom |ml∧bl⊃mr∧br|)
(label missionary_driving_right)
;3
(axiom |mr∧br⊃ml∧bl|)
(label missionary_driving_left)
;4
(axiom |cl∧bl⊃cr∧br|)
(label cannibal_driving_right)
;5
(axiom |cr∧br⊃cl∧bl|)
(label cannibal_driving_left)
;6
(axiom |ml∧cl∧bl⊃mr∧cr∧br|)
(label missionary_and_cannibal_driving_right)
;7
(axiom |mr∧cr∧br⊃ml∧cl∧bl|)
(label missionary_and_cannibal_driving_left)
;8
(axiom |mr∧mr∧br⊃ml∧ml∧bl|)
(label two_missionarries_driving_left)
;9
(axiom |ml∧ml∧bl⊃mr∧mr∧br|)
(label two_missionaries_driving_right)
;10
(axiom |cr∧cr∧br⊃cl∧cl∧bl|)
(label two_cannibals_driving_left)
;11
(axiom |cl∧cl∧bl⊃cr∧cr∧br|)
(label two_cannibals_driving_right)
;12
(axiom |(ml∧ml∧cl∧cr∧cr)⊃converted|)
(label conversion)
;13
(axiom |(mr∧mr∧cr∧cl∧cl)⊃converted|)
(label conversion)
;14
(axiom |(ml∧ml∧ml∧cl∧cr)⊃converted|)
(label conversion)
;15
(axiom |(mr∧mr∧mr∧cr∧cl)⊃converted|)
(label conversion)
;16
(assume |ml∧ml∧ml∧cl∧cl∧cl∧bl|)
(label all_people_waiting_to_pass)
;first morning
(derive |(cl∧ml)∧(cl∧ml)∧(cr∧mr)∧br| (16 6))
17. (DERIVE |(CL∧ML)∧(CL∧ML)∧(CR∧MR)∧BR|
(ALL_PEOPLE_WAITING_TO_PASS MISSIONARY_AND_CANNIBAL_DRIVING_RIGHT)
NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;first evening
(derive |(cl∧ml)∧(cl∧ml)∧(cl∧mr)∧bl| (17 5))
18. (DERIVE |(CL∧ML)∧(CL∧ML)∧(CL∧MR)∧BL| (17 CANNIBAL_DRIVING_LEFT) NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;second morning
(derive |(cl∧mr)∧(cl∧mr)∧(cl∧mr)∧br| (18 9))
19. (DERIVE |(CL∧MR)∧(CL∧MR)∧(CL∧MR)∧BR| (18 TWO_MISSIONARIES_DRIVING_RIGHT)
NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;second evening
(derive |(cl∧ml)∧(cl∧mr)∧(cl∧mr)∧bl| (19 3))
20. (DERIVE |(CL∧ML)∧(CL∧MR)∧(CL∧MR)∧BL| (19 MISSIONARY_DRIVING_LEFT) NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;third morning
(derive |(cl∧ml)∧(cr∧mr)∧(cr∧mr)∧br| (20 11))
21. (DERIVE |(CL∧ML)∧(CR∧MR)∧(CR∧MR)∧BR| (20 TWO_CANNIBALS_DRIVING_RIGHT)
NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;third evening
(derive |(cl∧ml)∧(cl∧ml)∧(cr∧mr)∧bl| (21 7))
22. (DERIVE |(CL∧ML)∧(CL∧ML)∧(CR∧MR)∧BL|
(21 MISSIONARY_AND_CANNIBAL_DRIVING_LEFT) NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;fourth morning
(derive |(cr∧ml)∧(cr∧ml)∧(cr∧mr)∧br| (22 11))
23. (DERIVE |(CR∧ML)∧(CR∧ML)∧(CR∧MR)∧BR| (22 TWO_CANNIBALS_DRIVING_RIGHT)
NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;fourth evening
(derive |(cr∧ml)∧(cr∧ml)∧(cr∧ml)∧bl| (23 3))
24. (DERIVE |(CR∧ML)∧(CR∧ML)∧(CR∧ML)∧BL| (23 MISSIONARY_DRIVING_LEFT) NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;fifth morning
(derive |(cr∧ml)∧(cr∧mr)∧(cr∧mr)∧br| (24 9))
25. (DERIVE |(CR∧ML)∧(CR∧MR)∧(CR∧MR)∧BR| (24 TWO_MISSIONARIES_DRIVING_RIGHT)
NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;fifth evening
(derive |(cr∧ml)∧(cr∧ml)∧(cr∧mr)∧bl| (25 3))
26. (DERIVE |(CR∧ML)∧(CR∧ML)∧(CR∧MR)∧BL| (25 MISSIONARY_DRIVING_LEFT) NIL)
(derive |converted| (conversion *))
; failed to derive
CONVERTED
;sixth morning
(derive |(cr∧mr)∧(cr∧mr)∧(cr∧mr)∧br| (26 9))
27. (DERIVE |(CR∧MR)∧(CR∧MR)∧(CR∧MR)∧BR| (26 TWO_MISSIONARIES_DRIVING_RIGHT)
NIL)
(label all_people_passed)
con-Dec-89 0827 beeson@ucscd.UCSC.EDU autism and common-sense reasoning
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 21 Dec 89 08:27:30 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA18151; Thu, 21 Dec 89 08:28:32 -0800
Date: Thu, 21 Dec 89 08:28:32 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8912211628.AA18151@ucscd.UCSC.EDU>
To: jmc@sail.stanford.edu, val@sail.stanford.edu
Subject: autism and common-sense reasoning
I mailed you (both) an article in which it's claimed that the
cause of autism is an inability to carry out certain kinds of
reasoning abt the beliefs of others. It's also claimed (but the
evidence is weaker) that this is probably due to brain damage,
from which it follows that a specific part of the brain is responsible
for reasoning of the kind autistic people can't do. I would be delighted
to heayour comments on the article in the light of years of research
on formalizing reasoning. To ask a particular question: can a formal
model of autistic cognition be given by modifying one or more models
of ordinary cognition?
∂21-Dec-89 1047 MPS
Call George Joseph 856-8877
∂21-Dec-89 1209 phil@ub.d.umn.edu Philosophy and Artificial Intelligence
Received: from ub.d.umn.edu ([131.212.32.6]) by SAIL.Stanford.EDU with TCP; 21 Dec 89 12:09:27 PST
Received: by ub.d.umn.edu (5.59/UMD-891211)
id AA16096; Thu, 21 Dec 89 14:09:12 CST
From: phil@ub.d.umn.edu (Philosophy Dept)
Message-Id: <8912212009.AA16096@ub.d.umn.edu>
Subject: Philosophy and Artificial Intelligence
To: AI_and_Phil@ub.d.umn.edu
Date: Thu, 21 Dec 89 14:09:10 CDT
Cc: phil@ub.d.umn.edu
X-Mailer: Elm [version 2.1 PL1]
A member of the editorial board advised me yesterday that an unusual but
presumably temporary situation has arisen where there appear to be two
journals with different publishers, editors, boards, etc., but the same
name:
(1) the Kluwer journal of which I am the editor and Bill Rapaport is the
book review editor, whose editorial board currently includes Jon Barwise,
Robert Cummins, Jerry Fodor, Jaakko Hintikka, David Israel, Henry Kyburg,
John McCarthy, Donald Nute, Barry Richards, John Searle, Stephen Stich,
and Terry Winograd; and,
(2) a Bergverlag journal being edited by Kazim Zadeh and co-edited by Jay
Garfield and Donald Nute, whose editorial board includes Radu Bogdan, Jer-
ry Fodor, Dov Gabbay, Antony Galton, Paul Gochet, Richard Grandy, Deborah
Johnson, Robert Kirk, Henry Kyburg, Dan Lloyd, Christopher Maloney, Peter
Millican, James Moor, R. J. Nelson, Zenon Pylyshyn, Mark Rollins, Joseph
Sneed, Edward Stabler, and Paul Thagard.
Kluwer and I were aware that Zadeh was promoting a journal with the name
"Artificial Intelligence and Philosophy" and a substantially different
editorial focus than "Philosophy and Artificial Intelligence". If Berg-
verlag now has a claim to "Philosophy and Artificial Intelligence", then
I shall recommend to Kluwer that we adopt "Artificial Intelligence and
Philosophy" instead. Apart from the short-term confusion that these cir-
cumstances may generate, however, the long-term prospects for our journal
are excellent. We are grateful that you belong to our editorial board.
We remain firmly committed to this project.
James H. Fetzer and
Martin Scrivener, Kluwer
∂21-Dec-89 1418 shankar@argon.csl.sri.com thesis
Received: from argon.csl.sri.com by SAIL.Stanford.EDU with TCP; 21 Dec 89 14:18:17 PST
Received: by argon.csl.sri.com at Thu, 21 Dec 89 14:16:03 -0800.
(5.61.11/XIDA-1.2.8.27) id AA07273 for JMC@SAIL.Stanford.EDU
Date: Thu, 21 Dec 89 14:16:03 -0800
From: Natarajan Shankar <shankar@csl.sri.com>
Message-Id: <8912212216.AA07273@argon.csl.sri.com>
To: JMC@SAIL.Stanford.EDU
In-Reply-To: John McCarthy's message of 21 Dec 89 1152 PST <uNZKg@SAIL.Stanford.EDU>
Subject: thesis
John, Here's the reference for the dissertation. There is a short paper which
appeared in the Proceedings of the Logic, philosophy, and methodology of
science congress, Moscow 1987. I can't find a good reference for it though
since I haven't seen the book version of it.
