perm filename S86.IN[LET,JMC] blob
sn#820361 filedate 1986-07-02 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00685 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00073 00002 ∂01-Apr-86 0414 wachsmuth%gvaic2.DEC@decwrl.DEC.COM Situation Calculus Planning in blocks & related worlds.
C00075 00003 ∂01-Apr-86 1320 vijay@ernie.berkeley.edu AI DISC: DREYFUS
C00094 00004 ∂01-Apr-86 1325 vijay@ernie.berkeley.edu AI DISC: Winograd Position
C00105 00005 ∂01-Apr-86 1539 VAL Common Sense and Non-Monotonic Reasoning Seminar
C00106 00006 ∂01-Apr-86 1604 ME Prancing Pony Bill
C00108 00007 ∂02-Apr-86 1108 RA fellowship letter
C00109 00008 ∂02-Apr-86 1413 LES Qlisp and Parallel Processors
C00112 00009 ∂02-Apr-86 2024 CLT
C00113 00010 ∂03-Apr-86 1050 SJM you presence
C00114 00011 ∂03-Apr-86 1238 TRACZ@SU-SIERRA.ARPA What do you think?
C00120 00012 ∂03-Apr-86 1348 RA trip to Austin
C00121 00013 ∂03-Apr-86 1358 RA VTSS course
C00122 00014 ∂03-Apr-86 1431 RA not feeling well
C00123 00015 ∂03-Apr-86 1512 KTRACY@SU-SUSHI.ARPA VTSS160 Grade?
C00124 00016 ∂03-Apr-86 1517 vijay@ernie.berkeley.edu
C00147 00017 ∂03-Apr-86 1522 vijay@ernie.berkeley.edu Please acknowledge receipt ...
C00148 00018 ∂03-Apr-86 1611 LES
C00149 00019 ∂03-Apr-86 1727 ROSENSCHEIN@SUMEX-AIM.ARPA "Any fool" reference
C00151 00020 ∂03-Apr-86 1847 LES Qlisp Task Description
C00156 00021 ∂03-Apr-86 1930 LES Qlisp Review
C00157 00022 ∂03-Apr-86 1953 RPG
C00158 00023 ∂03-Apr-86 1954 RPG Milestones
C00159 00024 ∂03-Apr-86 2017 LES EBOS budget
C00160 00025 ∂03-Apr-86 2337 LES EBOS Budget
C00167 00026 ∂04-Apr-86 0158 HIRSH@SUMEX-AIM.ARPA Re: AI Search
C00169 00027 ∂04-Apr-86 1111 LES reply to message
C00170 00028 ∂04-Apr-86 1204 VAL Common Sense and Non-Monotonic Reasoning Seminar
C00172 00029 ∂04-Apr-86 1535 SJM women2[1,sjm]
C00173 00030 ∂04-Apr-86 1658 vijay@ernie.berkeley.edu Re: Charniak
C00175 00031 ∂04-Apr-86 1847 LES Alliant 5th Proposal
C00176 00032 ∂04-Apr-86 1937 RA slides
C00177 00033 ∂05-Apr-86 0236 bibel%germany.csnet@CSNET-RELAY.ARPA [CSNET-RELAY Memo Service: Failed mail (msg.a027248)]
C00183 00034 ∂05-Apr-86 1527 NILSSON@SU-SCORE.ARPA Latombe
C00185 00035 ∂05-Apr-86 1626 NILSSON@SU-SCORE.ARPA AI Prospects and Visits
C00189 00036 ∂07-Apr-86 0506 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Your visit
C00191 00037 ∂07-Apr-86 0623 somewhere!ito@aoba.tohoku.junet
C00195 00038 ∂07-Apr-86 1359 LES Qlisp Task Description
C00207 00039 ∂07-Apr-86 1522 CLT rev1(u,v*w)=rev1(u,v)*w
C00211 00040 ∂08-Apr-86 0652 BMOORE@SRI-AI.ARPA Raphael's paper on knowledge
C00213 00041 ∂08-Apr-86 0911 CLT
C00214 00042 ∂08-Apr-86 1556 LES Symbolics Meeting
C00215 00043 ∂08-Apr-86 1608 CONTRERAS@SU-SCORE.ARPA message
C00216 00044 ∂09-Apr-86 0947 AAAI-OFFICE@SUMEX-AIM.ARPA Program Committee Reception
C00218 00045 ∂09-Apr-86 1033 vijay@ernie.berkeley.edu AI DISC: Douglas Hofstadter
C00246 00046 ∂09-Apr-86 1129 KOHEN@SU-SUSHI.ARPA Errata for Lisp Programming & Proving
C00248 00047 ∂09-Apr-86 1134 RA finals CS306
C00249 00048 ∂09-Apr-86 1206 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Re: visit
C00250 00049 ∂09-Apr-86 1538 YAO@SU-SCORE.ARPA admission of R. Zabih
C00252 00050 ∂09-Apr-86 1725 CLT Zabih
C00253 00051 ∂10-Apr-86 0720 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Your visit
C00255 00052 ∂10-Apr-86 1138 VAL Today's Non-Monotonic Seminar: Reminder
C00257 00053 ∂10-Apr-86 1416 SJG pointer needed
C00258 00054 ∂10-Apr-86 1550 HIRSH@SUMEX-AIM.ARPA Mitchell visit
C00260 00055 ∂10-Apr-86 1655 LES Request for DARPA lecture
C00262 00056 ∂10-Apr-86 1749 VAL Common Sense and Non-Monotonic Reasoning Seminar
C00265 00057 ∂11-Apr-86 0035 LES The next contest . . .
C00266 00058 ∂11-Apr-86 0410 gideon%edai.edinburgh.ac.uk@Cs.Ucl.AC.UK Re: Your visit
C00267 00059 ∂11-Apr-86 0951 UNIETIS@SU-SUSHI.ARPA CS 306
C00269 00060 ∂11-Apr-86 1122 guibas@decwrl.DEC.COM Miklos Simonovits
C00271 00061 ∂11-Apr-86 1402 CLT japan collaboration
C00273 00062 ∂11-Apr-86 1737 vijay@ernie.berkeley.edu
C00276 00063 ∂12-Apr-86 0503 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Your visit
C00278 00064 ∂12-Apr-86 1452 RDZ@MC.LCS.MIT.EDU Graduate School Paperwork
C00280 00065 ∂13-Apr-86 0826 CHEADLE@SU-SCORE.ARPA Re: Graduate School Paperwork
C00282 00066 ∂14-Apr-86 0034 Carnese@SRI-KL.ARPA talk at May CPSR meeting
C00283 00067 ∂14-Apr-86 0546 PETTY@RED.RUTGERS.EDU 1986-April-technical mailing
C00287 00068 ∂14-Apr-86 0850 WINOGRAD@SU-CSLI.ARPA Re: policy
C00289 00069 ∂14-Apr-86 1332 WINOGRAD@SU-CSLI.ARPA Fellowships in AI←
C00291 00070 ∂14-Apr-86 1525 tgd%oregon-state.csnet@CSNET-RELAY.ARPA Request for funding: Knowledge Compilation Workshop
C00293 00071 ∂15-Apr-86 1004 RA inactive files
C00294 00072 ∂15-Apr-86 1529 RA IBM Almaden dedication ceremony
C00295 00073 ∂15-Apr-86 2231 HIRSH@SUMEX-AIM.ARPA Mitchell meeting times
C00297 00074 ∂15-Apr-86 2331 vijay@ernie.berkeley.edu AI Disc: Joseph Weizenbaum
C00309 00075 ∂16-Apr-86 1129 RA Meeting Tom Mitchell
C00310 00076 ∂16-Apr-86 1204 RA telex from Milan
C00311 00077 ∂16-Apr-86 1419 LES DARPA $ spending
C00324 00078 ∂16-Apr-86 1507 binford@su-whitney.arpa DARPA $ spending
C00325 00079 ∂16-Apr-86 1624 CLOUTIER@SU-SIERRA.ARPA NAE Nomination Forms
C00326 00080 ∂16-Apr-86 1742 VAL Reminder: Common Sense and Non-Monotonic Reasoning Seminar
C00329 00081 ∂16-Apr-86 1808 lantz@su-gregorio.arpa my Sun's
C00330 00082 ∂16-Apr-86 2229 vijay@ernie.berkeley.edu AI DISC: David Rumelhart
C00342 00083 ∂17-Apr-86 0327 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Paper
C00344 00084 ∂17-Apr-86 1414 RA Mr. Baltes
C00345 00085 ∂17-Apr-86 1622 LES DARPA Equipment Finish (?)
C00347 00086 ∂17-Apr-86 1653 WIEDERHOLD@SU-SCORE.ARPA Industrial lecturers
C00349 00087 ∂17-Apr-86 1657 gnelson@decwrl.DEC.COM Industrial Lecturship
C00350 00088 ∂17-Apr-86 1758 LES Possible time change
C00351 00089 ∂18-Apr-86 0449 BONNIE%upenn.csnet@CSNET-RELAY.ARPA AI Fellowships
C00352 00090 ∂18-Apr-86 0815 RICHARDSON@SU-SCORE.ARPA AI Prospects and Visits
C00354 00091 ∂18-Apr-86 1031 RA AI prospects
C00355 00092 ∂18-Apr-86 1328 binford@su-whitney.arpa Possible time change
C00356 00093 ∂18-Apr-86 1347 LES Meeting Confirmation
C00357 00094 ∂18-Apr-86 1443 RICHARDSON@SU-SCORE.ARPA Etherington/AI Prospect
C00359 00095 ∂18-Apr-86 1457 VARDI@IBM.COM
C00360 00096 ∂18-Apr-86 1900 LES DARPA update
C00371 00097 ∂18-Apr-86 1908 vijay@ernie.berkeley.edu AI DISC: NOTE
C00372 00098 ∂20-Apr-86 0000 JMC
C00373 00099 ∂20-Apr-86 1951 RPG Invitation
C00375 00100 ∂21-Apr-86 0825 RICHARDSON@SU-SCORE.ARPA AI Search Committee
C00376 00101 ∂21-Apr-86 1020 guibas@decwrl.DEC.COM re: Miklos Simonovits
C00378 00102 ∂21-Apr-86 1118 gnelson@decwrl.DEC.COM re: Industrial Lecturship
C00380 00103 ∂21-Apr-86 1338 RICHARDSON@SU-SCORE.ARPA Yoav Shoham
C00382 00104 ∂21-Apr-86 1540 FEIGENBAUM@SUMEX-AIM.ARPA [Ellie Engelmore <EEngELmore@SUMEX-AIM.ARPA>: SJ Mercury/Book Review request]
C00385 00105 ∂21-Apr-86 1545 FEIGENBAUM@SUMEX-AIM.ARPA [Anne Richardson <RICHARDSON@SU-SCORE.ARPA>: Faculty Candidates]
C00388 00106 ∂21-Apr-86 1606 FEIGENBAUM@SUMEX-AIM.ARPA re: [Ellie Engelmore <EEngELmore@SUMEX-AIM.ARPA>: SJ Mercury/Book Review request]
C00390 00107 ∂21-Apr-86 2248 HST 30 years conference
C00391 00108 ∂22-Apr-86 0906 WINOGRAD@SU-CSLI.ARPA PhD Minor requirement
C00402 00109 ∂22-Apr-86 1011 KARP@SUMEX-AIM.ARPA Re: PhD Minor requirement
C00403 00110 ∂22-Apr-86 1130 nttlab!NTT-20!Goto@su-shasta.arpa Common Business Communication Language
C00406 00111 ∂22-Apr-86 1249 LES re: The next contest . . .
C00407 00112 ∂22-Apr-86 1306 VAL Common Sense and Non-Monotonic Reasoning Seminar
C00410 00113 ∂22-Apr-86 1454 pratt@su-navajo.arpa PhD Minor requirement
C00411 00114 ∂22-Apr-86 1607 ME net adr
C00412 00115 ∂22-Apr-86 1616 ME Lathrop
C00414 00116 ∂22-Apr-86 2028 GENESERETH@SUMEX-AIM.ARPA Re: Yoav Shoham
C00416 00117 ∂22-Apr-86 2259 CLT books to by re ebos
C00417 00118 ∂23-Apr-86 0803 RICHARDSON@SU-SCORE.ARPA AI Prospect David Etherington
C00418 00119 ∂23-Apr-86 0803 RICHARDSON@SU-SCORE.ARPA AI Prospective/Shoham
C00419 00120 ∂23-Apr-86 0900 JMC
C00420 00121 ∂23-Apr-86 0912 CLT calendar item
C00421 00122 ∂23-Apr-86 1123 SJM Kugel's address
C00422 00123 ∂23-Apr-86 1209 CLT japan
C00423 00124 ∂23-Apr-86 2053 RESTIVO@SU-SCORE.ARPA Review of Sophist Dreyfus
C00538 00125 ∂24-Apr-86 0032 Carnese@SRI-KL.ARPA Re: May 14 CPSR meeting
C00539 00126 ∂24-Apr-86 0841 RA receipt
C00540 00127 ∂24-Apr-86 0936 SJG re: pointer needed
C00541 00128 ∂24-Apr-86 0944 RPG F-111
C00542 00129 ∂24-Apr-86 0959 SJG re: pointer needed
C00543 00130 ∂24-Apr-86 1129 CHEADLE@SU-SCORE.ARPA Gray Tuesday
C00547 00131 ∂24-Apr-86 1152 CHEADLE@SU-SCORE.ARPA Gray Tuesday
C00552 00132 ∂24-Apr-86 1200 RICHARDSON@SU-SCORE.ARPA AI Candidate - Batali
C00554 00133 ∂24-Apr-86 1346 BRONSTEIN@SU-SCORE.ARPA Re: We should give diplomacy a chance in Libya.
C00556 00134 ∂24-Apr-86 1645 NILSSON@SU-SCORE.ARPA Re: Batali
C00557 00135 ∂24-Apr-86 1720 EENGELMORE@SUMEX-AIM.ARPA Rozak Book Review
C00559 00136 ∂24-Apr-86 1811 Y.YDUJ@LOTS-B.STANFORD.EDU re: fun (from SAIL's BBOARD)
C00561 00137 ∂24-Apr-86 2030 JMC
C00562 00138 ∂25-Apr-86 0800 JMC
C00563 00139 ∂25-Apr-86 0953 RA taking the afternoon off
C00564 00140 ∂25-Apr-86 1146 RA leaving
C00565 00141 ∂25-Apr-86 1156 NILSSON@SU-SCORE.ARPA Etherington
C00567 00142 ∂25-Apr-86 1527 avg@su-aimvax.arpa Orals committee?
C00570 00143 ∂25-Apr-86 1634 RICHARDSON@SU-SCORE.ARPA Etherington
C00571 00144 ∂25-Apr-86 1639 RICHARDSON@SU-SCORE.ARPA Etherington
C00573 00145 ∂25-Apr-86 2016 vijay@ernie.berkeley.edu
C00574 00146 ∂25-Apr-86 2017 vijay@ernie.berkeley.edu AI DISC.
C00576 00147 ∂25-Apr-86 2016 vijay@ernie.berkeley.edu
C00577 00148 ∂26-Apr-86 1406 @SU-SCORE.ARPA:BUCHANAN@SUMEX-AIM.ARPA FYI -- maybe 55 but it will only take us 10 years!
C00587 00149 ∂27-Apr-86 1252 CLT
C00590 00150 ∂27-Apr-86 1836 shoham@YALE.ARPA Re: Chronological ignorance
C00594 00151 ∂28-Apr-86 1151 VARDI@IBM.COM
C00596 00152 ∂28-Apr-86 1206 lantz@su-gregorio.arpa Encore
C00598 00153 ∂28-Apr-86 1207 RA lunch
C00599 00154 ∂28-Apr-86 1416 VAL re: Chronological ignorance
C00601 00155 ∂28-Apr-86 1613 VAL re: Chronological ignorance
C00603 00156 ∂28-Apr-86 2000 SJM
C00604 00157 ∂29-Apr-86 0933 CLT calendar item
C00605 00158 ∂29-Apr-86 0957 RICHARDSON@SU-SCORE.ARPA Yoav Shoham
C00606 00159 ∂29-Apr-86 0958 RA library
C00607 00160 ∂29-Apr-86 1057 SJM computer failure
C00608 00161 ∂29-Apr-86 1107 KOLKOWITZ@SU-SCORE.ARPA
C00610 00162 ∂29-Apr-86 1200 JMC
C00611 00163 ∂29-Apr-86 1204 LLW@S1-A.ARPA Pournelle's Affair
C00613 00164 ∂29-Apr-86 1215 GSMITH@SUMEX-AIM.ARPA AI Qualifying Exams, 5/20 and 5/23
C00616 00165 ∂29-Apr-86 1312 CLT SAIL lines
C00617 00166 ∂29-Apr-86 1336 RA John Kwapisz
C00618 00167 ∂29-Apr-86 1407 VAL Shoham's paper
C00620 00168 ∂29-Apr-86 1424 TAJNAI@SU-SCORE.ARPA Dr. Peled of IBM
C00622 00169 ∂29-Apr-86 1519 RA Sally Tuttle
C00623 00170 ∂29-Apr-86 2000 SJM
C00624 00171 ∂30-Apr-86 0114 HST lisp standardization and conference
C00625 00172 ∂30-Apr-86 0925 RICHARDSON@SU-SCORE.ARPA Yoav Shoham
C00627 00173 ∂30-Apr-86 1034 NILSSON@SU-SCORE.ARPA Re: AI Qualifying Exams, 5/20 and 5/23
C00629 00174 ∂30-Apr-86 1131 LES Asking for the RT (again)
C00630 00175 ∂30-Apr-86 1151 VAL Shoham
C00631 00176 ∂30-Apr-86 1341 RPG Rumor
C00632 00177 ∂30-Apr-86 1500 JMC
C00633 00178 ∂30-Apr-86 1526 RA coming in late
C00634 00179 ∂30-Apr-86 1509 SJM produc
C00635 00180 ∂30-Apr-86 1538 CLT to try kcl
C00636 00181 ∂30-Apr-86 1542 kolk@su-carmel.arpa ibmrtpc1
C00637 00182 ∂30-Apr-86 1646 VAL Reminder: Non-Monotonic Reasoning Seminar
C00640 00183 ∂30-Apr-86 1824 vijay@ernie.berkeley.edu Re: are you there?
C00642 00184 ∂30-Apr-86 1857 vijay@ernie.berkeley.edu AI DISC: Hofstadter's reply
C00664 00185 ∂30-Apr-86 1941 vijay@ernie.berkeley.edu AI DISC: Charniak's Response
C00681 00186 ∂30-Apr-86 2000 JMC
C00682 00187 ∂30-Apr-86 2000 JMC
C00683 00188 ∂30-Apr-86 2039 RWW kcl
C00684 00189 ∂30-Apr-86 2045 RWW fausto@sumex
C00685 00190 ∂30-Apr-86 2047 RWW mad
C00686 00191 ∂30-Apr-86 2225 UMA@IBM.COM
C00688 00192 ∂01-May-86 0102 HST lisp 30 yeArs
C00690 00193 ∂01-May-86 0900 JMC
C00691 00194 ∂01-May-86 1208 HST 30years lisp
C00692 00195 ∂01-May-86 1419 TAJNAI@SU-SCORE.ARPA Inference Corporation
C00693 00196 ∂01-May-86 1426 TAJNAI@SU-SCORE.ARPA re: Inference Corporation
C00695 00197 ∂01-May-86 1434 GIUNCHIGLIA@SUMEX-AIM.ARPA
C00696 00198 ∂01-May-86 1630 RA leaving
C00697 00199 ∂01-May-86 1642 LES
C00699 00200 ∂01-May-86 1731 VAL Etherington and Shoham
C00701 00201 ∂01-May-86 1801 NILSSON@SU-SCORE.ARPA
C00702 00202 ∂01-May-86 1827 NILSSON@SU-SCORE.ARPA cbcl
C00703 00203 ∂01-May-86 1953 JACOBS@su-sushi.arpa re: the undersecretary of idiocy (from SAIL's BBOARD)
C00705 00204 ∂01-May-86 2127 ME Prancing Pony Bill
C00707 00205 ∂02-May-86 0742 TOM@SU-SCORE.ARPA
C00709 00206 ∂02-May-86 0800 JMC
C00710 00207 ∂02-May-86 0840 RICHARDSON@SU-SCORE.ARPA John Batali
C00713 00208 ∂02-May-86 0847 TAJNAI@SU-SCORE.ARPA re: Dr. Peled of IBM
C00715 00209 ∂02-May-86 0915 NILSSON@SU-SCORE.ARPA re: cbcl
C00716 00210 ∂02-May-86 1206 MOGUL@SU-SCORE.ARPA re: the undersecretary of idiocy (from SAIL's BBOARD)
C00720 00211 ∂02-May-86 1134 SJM carolyn
C00722 00212 ∂02-May-86 1531 MOGUL@SU-SCORE.ARPA re: Purge of dissident scientists? (from SAIL's BBOARD)
C00725 00213 ∂03-May-86 1121 HALPERN@IBM.COM
C00734 00214 ∂04-May-86 2146 PMF Weekend Plans Shift
C00736 00215 ∂04-May-86 2227 LES Facilities Committee Meetings of 4/23/86, etc.