I've also included the reference to Art Quaife's work where he derives the
incompleteness theorems starting from the Loeb derivability conditions using
the Argonne resolution prover. There is also the work of Sieg, Marinov, and
Marinov from the Computer-assisted instruction at Stanford volume. I will be
mailing you a copy of the NYT review of Penrose's book. I haven't been able to
find Penrose's book in a bookstore yet.
With regard to the dialogue you outlined between a machine intelligence and a
human skeptic, the computer could ask the human being whether she believed in
the laws of first order logic, then further ask the human whether she believed
the axioms of Robinson arithmetic (Shoenfield Chap. 2) or its analogue in
hereditarily finite set theory which I used (the formal theory I used had
induction scheme, but it is dispensable in favor of a single axiom). From
this, the computer can proceed to plug in the proof which I did, whereupon, no
matter what else the human believes (i.e., any finite number of additional
axioms), the incompleteness of the human follows. The possibilities for
amusing dialogues around this subject are limitless. One such appears in
Smullyan's "5000BC", which I'll email you later.
Wang's book on Goedel has some interesting opinions by Goedel on these issues.
My impression is that he felt that on issues like the truth or falsity of the
continuum hypothesis, whereas a formal system would by neutral, a human might
be able to intuit reasons for believing either CH or (not CH).
Cheers,
Shankar
@phdthesis{Shankar86,
Key="Shankar",
Author="N. Shankar",
Title="Proof-checking Metamathematics",
school="Computer Science Department, The University of Texas at
Austin",
Note= {Also available as a Computational Logic Inc. Technical Report
No. 9}
Year="1986"}
@article{Quaife88,
Key = {Quaife88},
Author = {A. Quaife},
Title = {Automated Proofs of L\"ob's Theorem and G\"odel's Two
Incompleteness Theorems},
Journal = {Journal of Automated Reasoning},
Year = {1988},
Volume = {4},
Number = {2},
Pages = {219--231}
Month = {June}
}
∂21-Dec-89 1423 beeson@ucscd.UCSC.EDU rationalizing what we are going to do anyway
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 21 Dec 89 14:23:10 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA22729; Thu, 21 Dec 89 14:24:08 -0800
Date: Thu, 21 Dec 89 14:24:08 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8912212224.AA22729@ucscd.UCSC.EDU>
To: JMC@SAIL.Stanford.EDU, VAL@SAIL.Stanford.EDU, beeson@ucscd.UCSC.EDU
Subject: rationalizing what we are going to do anyway
This phrase in your message reminds me of some research a
friend of mine told me about re "persistence of beliefs".
Subjects are given (fake) suicide notes and told that some are
real and some are fake. They are asked to discriminate the real
from the fake. They are given praise for their success (from
a prepared script, but they don't know it.) Then they are tested
to determine whether they believe they are good at the task.
Of course they do. Then they are told the whole story of the
experiment, that all the notes are fake, and shown the script
from which the praise was read. Then they are re-tested whether
they believe they are good at the task. Result: NO CHANGE: they
still believe they are good at telling real from fake suicide notes.
When asked why they always have some rationalization: my cousin
tried to commit suicide once, etc.
∂22-Dec-89 1039 VAL
Are we meeting for lunch today?
∂22-Dec-89 1110 Mailer re: Castro?
To: JMC@SAIL.Stanford.EDU, su-etc@SAIL.Stanford.EDU
From: Dick Gabriel <RPG@SAIL.Stanford.EDU>
[In reply to message from JMC rcvd 22-Dec-89 07:13-PT.]
I couldn't figure out what characteristics the countries of Eastern Europe
and Cuba share that John thought have any bearing on this question.
-rpg-
∂22-Dec-89 1145 MPS interview
David Friedman, Discovery Magazine, would like to telephone
interview you about Doug Lenat.
617 734 8772
∂22-Dec-89 1326 weening@Gang-of-Four.Stanford.EDU
Received: from Gang-of-Four.Stanford.EDU by SAIL.Stanford.EDU with TCP; 22 Dec 89 13:25:56 PST
Received: by Gang-of-Four.Stanford.EDU (5.61/inc-1.0)
id AA10827; Fri, 22 Dec 89 13:26:47 -0800
Date: Fri, 22 Dec 89 13:26:47 -0800
From: weening@Gang-of-Four.Stanford.EDU (Joe Weening)
Message-Id: <8912222126.AA10827@Gang-of-Four.Stanford.EDU>
To: jmc@sail.Stanford.EDU
>From: bks@alfa.berkeley.edu (Brad Sherman)
Newsgroups: sci.crypt
Subject: Quantum communication thwarts eavesdroppers
Date: 22 Dec 89 19:14:22 GMT
Reply-To: bks@alfa.berkeley.edu (Brad Sherman)
Organization: University of California, Berkeley
>From _New_Scientist_ 9 December, 1989, pp.25-26 (without permission)
Byline: David Deutsch
Quantum Communication Thwarts Eavesdroppers
Researchers at IBM's Thomas J. Watson Laboratory in Yorktown Heights,
New York, have built a device for sending information between two
parties with absolute security from eavesdropping. The device uses
a form of coding called "Quantum Public Key Distribution["] (QPKD),
which exploits quantum effects. It cannot be implemented on any
existing computer because of the intrinsic limitations of classical
(that is, non-quantum) information processing.
The device, built by Charles Bennett and John Smolin ... uses very
faint flashes of light to transmit messaes over an unprotected
communications channel. It allows the sender and receiver to agree
on a code without ever meeting in person. If an eavesdropper tries
to monitor the signal in transit, the uncertainty principle ...
ensures that the signal is disturbed in such a way that the sender
and receiver are alerted.
... [Long description of the use of polarized photons and the public
key mechanism omitted to prevent C.T.syndrome --RTFA for details] ...
The prototype QPKD machine is capable of transmitting its secret
keys over a distance of 50 centimetres. Plans are in hand for an
improved model using optical fibres that would have a range of
tens or perhaps hundreds of metres. Relay stations cannot be
used because they would have to measure the photons, lose quantum
coherence, and so be vulnerable to eavesdropping. So it may be
some time before quantum cryptographic techonolgy becomes widely
usefule in real applications.
The lasting significance of QPKD will be for the foundations of
computer science. The Turing Machine is no longer a universal
model for practical computations.
------
Brad Sherman(bks@alfa.berkeley.edu)
∂22-Dec-89 1448 MPS MERRY OLE' CHRISTMAS
To: JMC, CLT, VAL
M
E E
R R R
R R R R
Y Y Y Y Y
C C C C C C
H H H H H H H
R R R R R R R R
I I I I I I I I I
S S S S S S S S S S
T T T T T T T T T T T
M M M M M M M M M M M M
A A A A A A A A A A A A A
S S S S S S S S S S S S S S
| |
| |
I want to wish all of you a very joyous holiday season and prosperous new
year.
See you next year.
Pat
∂22-Dec-89 1535 shankar@argon.csl.sri.com Thanks!
Received: from argon.csl.sri.com by SAIL.Stanford.EDU with TCP; 22 Dec 89 15:35:36 PST
Received: by argon.csl.sri.com at Fri, 22 Dec 89 15:33:20 -0800.
(5.61.11/XIDA-1.2.8.27) id AA07995 for JMC@SAIL.Stanford.EDU
Date: Fri, 22 Dec 89 15:33:20 -0800
From: Natarajan Shankar <shankar@csl.sri.com>
Message-Id: <8912222333.AA07995@argon.csl.sri.com>
To: JMC@SAIL.Stanford.EDU
In-Reply-To: John McCarthy's message of 22 Dec 89 0240 PST <SO2FS@SAIL.Stanford.EDU>
Subject: Thanks!
John,
Merry Christmas and Happy New Decade to you, Carolyn, and Timothy.
I'm deeply grateful to you for letting me teach CS306 (and for much else). I
don't know if I did it full justice, but I did try to teach a few novelties
like the continuation-passing style, and continuation-based program
transformation. Joe Weening suggested that I might want to make a tech. report
out of my handouts. There is only so much one can do in a quarter, though. I
would enjoy teaching a more advanced course such as CS350, if the need ever
arises. The students seem interested in such a course.
Warm regards,
Shankar
∂26-Dec-89 0527 jsl%pres_res.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU E-mail
Received: from ROCKY2.ROCKEFELLER.EDU ([129.85.1.2]) by SAIL.Stanford.EDU with TCP; 26 Dec 89 05:27:31 PST
Received: from [129.85.1.242] by ROCKY2.ROCKEFELLER.EDU (5.61/1.34)
id AA01329; Tue, 26 Dec 89 08:29:05 -0500
Received: by pres_res.rockefeller.edu (4.0/SMI-4.0)
id AA23150; Tue, 26 Dec 89 08:31:04 EST
Message-Id: <8912261331.AA23150@pres_res.rockefeller.edu>
To: (John McCarthy) jmc@sail.stanford.edu
Reply-To: (J. Lederberg)lederberg@ROCKY2.ROCKEFELLER.EDU
Subject: E-mail
Date: Tue, 26 Dec 89 08:31:00 -0500
From: jsl%pres_res.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU
Hi John:
NY Times today refers to an article you wrote last summer
deploring fax vs. email. Could you msg. that to me?
Greetings,
Josh
∂26-Dec-89 1056 jsl%casp1.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU Re: Final version of Networks considered harmful.
Received: from ROCKY2.ROCKEFELLER.EDU by SAIL.Stanford.EDU with TCP; 26 Dec 89 10:56:30 PST
Received: from CASP1.ROCKEFELLER.EDU by ROCKY2.ROCKEFELLER.EDU (5.61/1.34)
id AA02645; Tue, 26 Dec 89 13:58:01 -0500
Received: by casp1.rockefeller.edu (4.0/SMI-4.0)
id AA07233; Tue, 26 Dec 89 13:54:21 EST
Message-Id: <8912261854.AA07233@casp1.rockefeller.edu>
To: jmc@sail.stanford.edu
Subject: Re: Final version of Networks considered harmful.
In-Reply-To: Your message of 26 Dec 89 10:08:00 -0800.
Date: Tue, 26 Dec 89 13:54:20 -0500
From: jsl%casp1.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU
Dear John--
Thank you for your prompt reply.
I'll copy it to Bob Kahn -- his "Foundation" 'National Research
Initiatives is deeply concerned about the same matters.
--
Have you received a copy of the "Collaboratory" workshop report
that Bill Wulf (NSF) sponsored here last spring? If not, I'll send
you. There are some graphics we haven't got a standard to send you --
though in fact Larry Rosenberg is posting a POSTSCRIPT version of that
too.
Josh
{Your polemic made me nostalgic -- for you}.
---------------------------------------
P.S.
QUOTE:{
A number of people have asked me about my future plans. As you
probably know, Dr. David Baltimore will succeed me as president
of the Rockefeller University on July 1, 1990, when I reach
mandatory retirement age (65). I will remain on the faculty with the
title of University Professor, and am now planning the details of
my research programs. I have no doubt they will embrace the same
areas as engaged me at Stanford, namely
1). Microbial Genetics
2). Theory formation from a computer science, historical and
philosophical perspective.
3). Science and public policy.
Joshua Lederberg
∂26-Dec-89 1213 jsl%casp1.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU Re: Final version of Networks considered harmful.
Received: from ROCKY2.ROCKEFELLER.EDU by SAIL.Stanford.EDU with TCP; 26 Dec 89 12:13:21 PST
Received: from CASP1.ROCKEFELLER.EDU by ROCKY2.ROCKEFELLER.EDU (5.61/1.34)
id AA03002; Tue, 26 Dec 89 15:14:45 -0500
Received: by casp1.rockefeller.edu (4.0/SMI-4.0)
id AA07322; Tue, 26 Dec 89 15:10:56 EST
Message-Id: <8912262010.AA07322@casp1.rockefeller.edu>
To: jmc@sail.stanford.edu
Subject: Re: Final version of Networks considered harmful.
In-Reply-To: Your message of 26 Dec 89 11:04:00 -0800.
<cQ9zy@SAIL.Stanford.EDU>
Date: Tue, 26 Dec 89 15:10:54 -0500
From: jsl%casp1.ROCKEFELLER.EDU@ROCKY2.ROCKEFELLER.EDU
Here's the text of collaboratory, with troff marks that are either
useful or not too intrusive.
Sure -- I'd be glad to see you; or have lunch or supper with us.
Josh
---------------
Hard copy follows.
.ig
Changes 9/7/89 phoned by Keith U.
.RU
Dr. William Wulf
Directorate for Computer and Information Science
National Science Foundation
Washington, DC.
Dear Bill:
.fi
It is our great pleasure to hand over to you this report
of our workshop on the collaboratory.
It would be hard to overestimate the enthusiasm and labor
that went into it on the part of so many participants, and for
that reason it is difficult to single out the individuals who
deserve special credit for their efforts during and after the
workshop. I will mention that after we folded the tents, Barry
Leiner continued with a heroic effort at completing the script;
and Richard Rhodes did excellent service in making it still more
readable to a wider audience.
We hope it responds to the need for which the workshop was
called, and that you will enjoy the product.
\h'|4.5i'Yours sincerely,
.sp 3
.nf
.po 0.5i
.in 0.75i
Joshua Lederberg
President
The Rockefeller University
New York NY 10021-6399
212:570-8080
jsl@rocky2.rockefeller.edu
Fax: 212:570-8651
.sp -7
.in +3i
Keith Uncapher
Associate Dean, Info. Sci
University of Southern California
University Park - OHE 330A
Los Angeles, California 90089-1454
213:743-3143
uncapher@venera.isi.edu
.SH
.nf
.ps 18
.vs 20
.ft B
.sp 4i
.ce 2
Towards a National Collaboratory
.sp 0.5i
.ft R
.ps 14
.vs 16
Report of an Invitational Workshop
.ft B
.ps 12
.vs 16
.bp
.ce 6
.nf
.sp 1i
.ft B
Towards a National Collaboratory
.ft R
Report of an Invitational Workshop
At The Rockefeller University
March 17-18, 1989
.ft I
Joshua Lederberg & Keith Uncapher
.ft R
co-chairs
.sp 1i
.fi
.ce
.ft B
Abstract
.ft R
.ad
.hy
Some of the most pressing scientific challenges facing the
United States and the world can be met only through remote interaction
with instruments, colleagues and data. At the request of Dr. William
Wulf, Director, NSF Directorate for Computer and Information Science
and Engineering, an invitational workshop convened at Rockefeller
University on March 17-18, 1989. The workshop developed recommendations
for a research agenda leading to a National Collaboratory, a resource
that would use networking and computer technology to support remote
interaction. By providing the first technological base specifically
created to support collaboration independent of distance and by
increasing productive access to scarce and expensive national
scientific assets, the National Collaboratory will significantly
increase the productivity of science and engineering, accelerate
the pace of discovery and amplify the capabilities of human intellect.
The workshop endorsed the concept of the National
Collaboratory enthusiastically. It recommended a three-fold research
agenda that would repeat itself in cycles of design, implementation
and testing to evolve a Collaboratory of increasing value to
science and to the nation. The first direction concerns system
architecture and integration and would examine ways to allow people
and machines to use the Collaboratory's component technologies most
effectively. The second would evolve the tools and technologies
themselves. The third would develop user-oriented testbeds to
validate both technology and organization.
.bp 0
.sp 0.5i
.ce
.ft B
Introduction
.ft R
Remote interaction is becoming a necessity in many areas of
science. Interaction with remote instrumentation will be necessary
because instruments are inaccessible (such as the space telescope) or
operate in an environment hostile to people (such as unmanned deep
ocean vehicles). Remote interaction with colleagues will be necessary
whenever the appropriate mix of talents to address an
interdisciplinary problem is not collocated anywhere. Remote
interaction with data will be necessary when the data are too
vast to be replicated and managed at a single location (such as the
global seismic database). Some of the most pressing scientific
challenges facing us, such as that of global change, are inherently
distributed and exhibit all of these properties: remote interaction
with instruments, colleagues, and data is essential to solving them.
Although the technology to support these remote interactions is
advancing rapidly, a very high-speed national
network is needed as a foundation for research and education
in a Collaboratory. Such a network is an essential part of a national
information infrastructure. Supercomputers and high speed
workstations are becoming ubiquitous. Open systems standards are
making these facilities interoperable and accessible to the research
community. Integrating these technologies into an infrastructure to
support scientific research could significantly improve its pace and
quality. Large scientific projects will work productively over long
durations and distances. New collaborations will arise because
distance will no longer constrain carrying out tasks or sharing data.
Data streams produced by sensors and instruments around the world
will be aggregated and the new findings distributed across the
research community.
Recognizing the immense potential provided by the
Collaboratory, an invitational workshop was held at Rockefeller
University on March 17-18, 1989 at the request of Dr. William Wulf
(Director, NSF Directorate for Computer and Information Science and
Engineering). Chaired by Drs. Joshua Lederberg and Keith Uncapher,
the purpose of the workshop was to develop recommendations for a
research agenda that would lead to the National Collaboratory.
This report presents the results of that workshop. Section II
provides a glimpse of the impact that the Collaboratory can have on
future scientific research. Section III discusses the functions that a
Collaboratory must provide to achieve its potential. Section IV
discusses a research agenda that would move towards the
Collaboratory. Section V recommends a set of steps that NSF should
take to make the Collaboratory a reality. Appendix A is Dr. Wulf's
charge to the workshop; B provides a list
of workshop attendees and C provides the workshop
agenda.
This report provides only a summary of the workshop
discussions. A wealth of draft material was developed in the process
of preparing and documenting the workshop. For those interested,
the appropriate workshop session chairmen may be contacted for
more details in those respective areas.
.ne 4
.ce
.ft B
.SH
I. Science and Support for Collaboration
.ft R
Today's scientific challenges can be characterized by the nature
of the phenomena to be understood and the resources available to
understand them. More than ever before, science focuses on
phenomena that are remote and inaccessible, are inherently
distributed across space and time, and are conceptually and
computationally complex. Deep space and the deep oceans, for
instance, afford remote and inaccessible phenomena. The first
challenge is making the phenomena accessible by bringing data
closer to the scientists. Global change phenomena, for instance, are
inherently distributed across space and time. The first challenge is
constructing and managing geographically and temporally
distributed data bases. Many phenomena are conceptually and
computationally complex, generating large, heterogeneous data
streams and requiring large, multidisciplinary research teams. The
first challenge is organizing data about the phenomena so that
scientists can build on rather than duplicate others' work.
To address such problems, science can bring to bear tools such
as instruments and facilities, information such as data bases and
libraries, and scientists themselves. Some of today's most important
scientific resources are in extremely short supply. There is only one
space telescope and a very limited number of cyclotrons, vessels in
the national oceanographic research fleet, scientific research
supercomputers, and so on. Furthermore, the expertise to best exploit
these technologies is in equally short supply. While these resources
are characterized by scarcity, scientific data bases and libraries are
characterized by huge and ever expanding volume. When data bases
are unwieldy, unmanageable, and inaccessible, scientists waste their
most important resource--their attention--trying to find out what is
already known.