C00762 00216 ∂05-May-86 0817 EPPLEY@SU-SCORE.ARPA Rutie
C00763 00217 ∂05-May-86 0900 JMC
C00764 00218 ∂05-May-86 0906 RICHARDSON@SU-SCORE.ARPA Jean-Claude Latombe
C00766 00219 ∂05-May-86 0930 CLT dinner with edwina and oliver
C00767 00220 ∂05-May-86 0949 CONTRERAS@SU-SCORE.ARPA message
C00768 00221 ∂05-May-86 0959 CLT dinner with edwina and oliver
C00769 00222 ∂05-May-86 1016 BIL@SU-SIERRA.ARPA Have you ever made a mistake?...
C00771 00223 ∂05-May-86 1023 VAL re: Comments on your paper conjecture settled
C00772 00224 ∂05-May-86 1347 RICHARDSON@SU-SCORE.ARPA Robotics Search Committee
C00774 00225 ∂05-May-86 1502 CLT
C00775 00226 ∂05-May-86 1531 LES Request for IBM RT PCs for Editor-based Operating System
C00778 00227 ∂05-May-86 1625 VAL Common Sense and Non-Monotonic Reasoning Seminar
C00780 00228 ∂05-May-86 1706 MOGUL@SU-SCORE.ARPA re: Purge of dissident scientists? (from SAIL's BBOARD)
C00783 00229 ∂05-May-86 1803 pjd@RIACS.ARPA Scientific Contributions of CS
C00790 00230 ∂05-May-86 2014 LES
C00804 00231 ∂05-May-86 2246 pjd@RIACS.ARPA Further Comments
C00814 00232 ∂06-May-86 0740 RICHARDSON@SU-SCORE.ARPA Principia Group
C00815 00233 ∂06-May-86 0956 RA John Kwapisz
C00816 00234 ∂06-May-86 1011 RA Re: hat and ring sizes and trip
C00817 00235 ∂06-May-86 1028 BURY@su-sushi.arpa CS 306 grade from Fall, remote
C00818 00236 ∂06-May-86 1040 NILSSON@SU-SCORE.ARPA Re: Batali
C00820 00237 ∂06-May-86 1040 NILSSON@SU-SCORE.ARPA
C00821 00238 ∂06-May-86 1119 RICHARDSON@SU-SCORE.ARPA re: Robotics Search Committee
C00822 00239 ∂06-May-86 1159 RICHARDSON@SU-SCORE.ARPA Robotics Search Committee/May 14
C00823 00240 ∂06-May-86 1325 RA John Kwapisz
C00824 00241 ∂06-May-86 1338 LES Facilities Budget Revision
C00826 00242 ∂06-May-86 1402 binford@su-whitney.arpa Jean-Claude Latombe
C00827 00243 ∂06-May-86 1405 binford@su-whitney.arpa Robotics Search Committee
C00828 00244 ∂06-May-86 1639 RICHARDSON@SU-SCORE.ARPA Robotics Search Committee/May 21
C00829 00245 ∂06-May-86 1831 binford@su-whitney.arpa Robotics Search Committee/May 14
C00831 00246 ∂06-May-86 2000 SJM
C00832 00247 ∂07-May-86 0656 JJW RA support for summer
C00833 00248 ∂07-May-86 0955 CLT rt unix upgrade
C00834 00249 ∂07-May-86 1114 NILSSON@SU-SCORE.ARPA [Tom Garvey <Garvey@SRI-AI.ARPA>: Danny Cohen visit]
C00836 00250 ∂07-May-86 1150 GIVAN@su-sushi.arpa summer work
C00838 00251 ∂07-May-86 1230 RA John Coke
C00839 00252 ∂07-May-86 1233 CLT dinner on Thursday.
C00840 00253 ∂07-May-86 1235 CLT calendar item
C00841 00254 ∂07-May-86 1332 LES Draft EBOS proposal
C00862 00255 ∂07-May-86 1413 RA Doug Spitz
C00863 00256 ∂07-May-86 1443 RICHARDSON@SU-SCORE.ARPA Follow-up Robotics Meeting/May 21
C00865 00257 ∂07-May-86 1533 CLT Remarks on Draft EBOS proposal
C00868 00258 ∂07-May-86 1810 LES re: Remarks on Draft EBOS proposal
C00870 00259 ∂07-May-86 1934 LES Datalink alternatives
C00874 00260 ∂08-May-86 0810 RICHARDSON@SU-SCORE.ARPA AI Faculty Candidate/Shankar
C00876 00261 ∂08-May-86 0810 RICHARDSON@SU-SCORE.ARPA AI Search Committee
C00878 00262 ∂08-May-86 0938 ullman@su-aimvax.arpa MS candidate Hirani
C00880 00263 ∂08-May-86 0954 VAL Reminder: Non-Monotonic Reasoning Seminar
C00882 00264 ∂08-May-86 1038 SJM good news v bad news
C00883 00265 ∂08-May-86 1054 RICHARDSON@SU-SCORE.ARPA AI Search/May 21
C00884 00266 ∂08-May-86 1054 NILSSON@SU-SCORE.ARPA re: [Tom Garvey <Garvey@SRI-AI.ARPA>: Danny Cohen visit]
C00885 00267 ∂08-May-86 1113 LES Alliant name
C00886 00268 ∂08-May-86 1150 CHEADLE@SU-SCORE.ARPA Conferral of MSAI diplomas
C00888 00269 ∂08-May-86 1224 knapp@gvax.cs.cornell.edu Report on scientific computing
C00891 00270 ∂08-May-86 1226 alliant!jat@EDDIE.MIT.EDU alliant system for q-lisp
C00894 00271 ∂08-May-86 1344 kolk@su-navajo.arpa Alliant name and address
C00895 00272 ∂08-May-86 1410 RINDFLEISCH@SUMEX-AIM.ARPA Request for Help
C00901 00273 ∂08-May-86 1623 RA leaving now
C00902 00274 ∂08-May-86 1750 LES Next meeting
C00903 00275 ∂08-May-86 2206 GOLUB@SU-SCORE.ARPA re: Letter to USSR
C00905 00276 ∂09-May-86 0708 HST scott
C00906 00277 ∂09-May-86 1050 RA transparencies
C00907 00278 ∂09-May-86 1243 jeh@gvax.cs.cornell.edu NSF report on computing
C00911 00279 ∂09-May-86 1435 RA doctor's appointment
C00912 00280 ∂09-May-86 1630 VAL pointwise circ'n paper
C00913 00281 ∂09-May-86 1650 ME bike locker unlocked with bike in it
C00914 00282 ∂11-May-86 1050 jeh@gvax.cs.cornell.edu re: NSF report on computing
C00916 00283 ∂11-May-86 1231 NILSSON@SU-SCORE.ARPA Jobs (Steve)
C00917 00284 ∂11-May-86 1351 NILSSON@SU-SCORE.ARPA AI search
C00921 00285 ∂11-May-86 2035 rms@PREP.AI.MIT.EDU I'm visiting
C00922 00286 ∂12-May-86 0832 RICHARDSON@SU-SCORE.ARPA Shankar
C00923 00287 ∂12-May-86 1229 LES DARPA Order received!
C00924 00288 ∂12-May-86 1444 CLT rt unix upgrade - revisited
C00925 00289 ∂13-May-86 1404 RA Conference on Computers and Mathematics, July 29, 1986
C00926 00290 ∂13-May-86 1407 RA Olivia Stuart, Time Mag.
C00927 00291 ∂13-May-86 1432 @SUMEX-AIM.ARPA:MRC@PANDA PDP-6 CONS instruction
C00929 00292 ∂13-May-86 1511 VAL Common Sense and Non-Monotonic Reasoning Seminar
C00932 00293 ∂13-May-86 1530 lantz@su-gregorio.arpa FYI
C00934 00294 ∂13-May-86 1533 RA Judith Shoolery
C00935 00295 ∂13-May-86 1548 GIVAN@su-sushi.arpa summer work
C00937 00296 ∂13-May-86 1550 VAL MCC
C00938 00297 ∂13-May-86 1803 JMC
C00939 00298 ∂13-May-86 1806 NILSSON@SU-SCORE.ARPA Shankar
C00941 00299 ∂13-May-86 1807 CLT shopping list
C00942 00300 ∂13-May-86 1810 JJW Re: PDP-6 CONS instruction
C00944 00301 ∂13-May-86 1819 CLT phone msg
C00945 00302 ∂13-May-86 2000 JMC
C00946 00303 ∂13-May-86 2017 @SUMEX-AIM.ARPA:MRC@PANDA Re: PDP-6 CONS instruction
C00948 00304 ∂14-May-86 0001 LES PDP-6 CONS instruction
C00950 00305 ∂14-May-86 0621 FAHLMAN@C.CS.CMU.EDU copyright holder
C00952 00306 ∂14-May-86 0922 RICHARDSON@SU-SCORE.ARPA Shankar
C00953 00307 ∂14-May-86 1000 JMC
C00954 00308 ∂14-May-86 1004 RICHARDSON@SU-SCORE.ARPA AI Search Committee/May 21
C00955 00309 ∂14-May-86 1001 RA John Hopcroft
C00956 00310 ∂14-May-86 1005 RA Susan Bishop, TV Ontario
C00957 00311 ∂14-May-86 1157 RA lunch
C00958 00312 ∂14-May-86 1350 RA Meeting with Shankar
C00959 00313 ∂14-May-86 1351 RA AI Search Committee
C00960 00314 ∂14-May-86 1452 NILSSON@SU-SCORE.ARPA Hooray!
C00963 00315 ∂14-May-86 1507 KTRACY@su-sushi.arpa Your book and my VTSS160 paper
C00965 00316 ∂14-May-86 1727 WINOGRAD@SU-CSLI.ARPA Meeting about syllabus and proposal for it
C00973 00317 ∂14-May-86 2000 JMC
C00974 00318 ∂15-May-86 0000 JMC
C00975 00319 ∂15-May-86 0131 HST lisp30
C00976 00320 ∂15-May-86 0929 GIVAN@su-sushi.arpa
C00977 00321 ∂15-May-86 1050 perlis@mimsy.umd.edu circumscriptive aspects
C00981 00322 ∂15-May-86 1059 GSMITH@SUMEX-AIM.ARPA AI Qual
C00983 00323 ∂15-May-86 1124 RA airline tickets
C00984 00324 ∂15-May-86 1147 VAL re: circumscriptive aspects
C00985 00325 ∂15-May-86 1328 GIVAN@su-sushi.arpa Re: reply to message
C00987 00326 ∂15-May-86 1359 RA Chuck Williams
C00988 00327 ∂15-May-86 1414 VAL Reminder: Non-Monotonic Reasoning Seminar
C00991 00328 ∂15-May-86 1621 RA leaving early
C00992 00329 ∂15-May-86 1730 CLT
C00993 00330 ∂15-May-86 1732 CLT
C00994 00331 ∂16-May-86 1200 jeh@gvax.cs.cornell.edu NSF report
C01003 00332 ∂16-May-86 1258 WINOGRAD@SU-CSLI.ARPA Meeting Tuesday at 2:15
C01004 00333 ∂16-May-86 1435 TAJNAI@SU-SCORE.ARPA Dr. Peled's visit
C01006 00334 ∂16-May-86 1625 SJM construction
C01007 00335 ∂16-May-86 1645 VAL pointwise circ'n paper
C01017 00336 ∂16-May-86 1746 RA Olivia Stuart, Time
C01018 00337 ∂16-May-86 1748 RA AI Search Committee
C01019 00338 ∂16-May-86 1808 NILSSON@SU-SCORE.ARPA [Brian Fromme <FROMME@SRI-WARBUCKS.ARPA>: [Tom Garvey <GARVEY@SRI-IU.ARPA>: [Pony Express Mailer]
C01023 00339 ∂17-May-86 1131 SJG should I sent this to BBoards about the uncertainty workshop?
C01026 00340 ∂17-May-86 1220 NILSSON@SU-SCORE.ARPA hopcroft report
C01027 00341 ∂17-May-86 1259 LES Alliant Schedule
C01028 00342 ∂17-May-86 1808 CLT shopping list
C01029 00343 ∂17-May-86 1818 ROMERO@SRI-AI.ARPA Symbolics Programmer
C01037 00344 ∂18-May-86 1111 WEISS@su-sushi.arpa re: Turning off call waiting
C01038 00345 ∂18-May-86 1703 IAM
C01039 00346 ∂18-May-86 2020 FRANK@SU-CSLI.ARPA re: Turning off call-waiting (from SAIL's BBOARD)
C01041 00347 ∂18-May-86 2032 SIEGMAN@SU-SIERRA.ARPA Call Forwarding
C01043 00348 ∂19-May-86 0849 RA Re: reference
C01044 00349 ∂19-May-86 0907 RA Expense Report (LA and Washington)
C01045 00350 ∂19-May-86 0918 CHEADLE@SU-SCORE.ARPA Ph.D. Committee meeting
C01046 00351 ∂19-May-86 1041 CRISPIN@SUMEX-AIM.ARPA re: Turning off call-waiting (from SAIL's BBOARD)
C01048 00352 ∂19-May-86 1056 RA TEX course
C01049 00353 ∂19-May-86 1224 RA Inference Board Meeting
C01050 00354 ∂19-May-86 1301 RA meeting
C01051 00355 ∂19-May-86 1322 VAL Common Sense and Non-Monotonic Reasoning Seminar
C01053 00356 ∂19-May-86 1348 LES EBOS Kickoff
C01054 00357 ∂19-May-86 1448 SJM Dorothy Parker
C01055 00358 ∂19-May-86 1516 RICHARDSON@SU-SCORE.ARPA AI Search Committee/May 21
C01056 00359 ∂19-May-86 1538 RA Re: TEX course
C01057 00360 ∂19-May-86 1658 RA Hertz receipt
C01058 00361 ∂19-May-86 1717 RA John Walecka
C01059 00362 ∂20-May-86 0923 CONTRERAS@SU-SCORE.ARPA message
C01060 00363 ∂20-May-86 1107 RA les
C01061 00364 ∂20-May-86 1113 RA Olivia Stuart
C01062 00365 ∂20-May-86 1637 RA Steve Jobs
C01063 00366 ∂20-May-86 1728 @SRI-AI.ARPA:LANSKY@SRI-WARBUCKS.ARPA YELLOW FORMS PLEASE!
C01075 00367 ∂20-May-86 1758 LANSKY@SRI-WARBUCKS.ARPA [Pony Express Mailer <Mailer@SRI-WARBUCKS.ARPA>:
C01088 00368 ∂21-May-86 0506 FEIGENBAUM@SUMEX-AIM.ARPA AI Search/Opinion
C01094 00369 ∂21-May-86 0900 JMC
C01095 00370 ∂21-May-86 1009 JJW Alliant
C01096 00371 ∂21-May-86 1021 UNIETIS@su-sushi.arpa cs306
C01098 00372 ∂21-May-86 1021 RICHARDSON@SU-SCORE.ARPA Robotics Search Committee
C01099 00373 ∂21-May-86 1146 RA
C01100 00374 ∂21-May-86 1443 VAL Mints' address
C01101 00375 ∂21-May-86 1509 WINOGRAD@SU-CSLI.ARPA PhD Program revision
C01112 00376 ∂21-May-86 1504 VAL mcc visit
C01113 00377 ∂21-May-86 1517 RA Jack Harper
C01114 00378 ∂21-May-86 1520 RA Academic costume
C01115 00379 ∂21-May-86 1558 KARP@SUMEX-AIM.ARPA Review of the comp reading list
C01117 00380 ∂21-May-86 1901 CLT shopping list
C01118 00381 ∂21-May-86 2000 JMC
C01119 00382 ∂22-May-86 0030 VAL Reminder: Non-Monotonic Reasoning Seminar
C01121 00383 ∂22-May-86 0609 F68349%BARILAN.BITNET@WISCVM.WISC.EDU Logician
C01127 00384 ∂22-May-86 0800 JMC
C01128 00385 ∂22-May-86 0900 JMC
C01129 00386 ∂22-May-86 1006 CHEADLE@SU-SCORE.ARPA Friday's AI Qual
C01131 00387 ∂22-May-86 1122 RA Jack Harper
C01132 00388 ∂22-May-86 1200 JMC
C01133 00389 ∂22-May-86 1254 GRP Files on IBM RT
C01134 00390 ∂22-May-86 1331 LES Qlisp task status
C01136 00391 ∂22-May-86 1550 VAL Perlis on Circ. with sets
C01137 00392 ∂22-May-86 1701 SJM book
C01138 00393 ∂22-May-86 2041 AN02@A.CS.CMU.EDU Re: reference to GPS
C01144 00394 ∂23-May-86 0130 GRP Unix system calls
C01145 00395 ∂23-May-86 0131 GRP files as active objects
C01146 00396 ∂23-May-86 0609 jeh@gvax.cs.cornell.edu Memo about the NSF report on scientific computing
C01148 00397 ∂23-May-86 0950 lantz@su-gregorio.arpa comp
C01152 00398 ∂23-May-86 1013 G.GINNIE@LOTS-B.STANFORD.EDU Party Corrections
C01154 00399 ∂23-May-86 1016 G.GINNIE@LOTS-B.STANFORD.EDU *PARTY*
C01158 00400 ∂23-May-86 1515 SJM Saturday schedule
C01159 00401 ∂23-May-86 1627 KARP@SUMEX-AIM.ARPA Comp readings library
C01161 00402 ∂23-May-86 1624 VAL pointwise circ'n paper
C01162 00403 ∂23-May-86 1636 SJM saturday schedule
C01163 00404 ∂24-May-86 1437 JIML@SU-CSLI.ARPA re: Turning off call-waiting (from SAIL's BBOARD)
C01164 00405 ∂24-May-86 1444 LES Mason trip request
C01165 00406 ∂24-May-86 1455 LES
C01167 00407 ∂24-May-86 1458 IAM
C01168 00408 ∂25-May-86 0827 GZT.TDF%OZ.AI.MIT.EDU@AI.AI.MIT.EDU
C01170 00409 ∂25-May-86 0900 JMC
C01171 00410 ∂25-May-86 0907 GZT.TDF%OZ.AI.MIT.EDU@AI.AI.MIT.EDU reply to message
C01173 00411 ∂25-May-86 1832 vijay@ernie.Berkeley.EDU Re: any sign of life?
C01175 00412 ∂25-May-86 1838 vijay@ernie.Berkeley.EDU Correction...
C01176 00413 ∂25-May-86 1849 vijay@ernie.Berkeley.EDU Weizenbaum's stage 3 comments
C01197 00414 ∂25-May-86 1851 vijay@ernie.Berkeley.EDU Dreyfus's stage 3 comments
C01204 00415 ∂25-May-86 1859 vijay@ernie.Berkeley.EDU Papert's name...
C01206 00416 ∂26-May-86 0900 JMC
C01207 00417 ∂26-May-86 1124 NILSSON@SU-SCORE.ARPA Tom Mitchell
C01210 00418 ∂26-May-86 1732 vijay@ernie.Berkeley.EDU Replies....