Team Science
Scientific research is conducted in diverse styles, with tools
ranging from a test tube or a hand-held calculator to a space station
or a giant accelerator. Much of it remains fiercely individualistic,
with a single scientist displaying a high degree of creativity and
versatility in the completion of a project from inception to tool-
building, data collection, and theoretical analysis. However, even the
isolated investigator works within a social definition of discovery, if
only to demarcate what is "new" knowledge. Furthermore, what is
discovered must meet some canonical criteria of significance. All this
implies a community of scientists and science as inherently a social
enterprise. Its practice often requires highly organized teamwork in
the coordination of highly specialized skills from diverse disciplines.
.bp
Figure 1. Epicycles of scientific discovery
.bp
The scientific process can be described graphically as an
iterative process shown in Figure 1. While this figure depicts the
phases of the scientific process of individual investigators, it is
equally useful to describe the process of team science -- the
collaboration among scientists working on a common problem. Each
node of the figure can be thought of as a step in the process of group
scientific research. This may be the tacit collaboration of minds
reached by searching and reading the literature. It may be the
explicit bilateral exchange of ideas, reagents, tips, people, other
resources. It may be a formal collaboration over extended periods of
time, with contractually agreed division of labor and allocation of
credit.
.ne 4
Technology Development
New technologies for managing and exchanging information are
emerging just in time to match the ever-growing complexity of
scientific inquiry. The central technology is the fusion of computer
and electronic communications that gives us electronic
mail, teleconferencing, widely accessible scientific databases, ports to
supercomputers, and access to remote computing and experimental
facilities. Some styles of experimentation are absolutely dependent
on physical collaboration; others have more modest instrumental
needs -- but none can avoid the need to remain current about the
rapidly changing conceptual streams of contemporary science. The
complexities of science are outstripping the capability of any single
institution to embrace the full range of pertinent expertise in any
area. We are therefore becoming more sensitive to the needs and
opportunities to facilitate collaboration, at every level, among
investigators at widely separated laboratories. \fIThe goal is to build no
less than a distributed intelligence, fully and seamlessly networked, with
fully supported computational assistance designed to accelerate the pace
and quality of discourse, and a broadening of the awareness of discovery:
in a word, a Collaboratory.\fR
.ce
.ft B
II. Expected Impact
.ft R
The National Collaboratory has the potential of greatly
amplifying future scientific research. It can promote and support
collaborations between multi-disciplinary teams of scientists working
on the great challenges facing the nation and the world. It can
enhance the use of large and inaccessible experimental facilities. It
can promote rapid exchange of information To illustrate this
potential, we now discuss several examples of national scientific
programs where the Collaboratory could have major impact.
Global Change (NSF, NASA, USGS, DOE)
The FCCSET Committee on Earth Sciences has recently issued a
report presenting an initial strategy for comprehensive, long-term
U.S. Global Change Research Program. The goal of the Program is to
provide a sound scientific basis for developing national and
international policy on global change issues. The scientific objectives
of the Program are to monitor, understand, and ultimately predict
global change. This interdisciplinary program contains seven
integrated science elements: 1) Bio-geochemical Dynamics, 2)
Ecological Systems and Dynamics, 3) Climate and Hydrologic Systems,
4) Human Interactions, 5) Earth System History, 6) Solid Earth
Processes, and 7) Solar Influences.
A high performance computing and communications national
infrastructure will be required to carry out this program. Earth
sensors generating data rates of hundreds of megabits per second
must be fed into combined databases for global analysis.
Supercomputers must be used to model and predict behavior, to
analyze observational data, and to compare models to observed data.
Powerful graphic workstations must be used to assist scientists in
visualizing global environmental processes. Advanced collaboration
technologies will be needed to help coordinate the gathering of
sensor data, the analysis of that data, the comparison of results
across disciplines, and the conversion of data to information to
knowledge. High performance networks will be required to connect
scientists, supercomputers, sensors, and databases. Furthermore,
these networks must connect scientists from a variety of disciplines
and organizations, and provide them access to resources managed by
a variety of institutions around the earth.
Collaboration between instruments is required to achieve
coordinated measurements of specific areas and/or events.
Simultaneous data collection from such instruments must take place
to support future analysis, and this analysis will require databases
spanning many years. The sensors involved will generate very large
data rates and therefore methods will be required for dealing with
massive datasets from different types of sensors. Multiple disciplines
must be involved, and this will require support for collaboration
between a wide variety of scientists from such disciplines.
Human Genome (NSF, NIH, DOE)
The goal of the human genome project is the mapping and
eventual sequencing of the human genome. The human genome
consists of a string of four different purine and pyrimidine
nucleotide bases strung together to form the genetic code which
dictates the characteristics of each individual. Only a very small
fraction of the total genome has presently been mapped and even
less has been sequenced. The human genome is slightly over three
billion nucleotide pairs in size, and it is estimated that at the current
time, the cost of sequencing is approximately one dollar per
nucleotide pair. There is world-wide interest in pursuing this project
and there is some "national pride" behind some nations' efforts, yet
no single nation can afford the total expenditure. The project is
information intensive, and the need for sharing that information is
great. The entire project includes far more than just mapping and
sequencing, it also includes information on the structure, function,
expression and interaction of the entire genome. Considering that
there are probably on the order of 100,000 functional genes in the
human, the potential for data collection, storage and sharing is
enormous. It is especially demanding in the context of the eventual
use of some of these data in medical practice and the demands this
will place on transmission and formatting.
One aspect of the project presents special problems. There will
be hundreds, or perhaps thousands, of different laboratories around
the world sequencing portions of the genome on a daily basis. With
the expected advances in sequencing technology it is anticipated that
hundreds of thousands of bases a day may be determined. Workers
will not always be certain of the map location of the DNA fragment
they are sequencing, while it may be of great importance for
understanding a particular disease. Initial strings will be signatures
that can help relate one fragment to another and to the emerging
map. It is essential that this sequencing "network" be transformed
into a distributed data generation system so that the results from
one group can be entered into an international "data base" of
sequences in REAL TIME, with a method of feedback enabling the
investigators to know if they have begun sequencing a region
previously done by another group, and the extent that contiguous
regions have been covered. There are a number of new resources
required for this application, including shared network access to a
database that is "intelligent" enough to always be comparing strings
of the incoming data with the existing data base and then is able to
provide this comparison back to the investigator.
Parallel Processing (NSF, DARPA, NASA, DOE)
Many problems in science and other fields such as aircraft
engineering require massive increases in computing horsepower. For
instance, design of the National Aerospace Plane and other advanced
aerospace vehicles requires the ability to "fly the plane in the
computer" as conventional prototyping techniques are too expensive
in both time and material. To do this requires extremely large fluid
dynamic simulations. This and similar requirements are driving a
major initiative to develop high performance computing. Much of this
thrust is in developing new parallel computing architectures and the
means for applying these architectures to the large computational
problems.
Because the community doing research on parallel computing
already needs access to remote facilities, namely the new parallel
machines, this area represents an opportunity for improving the
effectiveness of the research. Exchange of software, the ability to
remotely test and demonstrate new algorithms, and the ability to tie
different architectures together over the network all are examples of
the promise that the Collaboratory holds for the parallel processing
research community.
.ne 5
.ce
.ft B
III. The Functional Collaboratory
.ft R
What is the Collaboratory? As we use the term here, it is the
combination of technology, tools and infrastructure that allow
scientists to work with remote facilities (co-laboratory) and each
other (collaborat-ory) as if they were co-located and effectively
interfaced.
It is important to recognize that scientists have always
collaborated. For one thing, an implicit collaboration occurs through
the published literature. Conferences result in face-to-face
discussions. Experimental facilities are designed through teams.
Papers are co-authored. Scientists share students, critical reagents
and cell lines. The norm of mutual scientific criticism is an intense
form of intellectual collaboration, however antagonistic it may
appear. Besides its cognitive utility, criticism is also indispensable to
a rational system for the allocation of resources, tenured positions,
research funds, and facilities.
Furthermore, as science becomes more multi-disciplinary and
requires large expensive experimental facilities, there is an
increasing need to gain access to remote databases, computing, and
experimental resources. It is not possible to duplicate these facilities
in the immediate proximity of the scientist. The Collaboratory, by
supporting the use of remote instruments and other experimental
facilities, would allow a scientist to access such facilities from their
home institution, greatly expediting such scientific research.
The Collaboratory will provide seamless access to colleagues,
instruments, data, information, and knowledge.
Functional Description
In Figure 1 above, we saw a description of the scientific
process. Each one of the phases of that process can and should be
supported through the Collaboratory. Figure 2 depicts a functional
description of the National Collaboratory, showing how each of the
research functions of Figure 1 might be supported through technical
capabilities in the Collaboratory.
For example, one critical function to doing team scientific
research is project organization and management. Support of this
part of the process requires a variety of functions within the
Collaboratory. For example, there is a need to support coordination of
action, joint design, and resource scheduling. The Collaboratory
incorporates a wide set of tools and capabilities integrated to provide
an effective method for scientists to work with facilities and other
scientists in the conduct of their research. Examples of such tools are
electronic mail with its supporting services such as directories, tools
to support structured discussions, a digital library with appropriate
search mechanisms, user education and training tools, real-time
computer supported multi-media teleconferencing, a remote
experiment scheduler, and so on. These capabilities are in turn
supported by a number of enabling technologies, such as networks,
advanced human-machine interfaces, high resolution displays, and
video compression techniques. Finally, underlying the Collaboratory
paradigm must be an understanding of how groups work together as well
as offering seamless new technologies to enhance the availability of
old knowledge, permit scientists new means of accelerating the pace
of discovery and support the amplification of human intellectual capability.