C01211 00419 ∂26-May-86 1800 JMC
C01212 00420 ∂27-May-86 0700 JMC
C01213 00421 ∂27-May-86 0922 JMC
C01214 00422 ∂27-May-86 1006 CONTRERAS@SU-SCORE.ARPA message
C01216 00423 ∂27-May-86 1030 JMC
C01217 00424 ∂27-May-86 1123 RA References
C01218 00425 ∂27-May-86 1205 GRP Reminder: next meeting June 10 at 2pm
C01219 00426 ∂27-May-86 1209 JJW Alliant now on Ethernet
C01222 00427 ∂27-May-86 1349 RA Jack Harper
C01223 00428 ∂27-May-86 1444 RA [Reply to message recvd: 27 May 86 14:41 Pacific Time]
C01224 00429 ∂27-May-86 1609 RA New seats
C01225 00430 ∂27-May-86 1619 VAL Common Sense and Non-Monotonic Reasoning Seminar
C01228 00431 ∂27-May-86 1648 LES
C01230 00432 ∂27-May-86 1713 GRP editor ramblings
C01231 00433 ∂28-May-86 0102 LES Facilities Committee Meeting Today
C01236 00434 ∂28-May-86 0836 ROSENBLOOM@SUMEX-AIM.ARPA Re: typo
C01237 00435 ∂28-May-86 0900 JMC
C01238 00436 ∂28-May-86 1011 RA Academic Costume
C01239 00437 ∂28-May-86 1033 RA Re: Academic Costume
C01240 00438 ∂28-May-86 1103 tgd%oregon-state.csnet@CSNET-RELAY.ARPA Proposal for "Knowledge Compilation" Workshop
C01259 00439 ∂28-May-86 1204 SINGER@su-sushi.arpa maternity leave
C01260 00440 ∂28-May-86 1413 LANSKY@SRI-WARBUCKS.ARPA PLEASE RESPOND IMMEDIATELY
C01272 00441 ∂28-May-86 1627 vijay@ernie.Berkeley.EDU
C01274 00442 ∂29-May-86 1210 VAL Reminder: Non-Monotonic Reasoning Seminar
C01277 00443 ∂29-May-86 1321 WALDINGER@SRI-AI.ARPA circumscription
C01279 00444 ∂29-May-86 1339 lantz@su-gregorio.arpa FYI: Changes to the Proposed systems (hardware) syllabus
C01282 00445 ∂29-May-86 1345 lantz@su-gregorio.arpa prod...
C01287 00446 ∂29-May-86 1445 VAL re: circumscription
C01289 00447 ∂30-May-86 1237 KIRSH%OZ.AI.MIT.EDU@MC.LCS.MIT.EDU [kirsh: AI Workshop]
C01295 00448 ∂30-May-86 1609 RA Msg. from Feigenbaum's office
C01296 00449 ∂31-May-86 0548 nttlab!masahiko@su-shasta.arpa my visit to Stanford
C01299 00450 ∂31-May-86 1234 vijay@ernie.Berkeley.EDU AI DISC: Daniel Bobrow
C01317 00451 ∂01-Jun-86 1213 REM@IMSSS Prize for Go program
C01324 00452 ∂01-Jun-86 1457 LES DARPA Update
C01330 00453 ∂02-Jun-86 1512 FEIGENBAUM@SUMEX-AIM.ARPA please respond!
C01332 00454 ∂03-Jun-86 1416 hitson@su-pescadero.arpa CSD-CFC possible action items...
C01337 00455 ∂03-Jun-86 1445 ME Prancing Pony Bill
C01339 00456 ∂03-Jun-86 1740 JJW Gang-of-Four
C01341 00457 ∂05-Jun-86 1408 NILSSON@SU-SCORE.ARPA [Tom Strat <STRAT@SRI-AI.ARPA>:]
C01345 00458 ∂05-Jun-86 1424 LES Minutes of Facilities Committee Meeting on 5/28
C01356 00459 ∂05-Jun-86 1741 LES Pucci visit
C01357 00460 ∂05-Jun-86 1957 SJG busy Friday evening? (6/6)
C01358 00461 ∂06-Jun-86 0859 CONTRERAS@SU-SCORE.ARPA message
C01359 00462 ∂06-Jun-86 1043 RA Chiang
C01360 00463 ∂06-Jun-86 1400 RPG NEXT Evaluation
C01374 00464 ∂06-Jun-86 1533 RA
C01375 00465 ∂06-Jun-86 1557 RINDFLEISCH@SUMEX-AIM.ARPA Re: [Tom Strat <STRAT@SRI-AI.ARPA>:]
C01377 00466 ∂07-Jun-86 1759 SJG re: busy Friday evening? (6/6)
C01378 00467 ∂07-Jun-86 1759 SJG re: busy Friday evening? (6/6)
C01379 00468 ∂08-Jun-86 1153 FEIGENBAUM@SUMEX-AIM.ARPA re: please respond!
C01380 00469 ∂09-Jun-86 0007 100 (from: jmc on TTY44)
C01381 00470 ∂09-Jun-86 0814 JMC
C01382 00471 ∂09-Jun-86 0842 SJM Good News/Bad News
C01383 00472 ∂09-Jun-86 0853 RA check
C01384 00473 ∂09-Jun-86 0901 VAL Waldinger on McDermott's "Critique"
C01388 00474 ∂09-Jun-86 0907 RA telex from Barbara Kuhn
C01389 00475 ∂09-Jun-86 1000 JMC
C01390 00476 ∂09-Jun-86 1041 RA Please call
C01391 00477 ∂09-Jun-86 1041 OR.LUSTIG@su-sierra.arpa Re: I always thought musicians were fools.
C01393 00478 ∂09-Jun-86 1058 GRP Tomorrow's meeting
C01394 00479 ∂09-Jun-86 1130 BARWISE@SU-CSLI.ARPA Invitation to Lunch
C01396 00480 ∂09-Jun-86 1144 RA Charles Wallace, NSF
C01397 00481 ∂09-Jun-86 1306 RA John Hopcroft
C01398 00482 ∂09-Jun-86 1351 SJM
C01399 00483 ∂09-Jun-86 1517 RA Trip to LA
C01400 00484 ∂09-Jun-86 1520 RA ebos meeting
C01401 00485 ∂09-Jun-86 1637 RA Spektrum bill
C01402 00486 ∂09-Jun-86 1755 LES EBOS Meeting
C01403 00487 ∂09-Jun-86 1839 CLT EBOS meeting
C01404 00488 ∂09-Jun-86 1846 LES EBOS meeting
C01405 00489 ∂09-Jun-86 1855 vijay@ernie.Berkeley.EDU AI DISC: Daniel Bobrow Response to Joseph Weizenbaum
C01409 00490 ∂09-Jun-86 1958 CLT calendar item
C01410 00491 ∂09-Jun-86 2000 JMC
C01411 00492 ∂09-Jun-86 2243 somewhere!nobuo@nttyrl.ntt.junet Common Business Communication Language
C01413 00493 ∂10-Jun-86 0729 Mailer failed mail returned
C01414 00494 ∂10-Jun-86 0930 JMC
C01415 00495 ∂10-Jun-86 0953 CONTRERAS@SU-SCORE.ARPA message
C01416 00496 ∂10-Jun-86 1001 MDD Ernie Davis
C01417 00497 ∂10-Jun-86 1116 RA Ebos meeting
C01418 00498 ∂10-Jun-86 1121 TAJNAI@SU-SCORE.ARPA
C01419 00499 ∂10-Jun-86 1426 CSL.GERLACH@su-sierra.arpa CSL fac cand--MacQueen
C01422 00500 ∂10-Jun-86 1559 RA Kathleen Martin
C01423 00501 ∂10-Jun-86 1616 SJM where?
C01424 00502 ∂10-Jun-86 1657 G.GORIN@LOTS-A confirm meeting
C01425 00503 ∂10-Jun-86 1724 CLT calendar item correction
C01426 00504 ∂10-Jun-86 PALLAS@su-sushi.arpa 10-Jun-86 JMC re: Evaluation of CS Profs @ Stanford (from SAIL's BBOARD)
C01429 00505 ∂11-Jun-86 0946 davism@nyu-csd1.arpa
C01432 00506 ∂11-Jun-86 1000 JMC
C01433 00507 ∂11-Jun-86 1021 RA Invitation to 2nd World Basque Congress
C01434 00508 ∂11-Jun-86 1030 VAL Boyer's problem
C01436 00509 ∂11-Jun-86 1105 SJM bday essay
C01437 00510 ∂11-Jun-86 1157 GIVAN@su-sushi.arpa Starting summer work
C01439 00511 ∂11-Jun-86 2003 LES Last (?) shot at Qlisp
C01462 00512 ∂11-Jun-86 2016 LES Oops
C01464 00513 ∂12-Jun-86 1112 RA Dr. Perris
C01465 00514 ∂12-Jun-86 1212 LANSKY@SRI-WARBUCKS.ARPA timberline workshop
C01476 00515 ∂12-Jun-86 1215 SJM sjm's movements
C01477 00516 ∂12-Jun-86 1322 CL.SHANKAR@R20.UTEXAS.EDU Re: possible research associate position
C01479 00517 ∂12-Jun-86 1427 RWW
C01480 00518 ∂12-Jun-86 1418 LES EBOS progress
C01482 00519 ∂12-Jun-86 1446 LES Alliant
C01484 00520 ∂12-Jun-86 1548 john@lbl-csam.ARPA mail probably intended for you
C01490 00521 ∂12-Jun-86 1649 CLT
C01491 00522 ∂12-Jun-86 1747 @MIT-MULTICS.ARPA,@UMich-MTS.Mailnet:Jeff←Pelletier@UQV-MTS
C01495 00523 ∂12-Jun-86 1828 "Lwd at vanderbilt%vanderbilt.csnet"@CSNET-RELAY.ARPA Why no AI?
C01498 00524 ∂12-Jun-86 2000 SJM
C01499 00525 ∂13-Jun-86 0700 JMC
C01500 00526 ∂13-Jun-86 0904 CSL.GERLACH@su-sierra.arpa re: CSL fac cand--MacQueen
C01502 00527 ∂13-Jun-86 0958 RPG Concurrency Features
C01503 00528 ∂13-Jun-86 1124 VAL Przymusinski's paper
C01504 00529 ∂13-Jun-86 1258 CLT japan collaboration -- time extension
C01505 00530 ∂13-Jun-86 1606 AAAI-OFFICE@SUMEX-AIM.ARPA Strategic Planning Meeting
C01507 00531 ∂13-Jun-86 1641 VAL reply to message
C01508 00532 ∂13-Jun-86 1906 LES Alliant etc.
C01510 00533 ∂13-Jun-86 1940 OTA-CIT@USC-ISI.ARPA Re: [John McCarthy <JMC@SU-AI.ARPA>: misdirected mail from Earl Dowdy and Jim Dray ]
C01512 00534 ∂14-Jun-86 1613 HK.WFM@Lindy NEXT
C01516 00535 ∂14-Jun-86 1625 HK.WFM@Lindy re: NEXT
C01517 00536 ∂15-Jun-86 2000 SJM
C01519 00537 ∂16-Jun-86 0004 LES re: Interpress
C01520 00538 ∂16-Jun-86 0600 JMC
C01521 00539 ∂16-Jun-86 0911 BSCOTT@SU-SCORE.ARPA Summer Supplemental Research Salaries
C01523 00540 ∂16-Jun-86 0914 CLT Cate
C01524 00541 ∂16-Jun-86 0929 CSL.GERLACH@su-sierra.arpa [dbm.sola%btl.csnet@CSNET-RELAY.ARPA: Re: Abstract]
C01527 00542 ∂16-Jun-86 0936 RPG Common Lisp
C01528 00543 ∂16-Jun-86 1012 SJM objectivity
C01529 00544 ∂16-Jun-86 1035 RA George Marotta
C01530 00545 ∂16-Jun-86 1136 RA Bob Givan
C01532 00546 ∂16-Jun-86 1152 davism@nyu-csd1.arpa
C01533 00547 ∂16-Jun-86 1427 RA Danzig
C01534 00548 ∂16-Jun-86 1621 GIVAN@su-sushi.arpa me
C01535 00549 ∂16-Jun-86 1942 FAHLMAN@C.CS.CMU.EDU further development of Lisp
C01538 00550 ∂16-Jun-86 2250 hector%utai.toronto.edu@CSNET-RELAY.ARPA A proposal
C01547 00551 ∂17-Jun-86 0859 DLW@ELEPHANT-BUTTE.SCRC.Symbolics.COM further development of Lisp
C01551 00552 ∂17-Jun-86 0921 VAL re: A proposal
C01552 00553 ∂17-Jun-86 1000 JMC
C01553 00554 ∂17-Jun-86 1059 GRP Reminder: meeting today at 2pm
C01554 00555 ∂17-Jun-86 1439 RA George Morotta
C01555 00556 ∂17-Jun-86 1518 RA reservation
C01556 00557 ∂17-Jun-86 1538 CLT alliant
C01559 00558 ∂17-Jun-86 1703 LES Alliant acceptance
C01560 00559 ∂17-Jun-86 1856 CWR
C01561 00560 ∂17-Jun-86 2108 CLT Alliant acceptance
C01562 00561 ∂18-Jun-86 0700 JMC
C01563 00562 ∂18-Jun-86 1142 RA leaving for lunch
C01564 00563 ∂18-Jun-86 1327 SJM reservations
C01565 00564 ∂18-Jun-86 1659 RA Urgent from John Kwapisz
C01566 00565 ∂18-Jun-86 1728 CLANCEY@SUMEX-AIM.ARPA Re: two proposed workshops
C01571 00566 ∂18-Jun-86 2000 SJM
C01572 00567 ∂18-Jun-86 2042 GENESERETH@SUMEX-AIM.ARPA
C01573 00568 ∂18-Jun-86 2147 SF@SU-CSLI.ARPA Re: possible Freiling visit
C01575 00569 ∂19-Jun-86 0605 CLT japan collaboration
C01576 00570 ∂19-Jun-86 0845 VAL blocks
C01577 00571 ∂19-Jun-86 0854 VAL
C01578 00572 ∂19-Jun-86 1003 RPG Weening
C01579 00573 ∂19-Jun-86 1039 RA Kate Kelly, Sapiens Software
C01580 00574 ∂19-Jun-86 1050 RA my replacement
C01581 00575 ∂19-Jun-86 1143 SJG busy Saturday night?
C01582 00576 ∂19-Jun-86 1422 VAL Recursive negation
C01585 00577 ∂19-Jun-86 1449 CHEADLE@SU-SCORE.ARPA progress?
C01587 00578 ∂19-Jun-86 1452 JJW Re: progress?
C01588 00579 ∂19-Jun-86 1703 SJM space salvage
C01589 00580 ∂20-Jun-86 0126 SJM
C01590 00581 ∂20-Jun-86 0135 GIVAN@su-sushi.arpa locating Bob
C01591 00582 ∂20-Jun-86 0136 @SU-SCORE.ARPA:shoham@YALE.ARPA Arrangements
C01592 00583 ∂20-Jun-86 0137 avg@su-aimvax.arpa Re: Recursive negation
C01595 00584 ∂20-Jun-86 0148 WINOGRAD@SU-CSLI.ARPA Chinese visitor
C01598 00585 ∂20-Jun-86 0154 SLEEMAN@SUMEX-AIM.ARPA AAAI-sponsored scientific meeting?
C01601 00586 ∂20-Jun-86 0335 a32275%ccut.u-tokyo.junet%utokyo-relay.csnet@CSNET-RELAY.ARPA US-Japan collaboration
C01606 00587 ∂20-Jun-86 0956 OR.DANTZIG@su-sierra.arpa Re: Nunez
C01607 00588 ∂20-Jun-86 1206 RA Thelma, Inference
C01608 00589 ∂20-Jun-86 1351 hpm@rover.ri.cmu.edu robot vehicle update
C01610 00590 ∂20-Jun-86 1355 WINOGRAD@SU-CSLI.ARPA Comp systems Syllabus
C01615 00591 ∂20-Jun-86 1429 WINOGRAD@SU-CSLI.ARPA Comp Theory Syllabus
C01620 00592 ∂20-Jun-86 1439 WINOGRAD@SU-CSLI.ARPA Applicatons and techniques Syllabus
C01624 00593 ∂20-Jun-86 1442 WINOGRAD@SU-CSLI.ARPA Courses for Comp Syllabus
C01627 00594 ∂20-Jun-86 1437 WIEDERHOLD@SUMEX-AIM.ARPA John Sowa industrial lectureship
C01628 00595 ∂20-Jun-86 1509 lantz@su-gregorio.arpa Re: Applicatons and techniques Syllabus
C01630 00596 ∂20-Jun-86 1514 lantz@su-gregorio.arpa Re: Courses for Comp Syllabus
C01632 00597 ∂20-Jun-86 1514 SJG tomorrow evening
C01633 00598 ∂20-Jun-86 1447 WINOGRAD@SU-CSLI.ARPA
C01636 00599 ∂20-Jun-86 1654 WINOGRAD@SU-CSLI.ARPA Teaching specialization
C01639 00600 ∂20-Jun-86 1716 SJM other things you could mess with
C01640 00601 ∂20-Jun-86 1721 VAL
C01641 00602 ∂20-Jun-86 1727 berglund@su-pescadero.arpa Re: Teaching specialization
C01643 00603 ∂20-Jun-86 2344 SF@SU-CSLI.ARPA re: possible Freiling visit
C01644 00604 ∂21-Jun-86 0750 lantz@su-gregorio.arpa Re: Teaching specialization
C01646 00605 ∂21-Jun-86 0937 SLEEMAN@SUMEX-AIM.ARPA re: AAAI-sponsored scientific meeting?
C01647 00606 ∂21-Jun-86 1213 SJM
C01648 00607 ∂21-Jun-86 1341 SJM kitty michaelson
C01649 00608 ∂21-Jun-86 1741 RESTIVO@SU-SCORE.ARPA Alliant
C01650 00609 ∂21-Jun-86 1828 OTA-CIT@USC-ISI.ARPA First discussion topic
C01655 00610 ∂22-Jun-86 1646 @SU-SCORE.ARPA:shoham@YALE.ARPA historical question
C01657 00611 ∂23-Jun-86 0948 AI.ELLIE@MCC.COM visit to MCC
C01658 00612 ∂23-Jun-86 1000 JMC
C01659 00613 ∂23-Jun-86 1148 RA John Nafeh
C01661 00614 ∂23-Jun-86 1339 ME hosts
C01663 00615 ∂23-Jun-86 1343 AI.ELLIE@MCC.COM
C01664 00616 ∂23-Jun-86 1405 davism@nyu-csd1.arpa
C01667 00617 ∂23-Jun-86 1423 VAL
C01668 00618 ∂23-Jun-86 1435 SJG who should handle this?
C01669 00619 ∂23-Jun-86 1437 RA Frank, Dina Bolla
C01670 00620 ∂23-Jun-86 1618 GRP Reminder: meeting at 11 tomorrow
C01671 00621 ∂23-Jun-86 1651 WINOGRAD@SU-CSLI.ARPA Going, going, ...