In summary, Figure 2 attempts to portray the National
Collaboratory functionality and its relationship to the scientific
process. The Collaboratory is not a single part of Figure 2, but rather
is the Figure as a whole and more.
Assumptions
In the next section, we provide recommendations as to a
research agenda for achieving this Collaboratory. These
recommendations are based on several assumptions. First of all, we
assume that the underlying communications network provides a high
level of minimum capability. The initial stage of the Collaboratory
requires and assumes a national network based at a minimum on T1
communication lines (1.544 Mbps) with sufficient backbone
bandwidth (probably at least T3 or 45 Mbps) so that the anticipated
load can be supported without undue delays. As enhanced
capabilities are developed incorporating such technologies as high
quality video, the required bandwidth will increase, eventually to
approximately 1 Gbps. We note that such a national networking
resource is being planned through the activities of the Federal
Coordinating Council on Science, Engineering and Technology.
A FCCSET report, "The federal high-performance computing
program", is being distributed by the Office of Science and
technology Policy. The executive branch
and Congress have proposed support for a national netwrok
initiative.
The second assumption we make is that the computing system
available to the scientist is of relatively high capability. It will consist
of workstations having, at a minimum, a 10 Mips processor speed, 10
Mbytes memory, and 1000 x 1000 pixel color display. These
workstations will be coupled to high performance computing with
processor power of gigaflops and beyond.
Last but not least, we assume that the infrastructure we
describe above is made available to the users/scientists. That is, we
assume that through federal, state, private, and university
initiatives, the computing and communication infrastructure
represented by the capabilities of the above two paragraphs is in fact
available to the scientists.
.ne 5
.ce
.ft B
IV. Research Agenda
.ft R
As we saw above, there is tremendous potential payoff to a sponsored
research initiative targeted at a National Collaboratory. Such an
initiative would focus on developing and demonstrating the technologies
required to make the National Collaboratory a reality.
A three-fold research approach is recommended. The first
focuses on system architecture and integration, and would be aimed
at the system issues that allow people and machines to most
effectively utilize the various technologies that are the components
of the Collaboratory. The second component is a research program
expected to evolve the underlying technologies and tools themselves.
Last but not least, user-oriented testbeds coupled with better
theories about the processes of collaboration are required to validate
both the technical approaches and the overall system components,
allowing understanding of the requirements on such a system and
the role that the various technological components play in the overall
design. This three-fold approach allows an iterative cycle of design,
implementation, and testing to evolve the most useful Collaboratory
for the scientists and for the nation.
System Architecture and Integration
Much of the research required to bring the Collaboratory into
existence is in the area of system architecture and integration. The
Collaboratory is more than a set of tools. It is a functional capability
to improve scientific effectiveness by taking advantage of a broad set
of resources, including but not limited to remote facilities and other
scientists. Understanding the appropriate system architecture, where
such an architecture includes not only the underlying tools but also
the people that are to use those tools, requires a dedicated effort.
Available Technologies
Fortunately, there are a variety of tools already available, both
from the research community and the commercial sector. The
required research into system architecture can therefore begin by
addressing the appropriate system to take advantage of existing
tools. It should focus on the interfaces between such tools, the
enhancements and modifications required to make those tools fully
accessible and usable by the scientific community, and the degree to
which the tools work together to support scientific research.
Examples of available tools include the following:
* Electronic Mail. Fundamental to team efforts is the ability to
communicate rapidly and reliably. Electronic mail provides this
function in a way unequalled by other forms of communications. While
electronic mail technologies need enhancements in the areas of
interoperability (there are too many incompatible systems),
graphics capability, privacy, and user support (such as white
pages services), this technology has been available and proven
over the last two decades.
* Electronic File Transfer. As teams undertake cooperative
development and analysis activities, they need to share results.
Many of these results are represented as computer files. The
technology to accomplish file transfer is well proven.
* Remote Access and Control. Operation of remote facilities
and instruments will require the ability for a user to interact
with the instrument computers. Remote login has been available
and proven over the last two decades. However, enhancements to
assure appropriate access control and authentication will be
required for safety and security reasons.
* Shared Files. While the technology to exchange files is well
proven, the ability to share information through such files is
still quite crude. The only common information representation
widely accepted today is ascii text representation (and even there,
such facilities as tab settings sometimes give problems.)
Standard file representations for higher level functionality
(such as graphics and research data representations) are
available and are being used in limited communities, but need
to be more widely standardized and adopted.
* Database Access. Related to the ability to share files is the
ability to store and retrieve data from shared databases. Again,
as in shared files, the technology and standards to accomplish
this has been demonstrated in the research community and in
limited user communities, but further work is needed to insure
widespread standardization and adoption.
* Access Control and Authentication. It is critical that,
as teams share information and provide for shared control of
resources, appropriate security mechanisms be provided. Crude
forms of such technologies have been demonstrated and could be
adopted while further research is conducted in this area.
* Multimedia Mail. While electronic mail in its current form
(text) provides highly useful functionality, team activities
require integration of other media, such as graphics, sound, spread
sheets, and scanned images. We already have seen the demonstration
of the utility of such media through the widespread use of fax.
Multimedia mail is available already through a number of proprietary
products and limited demonstrations of interoperability have
been accomplished. However, widespread adoption of
multimedia mail will require the emergence of a powerful
mechanism for interoperability.
* Structured Interaction Support. In recent years, several tools
have been created that are aimed at supporting the process of
collaboration and teamwork, including tools for video and multi-
media teleconferencing, tools for structuring conversations,
and tools for sharing information. Many of these tools have been
developed in a proprietary architecture, targeted at specific
machines and operating systems. None of these tools has yet seen
widespread use by practicing scientists. To make them genuinely
useful, they must demonstrably improve current practice and be
easily accessible. Widespread use in the communities above
will require porting these systems to an open architecture.
* Simulation of Instruments. Prototyping instruments, while
ultimately necessary, is an expensive process. Advances in
computing technology permits these prototypes to be done, at
least in the initial stages, through computer simulation. By having
totally compatible hardware and software instantiations of the
instruments, remote debugging of the instrument software both in
the design phases and even during operations can be facilitated.
Furthermore, once a complicated instrument , smart or conventional,
can be simulated, it can then be distributed (by electronic
transmission or otherwise) to a large number of scientists. This
can allow multiple scientists to work on the development and
evolution of shared instruments, or to conduct a large simulated
distributed experiment (e.g. in robotics research.)
Thus, we recommend a research program that investigates the
integration of an appropriately selected set of tools to provide
Collaboratory functionality to users. The interface of the users to the
system must be addressed as part of this research program.
Integrating Technologies
Integration of the technologies described above into a National
Collaboratory will require development of higher level functions
along the lines described above in the section on the Functional
Collaboratory. A method for accomplishing this that has worked well
in the networking community is to assemble teams of researchers
working on the related areas and have them address the required
functionality. This would involve integration of both the available
technologies described above as well as the advanced technologies
described below as they become sufficiently mature, and would
allow the evolution and enhancement of the tools to take place in a
context of user feedback. In the following we describe several such
integration teams.
* Digital Instrumentation. This activity would be dedicated to
theories and technologies for remote and multi-user instrumentation.
It would connect with one or more scientific communities where
necessity or cost requires a separation between scientists and
their instruments. Research could include not only experimentation
with communication and control technology, but also with the
principled design for effective human-instrument interfaces when
there are inherent delays and scheduling problems. One vector of
research could involve experiments and technology for the control
of instrumentation through cooperating agents, where some of the
agents could be human and some could be automated. Such interfaces
would need to enable real-time communication as well as real-time
control and feedback from instruments.
* Multi-media Meetings. This would focus on experiments on
collaboration in meetings, such as might take place between
scientists working together at a distance. Since high-bandwidth
transmission can otherwise become a dominating cost, a key factor
for this capability would be some economic means of providing very
high-bandwidth transmission economically between meeting places.
An example of a site would be a university with a highly distributed
campus, where cross-disciplinary interactions across departments
is inhibited by the physical separations. Alternatively, the center
might be made of several related academic departments from
universities is a single city or region. Technology could include
the development of means for remote video, shared computational
whiteboards, and high-quality audio. Social experiments across
the "distributed center" could study the effects of the technology
on interpersonal argumentation at a distance and on the extent to
which collaborations were enhanced or enabled.
* Digital Mail. Electronic mail is one of the most ubiquitous
and still developing collaboration technologies. Because of its
relatively low cost, it has the potential of being able to reach
relatively large numbers of people with existing networks and
workstations. Nonetheless, electronic mail has developed haphazardly.
For example, the current electronic mail networks lack provisions for
being able to send a verified and trustable "electronic check," and
thereby, provide an inadequate basis for linking "value-added"
services by electronic mail. Electronic mail systems also use
inadequate addressing mechanisms, so that current networks often
fail for inappropriate reasons in delivering mail. No comprehensive
yellow-page services exist in the network. A team focussed on
digital mail would explore technical extensions and social issues
surrounding the use of electronic mail.
* Scientific Reference Service. This would be a service to
provide expertise and a network to answer tough questions that
arise in current scientific work and are not readily answered at
a home institution -- a "Who Knows ... " service, using literature
access as route to solution. These are intelligent agents, drawing
first on human, then on artificial intelligence.
* Digital Journals and Peer Review. This activity would be
dedicated to experiments on peer review and commentary. To date,
this kind of process is most prevalent in the computer science
and engineering communities, which use "electronic technical
reports" as an intermediate form prior to formal publication.