C01676 00622 ∂23-Jun-86 1758 LES Proposed Meeting
C01677 00623 ∂23-Jun-86 1831 AI.BOYER@MCC.COM Gnu & Editor based operating system
C01680 00624 ∂23-Jun-86 2000 JMC
C01681 00625 ∂23-Jun-86 2221 SF@SU-CSLI.ARPA Freiling visit
C01682 00626 ∂23-Jun-86 2237 ME domain names
C01683 00627 ∂23-Jun-86 2250 ME domains
C01685 00628 ∂24-Jun-86 0110 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Paper on circumscription and induction
C01687 00629 ∂24-Jun-86 0845 CSL.GERLACH@su-sierra.arpa Macqueen impressions
C01689 00630 ∂24-Jun-86 1202 G.GORIN@LOTS-B.STANFORD.EDU re: proposed meeting
C01690 00631 ∂24-Jun-86 1225 CLT meeting
C01691 00632 ∂24-Jun-86 1352 LES re: proposed meeting
C01692 00633 ∂24-Jun-86 1450 KARP@SUMEX-AIM.ARPA PhD program
C01694 00634 ∂24-Jun-86 1503 berglund@su-pescadero.arpa Uproar
C01695 00635 ∂24-Jun-86 1609 RA Trip to Portland
C01696 00636 ∂24-Jun-86 1734 WINOGRAD@SU-CSLI.ARPA [Andy Freeman <ANDY@su-sushi.arpa>: PhD changes]
C01702 00637 ∂24-Jun-86 1746 WINOGRAD@SU-CSLI.ARPA [Andy Freeman <ANDY@su-sushi.arpa>: Re: New Proposals for the Ph.D. Program]
C01706 00638 ∂24-Jun-86 1748 WINOGRAD@SU-CSLI.ARPA
C01712 00639 ∂24-Jun-86 1836 LES Alliant visit
C01713 00640 ∂24-Jun-86 1951 LES Facilities Update
C01716 00641 ∂24-Jun-86 2019 lantz@su-gregorio.arpa
C01720 00642 ∂24-Jun-86 2135 a32275%ccut.u-tokyo.junet%utokyo-relay.csnet@CSNET-RELAY.ARPA re: US-Japan collaboration
C01722 00643 ∂25-Jun-86 0000 JMC
C01723 00644 ∂25-Jun-86 0024 HST 1st idea
C01726 00645 ∂25-Jun-86 0039 HST first idea for an announcement and call for proposals:
C01729 00646 ∂25-Jun-86 0052 HST possible speakers for the 30-years-lisp conference
C01732 00647 ∂25-Jun-86 0739 TOM@SU-SCORE.ARPA Re: Facilities Update
C01734 00648 ∂25-Jun-86 1101 WINOGRAD@SU-CSLI.ARPA [Andy Freeman <ANDY@su-sushi.arpa>: Re: New Proposals for the Ph.D. Program]
C01743 00649 ∂25-Jun-86 1128 hitson@su-pescadero.arpa Re: Facilities Update
C01747 00650 ∂25-Jun-86 1257 mcdermott-drew@yale.ARPA Reply to your message
C01750 00651 ∂25-Jun-86 1321 LES re: Facilities Update
C01751 00652 ∂25-Jun-86 1411 BOSACK@SU-SCORE.ARPA Re: Facilities Update
C01753 00653 ∂25-Jun-86 1422 AI.ELLIE@MCC.COM Abstract - McCarthy's Thursday talk
C01756 00654 ∂25-Jun-86 1514 WINOGRAD@SU-CSLI.ARPA Re: phdcom
C01758 00655 ∂25-Jun-86 1730 WINOGRAD@SU-CSLI.ARPA New draft and meeting tomorrow at 11
C01762 00656 ∂25-Jun-86 1734 WINOGRAD@SU-CSLI.ARPA [Terry Winograd <WINOGRAD@SU-CSLI.ARPA>: New version]
C01764 00657 ∂25-Jun-86 1817 RA David Kirsh
C01765 00658 ∂26-Jun-86 0002 lantz@su-gregorio.arpa Re: phdcom
C01767 00659 ∂26-Jun-86 0108 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK re: Paper on circumscription and induction
C01769 00660 ∂26-Jun-86 0600 JMC
C01770 00661 ∂26-Jun-86 1357 ME MCC.COM
C01771 00662 ∂26-Jun-86 1423 WINOGRAD@SU-CSLI.ARPA Qual areas
C01773 00663 ∂26-Jun-86 1610 WINOGRAD@SU-CSLI.ARPA Last changes
C01775 00664 ∂26-Jun-86 2049 lantz@su-gregorio.arpa Re: Last changes
C01777 00665 ∂26-Jun-86 2326 nttlab!nttyrl!nobuo@su-shasta.arpa Common Business Communication Langauge
C01780 00666 ∂27-Jun-86 0913 ROSENBLOOM@SUMEX-AIM.ARPA Re: Last changes
C01783 00667 ∂27-Jun-86 0952 TAL call todorovich
C01784 00668 ∂27-Jun-86 1101 langley@ICSE.UCI.EDU machine learning workshop
C01788 00669 ∂27-Jun-86 1317 AI.WOODY@MCC.COM Hi
C01789 00670 ∂27-Jun-86 1340 OTA-CIT@USC-ISI.ARPA Split in content of AI
C01792 00671 ∂27-Jun-86 1526 SJM UK trip
C01794 00672 ∂27-Jun-86 1656 langley@CIP.UCI.EDU Re: machine learning workshop
C01796 00673 ∂28-Jun-86 1524 NILSSON@SU-SCORE.ARPA Re: First discussion topic
C01804 00674 ∂29-Jun-86 0457 100 (from: jmc on TTY44)
C01805 00675 ∂29-Jun-86 1300 CRISPIN@SUMEX-AIM.ARPA re: Contra aid (from SAIL's BBOARD)
C01808 00676 ∂29-Jun-86 2123 ULLMAN@SU-SCORE.ARPA Database reading
C01810 00677 ∂29-Jun-86 2202 ullman@su-aimvax.arpa Re: Recursive negation
C01813 00678 ∂30-Jun-86 0937 OTA-CIT@USC-ISI.ARPA Second Discussion Topic
C01818 00679 ∂30-Jun-86 0948 CONTRERAS@SU-SCORE.ARPA message
C01819 00680 ∂30-Jun-86 1054 TAL Chris Freiling called
C01820 00681 ∂30-Jun-86 1134 KARP@SUMEX-AIM.ARPA test
C01821 00682 ∂30-Jun-86 1530 CLT WICS
C01823 00683 ∂30-Jun-86 1821 G.MDP@SU-SCORE.ARPA Re: Anti-American punctuation
C01824 00684 ∂30-Jun-86 1846 Bobrow.pa@Xerox.COM Re: Lifschitz for editorial board
C01825 00685 ∂30-Jun-86 1846 GRP graphics on apa16
C01828 ENDMK
C⊗;
∂01-Apr-86 0414 wachsmuth%gvaic2.DEC@decwrl.DEC.COM Situation Calculus Planning in blocks & related worlds.
Received: from DECWRL.DEC.COM by SU-AI.ARPA with TCP; 1 Apr 86 04:14:21 PST
Received: from DEC-RHEA.ARPA (dec-rhea) by decwrl.DEC.COM (4.22.03/4.7.34)
id AA16202; Tue, 1 Apr 86 03:35:32 pst
Message-Id: <8604011135.AA16202@decwrl.DEC.COM>
Date: Tuesday, 1 Apr 1986 03:31:11-PST
From: wachsmuth%gvaic2.DEC@decwrl.DEC.COM (Markus L. Wachsmuth)
To: JMC@su-ai.ARPA, wachsmuth%gvaic2.DEC@decwrl.DEC.COM
Subject: Situation Calculus Planning in blocks & related worlds.
Dr. McCarthy,
Could you please have the following papers sent to me:
* SITUATION CALCULUS PLANNING IN BLOCKS AND RELATED WORLDS.
* COMMON SENSE AND NON-MONOTONIC REASONING
Thank you very much.
Sincerely,
Markus Wachsmuth
Digital Equipment Corp. (Europe)
43 rte de Prevessin
CH-1217 GENEVA
SWITZERLAND
P.S. The papers' titles may be incorrect, but the topics are what matters.
∂01-Apr-86 1320 vijay@ernie.berkeley.edu AI DISC: DREYFUS
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 1 Apr 86 13:20:32 PST
Received: by ernie.berkeley.edu (5.45/1.9)
id AA27558; Tue, 1 Apr 86 13:20:57 PST
Date: Tue, 1 Apr 86 13:20:57 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604012120.AA27558@ernie.berkeley.edu>
To: jmc@su-ai.arpa
Subject: AI DISC: DREYFUS
Hello Dr. McCarthy,
You might have seen a version of the following given by Dreyfus,
however, both Stuart and Hubert Dreyfus say that as a starting
point - it articulates their ideas most clearly.
------------------------------------------------------------------
CONVENTIONAL AI: A DEGENERATING RESEARCH PROGRAM
Looking back over 30 years, the field of conventional
rule-based AI appears more and more to be a perfect example
of what Imre Lakatos has called a degenerating research pro-
gram.[1] AI began auspiciously with Newell and Simon's work
at RAND. In retrospect, we see we failed to appreciate the
importance of this early work. Newell and Simon proved that
computers could do more than calculations. They demonstrated
that the symbols computers manipulate could stand for any-
thing, including features of the real world, and programs
could be used as rules for relating these features, so that
computers acting as logic machines could be used to simulate
certain important aspects of intelligence. Thus the
information-processing model of the mind was born. By 1970
AI, using symbolic representations, had turned into a flour-
ishing research program. Marvin Minsky, head of the M.I.T.
program, predicted that "within a generation the problem of
creating `artificial intelligence' will be substantially
solved."[2]
Then, rather suddenly, the field ran into unexpected
difficulties. The trouble started, as far as we can tell,
with the failure of attempts to program children's story
understanding. It turned out to be much harder than one
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [1] Imre Lakatos, Philosophical Papers, ed. John Wor-
rall, Cambridge University Press, 1978.
$9 [2] Marvin Minsky, Computation: Finite and Infinite
Machines, Prentice Hall, 1967, p. 2.
- 2 -
expected to formulate the required theory of common sense.
It was not, as Minsky had hoped, just a question of catalo-
guing a few hundred thousand facts. The common sense
knowledge problem became the center of concern. Minsky's
mood changed completely in the course of fifteen years. He
told a reporter: "the AI problem is one of the hardest sci-
ence has ever undertaken."[3]
Related problems were also noted although not often
seen as related. Cognitivists discovered the importance of
images and prototypes in human understanding and logic
machines turned out to be very poor at dealing with either
of them. Gradually most researchers have become convinced
that human beings form images and compare them by means of
holistic processes quite different from the logical opera-
tions computers perform on descriptions.[4] Some AI workers
hope for help from parallel processors, machines that can do
many things at once and hence can make millions of infer-
ences per second, but if human image processing operates on
holistic representations that are not descriptions and
relates these representations in other than rule-like ways,
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [3] Gina Kolata, "How Can Computers Get Common
Sense?", Science, Vol. 217, 24 September 1982, p. 1237.
$9 [4] For an account of the experiments which show that
human beings can actually rotate, scan, and otherwise
use images, and the unsuccessful attempts to understand
these capacities in terms of programs which use
features and rules, see Imagery, Ned Block, ed., M.I.T.
Press/Bradford Books, 1981. Also Ned Block, "Mental
Pictures and Cognitive Science," The Philosophical Re-
view, Oct. 1983, pp. 499-541.
- 3 -
this appeal to parallel processing misses the point. The
point is that human beings are able to form and compare
their images in a way that cannot be captured by any number
of procedures that operate on symbolic descriptions.
Another human capacity which computers functioning as
analytic engines cannot copy is the ability to recognize the
similarity between whole images. Recognizing two patterns as
similar, which seems to be a direct process for human
beings, is for a logic machine a complicated process of
first defining each pattern in terms of objective features
and then determining whether, by some objective criterion,
the set of features defining one pattern match the features
defining the other pattern.
As we see it, all AI's problems are versions of one
basic problem. Current AI is based on the idea which has
been around in philosophy since Descartes, that all under-
standing consists in forming and using appropriate represen-
tations. In conventional AI these have been assumed to be
symbolic descriptions. So common sense understanding has to
be understood as some vast body of propositions, beliefs,
rules, facts and procedures. AI's failure to come up with
the appropriate symbolic descriptions is called the common
sense knowledge problem. As thus formulated this problem
has so far turned out to be insoluble, and we predict it
will never be solved.
What hides this impasse is the conviction that the
- 4 -
common sense knowledge problem must be solvable since human
beings have obviously solved it. But human beings may not
normally use common sense knowledge at all. What common
sense understanding amounts to might well be everyday know-
how. By know-how we do not mean procedural rules, but know-
ing what to do in a vast number of special cases. For exam-
ple, common sense physics has turned out to be extremely
hard to spell out in a set of facts and rules. When one
tries, one either requires more common sense to understand
the facts and rules one finds or else one produces formulas
of such complexity that it seems highly unlikely they are in
a child's mind.
Theoretical physics also requires background skills
which may not be formalizable, but the domain itself can be
described by abstract laws that make no reference to
specific cases. AI researchers conclude that common sense
physics too must be expressible as a set of abstract princi-
ples. But it just may be that the problem of finding a
theory of common sense physics is insoluble. By playing
almost endlessly with all sorts of liquids and solids for
several years the child may simply have built up a repertory
of prototypical cases of solids, liquids, etc. and typical
skilled response to their typical behavior in typical cir-
cumstances. There may be no theory of common sense physics
more simple than a list of all such typical cases and even
such a list is useless without a similarity-recognition
ability. If this is indeed the case, and only further
- 5 -
research will give us an answer, we could understand the
initial success and eventual failure of AI. It would seem
that AI techniques should work in isolated domains but fail
in areas such as natural language understanding, speech
recognition, story understanding, and learning where the
structure of the problem mirrors the structure of our every-
day physical and social world.
In 1979 we predicted stagnation for AI, but also
predicted the success of programs called expert systems
which attempted to produce intelligent behavior in domains
such as medical diagnosis and spectrograph analysis which
are completely cut off from everyday common sense. Now we
think we were uncharacteristically over-optimistic concern-
ing the future of intelligent logic machines. It has turned
out that, except in certain structured domains where what
constitutes the relevant facts and how these facts are
changed by decisions is known objectively, no expert system
based on rules extracted by questioning experts does as well
as the experts themselves, even though the computer is pro-
cessing with incredible speed and unerring accuracy what are
supposed to be the experts' rules.
In our just published book Mind Over Machine we attempt
to explain this surprising development. We argue that
beginners in a domain are given principles to follow, but
most domains in which human beings acquire skills and
achieve expertise are, like everyday physics, domains which
- 6 -
do not lend themselves to being understood at an expert
level in terms of principles.[5] Therefore experts, as even
Edward Feigenbaum has noted, are never satisfied with gen-
eral principles but prefer to think of their field of exper-
tise as a huge set of special cases.[6] No wonder expert
systems based on principles abstracted from experts do not,
in unstructured domains, capture those experts' expertise
and so never do as well as the experts themselves.
We still think, as we did in 1965, that someday comput-
ers may be intelligent just as one day the alchemists' dream
of transmuting lead into gold came true. AI may be
achieved, however, only after researchers give up the idea
of finding a local symbolic representation of high-order
macrostructural features describing the world and turn
instead to some sort of microstructural distributed, holis-
tic representation that is directly amenable to association,
generalization and completion. If this is, indeed, the
direction AI should go, it will be aided by the massively
parallel machines on the horizon, but not because parallel
machines can make millions of inferences per second, but
because faster, more parallel architecture can better imple-
ment the kind of neurally inspired processing that does not
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [5] Hubert L. Dreyfus and Stuart E. Dreyfus, Mind
over Machine, Free Press/Macmillan (1986).
$9 [6] Edward A. Feigenbaum and Pamela McCorduck, The
Fifth Generation, Artificial Intelligence and Japan's
Computer Challenge to the World, Addison-Wesley Pub-
lishing Company, 1983, p. 82.
- 7 -
use macrostructural representations of rules and features at
all.[7]
Hubert L. Dreyfus and Stuart E. Dreyfus
University of California, Berkeley
$9←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [7] See for example D. Rumelhart and J. McClelland,
Parallel Distributed Processing: Explorations in the
Microstructure of Cognition, MIT Press/ Bradford Books,
1986.
P.S. Please acknowlege receipt of this mail.
∂01-Apr-86 1325 vijay@ernie.berkeley.edu AI DISC: Winograd Position
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 1 Apr 86 13:25:24 PST
Received: by ernie.berkeley.edu (5.45/1.9)
id AA27764; Tue, 1 Apr 86 13:25:53 PST
Date: Tue, 1 Apr 86 13:25:53 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604012125.AA27764@ernie.berkeley.edu>
To: jmc@su-ai.arpa
Subject: AI DISC: Winograd Position
The best thing to do in a short position paper is to put forth some
clear and probably controversial assertions, without giving elaborate
motivations and justifications contrasting them with other ways of
understanding. These fuller discussions appear at length in my recent
book with Fernando Flores, Understanding Computers and Cognition.
1. In characterizing AI, there are two very different starting points.
We can take it as the general enterprise of developing intelligent
artifacts (by any physical means whatsoever), or as the expression of a
coherent methodology and theory.
2. To the extent that AI means "anything anyone might invent that shows
intelligence," discussion belongs in the realm of science fiction, since
there is little concrete to be said. To the extent we are talking about
what people have really done in AI, there is a strong coherent ideology,
variously labelled the "computational paradigm," "cognitive paradigm,"
"physical symbol system hypothesis," etc. Most of the existing AI
enterprise operates within it (including, though to a somewhat lesser
extent, the current work on connectionism).
3. The cognitive symbol-processing approach will have useful
applications, but these will not be as widespread or significant as
proponents claim. In general, those tasks that are closer to
"puzzle-solving" will be best covered, and those closer to "common
sense" and "ordinary understanding" will remain unmechanized. This
applies not only to existing technology, but to any of the foreseeable
improvements following in the general scientific direction that is being
pursued (including "massively" parallel machines, nonmonotonic
reasoning, etc., etc.).
4. I am not so concerned with the danger that attempts to fully
duplicate human intelligence will fail (as long as people don't bank to
heavily on optimistic predictions), but rather that the enterprise has
an effect of redefining intelligence---of shaping human understanding of
what is to count as "intelligent." In particular, AI is based on a
"rationalistic" account of human thought and language, which focusses on
systematic reasoning based on symbolic representations within an
explicitly articulated domain of features. This approach has important
uses, but systematically undervalues other aspects of intelligent human
action, both in the individual and within a tradition. Emphasis on
rationalism is not new to AI, having a long history in Western thought
(beginning with Plato, expressed more thoroughly by Descartes).
Computers (and AI in particular) give it a powerful operational form.
5. A healthy skepticism about AI (and the rationalistic orientation in
general) is needed as a guide for design of computer systems that make
sense. We are easily seduced by the image of the "thinking machine"
into claiming that the problems of designing and working with computer
technology will be solved when the machines get smart enough. The Fifth
Generation hoopla (both the Japanese original report and later books and
responses) is an egregious example of this fallacy. The phenomena of
"computerization" (in its pejorative sense) derive from the
reorganization of social systems to fit the properties of particular
computer implementations. It will not be prevented by having "smart"
machines, and in fact is accelerated by advocating the use of computers
in less structured areas of human life and society.
6. My major interest lies in research (both theoretical and applied)
that will support the development of technology to provide the
advantages of using computers while anticipating and avoiding negative
effects on people's work and lives. The rationalistic tradition does
not provide a sufficient basis for this design, since it takes as its
starting point an impoverished account of what people do. A new
starting point will come from an understanding of the phenomenology of
human communication and use of technology. We can draw on the
philosophical tradition of phenomenology, and its insights can be given
concrete operational meaning in the context of design.
7. It is often claimed that concerns of "social impact" should be left
to the political process, or perhaps to engineers who are directly
developing products, but should be ignored in pursuing "pure science."
These (often self-serving) claims are based on a rationalistic (and
narrow) understanding of science as a human enterprise. They might be
true for some idealized scientist living self-sufficiently and
incommunicado on an isolated island, but are irrelevant to the real
world. The continuing enterprise of any science depends on a public
consensus that supports the allocation of resources to it. This
consensus is maintained by a process of publication and "education" in
which the ideology of the science is promulgated and justified. As
members of the "AI community" we all participate in this, through
writing, talking, and teaching.
8. AI scientists and engineers have a responsibility to take their work
seriously---to recognize that both their inventions and their words have
a serious effect and to consider the effects consciously. The issue
isn't censorship, but positive action. It is useless to try to label
work that "shouldn't be done," but instead we can use our knowledge and
status to advance the things that "should be done," rather than just
those that "can be done." I anticipate a gradual shift of effort and
emphasis within the field as we go beyond the the early science-fiction
dreams that motivated the field, and look at directions for new research
(including theoretical research) that better deals with the realities of
human society. In particular, computers (using AI techniques) will be
understood in terms of the complex and productive ways in which they can
serve as a medium for human-to-human communication, rather than being
personified as surrogate people.
-TERRY WINOGRAD
-------
P.S. Please acknowledge receipt of this mail --- Thanks.
∂01-Apr-86 1539 VAL Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
The Monkey, the Banana, and the Bomb:
Information-Gathering in a Situational Logic
Richard Waldinger
SRI International
Thursday, April 3, 4pm
MJH 252
Abstract
In formulating a plan, certain actions are of value only
to increase the agent's knowledge about the environment.
We consider how such plans can be constructed within a
situational logic.
∂01-Apr-86 1604 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 1 April 1986
Previous Balance 14.57
Payment(s) 14.57 (check 3/16/86)
-------
Current Charges 6.00 (bicycle lockers)
0.30 (vending machine)
-------
TOTAL AMOUNT DUE 6.30
Please deliver payments to Debbie Woodward, room 040, Jacks Hall.
Make checks payable to: STANFORD UNIVERSITY.
To ensure proper crediting, please include your Pony account name on your check.
Note: Payment recordation takes up to three weeks after delivery of a payment
(but not beyond the next billing date).
Bills are payable upon presentation. Interest of 1.5% per month will be
charged on balances remaining unpaid 25 days after bill date above.
∂02-Apr-86 1108 RA fellowship letter
I put a hard copy of the list I got from Claudia on your desk. Do you
want me to write a personal letter to every one on the list, or do you want
a general letter addressed to Dear Colleagues. If the latter, what should
the address say, or do you want it simply to start with Dear Colleagues with
no addressee.
∂02-Apr-86 1413 LES Qlisp and Parallel Processors
To: JMC, CLT, RPG, JJW
I reached Steve Squires on Monday afternoon and got confirmation that
release of the DARPA Order for Qlisp is imminent and that they need
another brief task description with milestones and another budget (sigh).
The latter will have to go through the Stanford mill. I am preparing
them.
I also pressed him on the machine selection question. He again reviewed
his grand plan and added a few embellishments: after the Multimax gets
MIPS-X and Mach-1 added to it, it will be repackaged and put into
satellites, providing computational support for S.D.I. No problem.
I too occasionally smoke dope.
Squires went on to say that he understood that there is local interest in
getting three kinds of machines: Sequent, Encore and Alliant and that as
long as there was at least one Encore locally, to permit later transplants
as seems appropriate, that it looks OK to him. He suggested I recheck
Hennessy's plans.
I reached Hennessy on Tuesday and learned that he had turned around and
decided to buy a Multimax and that Luckham is definitely getting a Sequent
system. (This was after Hennessy flamed at me earlier for even
considering Encore.) It is clear, then, that we are free to choose
whichever machine we wish. Based on earlier discussions, it appears to me
that we should go ahead with the Alliant purchase. I am preparing
documentation in support of that plan.
Comments?
∂02-Apr-86 2024 CLT
To: JMC, CLT
how about meeting with joe 3pm sat?
∂03-Apr-86 1050 SJM you presence
To the elusive Dr. Fu Manchu--
You cannot elude me for very much longer. I am on
your trail, I and my bumbling companion. No matter what disguises
you assume, what subterfuges you employ, what low dens you conceal
yourself in, I will find you. Why not give yourself up now, bowing
to the inevitable, and sparing the world the frightful devastation
that must otherwise ensue?
Nayland `Susie' Smith
∂03-Apr-86 1238 TRACZ@SU-SIERRA.ARPA What do you think?
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 3 Apr 86 12:37:54 PST
Date: Thu 3 Apr 86 12:37:05-PST
From: Will Tracz <TRACZ@SU-SIERRA.ARPA>
Subject: What do you think?
To: jmc@SU-AI.ARPA
cc: tracz@SU-SIERRA.ARPA
Professor McCarthy: I have always respected your contributions to
the bulletin board, and am hopeful that you would take a few minutes
to share your opinions on the future directions of computer architecture
and software.
As a blind optimist, I placed the following message on the bboard, and
only received 2 replys. I take this as an indication of the technical
communities lack of interest, or insecurity in expressing oneself on
such a question.
I would be interested in knowing your "top 5" areas of research and
development, which will have the most profound effect in the future.
Thank you for your time and attention. Enclosed is a copy of my
origional question.
Will Tracz
------------------------------------------------------------------------
I had a paper accepted at a workshop on Future Directions in Computer
Architecture and Software. As a result, I began having this deep personal
reflection and examination on what I thought were the driving factors that
would affect the future. If you have the time and the mind, I would value
your opinion and any comments you have on the following:
How would you classify each of the following research areas and rank them on
order of importance? Where are the bottlenecks, and what effect would a
technological breakthrough have?
The categories are:
1) Emerging Technology - a relatively young technology which is still
being defined.
2) Evolving Technology - somewhat well defined and understood, but
going through rapid changes.
3) Mature Technology - well defined, well understood with very little
new advances made each year, yet still a
worthwhile topic of research.
4) Stagnant Technology - an old, well defined technology which has
little potential for significant new advances.
The subject areas are:
1) VLSI Design
2) CAD Tools
3) Data Bases
4) Expert Systems
5) Machine Learning
6) Artificial Intelligence
7) Software Engineering
8) Compilers
9) Operating Systems
10) Distributed Processing
11) Programming Languages
12) Microprogramming
13) Computer Architecture
14) Man Machine Interfaces
15) Computer Graphics
16) Super Computers
17) Personal Computers
18) Networking
19) CD/Optical Disks
20) etc.
I realize that this is a non-trivial question, which is a little vague at
best, but I thought it was worth asking, and, again, any comments you may have
would be appreciated. I will compile the results of the survey and post
them at the end of the month. I also will be presenting a workshop summary
as part of the EE 385B Seminar series if anyone is interesting in finding
out what transpires there (May 5-7, Charleston SC).
Thank you for your time and attention.
Will Tracz - your friendly used program salesman
-------
∂03-Apr-86 1348 RA trip to Austin
On your coming trip to Austin you have first class and Vladimir has coach.
Would you like to move to coach so that you can sit with Val?
Please let me know so I can call Franklin back.
Thanks,
∂03-Apr-86 1358 RA VTSS course
You said yesterday that you did not get your VTSS grade sheet, is this indeed
the case? I looked in your in box but did not find it.
∂03-Apr-86 1431 RA not feeling well
I don't feel well and will go home shortly.
∂03-Apr-86 1512 KTRACY@SU-SUSHI.ARPA VTSS160 Grade?
Received: from [36.36.0.196] by SU-AI.ARPA with TCP; 3 Apr 86 15:10:47 PST
Date: Thu 3 Apr 86 15:11:44-PST
From: Kim W. Tracy <KTRACY@SU-SUSHI.ARPA>
Subject: VTSS160 Grade?
To: jmc@SU-AI.ARPA
Message-ID: <12195978670.38.KTRACY@SU-SUSHI.ARPA>
My grade for VTSS160 was not reported according to the people over in
the records office. Is there any need for concern on my part?
Kim Tracy
-------
∂03-Apr-86 1517 vijay@ernie.berkeley.edu
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 3 Apr 86 15:14:58 PST
Received: by ernie.berkeley.edu (5.45/1.9)
id AA28407; Thu, 3 Apr 86 15:15:33 PST
Date: Thu, 3 Apr 86 15:15:33 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604032315.AA28407@ernie.berkeley.edu>
To: jmc@su-ai.arpa
TURING THE CHINESE ROOM
John R. Searle
Since various garbled versions of my Chinese room argu-
ment continue to be current in the CS-AI community, I intend
first to set the record straight. Then I intend to review
the current state of the argument concerning strong AI.
Among other things, I am accused of holding the prepos-
terous view that somehow in principle, as a matter of logic,
only carbon-based or perhaps only neuronal-based substances
could have the sorts of thoughts and feelings that humans
and other animals have. I have repeatedly and explicitly
denounced this view. Indeed, I use a variation of the
Chinese room argument against it: simply imagine right now
that your head is opened up and inside is found not neurons
but something else, say, silicon chips. There are no purely
logical constraints that exclude any particular type of sub-
stance in advance.
My actual argument is very simple and can be set out in
a very few steps:
Definition 1. Strong AI is defined as the view that
the appropriately programmed digital computer with the right
inputs and outputs would thereby have a mind in exactly the
same sense that human beings have minds.
It is this view which I set out to refute.
Proposition 1. Programs are purely formal (i.e. syn-
tactical).
I take it this proposition needs no explanation for the
readers of this journal.
Proposition 2. Syntax is neither equivalent to nor
sufficient by itself for semantics.
I take it Proposition 2 is a conceptual or logical truth.
The point of the parable of the Chinese room is simply to
remind us of the truth of this rather obvious point: the man
in the room has all the syntax we can give him, but he does
not thereby acquire the relevant semantics. He still does
not understand Chinese.
It is worth pointing out that the distinction between
syntax and semantics is an absolutely foundational principle
behind modern logic, linguistics, and mathematics.
Proposition 3. Minds have mental contents (i.e. seman-
tic contents).
Now from these three propositions, it follows simply
that strong AI as defined is false. Specifically:
- 2 -
Conclusion 1 : Having a program -- any program by
itself -- is neither sufficient for nor equivalent to having
a mind.
Anyone who wishes to challenge my argument is going to
have to show at least that one of the three "axioms" is
false. It is very hard to see how anybody in the AI commun-
ity would want to challenge any of them. In particular, the
idea that the program is purely formal and the computer is a
formal symbol manipulating device is hardly something that I
need to teach workers in AI .
Once you appreciate the structure of the argument it is
easy to see that the standard replies to it in the strong AI
camp are simply irrelevant because they do not address them-
selves to the actual argument. Thus, for example, the "sys-
tems reply" (according to which `the room,' i.e. the whole
system, understands Chinese even though the man in the room,
i.e. the CPU, does not understand) simply misses the point.
The system has no way of getting from the syntax to the
semantics any more than the man does. The systems reply
cannot evade the sheer inexorability of the syntax/semantics
distinction. Which axioms does it wish to challenge? And
what grounds are being given for the challenge? The "robot
reply" (according to which if we put the system inside a
robot capable of causal interactions with the rest of the
world it would thereby acquire a semantics) simply concedes
that strong AI is false. It admits that syntax would not
be sufficient for semantics but insists that syntax plus
causation would produce a semantics. This involves a
separate mistake that I will come back to, but right now I
want to emphasize that none of the defenders of strong A I
-- a rather large group by the way -- has even begun to make
an effective challenge to any of the three principles I have
enunciated.
That is the argument against strong AI. It is that
simple. Anyone interested only in knowing if strong AI is
false can stop reading right here. But now out of this sim-
ple argument gives rise to a whole lot of other issues.
Some of them are a bit trickier, but I will keep the argu-
ment as simple as possible. As before, the "axioms" must be
obviously true and the derivations must be transparently
valid.
If creating a program is not sufficient for creating a
mind, what would be sufficient? What is the difference
between the relation that mental states have to brains and
the relation that programs have to their hardware implemen-
tations? What are the relations between mental processes
and brain processes anyhow? Well, obviously I am not going
to answer all of these questions in this short paper, but we
can learn a surprising amount by just reminding ourselves of
the logical consequences of what we know already.
- 3 -
One thing we know is this: quite specific neurophysio-
logical and neurobiological processes in the brain ←λc←λa←λu←λs←λe
those states, events, and processes that we think of as
specifically mental, both in humans and in the higher
animals. Of course the brain, like a computer, or for that
matter, like anything else, has a formal level (indeed many
formal levels) of description. But the ←λc←λa←λu←λs←λa←λl powers of the
brain by which it causes mental states have to do with
specific neurobiological features, specific electrochemical
properties of neurons, synapses, synaptic clefts, neuro-
transmitters, boutons, modules, and all the rest of it. We
can summarize this brute empirical fact about how nature
works as:
Proposition 4. Brains cause minds.
Let us think about this fact for a moment. The fact that a
system has mental states and that they are caused by neuro-
physiological processes has to be clearly distinguished from
the fact that a system that has mental states will charac-
teristically behave in certain ways. For a system might
have the mental states and still not behave appropriately
(if, say, the system is human and the motor nervous system
is interfered with in some way) and it might behave in a way
appropriate to having mental states without having any men-
tal states (if, say, a machine is set up to simulate the
input-output functions of the human system without having
the appropriate mental states -- in a familiar example, the
system might emit the right answers to the right questions
in Chinese and still not understand a word of Chinese.) So
the claim that Brains Cause Minds is not to be confused with
the claim that Minds Cause Behavior. Both are true. But
the claim that brains cause minds is a claim about the "bot-
tom up" powers of the brain. It is a summary of the claim
that lower level neurophysiological processes cause, e.g.,
thoughts and feelings. So far it says nothing at all about
external behavior. Just to keep the distinction straight,
let us write this separate proposition as:
Proposition 5. Minds cause behavior.
Now with P. 5, unlike P. 4, we are not talking about bottom
up forms of causation. We are simply summarizing such facts
as that my pain causes me to say "Ouch," my thirst causes me
to drink beer, etc.
From P. 4 and P. 5 by transitivity of causation, we
can infer
Conclusion 2. Brains cause behavior.
But now with the clear distinction between P. 4 & P. 5
and the observation that the input-output relations of human
beings are mediated by mental states, we can see the real
- 4 -
power and implications of P. 4. The causal powers of the
brain consist not merely in the fact stated by C. 2, that
brains causes it to be the case that in response to certain
stimuli a person will emit certain outputs (e.g. someone
pinches me and I say "Ouch"). The claim is rather that
specific biochemical features of the brain by bottom-up
forms of causation cause all of our mental phenomena includ-
ing those mental phenomena that mediate input-output rela-
tions, i.e. those mental phenomena that cause behavior.
(E.g., when someone pinches me and I say "Ouch" it is
because I feel a pain, and the sensation of pain is caused
by neuron firings in the thalamus and the somato sensory
cortex.)
We have then a clear distinction between the causal
powers of the brain to produce mental states and the causal
powers of the brain (together with the rest of the nervous
system) to produce input-output relations. I certainly have
not demonstrated that P. 4 is true, but I take it that its
truth is demonstrated by the past century of neurobiology.
And in any case, does anyone really doubt it? Does anyone
really doubt that all of our mental states are caused by
low level (e.g.neuronal) processes in the brain? Now from
P. 4, it follows trivially that
Conclusion 3. Any system capable of causing minds
would have to have causal powers equivalent to the bottom-up
causal powers of brains.
This is a trivial consequence of P. 4. Conclusion 3 does
not tell us anything about how those causal powers have to
be realized. As far as logical possibility is concerned
they could be realized, as I have pointed out on numerous
occasions, in green slime, silicon chips, vacuum tubes, or
for that matter, old beers cans. I have also claimed that,
as a matter of empirical fact, the probabilities that beer
cans, silicon chips, etc. have the same causal powers as
neurons is, roughly speaking, zero. The chances that chemi-
cal properties of silicon chips will be equal in their
bottom-up causal powers to the properties of neurons is
about as great as the chances that silicon chips will be
able to perform photosynthesis, lactation, digestion, or any
other specifically biological process. However, as I have
said repeatedly, that is an empirical claim on my part, not
something to be established by philosophical argument alone.
But, once again, does anyone in AI really question it? Is
there someone in AI so totally innocent of biological
knowledge that he thinks that the specfic biochemical powers
of human nervous systems can be duplicated in silicon chips
(transistors, vacuum tubes -- you name it)? Frankly, I
doubt it. I think the underlying mistake comes not from
ignorance but from confusion: the confusion is to suppose
that the same input-output function implies the presence of
the same bottom up causation. This view is enshrined in the
- 5 -
Turing test, but a moment's reflection is sufficient to show
that it is false. For example, at an appropriate level of
description an electrical engine can have the same input-
output function as a gasoline engine -- it can be designed
to respond in the same way to the same commands -- but it
works on completely different internal principles. Analo-
gously a system might pass the Turing test perfectly, it
might have the same information processing input-output
functions as those of a human being and still not have any
inner psychology whatever. It might be a total zombie.
We can now see what was wrong with the robot reply. It
had the wrong level of causation. The presence of input-
output causation that would enable a robot to function in
the world ←λi←λm←λp←λl←λi←λe←λs ←λn←λo←λt←λh←λi←λn←λg ←λw←λh←λa←λt←λe←λv←λe←λr about the presence of
bottom-up causation that would produce mental states.
Now from these elementary considerations, we can derive
two further conclusions.
Conclusion 4. The way that brains cause minds cannot
be solely in virtue of instantiating a computer program.
This conclusion follows from Proposition 4 and Conclusion 1,
that is, from the fact that brains do cause minds, and the
fact that programs are not enough, we can derive Conclusion
4.
Conclusion 5. Any artifact that we design, any system
that is created artifically for the purpose of creating
minds, could not do it solely in virtue of instantiating a
computer program , but would have to have causal powers
equivalent to the bottom-up causal powers of the brain.
This conclusion follows from Conclusions 1 and 3.
Now in all of the vast amount of literature that has
grown up around the Chinese room argument, I cannot see that
any of my critics have ever faced up to the sheer logical
structure of the argument. Which of its axioms do they wish
to deny? Which steps in the derivation do they wish to
challenge? What they have done rather, like Hofstatder and
Dennett, is persistently misquote me or attribute views to
me which are not only views I do not hold, but views which
I have explicitly denied. I am prepared to keep winning
this same argument over and over again, because its steps
are so simple and obvious, and its "assumptions" can hardly
be challenged by anybody who accepts the modern conception
of computation and indeed our modern scientific world view.
It can no longer be doubted that the classical concep-
tion of AI, the view that I have called strong AI, is pretty
much obviously false and rests on very simple mistakes. The
question then arises, if strong AI is false what ought AI to
be doing ? What is a reasonable research project for weak
AI? That is a topic for another paper.
-------
-John R. Searle
∂03-Apr-86 1522 vijay@ernie.berkeley.edu Please acknowledge receipt ...
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 3 Apr 86 15:22:12 PST
Received: by ernie.berkeley.edu (5.45/1.9)
id AA28692; Thu, 3 Apr 86 15:22:57 PST
Date: Thu, 3 Apr 86 15:22:57 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604032322.AA28692@ernie.berkeley.edu>
To: jmc@su-ai.arpa
Subject: Please acknowledge receipt ...
... of Searle's stage 1 paper.
Thanks,
VR
∂03-Apr-86 1611 LES
To: JMC, CLT
∂03-Apr-86 1542 RPG
∂03-Apr-86 1455 LES One more time
JMC would like to have a 4-way discussion with you, me and Carolyn sometime
tomorrow. Telephone will probably suffice. When is a good time?
Most any time would be ok. 1pm?
-rpg-
∂03-Apr-86 1727 ROSENSCHEIN@SUMEX-AIM.ARPA "Any fool" reference
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 3 Apr 86 17:26:59 PST
Date: Thu 3 Apr 86 17:28:30-PST
From: Jeff Rosenschein <ROSENSCHEIN@SUMEX-AIM.ARPA>
Subject: "Any fool" reference
To: jmc@SU-AI.ARPA
Message-ID: <12196003568.60.ROSENSCHEIN@SUMEX-AIM.ARPA>
I got the reference to the "any fool" paper from IJCAI77; thanks.
When I checked the proceedings, though, I saw that the text of the
paper doesn't appear, the name simply appears in the table of contents
(with the page listed as ***). Is there another reference that I can
use?
--Jeff Rosenschein
-------
∂03-Apr-86 1847 LES Qlisp Task Description
To: JMC, RPG, CLT
[We need a task description for the forthcoming Qlisp project, including
milestones. Here is a straw-man. We should, of course, set milestones
that we can't possibly miss, Murphy's Law being what it is. Comments are
solicited. -Les]
An extension of Common Lisp for parallel processing, called Qlisp, is to
be developed using queue-based multi-processing. It will have the
following features.
o Each processor can address the whole of memory, and a processor may
execute programs anywhere in memory on data located anywhere in memory.
o The programmer indicates when parallelism is possible by using parallel
constructions in the source language, which is an extension of Lisp.
o When a program comes to a statement allowing parallelism and decides
(according to the computed value of an allow-parallelism parameter
in the statement) that parallelism is to be invoked, it adds a
collection of tasks to a queue and starts on the first.
o When a processor completes a task it goes to the queue for its next task.
Basing parallelism on run-time queues means that a program isn't written
or compiled for any specific number of processors. The number available
can even change during the course of a computation. Tasks need not be of
similar length, because a processor finishing a short task merely takes
another from the queue.
This project will involve a collaboration between Stanford University, and
Lucid, Inc. An initial task of 18 months duration will produce a
functioning Qlisp system, documentation and debugging tools at a cost of
$1,943,464. Specific milestones are as follows.
1 June 1986: Work begins.
1 Dec. 1986: Have a monoprocessor version of Common Lisp working on the
chosen parallel computer.
1 June 1987: Preliminary version of Qlisp working, with preliminary
documentation.
30 Nov. 1987: Improved version of Qlisp completed together with documentation.
Test results of sample problems run on the parallel computer documented.
This is planned as a step toward the longer range goal of demonstrating
symbolic computation in a parallel computing system.
∂03-Apr-86 1930 LES Qlisp Review
To: RPG
CC: JMC, CLT
On second thought, JMC would prefer to have a face-to-face meeting at,
say, 1:45pm if possible to review the machine decision and the task
statement, including milestones.
∂03-Apr-86 1953 RPG
To: LES
CC: JMC, CLT
∂03-Apr-86 1930 LES Qlisp Review
To: RPG
CC: JMC, CLT
On second thought, JMC would prefer to have a face-to-face meeting at,
say, 1:45pm if possible to review the machine decision and the task
statement, including milestones.
Ok, I'll see you at 1:45.
-rpg-
∂03-Apr-86 1954 RPG Milestones
To: JMC, LES, CLT
The full common lisp is no problem for me. I believe that to get the
whole Common Lisp is as easy as getting any part of it. I further believe that
Squires can have an easier go of it if he can say that *this* project,
at least, is doing full Common Lisp.
-rpg-
∂03-Apr-86 2017 LES EBOS budget
If we use $100k as the upper bound and 1 year as the initial duration of
the EBOS project, then there appears to be no solution. In fact, if we
cut out everyone but Greep and 5% of a secretary's time, I get a one-year
budget of $135,416.
Either we need to get more money or cut the performance time.
∂03-Apr-86 2337 LES EBOS Budget
Here 'tis. I neglected to mention that it includes one part-time
student, as required unless special dispensation is obtained.
--------------------------
Stanford University proposal to IBM
for
Editor-based Operating System
One year beginnine 1 June 1986.