Technology to support such electronic journals could serve one or
more communities with services for logging documents, logging
comments on documents, and offering digital document retrieval.
An example of a "community" would be all of the engineering and
computer science departments in a set of universities. Many technical
approaches would be possible. One would be to have document requests
and comments serviced by electronic mail, document transmission by
fax, and document storage by image databases. Other technical
approaches could support collaborative writing, including
experimentation with different modes for commentary, editing, and
document exchange. Social experiments (see below) could determine
salient effects on the perceived qualities of number of reviews,
and effects on peer group size.
* Digital Library. The digital library would contain documents of
all sorts, including software, video, and other "unusual forms"
in a distributed electronic database. There are many variations
on a digital library. One variation would explore a broad range
of concepts for collaboration and electronic publishing. Technical
vectors could include experimentation with multi-media and hypertext,
and development of techniques for integrating services over distributed
libraries. Social vectors could include experimentation with policies
and mechanisms for citation norms and the pricing, collection, and
distribution of royalties.
Another component of the digital library is techniques for
discovery through digital search. Several variations on a theme
would lead to a single technical agenda. Key technologies for this
would include specialized hardware for scanning and comparing
strings of digital data, specialized hardware for storing documents,
and specialized technology for scanning and recognizing the contents
of documents. A major part of the document base could be a set of
human-developed documents, for which there is special reason to
expect that the documents lend themselves to large scale linguistic
analysis or comparison. Other "documents" could include linearly-
encoded scientific data, such as sequences of linear molecules. Social
questions around such a center would include an analysis of the
ways that the automated searching of the document base enabled
collaboration or discovery. The Digital Library is likely to be
the key to valuable old knowledge, and new knowledge so vital to
the scientific process.
.ne 4
Advanced Technologies
The development and evolution of the capabilities of the
National Collaboratory described above will require the availability
of new and advanced tools not in existence today. This component of
the program represents the research into underlying technologies.
We describe some examples of potential research areas, recognizing
that ultimately the new ideas are likely to come from the research
community itself.
* Hypermedia Conversation Support. We saw above that multimedia
electronic mail and conferencing along with support for structured
interactions has already been demonstrated. Coupled with hypermedia
tools (such as Hypercard and Hyperties), there is a possibility of
providing support for team interactions. A hypermedia database can
be created and maintained that tracks the team interactions and
provides an ongoing record of the interactions. This would provide
a long-term record of design decisions, operational problems and
corrections, and research approaches.
* Intelligent Agents. New ideas in distributed systems coupled
with high bandwidth networks makes possible the increased use of
distributed processes acting on behalf of the user. In particular,
we think of each entity in the Collaboratory (e.g. scientist,
instrument, database, computing resource) as having an intelligent
agent between it and the network. These intelligent agents are
responsible for acting on behalf of the attached entity, negotiating
with other agents to carry out out the assigned tasks. They are
also responsible for interacting with their attached entity to
provide the required interface. An example of such an agent is the
"Knowbot" described by Cerf and Kahn in their work on a digital
library. Such a Knowbot would conduct a search of the distributed
digital library to find the desired information. Other applications
of such intelligent agents are the scheduling of coordinated
experiments using multiple instruments/sensors and joint design
activities. The use of intelligent user agents would greatly
increase the utility of the Collaboratory.
* Interoperable Data Description. As science becomes more
multidisciplinary, there is an increasing need for access to
data derived from different sensors and other sources. For users
to be able to take advantage of such data, the data must be
described in a manner that is understandable to the analysis system.
Yet, it may not be possible to use a common data format across
disciplines. Research is required into methodologies for describing
data in an interoperable manner that recognizes the unique
requirements of the individual data. For example, research
into global environmental change will require the understanding of
data from atmospheric, space, and ground sensors as well as outputs
of computer models.
* Information Fusion. Not only does data need to be manipulated,
it also must be understood. Techniques must be developed for
understanding data from heterogeneous sources so that the
information can be integrated into an understanding of the
overall problem. For example, gene sequence information must
be integrated with clinical observations to understand the role
of various genetic information.
* Smart Agents for the Design of Experiments. As instruments
become more complex and expensive, are used in multi-sensor
experiments, and are shared by multiple investigators, the
design of the experiments become more complex. Standard
collaboration tools could facilitate the design of such
experiments, but in addition, tools specifically aimed at the planning,
scheduling, coordination, and operation design of the experiments
must be developed. These tools should themselves incorporate
sufficient intelligence to be a real "assistant" to the scientists and
engineers doing the experiment design. As an example of such a
facility, the DARPA/NSF-sponsored MOSIS facility for sharing and
brokering VLSI manufacturing and design is a cost-effective way of
meeting the needs of many people in a timely manner.
* Smart Data Gathering. Instruments are becoming capable of
generating more and more data (e.g. the SAR and HiRIS being
planned for future NASA science missions.) This taxes the
communication and data archiving system. In addition, many
of the instruments are capable of having parameters (such
a direction of look and spectral band) adjusted under computer
control. Many experiments only provide a single "shot" at
gathering the data (e.g. deep space probes, experiments where
the underlying conditions change). Incorporation of intelligence
into the instruments allows the possibility of "self-directed"
data gathering, with the instrument itself deciding when
data is significant and should be transmitted, setting parameters
based on local feedback, and doing preliminary data reduction. This
can lead to both reduction in communications and archiving
requirements and better scientific data.
Technology Utilization
Many of the critical issues needing investigation to make the
Collaboratory a reality involve the relationship between the
technologies and the way that scientists do and will conduct their
research. Understanding these issues will involve a partnership
between users and technologists to explore the utility of various
technical approaches and the methods by which the user community
can take advantage of the emerging technologies.
This research area involves establishing appropriate testbeds
to understand the impact of the available technologies coupled with a
research program into the underlying mechanisms of collaboration
itself.
User Testbeds
Historically, many new technologies have been left dangling at
the end of the research cycle waiting to be adopted by some user
community or integrated into commercial products. Since the
potential users have had little experience with such technologies,
they have little basis for asking for those tools.
One approach that has proved to be successful is the user-
oriented rapid-prototyping testbed. The concept is a partnership
between users and developers to integrate a prototype information
system tailored to the user's environment. The users then conduct
their activities using these prototype tools. As a result, the users
better understand their requirements and the potential technologies
to solve their problems, and the developers better understand the
desirable directions for future technology enhancements and
upgrades.
To be successful, such testbeds must have certain critical
attributes. First and foremost, they must represent a partnership
between users who see the potential for the Collaboratory improving
their scientific research and technologists who are interested in
working with such a user community. Second, they need to be of a
size that is sufficient to explore the impact of the Collaboratory on
the scientific research (which means a team of geographically
dispersed users working with laboratory facilities). However, they
also must be small enough to be manageable as a rapid-prototyping
environment (so, for example, issues such as software updates don't
become significant stumbling blocks in the prototype phase). Last but
not least, such testbeds must be provisioned with adequate
infrastructure (e.g. networking, workstations, etc.) so that the
prototype provided to the user does in fact represent the future
potential. Note that it is critical that those working on developing the
technologies be provisioned at any stage with the next stage of
infrastructure so as to act as a leading edge for the technology
insertion.
The scientific applications discussed above represent significant
opportunities for such testbeds. In each case (Global Environmental
Research, Human Genome Project, etc.), there is a sizable research
community being formed, driven by the required scientific research,
that needs the Collaboratory to carry out its work.
Collaboration Mechanisms
In order to develop useful tools for supporting scientific
collaboration, an important first step is to better understand the
processes of collaboration itself: How do people collaborate and
coordinate their work today-with or without computer technology?
What are the needs for effective collaboration that the new
technology might help fill? How might new technology allow us to
collaborate in different ways?
Understanding the processes of collaboration and coordination
will, of necessity, be a thoroughly multidisciplinary enterprise. It will
draw upon and contribute to a variety of existing disciplines
including computer science, sociology, psychology, economics,
anthropology, organization theory, and management science.
Some of the work to be done is observational. In order to
better understand how scientific collaboration works today, we need
to observe it more closely using methodologies such as field studies,
surveys, and archival analyses. There is also an important place for
experimental work. As new kinds of coordination technologies
become available, it will be essential to study experimental versions
of such systems in ways that can rapidly influence the design of
successive versions. In some cases, it will be useful to systematically
experiment with alternative versions of specific technologies. Thus, a
sizable portion of this research would be done in the context of the
user testbeds described above. Finally, it will be crucial to synthesize
previous work into new theories about how collaboration and
coordination occur and how technology can help.
Examples of the research questions needing to be addressed
are the following:
What are the basic processes involved in coordination? For
example, are there basic processes (like resource allocation,
task assignment, and information sharing) that are
fundamental to all coordination?
What structures are possible for carrying out these processes?
What are the relative advantages and disadvantages of these
different structures? For instance, are there conditions under
which "networked organizations" with lots of lateral (as
opposed to vertical) communication lead to more rapid
adaptation than hierarchical organizations? Do market-like
coordination structures provide greater scope for individual
autonomy and initiative than centralized hierarchical
structures? What incentives do different structures provide for
individuals to participate and contribute to common goals?
How is collaboration among scientists different from any other
kind of collaboration. It is commonly believed, for example,
that scientists have more individual discretion than many other
kinds of workers in choosing the goals, collaborators, and times
for their work. Whether this belief is true or not, it is clear that
there are characteristic processes involved in the conduct of
science and that there is a very highly developed set of social
and organizational structures to coordinate these scientific
activities.
How might the use of new collaboration technologies affect
incentive structures for the conduct of science? For example, if
new communication technologies make it feasible to include
much larger numbers of people in the writing and critiquing of
a scientific paper, how should all these contributions be
acknowledged? As another example, if voluminous shared data
bases (such as a map of the human genome) become critical in
the development of certain sciences, what will be the
incentives for individual investigators to contribute to such
databases? For instance, when scientists are being evaluated
for promotion, how will these contributions be weighted
relative to journal publications or other professional activities?
How, if at all, is the social structure of science affected by
intensive use of communication technology for remote
collaboration and resource sharing? For instance: Does the role
of gatekeepers change? (Gatekeepers include journal editors,
awards committees, and proposal review committees.) Do
formerly "peripheral" scientists become more central?
("Peripheral" scientists are outside the status and prestige
centers- perhaps they are geographically isolated or working
on unpopular problems or using unconventional methods.) Do
scientists' local neighborhoods change? Despite the fact that
science is a global enterprise and scientific publication is
directed at the world-wide community of scholars, much of a
scientist's working life is spent within his/her local
neighborhood.
To what degree is it possible to substitute capital (electronics)
for scientific labor. The real costs of the Collaboratory involve
the tradeoffs between capital and labor, not just the addition of
more capital to the equation. Especially the costs of the
network tend to be treated as a public good, while the labor is
not. It appears to be time, for instance, to simply document the
structure and use of the network system in the U.S. and the
costs of its maintenance and improvement. Another economic
issue on which we should at least gather the current wisdom is
how information is priced, and the impact of pricing regimes on
the production and consumption of information. For instance,
we have: (a) books, which are priced approximately by the
page, (b) journals which are priced by "what the traffic will
bear", (c) software, where use is licensed to a single machine
rather than on a more appropriate measure of usage, and (d)
data and scientific information, where there is little
relationship between price and either the cost of generating the
data or its value to the user. In some sense, we have a kind of
"social" market in science, with the achievement of prestige
eventually related to how publications are cited.
In summary, we expect the widespread use of new computer
and communication technologies to significantly affect the conduct of
science. In some cases, communication processes, social structures,
and reward systems may become quite different; in others, they may
be strikingly unchanged. Designing technologies to support these
processes effectively will require careful observation of how
scientific collaboration actually occurs (both with and without new
technology) and better theories about the nature of collaboration
itself.
.ne 5
.ft B
.ce
V. Conclusions and Recommendations
.ft R
By aggressively pursuing the research program outlined in the
above section, NSF could cause the Collaboratory to come into being,
thereby greatly improving national scientific and engineering
productivity. In pursuing such a research program, there are a
number of critical factors and issues to which attention must be paid.
Importance of Integration and User Testbedding
The Collaboratory is much more than just a set of tools. It is a
national computer-based infrastructure for scientific research.
Understanding the appropriate system architecture and the best
selection of tools will require a careful program of testing and
evaluation. The user-oriented rapid-prototyping testbed approach
recommended above requires a careful matching program of
technology integration.
Selection of User Testbed Communities
NSF/CISE should work in collaboration with other Government
agencies as well as offices within NSF to identify and carry forward
an aggressive rapid-prototyping testbed activity. It is essential that
such communities be enlisted in the testbedding process. There must
be immediate as well as long-term benefit to the scientific
community. Furthermore, the partners to CISE must be willing to
work with CISE in providing appropriate infrastructure to that
community to insure that it gets the full benefit of the Collaboratory.
Community Workshop
Identification of the appropriate approach to developing both
the prototype infrastructure and the user testbeds will require close
collaboration between the scientist users and those developing the
underlying technology. NSF should hold, early in the program, a
series of workshops aimed at identifying specific approaches to
integration and testbedding. These workshops should involve the
potential scientific users of the collaboratory, those developing and
integrating the technologies, and social and other scientists who
would be involved in understanding the impact of the Collaboratory.
Targeted Integrated System
The Collaboratory in its full and mature form will be a general
infrastructure supporting national (and international) scientific and
engineering research. In the early prototyping phase, though, while
testbeds are being used to understand the desired functions and
technologies, and perhaps more important, how users can take
advantage of those functions, it is critical that the integrated
prototypes be targeted at specific user communities. There will be an
underlying kernel of tools useful to everyone (e.g. electronic mail).
However, it is important in the testbedding phase to maintain
flexibility in the individual testbeds.
Draw on Experience of Other Agencies/Organizations
Other agencies have also been exploring advanced computer
and communications technologies and how they can support
communities of interest. For example, DARPA has historically had an
intensive effort in networking and advanced computing, and has
conducted user-oriented rapid-prototyping testbeds jointly with the
military services to evaluate the utility of those technologies in
supporting future command and control. NASA has an ongoing
program to evaluate similar technologies to support telescience (the
conduct of science using remote resources including other people).
NSF can gain considerable advantage by working closely with other
such programs in the conduct of the National Collaboratory research
agenda.
In closing, we wish to emphasize the importance attached to
this initiative. Attacking the great challenges facing the US and the
world will require taking full advantage of available technologies and
resources in the conduct of the research. There is no more valuable
resource in conducting scientific research than the scientists
themselves, and a program to establish the Collaboratory will allow
significant increases in scientific and engineering productivity, in the
pace of discovery, and in the amplification of human intellectual
capability. We wish to encourage the NSF to move forward on this
initiative with all due speed and stand willing to assist in whatever
way we can.
.bp
.nf
.ps 10
.vs 10
Appendix B - Workshop Attendees
.ta 2.2i
Daniel Atkins University of Michigan
Charles Brownstein National Science Foundation
Vinton Cerf Corporation for National Research Initiatives
Y.T. Chien National Science Foundation
Lynn Conway University of Michigan
David Farber University of Pennsylvania
Craig Fields Defense Advanced Research Projects Agency
Peter Freeman National Science Foundation
Frank Halasz Xerox Corporation
Robert Kahn Corporation for National Research Initiatives
David Kingsbury George Washington University
Joshua Lederberg (co-chair) The Rockefeller University
Barry Leiner Research Institute for Advanced Computer Science
Thomas Malone Massachusetts Institute of Technology
James Ostell National Library of Medicine
Jonathan Postel University of Southern California
Raj Reddy Carnegie Mellon University
Richard Rhodes Consultant
Thomas Rindfleisch Stanford University
Larry Rosenberg National Science Foundation
Herbert Schorr University of Southern California
Larry Smarr University of Illinois
Lee Sproull Carnegie Mellon University
Robert Sproull Carnegie Mellon University
Mark Stefik Xerox Corporation
Alvin Thaler National Science Foundation
Keith Uncapher (co-chair) University of Southern California
and Corporation for National Research Initiatives
Fred Weingarten Office of Technology Assessment
Harriet Zuckerman Columbia University
.ce
-------------------------
Appendix C - Workshop Agenda
.ta 0.6i 3.0i
.nf
Monday, August 17, 1989
Chairs:
0900 Plenary Lederberg
1400 Breakout Sessions
Digital Library Cerf
Enabling Technologies Kahn
Fundamentals of Collaboration Malone
1800 Plenary Lederberg/Uncapher
-----
Tuesday, March 18, 1989
0900 Breakout Sessions
Smart Instruments Leiner
Tools Stefik
Applications Kingsbury
1400 Plenary Lederberg/Uncapher
.ex
Figure 1: Epicycles of Scientific Discovery
From: J. Lederberg, Preface,\fI Excitement and Fascination of Science,\fR Vol. 3
Annual Reviews, Inc. (Palo Alto), 1989. Copyright Reserved.
Figure 2: The Functional Collaboratory
(Mark Stefik).
∂27-Dec-89 1033 eaf@sumex-aim.stanford.edu a merry christmas gift arrived
Received: from sumex-aim.stanford.edu by SAIL.Stanford.EDU with TCP; 27 Dec 89 10:33:00 PST
Received: by sumex-aim.stanford.edu (4.0/inc-1.0)
id AA12163; Wed, 27 Dec 89 10:32:24 PST
Date: Wed, 27 Dec 1989 10:32:24 PST
From: Edward A. Feigenbaum <eaf@sumex-aim.stanford.edu>
To: ksl@sumex-aim.stanford.edu
Cc: jmc@sail.stanford.edu, nilsson@tenaya.stanford.edu,
genesereth@cs.stanford.edu, feigenbaum@sumex-aim.stanford.edu
Subject: a merry christmas gift arrived
Message-Id: <CMM.0.88.630786744.eaf@sumex-aim.stanford.edu>
Just in time for christmas there came from the bindery new shiny copies of
the new Handbook of Artificial Intelligence, Volume IV. It is a huge volume
(700 pages!).
This volume is in a new style: authored survey articles. Congratulations and
thanks to the following Stanford authors: Penny Nii, Yumi Iwasaki, and
Bruce Buchanan (honorary Stanford person).
The book is dedicated to the memory of Dianne Kanerva, technical editor of
Volumes I-III, and a good friend to many of us.
Best holiday wishes,
Ed F.