Personnel
Prof. John McCarthy, Principal Investigator n/c
Tepper, Stephan, Research Programmer (100%) 38,000
-----, Student Research Assistant 12,600
(50% acad. yr., 100% summer)
-----, Secretary (5% time) 1,044
-------
Salary subtotals 51,644
Allowance for salary increases 3,099
(8% beginning 9/1/86)
-------
Salary totals 54,743
Staff benefits (25.4% till 9/1/86, 13,987
25.6% thereafter)
Travel (2 East Coast trips. @ $1000, 2,000
Computer time costs 4,000
Other direct costs 4,000
-------
Subtotal 78,730
Indirect costs (69% of direct costs 56,686
initially, 73% beginning 9/1/86,
-------
Total $ 135,416
∂04-Apr-86 0158 HIRSH@SUMEX-AIM.ARPA Re: AI Search
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 4 Apr 86 01:58:42 PST
Date: Fri 4 Apr 86 02:00:58-PST
From: Haym Hirsh <HIRSH@SUMEX-AIM.ARPA>
Subject: Re: AI Search
To: NILSSON@SU-SCORE.ARPA
cc: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA
In-Reply-To: <12195197505.17.NILSSON@SU-SCORE.ARPA>
Message-ID: <12196096860.11.HIRSH@SUMEX-AIM.ARPA>
From a BBoard msg announcing seminars at Berkeley: Batali will be out
this way next week. I believe this is the same guy, and that he ended
up "boxed" on the board (or if not, nearly so). Since he's in the
neighborhood, perhaps we should ask him by?
CS 298-13 Seminar
Faculty Candidate
John Batali
MIT - AI Lab
Recursive Self-Control:
A Computational Groundwork
for Rational Action
2:00 - 3:00 pm
Wednesday, 9 April 1986
597 Evans
-------
∂04-Apr-86 1111 LES reply to message
[In reply to message rcvd 04-Apr-86 01:46-PT.]
He is on the Facilities freebie list, as is Geoff Goodfellow and a bunch more.
∂04-Apr-86 1204 VAL Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
MULTI-VALUED LOGICS
Matt Ginsberg
Thursday, April 10, 4pm
MJH 252
Abstract: A great deal of recent theoretical work in inference has involved
extending classical logic in some way. I argue that these extensions
share two properties: firstly, the formal addition of truth values
encoding intermediate levels of validity between true (i.e., valid) and
false (i.e., invalid) and, secondly, the addition of truth values encoding
intermediate levels of certainty between true or false on the one hand
(complete information) and unknown (no information) on the other. Each
of these properties can be described by associating lattice structures
to the collection of truth values involved; this observation lead us
to describe a general framework of which both default logics and truth
maintenance systems are special cases.
∂04-Apr-86 1535 SJM women2[1,sjm]
Please don't forget to leave me word that you're finished with this
essay.
Susie
∂04-Apr-86 1658 vijay@ernie.berkeley.edu Re: Charniak
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 4 Apr 86 16:57:50 PST
Received: by ernie.berkeley.edu (5.45/1.9)
id AA18643; Fri, 4 Apr 86 16:58:34 PST
Date: Fri, 4 Apr 86 16:58:34 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604050058.AA18643@ernie.berkeley.edu>
To: JMC@su-ai.arpa
Subject: Re: Charniak
Hi,
Charniak's E-mail address is: ec%brown.csnet@csnet-relay.ARPA
I'll also take this opportunity to acknowledge receipt of
you're response to Winograd. If you could get the responses
to Dreyfus and Searle in before you leave, that would be
best - there is another paper about to be sent out from
Douglas Hofstadter. (You can probably get to that one
after you return).
-Enjoy Europe!
- Vijay
∂04-Apr-86 1847 LES Alliant 5th Proposal
To: JMC, CLT
. . . is in JMC's mailbox. It provides a minimum of 3 months of Test
and loaner processors bringing the complement to 8 for one year beginning
in about a year. Looks OK to me.
∂04-Apr-86 1937 RA slides
I found two files with slides in them. I put them, with other things
you might want to take care of before you leave, on your chair.
Rutie
-----
∂05-Apr-86 0236 bibel%germany.csnet@CSNET-RELAY.ARPA [CSNET-RELAY Memo Service: Failed mail (msg.a027248)]
Received: from CSNET-RELAY.ARPA by SU-AI.ARPA with TCP; 5 Apr 86 02:35:53 PST
Received: from germany by csnet-relay.csnet id ab10510; 5 Apr 86 5:35 EST
Date: Sat, 5 Apr 86 10:55:59 MET
From: bibel <bibel%germany.csnet@CSNET-RELAY.ARPA>
To: mccarthy%su-ai.arpa%germany.csnet@CSNET-RELAY.ARPA
Subject: [CSNET-RELAY Memo Service: Failed mail (msg.a027248)]
One more attempt:
----- Forwarded message # 1:
Via: csnet-relay;Date: Fri, 4 Apr 86 10:46:34 EST
From: CSNET-RELAY Memo Service (MMDF) <@csnet-relay.CSNET:mmdf@csnet-relay.arpa>
Subject: Failed mail (msg.a027248)
Sender: mmdf <@csnet-relay.CSNET:mmdf@csnet-relay.arpa>
To: bibel@germany
Your message could not be delivered to
'mcc@su-ai.ARPA (host: su-ai.arpa) (queue: smtp)' for the following
reason: ' I don't know anybody named mcc'
Your message follows:
Received: from germany by csnet-relay.csnet id a027248; 4 Apr 86 4:35 EST
Date: Fri, 4 Apr 86 9:53:04 MET
From: bibel <bibel@germany>
To: mcc%su-ai.arpa@germany.csnet
Subject: [CSNET-RELAY Memo Service: Failed mail (msg.a018361)]
----- Forwarded message # 1:
Via: csnet-relay;Date: Thu, 3 Apr 86 13:57:12 EST
From: CSNET-RELAY Memo Service (MMDF) <@csnet-relay.CSNET:mmdf@csnet-relay.arpa>
Subject: Failed mail (msg.a018361)
Sender: mmdf <@csnet-relay.CSNET:mmdf@csnet-relay.arpa>
To: bibel@germany
Your message could not be delivered to
'mcc@sail.ARPA (host: su-ai.arpa) (queue: smtp)' for the following
reason: ' I'm not host "SAIL.ARPA", in "RCPT TO:<mcc@sail.ARPA>"'
Your message follows:
Received: from germany by csnet-relay.csnet id a018361; 3 Apr 86 13:36 EST
Date: Thu, 3 Apr 86 19:52:31 MET
From: bibel <bibel@germany>
To: mcc%sail.arpa@germany.csnet
Subject: [CSNET-RELAY Memo Service: Failed mail (msg.ab05405)]
Since I did not recall correctly your e-mail address, this is a third
attempt.
Wolfgang
----- Forwarded message # 1:
Via: csnet-relay;Date: Mon, 31 Mar 86 5:37:43 EST
From: CSNET-RELAY Memo Service (MMDF) <@csnet-relay.CSNET:mmdf@csnet-relay.arpa>
Subject: Failed mail (msg.ab05405)
Sender: mmdf <@csnet-relay.CSNET:mmdf@csnet-relay.arpa>
To: bibel@germany
Your message could not be delivered to
'mccarthy@sumex-aim.ARPA (host: sumex-aim.arpa) (queue: smtp)' for the following
reason: ' No such local mailbox as "mccarthy", recipient rejected'
Your message follows:
Received: from germany by csnet-relay.csnet id ab05405; 31 Mar 86 5:34 EST
Date: Mon, 31 Mar 86 9:44:54 MET
From: bibel <bibel@germany>
To: mccarthy%sumex-aim.arpa@germany.csnet
Subject: visit
John,
Thanks again for your call. I am looking forward to your visit here in
Munich. If you like I pick you up at the Vierjahreszeiten hotel thursday
morning at any time which is convenient for you (At April 17). You may
ask the secretary of Rudolph Bayer to call me or my secretary and give
us the time you prefer.
Wolfgang
----- End of forwarded messages
----- End of forwarded messages
----- End of forwarded messages
∂05-Apr-86 1527 NILSSON@SU-SCORE.ARPA Latombe
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 5 Apr 86 15:27:26 PST
Date: Sat 5 Apr 86 15:28:40-PST
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Latombe
To: jmc@SU-AI.ARPA, cannon@SU-SIERRA.ARPA, reynolds@SU-SCORE.ARPA,
binford@SU-WHITNEY.ARPA
cc: nilsson@SU-SCORE.ARPA
Message-ID: <12196506040.17.NILSSON@SU-SCORE.ARPA>
Please save the date May 8 for a visit by Jean-Claude Latombe,
our prime candidate for the robotics position. I will schedule a
talk by Jean-Claude late in the afternoon, and arrange a lunch and
dinner for him. I'll ask Anne to arrange a meeting with him and the
committee and to set up appointments with individuals who want to talk
to him. -Nils
-------
∂05-Apr-86 1626 NILSSON@SU-SCORE.ARPA AI Prospects and Visits
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 5 Apr 86 16:26:30 PST
Date: Sat 5 Apr 86 16:27:41-PST
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: AI Prospects and Visits
To: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
grosof@SU-SCORE.ARPA, hirsh@SUMEX-AIM.ARPA
cc: nilsson@SU-SCORE.ARPA, bscott@SU-SCORE.ARPA
Message-ID: <12196516783.17.NILSSON@SU-SCORE.ARPA>
I've suggested the following dates and visits to prospective candidates
(I haven't heard back from Etherington yet):
Mon. Apr 21: Tom Mitchell
Mon. Apr 28 (Godel's birthday): Shankar
Thursday, May 1: Yoav Shoham
Monday, May 5: John Batali
This concentrates people maybe into a few days a bit more than we might
like, but we can gear up for it and be efficient. I suspect that these
dates won't all be convenient for all of us, but I hope they are ok for
most of us. (I will have to talk to Tom Mitchell quickly and then run
to catch a plane around noon on the date of his visit.)
Anne will keep us all informed when these dates are finally confirmed and
will want to make appointments for these people and for us as a committee
to see them. She'll also schedule them to give talks in the late afternoon
(probably) of the days they will be visiting.
-Nils
-------
∂07-Apr-86 0506 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Your visit
Received: from CS.UCL.AC.UK by SU-AI.ARPA with TCP; 7 Apr 86 05:05:44 PST
Received: from aiva.edinburgh.ac.uk by 44d.Cs.Ucl.AC.UK via Janet with NIFTP
id a000622; 7 Apr 86 12:50 BST
From: Alan Bundy <bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk>
Date: Mon, 7 Apr 86 12:08:30 GMT
Message-Id: <22414.8604071208@aiva.ed.ac.uk>
To: jmc <@cs.ucl.ac.uk,@cs.ucl.ac.uk:jmc@su-ai.arpa>
Subject: Your visit
Cc: bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk,
fjg%edai.edinburgh.ac.uk@cs.ucl.ac.uk
John
Is your visit still on for 15/16 April? Would you
like to give a seminar?
Alan
Alan Bundy <bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk>
∂07-Apr-86 0623 somewhere!ito@aoba.tohoku.junet
Received: from SU-SHASTA.ARPA by SU-AI.ARPA with TCP; 7 Apr 86 06:23:10 PST
Received: by su-shasta.arpa with TCP; Mon, 7 Apr 86 06:23:03 pst
Received: by ntt.junet (4.12/4.7JC-7) CHAOS with CHAOS-MAIL
id AA05102; Mon, 7 Apr 86 10:34:50 jst
Received: by ntt.junet (4.12/4.7JC-7) CHAOS with CHAOS-MAIL
id AA05076; Mon, 7 Apr 86 10:34:16 jst
Message-Id: <8604070134.AA05076@ntt.junet>
Received: by aoba.tohoku.junet (systemV.tohoku)
id AA05212; Sat, 5 Apr 86 23:30:55 JST
Date: Sat, 5 Apr 86 23:30:55 JST
From: ito@aoba.tohoku.junet (Takayasu ITO)
To: nttlab!nttlab!nttlab!Shasta!jmc@su-ai
Subject: On Lisp System
Dear Professor McCarthy:
I am sending this mail to you since I have heard that you have started
your project on Parallel Lisp System.
Last year my graduate students implemented a simple experimental parallel
processor based on MC68000. We constructed four MC68000's system successfully
,which works in 12.5MHz and have 0.5MB local memory and 4MB common memory at
present. We are expanding our system for eight MC68000's system.We are
designing parallel processor system based on MC68020's.
My graduate student and I are working on designing and implementing Parallel
Lisp System on these parallel processor system (and concurrent object oriented system).
My student have a good experience of writing Franzlisp interpreter on Unix
Workstation based on MC68000 and MC68020,using Assembler of MC68000.
This experience helps us to implement Parallel Lisp system.
At present our Lisp system automatically detect program fragment which can
be executed in parallel.
I appreciate if you would let me know the state of your project.
Professor M. Sato,formerly at University of Tokyo, was promoted as Full
Professor of Tohoku University on April 1st as you might know.
Please come to visit Sendai when you visit Japan next time.
Sincerely,Takayasu Ito(Sendai)
P.S.: My daughter was visiting my friend at Stanford to have experience to
stay at your country.She returned home on April 1st. She told me
that she could see you at campus while she was walking with Richards.
If you are going to attend IFIP at Dublin we may be able to see there.
∂07-Apr-86 1359 LES Qlisp Task Description
To: Squires@USC-ISI.ARPA, Pucci@USC-ISI.ARPA
CC: JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, BScott@SU-SCORE.ARPA
[Here is a proposed task description for the Qlisp project, which is to
be added to Contract N00039-84-C-0211, including a revised budget keyed
to a 1 June 1986 start. I invite comments on any problems that you might
see. Given that you already have a full proposal for this project, should
I put just the budget through the Stanford bureaucracy or should the text
of the task description also follow that route? -Les]
An extension of Common Lisp for parallel processing, called Qlisp, is to
be developed using queue-based multi-processing, as described in Gabriel &
McCarthy, "Queue-based Multiprocessor Lisp" in Proceedings of the 1984 ACM
Symposium on Lisp and Functional Programming, August 1984. It will have
the following features.
o Each processor can address the whole of memory, and a processor may
execute programs anywhere in memory on data located anywhere in memory.
o The programmer indicates when parallelism is possible by using parallel
constructions in the source language, which is an extension of Lisp.
o When a program comes to a statement allowing parallelism and decides
(according to the computed value of an allow-parallelism parameter
in the statement) that parallelism is to be invoked, it adds a
collection of tasks to a queue and starts on the first.
o When a processor completes a task it goes to the queue for its next task.
Basing parallelism on run-time queues means that a program isn't written
or compiled for any specific number of processors. The number available
can even change during the course of a computation. Tasks need not be of
similar length, because a processor finishing a short task merely takes
another from the queue.
This task covers the first phase of an expected three year program,
involving a collaboration between Stanford University, and Lucid, Inc.
This task of 18 months duration will produce a functioning Qlisp system,
documentation and debugging tools at a cost of $1,943,464. Assuming this
project starts on 1 June 1986, specific milestones are as follows.
1 Dec. 1986: Have a monoprocessor version of a subset of Common Lisp suitable
for parallel computation working on the chosen parallel computer.
Preliminary specification of Qlisp extensions to Common Lisp completed.
1 June 1987: Preliminary version of Qlisp working, with preliminary
documentation. Specification of test problems to be run will be complete.
30 Nov. 1987: Improved version of Qlisp completed together with documentation.
Test results of sample problems run on the parallel computer documented.
This is planned as a step toward the longer range goal of demonstrating
symbolic computation in a parallel computing system.
APPENDIX D -- Budget for 18 months from 1 June 1986 through 30 Nov. 1987
Stanford University proposal to DARPA
for
Qlisp on parallel processors
[Funds for the required parallel computer are NOT included.]
Personnel
Prof. John McCarthy, Principal Investigator 35,467
(15% acad. yr., 50% summer)
Lester Earnest, Senior Res. Assoc. (50%) 50,625
Carolyn Talcott, Research Associate (100%) 64,500
-----, Research Associate (100%) 61,500
-----, Computer Systems Analyst (100%) 60,000
-----, Computer Technician (25%) 12,000
Ian Mason, Student Research Assistant 20,160
(50% acad. yr., 100% summer)
-----, Student Research Assistant 20,160
(50% acad. yr., 100% summer)
-----, Student Research Assistant 7,560
(beg. 6/1/87; 50% acad. yr.,
100% summer)
-----, Student Research Assistant 7,560
(beg. 6/1/87; 50% acad. yr.,
100% summer)
-----, Secretary (50% time) 15,660
-------
Salary subtotals 355,192
Allowance for salary increases 28,415
(8% beginning 9/1/86,
16% beginning 9/1/87)
-------
Salary totals 383,607
Staff benefits (25.4% till 9/1/86, 98,459
25.6% till 9/1/87, then 26.2%)
Consultants (10 days/yr. @ $500) 7,500
Travel (8 East Coast trips/yr. @ $1000, 22,500
14 Western trips/yr. @ $500)
Computer maintenance 64,548
Computer time costs 40,000
Other direct costs 36,100
-------
Subtotal 652,714
Capital equipment
Ethernet workstation cluster (8 term.) 24,000
Ethernet page printer 11,937
Microwave link 30,000
Interfacing equipment & peripherals 12,000
-------
Subtotal 77,937
Lucid subcontract 722,600
Indirect costs (69% of direct costs 490,213
initially, 73% beginning 9/1/86,
excluding all capital equipment and
subcontract amounts over $25,000) -------
Total $1,943,464
Lucid
Project Cost Estimate
($000's)
Half-
Year 1 Year 2
Direct Labor (Sch. A) 109.2 56.7
Salary Related Expense @19.8% 21.6 11.2
-----------------
Salary Expense 130.8 67.9
Overhead @148% 193.6 100.5
Other Direct Project Costs
(Sch. B) 126.0 9.5
-----------------
450.4 177.9
Fee @15% 67.6 26.7
-----------------
Total Project Cost 518.0 204.6
Notes:
1. Project cost assumes 1 Sequent machine provided to Lucid at no additional
cost located at Lucid facilities for duration of project.
2. Salary Related Expense and overhead are provisional rates based on FY86
(year end June 30) business plan projections.
3. Direct Labor rates escalated at 10% per year.
Schedule A
Direct Labor
($000's)
Half-
Year 1 Year 2
Principal Investigator
man-months 3 1
rate per month 6.0 6.6
total dollars 18.0 6.6
Senior Scientist
man-months 12 6
rate per month 5.0 5.5
total dollars 60.0 33.0
Scientist
man-months 6 3
rate per month 4.2 4.6
total dollars 25.2 13.8
Technician
man-months 2 1
rate per month 3.0 3.3
total dollars 6.0 3.3
Total Man-Months 23 11
Total Dollars 109.2 56.7
Schedule B
Other Direct Project Costs
($000's)
Half-
Year 1 Year 2
Computer Equipment & Supplies
Symbolics 3670 108
Supplies @ 3K/yr 3 1.5
Stanford Communication 2 1
Travel & Subsistence
Air
2 trips to Washington, D.C. @1.5 ea.
1 trip to National Conference @1.5
4 trips to East Coast, technical workshop @1.5 ea.
Local travel @.5/yr 11 6
escalate @10%/yr
Printing and Reproduction @2K/yr 2 1
---------------
Total 126 9.5
∂07-Apr-86 1522 CLT rev1(u,v*w)=rev1(u,v)*w
BMP proves it trivially - see below
(NOTE-LIB "BOO.LIB" "BOO.LISP")
(DEFN APPEND (U V) (IF (LISTP U) (CONS (CAR U) (APPEND (CDR U) V)) V))
(DEFN REV (U V) (IF (LISTP U) (REV (CDR U) (CONS (CAR U) V)) V))
(PROVE-LEMMA REV-APPEND () (EQUAL (REV U (APPEND V W)) (APPEND (REV U V) W) ) )
Give the conjecture the name *1.
Let us appeal to the induction principle. The recursive terms
in the conjecture suggest three inductions. They merge into two
likely candidate inductions. However, only one is unflawed. We will
induct according to the following scheme:
(AND (IMPLIES (AND (LISTP U)
(P (CDR U) (CONS (CAR U) V) W))
(P U V W))
(IMPLIES (NOT (LISTP U)) (P U V W))).
Linear arithmetic and the lemma CDR-LESSP inform us that the measure
(COUNT U) decreases according to the well-founded relation LESSP in
each induction step of the scheme. Note, however, the inductive
instance chosen for V. The above induction scheme generates two new
conjectures:
Case 2. (IMPLIES (AND (LISTP U)
(EQUAL (REV (CDR U)
(APPEND (CONS (CAR U) V) W))
(APPEND (REV (CDR U) (CONS (CAR U) V))
W)))
(EQUAL (REV U (APPEND V W))
(APPEND (REV U V) W))),
which simplifies, opening up the definition of REV, to:
(IMPLIES (AND (LISTP U)
(EQUAL (REV (CDR U)
(APPEND (CONS (CAR U) V) W))
(APPEND (REV (CDR U) (CONS (CAR U) V))
W)))
(EQUAL (REV (CDR U)
(CONS (CAR U) (APPEND V W)))
(APPEND (REV (CDR U) (CONS (CAR U) V))
W))).
This further simplifies, applying CDR-CONS and CAR-CONS, and
expanding the definition of APPEND, to:
T.
Case 1. (IMPLIES (NOT (LISTP U))
(EQUAL (REV U (APPEND V W))
(APPEND (REV U V) W))),
which simplifies, unfolding the function REV, to:
T.
That finishes the proof of *1. Q.E.D.
[ 0.86700017 0.27499987 ]
REV-APPEND
∂08-Apr-86 0652 BMOORE@SRI-AI.ARPA Raphael's paper on knowledge
Received: from SRI-AI.ARPA by SU-AI.ARPA with TCP; 8 Apr 86 06:52:26 PST
Date: Tue 8 Apr 86 06:53:35-PST
From: Bob Moore <BMOORE@SRI-AI.ARPA>
Subject: Raphael's paper on knowledge
To: jmc@SU-AI.ARPA, fagin@IBM-SJ.ARPA
cc: bmoore@SRI-AI.ARPA
I found Bert Raphael's MIT AI memo "Introduction to the Calculus of
Knowledge". It is even earlier than I thought, dating from November
1961. Moreover, it refers to an even earlier memo by John, "A Calculus
of the Knowledge of Propositions". Do you think you might have a copy
of that one somewhere in your files, John?
I am sending both of you copies.
--Bob
-------
∂08-Apr-86 0911 CLT
julie 723-683?
∂08-Apr-86 1556 LES Symbolics Meeting
To: Facil@SU-AI.ARPA
Following up on initiatives by Tom Rindfleisch and Tom Binford,
representatives of Symbolics will appear at noon on Wednesday,
April 23 in MJH 220 (Chairman's Conference Room) to discuss pricing
policies on hardware and software maintenance. Open to all Committee
members who are interested.
∂08-Apr-86 1608 CONTRERAS@SU-SCORE.ARPA message
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 8 Apr 86 16:08:36 PST
Date: Tue 8 Apr 86 16:09:46-PST
From: Tina Contreras <CONTRERAS@SU-SCORE.ARPA>
Subject: message
To: JMC@SU-AI.ARPA
Message-ID: <12197299954.20.CONTRERAS@SU-SCORE.ARPA>
Barbara Clifford phoned, she said, the advisory council meeting will be
next month (May) 9th, 10th, and 11 at the Niven s in Tarzana. If you
should have any questions please call Barbara at 408 298-3269.
Tina
-------
∂09-Apr-86 0947 AAAI-OFFICE@SUMEX-AIM.ARPA Program Committee Reception
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 9 Apr 86 09:46:56 PST
Date: Wed 9 Apr 86 09:48:22-PST
From: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: Program Committee Reception
To: Nilsson@SU-SCORE.ARPA, JMC@SU-AI.ARPA, Fikes@USC-ECL.ARPA, Stefik@XEROX.COM,
Shortliffe@SUMEX-AIM.ARPA, Brown@XEROX.COM, Buchanan@SUMEX-AIM.ARPA,
Tenenbaum@SRI-KL.ARPA, Lerman@SRI-KL.ARPA, Engelmore@SUMEX-AIM.ARPA,
Feigenbaum@SUMEX-AIM.ARPA, Hart@SRI-AI.ARPA
cc: aaai-OFFICE@SUMEX-AIM.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Message-ID: <12197492667.33.AAAI-OFFICE@SUMEX-AIM.ARPA>
I'd like to invite you to this year's Program Committee Reception
on Friday, May 2, 7:00 p., in the courtyard at the Stanford
Park Hotel, 100 El Camino Real, Menlo Park.
Please join us!
Claudia
PS. Please RSVP -- thanks.
-------
∂09-Apr-86 1033 vijay@ernie.berkeley.edu AI DISC: Douglas Hofstadter
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 9 Apr 86 10:32:58 PST
Received: by ernie.berkeley.edu (5.45/1.9)
id AA11863; Wed, 9 Apr 86 10:33:32 PST
Date: Wed, 9 Apr 86 10:33:32 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604091833.AA11863@ernie.berkeley.edu>
To: jmc@su-ai.arpa
Subject: AI DISC: Douglas Hofstadter
**** Please acknowledge receipt of this mailing -Thanks, VR
Impressions of AI
=================
Douglas R. Hofstadter
University of Michigan
April, 1986
I recently went to see an exhibition of Winslow Homer paintings, one
of which was called "Dog on a Log". The title amused me, and I invented
variations on it, such as "Frog on a Dog on a Hog on a Log in a Bog". While
doing this, I was thinking about how people -- especially small children --
love these sorts of simple rhyme, and about how when you hear such things, the
rhyme just jumps out at you, louder than the image behind it. Then I thought
about how speech-understanding programs work, and it occurred to me that if I
pronounced such a phrase into a microphone, a top-notch speech-understanding
program could understand exactly what I had said, in terms of getting all the
words right and putting together the right meaning -- but it would not hear
the rhyme. Not at all. And this seemed to me downright eerie. What would
it be like to hear that phrase PERFECTLY, but not to hear the rhyme? It
would be unimaginably alien -- like being dead, was my feeling.
Many traditional AI people would empathize with my human reaction, and
would say that the program should be fixed up to be able to hear rhymes.
So they would add a "rhyme-detection module", in essence. This does not
seem to me to get you one iota closer to being "alive", however. What about,
say, alliteration? Of course you can add on an alliteration-detection module
as well, and so on ad infinitum. But how many people really believe that this
is how brains work?
When I entered AI some ten years ago, my definition of the field would
have run something like this: "AI is the attempt to understand what thinking
is by making theories, implementing them as computer programs, and learning
from their performance how to improve the theories." Although I wish this
were true, it is far from it. Nowadays, aside from this philosophy, there
are at least two other common views of what AI is: (1) "AI is `knowledge
engineering' -- the field in which one tries to implant expertise in
computers"; and (2) "AI is the branch of computer science concerned with making
computers do more and more things that, for humans, involve intelligence
-- but its practitioners don't care whether the computers accomplish their
results in the way that people do." Both views are highly pragmatic, having
more to do with engineering than with science. Admittedly, excellent minds
can become engrossed in either type of approach, but the concern with the
workings of the human mind has been almost lost, and I believe that the human
mind is much more complex and subtle than anything that people can think
up themselves. I am firmly convinced that AI must first and foremost be
a cognitive science -- a science of the human mind -- rather than a type
of engineering, if it is to make real progress. Artificial intelligence
should be artificial only in the sense that a model of the mind is not as
complex as a real mind.
The engineering approach to building intelligence is a totally open
competition with any idea welcome, whereas in a scientific quest for how minds
work, off-base theories would ideally be pruned by experiments revealing their
flaws. However, since the field is so new, people have not devoted enough
effort to figuring out what constitutes a good test of a theory or of a
program. For instance, I once watched demos of some programs at Yale having
a most baffling property: they used tremendously intricate machinery to
mimic wonderfully flexible acts of cognition -- but those programs worked
with ONLY ONE INPUT. I was reminded of my one visit as a small boy to
Disneyland, where I took thrilling rides down jungle rivers in boats --
but I was most disappointed to find out that these "boats" were actually
running on tracks, so that they had totally fixed routes. (I hope that Roger
Schank would appreciate this reminding-incident.) What should one make of
programs that are so complex, so fancy, so rich with insights -- yet so
brittle? I certainly don't know.
Everybody admits that AI programs don't have common sense yet. One of the
most frequently suggested remedies to this lacuna is that we just give programs
MORE KNOWLEDGE. So one ambitious on-going project is to translate all the
articles in a huge encyclopedia into a sophisticated knowledge-representation
language. I can see the appeal of such a project, but I think it has nothing
whatsoever to do with minds or commmon sense. Adding huge amounts of knowledge
to current "inference engines" is a brute-force approach to a very subtle
problem, and I believe it will fail.
By contrast, I believe that the essence of common sense can be brought out
most clearly in tiny domains, somewhat condescendingly called "toy domains" by
many AI researchers (I like the term). A common view is that AI has exhausted
toy domains and should leave them behind, just as children exhaust the interest
of building blocks and go on to bigger things. I think this is absurd. One
can find huge mysteries in the tiniest of domains, devoid of practically any
world knowledge at all (and for which having vast amounts or world knowledge
would not be of any help).
To my mind, the quality of a scientific project depends critically on
whether it has identified and focused on some central problem in a very clear
way. Toy domains once played (and still could play) the role, in AI, that
idealizations of all sorts have always played in physics. Physicists think
of gases in terms of colliding billiard balls, or solids as perfect lattices;
they think of point particles moving in perfectly straight lines in a perfect
vacuum; and so on. The exact analogue in AI of such idealizations is perhaps
not clear, but certainly it is not a program designed to be an expert in a
highly arcane discipline that practically no one in AI really understands.
Domains like that are so complex that even a weak program can do some things
that people have not thought of. And, although it sounds simple-minded,
I think people are impressed when computers sling around technical jargon in
smooth natural-language discourse, so that weak programs look more awesome
than they are. In a tiny domain, you can't get away with that kind of thing.
Well-chosen toy domains therefore provide much more stringent challenges for
AI than huge domains do.
The canonical toy domain is the MIT "blocks world", home of such classic
AI programs as Guzman's and Waltz's vision programs, Winston's arch-learning
program, Winograd's SHRDLU, and Sussman's HACKER, among others. For reasons
unclear to me, this domain has lost favor in the AI world. People seem to be
under the impression that no challenges of interest could any longer be framed
in such a "small" domain. And yet by no means could any of the above-mentioned
programs be said to have been completed. They all had major defects and acted
very differently from people. They weren't integrated with each other. Simply
producing a blocks-world program that smoothly integrated the skills of ALL
the above-mentioned programs (in improved form) would be a phenomenally hard
task, and, to my mind, a wonderful accomplishment for AI.
As another example of a rich toy domain, I suggest the world of
letterforms -- "a" through "z". Each letter can be written in uncountably
many different ways, and the alphabet as a whole can be designed in uncountably
many different styles. Each letter defines a category that overlaps and rivals
other categories in amazingly subtle ways. Current optical-character-reading
technology has produced useful devices that allow a computer to do pretty well
in reading letters in many styles, but it affords no insight whatsoever into
the fundamental questions of categories, category boundaries, and analogy
(the underpinnings of a uniform visual style). Such devices are as far from
human visual perception as the rhyme-deaf speech-understanding programs I
mentioned above are from human hearing.
The letterform world is an ideal universe in which to study high-level
vision, the interface of perception with categorization, the structure of
very fluid categories, complex associations, overlaps, and rivalries of
categories, and highly abstract analogies, all of which I believe are
at the core of thinking. These issues have only rarely been approached
in traditional AI work -- and yet they are, to me, the problem of mind
in a nutshell. Most of the current work on perception is focused on
low-level (modality-specific) aspects, but I think that the high-level
aspects -- where perception merges with cognition -- are where the greatest
challenge and interest for AI lie. Traditional AI, with its strong focus
on natural language and deduction, tends to presume that all items come
pre-categorized (labeled with words). In such situations, serial models
of thought can do impressive things. On the other hand, such situations
represent but a tiny fraction of what real organisms in the real world
confront in real time.
Recently, there has been a healthy swing away from the serial-cognition
thrust of traditional AI; this movement is often called "connectionism",
although I think "emergent mentality" might be better. Connectionism is
based on the idea that there are fundamental things that serial computers
have shown themselves to be terrible at, and that require a totally different
approach. Years ago, AI people were appalled if you suggested that perhaps
they should be paying attention to how the brain works. They felt that
whole premise of AI was that thought has nothing to do with hardware.
Connectionism is a kind of backlash to that philosophy, and has its roots
in a number of interesting places: neurology, statistical mechanics and
thermodynamics, perhaps automata theory, and so on.
Connectionism's thesis -- that cognition is a collective phenomenon, in
which symbols and representation are statistically emergent phenomena rather
than directly accessible structures and processes -- is a subtle one, which
many old-guard AI people find upsetting, perhaps even inconceivable. Their
resistance is very understandable to me, since I believed in the old view
of AI for a long time. The old view is pretty much a consequence of the
following tight chain of near-equalities:
computation = logic
logic = reasoning
reasoning = thinking
If you believe these premises (and certainly each contains a grain of truth),
then you will have a hard time rejecting the obvious logical conclusion, to
wit:
thinking = computation
This is the basis of standard, mainstream AI. More precisely, that thesis
might be spelled out this way: "Thinking can be implemented as software
that runs on a von-Neumann-type serial machine".
I disagree, in varying amounts, with all three of the "equations" that
underlie this thesis. Actually, the first one is all right, provided you
think of "logic" as meaning merely "digital logic" ("TTL logic"). The third
one is ridiculous, in my opinion, but at least I find it easy to say why: I
have come to believe that most of thinking has nothing to do with reasoning
at all, but with a kind of associationism in which each concept has a hard core
surrounded by a soft blur that cannot possibly be considered part of reasoning.
Typos, speech errors, and many other sorts of error are among the consequences
of that blur; so are clever jokes, beautiful analogies, and intuitive leaps
that result in scientific discovery. The trickiest equation of the three is
the middle one, where the left-hand side ("logic") means two very different
things at once: both "digital (TTL) logic" and "mathematical logic", while
the right-hand side ("reasoning") means "sensible thought pattterns". (This
confounding of meanings of one word is typical of the blurry quality of human
thought -- and in this case it leads to error.) This equation hides perhaps
the subtlest error; because AI people so revere reasoning and because computers
have such a beautifully "logical" internal architecture, one simply WANTS to
equate the two.
All connectionists find fault, for one reason or other, with this
chain of "equations", and believe that the individual micro-elements of
the substrate from which genuine thought (human or artificial) emerges need
not have any representational quality; that that quality attaches only to
large-scale emergent aspects of such a substrate, much as life is a property
of an organism but not of its individual molecules. I am very glad to see
such a school begin to flourish, and I expect many fundamental ideas to
emerge from connectionism over the next decade or two.
My only qualm about this new approach to modeling the mind is that
it so totally eschews AI's original vision of cognition as explicit serial
manipulation of meaning-carrying symbols that it will not provide a framework
in which to naturally address the epistemological questions that AI has been
trying to answer. The point is that AI has always attracted minds that are
interested in how MINDS work, not necessarily how BRAINS work. While it is
probable that mentality is a kind of emergent phenomenon, so that you have
to devote some time to studying its non-thinking micro-components, you have
to beware that you don't get so absorbed in the study of the substrate that
you totally forget about thinking itself.
Connectionist models run the risk of becoming huge simulations that
perform impressively but have little explanatory power. An analogy may
help. Astrophysicists have constructed impressive computer models of the
evolution of galaxies, with thousands of simulated stars interacting via
simulated gravitation. Some of the unexplained properties of galaxies have
been reproduced, such as the formation of spiral arms -- but no one really
understands why. This is a wonderful result, but an explanatory bridge
between the micro- and macro-levels has not been built.
Cognitive psychologists have in essence asked themselves "What is a
concept?" That is, to my mind, the single most important question that AI
ought to be trying to answer. Connectionism may provide an important part
of the answer to this question, but I do not think it will do so alone. Some
insights at the level of thoughts themselves, not just their substrate, will
be essential. Although in some ways, the following will be a shaky analogy,
I would like to suggest it, because it contains a grain of truth. The best
traditional AI (and cognitive psychology) is something like classical physics:
true on a large scale, false on a small scale. Connectionism is something
like quantum mechanics, the ultimate basis of all physics: true on all
scales, but in some sense "irrelevant" on a large scale. The problem
is therefore to link these two vastly different levels. In physics, the
"correspondence principle" says that the equations of quantum mechanics
must turn into their classical counterparts in the limit of large quantum
numbers. In that sense, a beautiful bridge is rigorously established between
the unfamiliar micro-world and the familiar macro-world. I would like to see
such a bridge established between connectionism and the study of cognition
itself, which includes traditional AI, cognitive psychology, linguistics,
and the philosophy of mind. The ultimate goal, to my mind, would be to
provide in this way a hierarchical explanation for such high-level theoretical
constructs as the id, the ego, and the superego (or something along those
lines), allowing us finally to locate in brains or machines the "soft
architecture of the soul".
AI is much more of a hodge-podge than I had expected it to be when I
entered the field. Because it is a fledgling science, and because so much
is unknown, a hundred flowers are blossoming -- but many of them will wilt
rapidly. No one can yet be sure that a particular program or a particular
school or a particular approach is really on the right track to explaining
minds, because most achievements, however grand they may appear on the
surface, are still only tiny fragments of a huge puzzle. It is vitally
important for AI people to resist grandiose claims, and to stress the
limitations of their models. When AI people learn to resist the allure
of complexity, I think they will have taken a major step forward.
∂09-Apr-86 1129 KOHEN@SU-SUSHI.ARPA Errata for Lisp Programming & Proving
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 9 Apr 86 11:29:32 PST
Date: Wed 9 Apr 86 11:30:30-PST
From: Abraham Kohen <KOHEN@SU-SUSHI.ARPA>
Subject: Errata for Lisp Programming & Proving
To: jmc@SU-AI.ARPA
Message-ID: <12197511258.19.KOHEN@SU-SUSHI.ARPA>
Errata:
LISP Programming & Proving @ 1984
(Version TeXed at 12:12pm September 19,1984)
p.37 equations (8)
fibb[n,k,m] <- if n = 0 then k else ....
↑
↑
reads "then k"
should be "then m"
(Symbols used above are approximations.)
(Been carrying this errata note in my pocket for some time now, so if
someone else has already noted this, then never mind.)
Abe Kohen
-------
∂09-Apr-86 1134 RA finals CS306
John,
Bob Givan is willing to grade your part of the CS306 finals; if this is ok with
you please let me know.
Thanks,
Rutie
------
∂09-Apr-86 1206 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Re: visit
Received: from CS.UCL.AC.UK by SU-AI.ARPA with TCP; 9 Apr 86 12:06:34 PST
Received: from aiva.edinburgh.ac.uk by 44d.Cs.Ucl.AC.UK via Janet with NIFTP
id a007215; 9 Apr 86 12:30 BST
From: Alan Bundy <bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk>
Date: Wed, 9 Apr 86 12:11:25 GMT
Message-Id: <7948.8604091211@aiva.ed.ac.uk>
To: JMC@su-ai.arpa
Subject: Re: visit
Cc: bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk
Great! Look forward to seeing you.
Alan
∂09-Apr-86 1538 YAO@SU-SCORE.ARPA admission of R. Zabih
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 9 Apr 86 15:38:17 PST
Date: Wed 9 Apr 86 15:39:21-PST
From: Andrew Yao <YAO@SU-SCORE.ARPA>
Subject: admission of R. Zabih
To: Cheadle@SU-SCORE.ARPA
cc: Nilsson@SU-SCORE.ARPA, JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, yao@SU-SCORE.ARPA
Message-ID: <12197556561.11.YAO@SU-SCORE.ARPA>
Victoria,
There is one more Ph.D. admission case we need to process. There is an
applicant, Ramin Zabih, who was late for our application deadline. Professor
John McCarthy would like to have him as a thesis student. Speaking for the
Admission Committee, I propose that we add Zabih to the list of the admitted.
Please start the process of sending him acceptance letter, information, etc,
as soon as you receive the approval of this proposal from Nils. You will
also get a confirmation message from John McCarthy to the effect that he
would like to have Zabih as a Ph.D. thesis student (Carolyn Talcott is
authorized by John to send a confirmation message for him).
Thanks.
-- Andy
-------
∂09-Apr-86 1725 CLT Zabih
To: Cheadle@SU-SCORE.ARPA
CC: JMC@SU-AI.ARPA
Hi,
This note is to confirm that JMC wants Ramin Zabih as a PhD thesis
student.