∂27-Dec-89 1334 bscott@Polya.Stanford.EDU Sabbatical Credit
Received: from Polya.Stanford.EDU by SAIL.Stanford.EDU with TCP; 27 Dec 89 13:34:09 PST
Received: by Polya.Stanford.EDU (5.61+IDA/25-eef) id AA28168; Wed, 27 Dec 89 13:29:56 -0800
Date: Wed, 27 Dec 1989 13:29:55 PST
From: Betty Scott <bscott@Polya.Stanford.EDU>
To: Cheriton@Pescadero.Stanford.EDU, Dill@Amadeus.Stanford.EDU,
EAF@SUMEX-AIM.Stanford.EDU, RWF@Sail.Stanford.EDU,
Genesereth@Polya.Stanford.EDU, Goldberg@Polya.Stanford.EDU,
Golub@Patience.Stanford.EDU, Guibas@SRC.DEC.COM,
AG@Amadeus.Stanford.EDU, Lam@Mojave.Stanford.EDU,
Latombe@Coyote.Stanford.EDU, ZM@Sail.Stanford.EDU,
JMC@Sail.Stanford.EDU, JCM@Polya.Stanford.EDU,
Rajeev@Polya.Stanford.EDU, Nilsson@Tenaya.Stanford.EDU,
Oliger@Pride.Stanford.EDU, Plotkin@Hudson.Stanford.EDU,
Pratt@Polya.Stanford.EDU, Shoham@Polya.Stanford.EDU,
Ullman@Nimbin.Stanford.EDU, Winograd@CSLI.Stanford.EDU
Cc: BScott@Polya.Stanford.EDU
Subject: Sabbatical Credit
Message-Id: <CMM.0.88.630797395.bscott@Polya.Stanford.EDU>
I have gone through all your files in order to summarize leave activity
and sabbatical credit from the time of your initial appointments, and have
sent this information to the SOE so that their records agree with ours. Rose
Ewing, who manages faculty affairs in the SOE, has entered all this informa-
tion into her on-line file, and has sent me a hard copy of each of your
records. I will send you a copy of your record sometime this week.
Betty
∂27-Dec-89 1603 JMC wrong number?
The book "Techno-bandits" by Melvern, et al, has the number HD38.7T43
printed in it inside. the Socrates listing (and therefore your home file)
has it listed as HD38.7M45.
∂28-Dec-89 1149 stager@Sunburn.Stanford.EDU CS499 grade sheet
Received: from Sunburn.Stanford.EDU by SAIL.Stanford.EDU with TCP; 28 Dec 89 11:49:02 PST
Received: by Sunburn.Stanford.EDU (5.61/25-eef) id AA09734; Thu, 28 Dec 89 11:49:28 -0800
Date: Thu, 28 Dec 1989 11:49:28 PST
From: "Claire E. Stager" <stager@Sunburn.Stanford.EDU>
To: jmc@sail
Subject: CS499 grade sheet
Message-Id: <CMM.0.88.630877768.stager@Sunburn.Stanford.EDU>
I sent a CS499 (Advanced Reading and Research) grade sheet your way a few
weeks ago. As I haven't received the completed sheet back yet, I'd
appreciate it if you could look around to see if you still have it. If
found, please return it to me asap (Claire Stager, CS-TAC box at MJH).
Thanks again.
Claire
∂28-Dec-89 1409 JMC dinner
The place that Whit & Mary had in mind is booked up and could not give us
a reservation after 5pm(which they said we could be a bit late for), if
this is too early, let me know, as the other possibility under consideration
is for Whit to cook dinner. It ought to be tonight though, becouse their
usual friday night affair is on for tomorrow night, and while you are
certainly always welcome to that, it is considered a separate issue from
tonights proposal. Also the Schroppel's will probably be joining us tonight.
∂28-Dec-89 1420 JMC cars
It seems to me that it might be a good idea to take my car over to Gary's
today so it does not have to be done at the crack of dawn tomorrow. How
about it?
∂28-Dec-89 1508 JMC books
Do you want me to hand enter books which are not on Socrates, but which do
have their Library of Congress #'s in them?
∂28-Dec-89 1529 JMC re: books
[In reply to message rcvd 28-Dec-89 15:08-PT.]
Yes on entering non-Socrates books. No on 5pm dinner. How about Saturday
night? I suppose the place is Chez TJ. Can we take the car over
this evening after they close and leave it there, either concealing the
key or pushing it under the door?
∂30-Dec-89 0857 beeson@ucscd.UCSC.EDU overall goals
Received: from ucscd.UCSC.EDU by SAIL.Stanford.EDU with TCP; 30 Dec 89 08:57:34 PST
Received: by ucscd.UCSC.EDU (5.61/1.34)
id AA26378; Sat, 30 Dec 89 08:58:43 -0800
Date: Sat, 30 Dec 89 08:58:43 -0800
From: beeson@ucscd.UCSC.EDU (20012000)
Message-Id: <8912301658.AA26378@ucscd.UCSC.EDU>
To: jmc@sail.stanford.edu
Subject: overall goals
The supplying of an "overall goal" is part of the universal appeal
of religion.
∂30-Dec-89 1818 freeman@uswest.com Re: I'm sending you
Received: from boulder.Colorado.EDU by SAIL.Stanford.EDU with TCP; 30 Dec 89 18:18:49 PST
Return-Path: <freeman@uswest.com>
Received: by boulder.Colorado.EDU (cu-hub.890824)
Received: by uswat.uswest.com (usw-hub.891227)
Received: by yuppieio.uswest.com (usw-generic.890623)
Date: Sat, 30 Dec 89 08:19:50 MST
From: Ed Freeman <freeman@uswest.com>
Message-Id: <8912301519.AA23535@yuppieio.uswest.com>
To: JMC@SAIL.Stanford.EDU
Subject: Re: I'm sending you
Cc: efreeman@uswest.com
Thanks John!
BTW: Do you know of anyone doing research on causal networks as a re
representation formalism for strategic planning type problems?? We have
a working prototype of a system, which is currently being field tested and
I would like to find out what others are doing in this area.
We are also currently working on the design of an "Agent Oriented Programming"
environment which will incorporate causal inference mechanisms into each
agent's repertoire of possible behaviors. We think that this will be a
relatively straightforward way of modeling an agent's "goal directed"
behavior as well as providing a common framework/language within which
an agent can "explain" its actions to others. If you are interested,
I will keep you posted...
Thanks again for the papers,
-ED-
∂31-Dec-89 0059 LES Imagen message
[In reply to message rcvd 31-Dec-89 00:56-PT.]
I take that that that came from Maple. Sorry, but I have no idea what
error message 88 is; Imagen customer service won't be accessible till
Tuesday, unfortunately.
∂31-Dec-89 1029 RIC@RML2.SRI.COM Re: Barnet article
Received: from rml2.sri.com by SAIL.Stanford.EDU with TCP; 31 Dec 89 10:29:18 PST
Date: Sun 31 Dec 89 10:30:27-PST
From: Ric Steinberger <RIC@RML2.SRI.COM>
Subject: Re: Barnet article
To: su-etc@SCORE.STANFORD.EDU, comments@KL.SRI.COM,
jmc@SAIL.STANFORD.EDU
Message-ID: <631132227.890000.RIC@RML2.SRI.COM>
Mail-System-Version: <VAX-MM(246)+TOPSLIB(136)@RML2.SRI.COM>
> From: John McCarthy <JMC@SAIL.Stanford.EDU>
> 2. Indeed the victory in the Cold War is not an American victory.
> Contrary to Barnet, it wasn't a German or Japanese victory either.
> We played the largest role in containing communism, but we didn't
> cause its collapse. What caused its collapse was disgust with
> its corruption, mendaciousness, incompetence and other facts
> by people in the communist countries including people making
> successful careers in the communist parties of these countries.
I'm glad you read the entire article. I think the points
Barnet (of the Institute for Policy Studies [regretibly The
New Yorker never provides any bibliographic details on their
authors]) makes are much more clearly made in 10 pages than
anything I could abstract in a few paragraphs.
I would like to briefly respond to point 2 above. I'm not
sure that Barnet is claiming anything other than the fact that
the conditions imposed on Germany and Japan in 1945 required
them not to "participate" in the Cold War and arms races that
have obsessed both the USA and the USSR for almost 45 years.
In other words, Germany and Japan are not victors in the Cold
War primarily because they were forbidden to fight in it.
They merely inherited a position of strength because excessive
resources in the USA and the USSR were spent trying to
counteract each other's every move.
While "corruption, mendaciousness, incompetence" were
important causes in the decline of communism, it should also
be noted that inefficiency, excessive central planning and
control, intolerance of non-Socialist ideas, supression of
individual ideas and initiatives, and attempting to chart a
political course of action without some cooperation from the
other major nations of the world also lead to the decline of
communism as a workable modern system of governance.
It should be pointed out that "corruption, mendaciousness,
[and] incompetence" are not native only to Socialist states of
the later half of the 20th century. These are afflictions of
both people and states that have lost their vision of a better
world and a better life for future generations. And the
decline of communism in no way guarantees that current forms
of 20th century capitalism, industrialism, or democracy have
an assured future.
-------
∂31-Dec-89 1702 RIC@RML2.SRI.COM Best wishes
Received: from rml2.sri.com by SAIL.Stanford.EDU with TCP; 31 Dec 89 17:01:51 PST
Date: Sun 31 Dec 89 17:02:56-PST
From: Ric Steinberger <RIC@RML2.SRI.COM>
Subject: Best wishes
To: jmc@SAIL.STANFORD.EDU
Message-ID: <631155776.170000.RIC@RML2.SRI.COM>
Mail-System-Version: <VAX-MM(246)+TOPSLIB(136)@RML2.SRI.COM>
John,
I've had a lot of fun "sparing" with you over the past year on
these bboards. Your postings and your responses to some of mine have
given me the chance to rethink some of my basic assumptions. You
have also challenged me in ways that forced me to figure out how to
*write* what I actually thought. Few people will acknowledge what a difficult
task this can be.
I wish you all the best in the coming year and look forward to
a continuing interchange of ideas with you and others in the su-etc
and comments community.
Regards,
Richard Steinberger
-------
∂31-Dec-89 1714 CLT Message for Timothy
I will be home as soon as I TEX my file.