Carolyn
∂10-Apr-86 0720 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Your visit
Received: from CS.UCL.AC.UK by SU-AI.ARPA with TCP; 10 Apr 86 07:19:56 PST
Received: from aiva.edinburgh.ac.uk by 44d.Cs.Ucl.AC.UK via Janet with NIFTP
id a002832; 10 Apr 86 11:43 BST
From: Alan Bundy <bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk>
Date: Thu, 10 Apr 86 10:32:05 GMT
Message-Id: <16676.8604101032@aiva.ed.ac.uk>
To: jmc <@cs.ucl.ac.uk,@cs.ucl.ac.uk:jmc@su-ai.arpa>
Subject: Your visit
Cc: bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk,
fjg%edai.edinburgh.ac.uk@cs.ucl.ac.uk,
gideon%edai.edinburgh.ac.uk@cs.ucl.ac.uk
John
We would be delighted if you could give a seminar when you come.
We have pencilled in 2pm on Tuesday 15th. Please choose whatever topic
you prefer, and please send a title and abstract if possible. Bernard Meltzer
is coming the next day and giving a seminar at 2pm Wed. You are invited to
lunch with Bernard myself and the seminar organiser, Gideon Sahar, on Wed
at 12.30pm. Look forward to seeing you.
Alan
∂10-Apr-86 1138 VAL Today's Non-Monotonic Seminar: Reminder
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
MULTI-VALUED LOGICS
Matt Ginsberg
Thursday, April 10, 4pm
MJH 252
Abstract: A great deal of recent theoretical work in inference has involved
extending classical logic in some way. I argue that these extensions
share two properties: firstly, the formal addition of truth values
encoding intermediate levels of validity between true (i.e., valid) and
false (i.e., invalid) and, secondly, the addition of truth values encoding
intermediate levels of certainty between true or false on the one hand
(complete information) and unknown (no information) on the other. Each
of these properties can be described by associating lattice structures
to the collection of truth values involved; this observation lead us
to describe a general framework of which both default logics and truth
maintenance systems are special cases.
∂10-Apr-86 1416 SJG pointer needed
Hi John:
Can you please tell me who coined the phrase "qualification problem" for
planning tasks, and give me a reference? Thanks much.
Matt
∂10-Apr-86 1550 HIRSH@SUMEX-AIM.ARPA Mitchell visit
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 10 Apr 86 15:49:47 PST
Date: Thu 10 Apr 86 15:50:33-PST
From: Haym Hirsh <HIRSH@SUMEX-AIM.ARPA>
Subject: Mitchell visit
To: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
nilsson@SU-SCORE.ARPA, grosof@SU-SCORE.ARPA, hirsh@SUMEX-AIM.ARPA
Message-ID: <12197820744.18.HIRSH@SUMEX-AIM.ARPA>
I am organizing Mitchell's schedule for his visit on Monday, 4/21,
instead of Anne. Could you let me know what constraints you have on
times to meet with him, and to make his talk? Thanks!
Haym
-------
∂10-Apr-86 1655 LES Request for DARPA lecture
To: JMC
CC: CLT
Bob Simpson of DARPA just called, trying to reach you. They are putting
together a series of lectures aimed at educating DARPA management and
invite you to talk on a subject of your choosing, such as perspectives on
language development or AI evolution.
Apparently the only person who has talked so far is Al Newell. They would
like to schedule you soon, e.g. Monday, April 21 if you could add it to
the end of your itinerary. For some reason they like to do it on Mondays,
so April 28 would be their next preference. Other days are possible.
Simpson can be reached by email: Simpson@isi
or phone: 202 694-5918.
∂10-Apr-86 1749 VAL Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
A THEORY OF DIAGNOSIS FROM FIRST PRINCIPLES
Raymond Reiter
Department of Computer Science, University of Toronto
and
The Canadian Institute for Advanced Research
Thursday, April 17, 4pm
MJH 252
ABSTRACT
Suppose given a description of a system, together with an
observation of the system's behaviour which conflicts with the way
the system is meant to behave. The diagnostic problem is to determine
those components of the system which, when assumed to be functioning
abnormally, will explain the discrepancy between the observed and
correct system behaviour.
We propose a general theory for this problem.The theory requires
only that the system be described in a suitable logic. Moreover, there
are many such suitable logics, e.g., first order, temporal, dynamic,
etc. As a result, the theory accomodates diagnostic reasoning in a wide
variety of practical settings, including digital and analogue circuits,
medicine, and database updates. The theory leads to an algorithm for
computing all diagnoses, and to various results concerning principles
of measurement for discriminating between competing diagnoses. Finally,
the theory reveals close connections between diagnostic reasoning and
non-monotonic reasoning.
∂11-Apr-86 0035 LES The next contest . . .
To: JMC, RPG, CLT, JJW
will be to choose a name for the Alliant computer, which if all goes
well will arrive in about 2 weeks.
∂11-Apr-86 0410 gideon%edai.edinburgh.ac.uk@Cs.Ucl.AC.UK Re: Your visit
Received: from CS.UCL.AC.UK by SU-AI.ARPA with TCP; 11 Apr 86 04:06:42 PST
Received: from edai.edinburgh.ac.uk by 44d.Cs.Ucl.AC.UK via Janet with NIFTP
id a000657; 11 Apr 86 10:39 BST
From: Gideon Sahar <gideon%edai.edinburgh.ac.uk@cs.ucl.ac.uk>
Date: Thu, 10 Apr 86 14:17:47 GMT
Message-Id: <20247.8604101417@edai.ed.ac.uk>
To: bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk,
jmc <@cs.ucl.ac.uk,@cs.ucl.ac.uk:jmc@su-ai.arpa>
Subject: Re: Your visit
Alan:
Jim's coming too ...
- Gideon
∂11-Apr-86 0951 UNIETIS@SU-SUSHI.ARPA CS 306
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 11 Apr 86 09:48:09 PST
Date: Fri 11 Apr 86 09:48:53-PST
From: David R. Unietis <UNIETIS@SU-SUSHI.ARPA>
Subject: CS 306
To: jmc@SU-AI.ARPA
cc: unietis@SU-SUSHI.ARPA
Message-ID: <12198017049.22.UNIETIS@SU-SUSHI.ARPA>
I was a student in your CS306 class last fall, and I have not yet received
a grade for the class (my final was in within the 8-day time period, but
after the deadline to get a grade that quarter). As I am an honors-coop
student, I will eventually need to get a grade for this class.
Thanks,
David Unietis (unietis@sushi)
-------
∂11-Apr-86 1122 guibas@decwrl.DEC.COM Miklos Simonovits
Received: from DECWRL.DEC.COM by SU-AI.ARPA with TCP; 11 Apr 86 11:21:14 PST
Received: from magic.ARPA (magic) by decwrl.DEC.COM (4.22.03/4.7.34)
id AA13233; Fri, 11 Apr 86 11:22:25 pst
Received: by magic.ARPA (4.12/4.7.34)
id AA02416; Fri, 11 Apr 86 11:24:00 pst
From: guibas@decwrl.DEC.COM (Leo Guibas)
Message-Id: <8604111924.AA02416@magic.ARPA>
Date: 11 Apr 1986 1123-PST (Friday)
To: jmc@sail
Cc: guibas@decwrl.DEC.COM
Subject: Miklos Simonovits
John,
I am heading the theory search committee this spring. A fellow who has
contacted us is Miklos Simonovits from the Mathematical Institute of the
Hungarian Academy of Sciences. He is very strong in combinaorics and
has recently been working in problems relating to parallel computation.
He is not interested in a permanent position but wishes to vist
Stanford for a quarter and teach. Nils suggested that we appoint him a
visiting professor for (say) winter quarter 1986-87 and referred me to
you. My committee favors this appointment.
I think Miklos would also help us make contact with the many other
excellent hungarians currently working in theoretical computer science.
L.
∂11-Apr-86 1402 CLT japan collaboration
To: JMC@SU-AI.ARPA, sf@SU-CSLI.ARPA, RWW@SU-AI.ARPA, JK@SU-AI.ARPA,
CG@SU-AI.ARPA, CLT@SU-AI.ARPA
The Japanese will be at Stanford again this summer during July and August.
Richard plans to go to Japan this May.
John and I will probably go next spring and Sol is thinking about
next spring too. Does anyone have any further plans?
As none of us has taken a trip to Japan yet, I checked with
the officals to see if it matters when the money is spent.
The word is that trips can be taken at anytime during the
grant period - which ends September 1987. Since two trips
are allowed, this still is a little short. They say there
should be no problem in getting a no-cost extension.
(We just need to write a letter to the project monitor
explaining why the extension is wanted.)
I propose that we request an extension to Sept 88.
Any comments?
∂11-Apr-86 1737 vijay@ernie.berkeley.edu
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 11 Apr 86 17:37:37 PST
Received: by ernie.berkeley.edu (5.45/1.12)
id AA01901; Fri, 11 Apr 86 17:38:20 PST
Date: Fri, 11 Apr 86 17:38:20 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604120138.AA01901@ernie.berkeley.edu>
To: jmc@su-ai.arpa
To Group 2 & "In the Middle" participants...
STAGE 1 PAPERS IN:
HUBERT & STUART DREYFUS
JOHN SEARLE
TERRY WINOGRAD
DOUGLAS HOFSTADTER
(TO BE SENT OUT: JOSEPH WEIZENBAUM, DAVID RUMELHART).
We have 4 very interesting papers in!! After reading
them, please send out your thoughts on them.. the authors, and others
are very interested to hear what you have to say. Don't worry about
length or extreme formality -- this is a discussion, so please
feel free to express your thoughts freely. We would like you
to get your responses to these papers in by friday of next week,
April 19.
Sincerely,
- VR RA JS DB
P.S. Remember, the sooner you get your responses in, the faster
you'll get reponses to them -- not only from the authours, but
from everyone. Weizenbaum and Rumelhart's papers will be on
their way soon.
Please acknowledge receipt of this mail. - Thanks.
∂12-Apr-86 0503 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Your visit
Received: from CS.UCL.AC.UK by SU-AI.ARPA with TCP; 12 Apr 86 05:03:30 PST
Received: from aiva.edinburgh.ac.uk by 44d.Cs.Ucl.AC.UK via Janet with NIFTP
id a001291; 11 Apr 86 17:43 BST
From: Alan Bundy <bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk>
Date: Fri, 11 Apr 86 17:23:58 GMT
Message-Id: <425.8604111723@aiva.ed.ac.uk>
To: jmc <@cs.ucl.ac.uk,@cs.ucl.ac.uk:jmc@su-ai.arpa>
Subject: Your visit
Cc: bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk,
fjg%edai.edinburgh.ac.uk@cs.ucl.ac.uk,
gideon%edai.edinburgh.ac.uk@cs.ucl.ac.uk,
tims%edai.edinburgh.ac.uk@cs.ucl.ac.uk
John
I am busy on Tuesday morning invigilating an exam, so
cannot see you until 12noon. Can you go to Forrest Hill first
where Tim Smithers is expecting you at about 9.30am? I will
see you at 12 and take you to lunch. Your seminar is at 2pm.
Are you fixed up for accomodation etc?
Alan
∂12-Apr-86 1452 RDZ@MC.LCS.MIT.EDU Graduate School Paperwork
Received: from MC.LCS.MIT.EDU by SU-AI.ARPA with TCP; 12 Apr 86 14:52:23 PST
Date: Sat, 12 Apr 86 17:54:30 EST
From: Ramin Zabih <RDZ@MC.LCS.MIT.EDU>
Subject: Graduate School Paperwork
To: Cheadle@SU-SCORE.ARPA
cc: JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, Nilsson@SU-SCORE.ARPA
Message-ID: <[MC.LCS.MIT.EDU].882569.860412.RDZ>
Hi,
Professor Nilsson just informed me that I would be admitted into the
PhD program at Stanford for next year. I intend to come to Stanford
after finishing up my Master's degree here at MIT. I believe,
however, that various paperwork needs to be taken care of first. In
particular, could you be so kind as to send me the informational
packet about Stanford that admitted students usually receive? My
address is:
Ramin Zabih
140 Hillside Road
Watertown, Mass., 02172
Thank you very much.
∂13-Apr-86 0826 CHEADLE@SU-SCORE.ARPA Re: Graduate School Paperwork
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 13 Apr 86 08:26:51 PST
Date: Sun 13 Apr 86 08:27:37-PST
From: Victoria Cheadle <CHEADLE@SU-SCORE.ARPA>
Subject: Re: Graduate School Paperwork
To: RDZ@MC.LCS.MIT.EDU
cc: nilsson@SU-SCORE.ARPA, jmc@SU-AI.ARPA
In-Reply-To: <[MC.LCS.MIT.EDU].882569.860412.RDZ>
Office: Margaret Jacks 258, 497-1519
Message-ID: <12198526542.8.CHEADLE@SU-SCORE.ARPA>
Ramin,
Yes, you will be receiving the first packet very soon.
There will be a couple of more sent over the ensuing months.
Please feel free to contact me if you have any other questions
or concerns. We look forward to seeing you in the Fall.
Victoria
-------
∂14-Apr-86 0034 Carnese@SRI-KL.ARPA talk at May CPSR meeting
Received: from SRI-KL.ARPA by SU-AI.ARPA with TCP; 14 Apr 86 00:33:54 PST
Date: Sun, 13 Apr 1986 21:42 PST
From: Carnese@SRI-KL
To: jmc at SU-AI.ARPA
Subject: talk at May CPSR meeting
Just wanted to remind you that the meeting is Wednesday, May 14, at 7:30.
Let me know if you want an overhead projector.
-- Dan
∂14-Apr-86 0546 PETTY@RED.RUTGERS.EDU 1986-April-technical mailing
Received: from RED.RUTGERS.EDU by SU-AI.ARPA with TCP; 14 Apr 86 05:46:20 PST
Date: 14 Apr 86 08:43:47 EST
From: PETTY@RED.RUTGERS.EDU
Subject: 1986-April-technical mailing
To: arpanet.mail: ;
cc: petty@RED.RUTGERS.EDU
Message-ID: <12198758860.27.PETTY@RED.RUTGERS.EDU>
Below is a list of our newest technical reports.
The abstracts for these are available for access via FTP with user account
<anonymous> with any password. The file name is:
<library>tecrpts-online.doc
If you wish to order copies of any of these reports please send mail via the
ARPANET to PETTY@RUTGERS. Thank you!!
[ ] DCS-TR-148 - (REVISED) -"PENALTY FORMULATION FOR ZERO-ONE
NONLINEAR PROGRAMMING", B. Kalantari and J.B. Rosen.
[ ] DCS-TR-169 - "TAXICABS AND LEAST-COST PARTITION ALGORITHMS",
T.J Marlowe, Jr.
[ ] DCS-TR-170 - "THE GENERAL CONSISTENT LABELING (OR CONSTRAINT
SATISFACTION) PROBLEM", B.A. Nadel.
[ ] DCS-TR-171 - "THREE CONSTRAINT SATISFACTION ALGORITHMS AND
THEIR COMPLEXITIES: SEARCH-ORDER DEPENDENT AND
EFFECTIVELY INSTANCE-SPECIFIC RESULTS", B.A. Nadel.
[ ] DCS-TR-172 - "INCREMENTAL UPDATING OF DEPTH-FIRST SEARCH TREES",
M.G. Sackrowitz.
[ ] DCS-TR-173 - "SIMPLIFIED STABLE MERGING TASKS", J. Salowe and
W. Steiger.
[ ] DCS-TR-174 - "DETERMINISTIC SELECTION IN 0(LOGLOG N) PARALLEL
TIME", M. Ajtai, J. Komlos, W.L. Steiger and
E. Szemeredi.
[ ] DCS-TR-175 - "FPOPT: A GLOBALLY OPTIMIZING COMPILER FOR FP",
J.S. Pendergrast and B.G. Ryder.
[ ] DCS-TR-176 - "WAVE PROPAGATION ANALYSIS OF THE BOX AND OTHER
IMPLICIT APPROXIMATIONS OF HYPERBOLIC EQUATIONS",
R. Vichnevetsky.
[ ] LCSR-TR-73 - "(REVISED) - QUADRATIC FUNCTIONS WITH EXPONENTIAL
NUMBER OF LOCAL MAXIMA", B. Kalantari.
[ ] LCSR-TR-75 - "ON THE EXISTENCE OF WEAK GREEDY MATCHING HEURISTICS",
M.D. Grigoriadis, B. Kalantari and C.Y. Lai.
[ ] LCSR-TR-76 - "A NEW CLASS OF HEURISTIC ALGORITHMS FOR WEIGHTED
PERFECT MATCHING", M.D. Grigoriadis and B. Kalantari.
[ ] LCSR-TR-77 - "A FAST ALGORITHM FOR TRUMMER'S PROBLEM", A.
Gerasoulis, M.D. Grigoriadis and Liping Sun.
-------
∂14-Apr-86 0850 WINOGRAD@SU-CSLI.ARPA Re: policy
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 14 Apr 86 08:50:06 PST
Date: Mon 14 Apr 86 08:47:22-PST
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Re: policy
To: CLT@SU-AI.ARPA
cc: WINOGRAD@SU-CSLI.ARPA, cheadle@SU-SCORE.ARPA, nilsson@SU-SCORE.ARPA,
jmc@SU-AI.ARPA
In-Reply-To: Message from "Carolyn Talcott <CLT@SU-AI.ARPA>" of Thu 10 Apr 86 09:40:00-PST
As far as I know, there is no departmental policy about when people
begin. The distinction is totally fuzzy for people who move up from
our own Masters. There are possible problems with starting midyear,
in terms of orientation, getting into the right data bases, etc. Are you
willing to take responsibility for making sure they happen? (Victoria
can help you know what needs doing, and it seems pointless to ask John
to pay attention to it).
As to course requirements, there are none, just number-of-unit requirements
(actually amount-of-tuition requirements) set by the university.
--t
-------
∂14-Apr-86 1332 WINOGRAD@SU-CSLI.ARPA Fellowships in AI←
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 14 Apr 86 12:18:41 PST
Date: Mon 14 Apr 86 12:15:49-PST
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Fellowships in AI←
To: jmc@SU-AI.ARPA
cc: nilsson@SU-SCORE.ARPA, genesereth@SUMEX-AIM.ARPA, buchanan@SUMEX-AIM.ARPA,
feigenbaum@SUMEX-AIM.ARPA, tob@SU-AI.ARPA, WINOGRAD@SU-CSLI.ARPA
John,
I just got your letter (I was surprised to get it from you in hardcopy
rather than on-line) about fellowships in AI. I think it is a great idea.
Moving funding away from specific projects (often development projects)
to supporting creative young people in an open-ended way seems like a
very promising direction. Are you looking for letters of support?
--t
-------
∂14-Apr-86 1525 tgd%oregon-state.csnet@CSNET-RELAY.ARPA Request for funding: Knowledge Compilation Workshop
Received: from CSNET-RELAY.ARPA by SU-AI.ARPA with TCP; 14 Apr 86 15:25:46 PST
Received: from oregon-state by csnet-relay.csnet id ae02024; 14 Apr 86 17:53 EST
Received: by orstcs.UUCP (4.12/6.3.ORST)
id AA08405; Fri, 11 Apr 86 12:11:37 pst
Date: Fri, 11 Apr 86 12:11:37 pst
From: Tom Dietterich <tgd%oregon-state.csnet@CSNET-RELAY.ARPA>
Posted-Date: Fri, 11 Apr 86 12:11:37 pst
To: jmc%sail@csnet-relay.arpa
Subject: Request for funding: Knowledge Compilation Workshop
Cc: bennett%sumex@csnet-relay.arpa, mostow%rutgers@csnet-relay.arpa,
tgd%oregon-state.csnet@CSNET-RELAY.ARPA
Dr. McCarthy,
Did you receive my proposal requesting AAAI funding for the Knowledge
Compilation Workshop? If so, what are our chances of getting AAAI funding?
Let me know if you need any other supporting documentation.
--Tom Dietterich (503) 754-4466
∂15-Apr-86 1004 RA inactive files
The inactive files are listed in the last page of peopin[1,jmc]. They are located
on top of the bookcase behind the door.
∂15-Apr-86 1529 RA IBM Almaden dedication ceremony
Jeannie Loui from IBM, (408) 927 1141, wanted to know whether you
intend to come to the ceremony
May 30, Fri. 9:30am. Please let me know.
∂15-Apr-86 2231 HIRSH@SUMEX-AIM.ARPA Mitchell meeting times
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 15 Apr 86 22:31:39 PST
Date: Tue 15 Apr 86 22:34:19-PST
From: Haym Hirsh <HIRSH@SUMEX-AIM.ARPA>
Subject: Mitchell meeting times
To: ra@SU-AI.ARPA
cc: hirsh@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA
Message-ID: <12199204968.15.HIRSH@SUMEX-AIM.ARPA>
I have scheduled McCarthy's meeting with Mitchell for Monday 4/21 1:30-2:00,
and the search committee (which McCarthy is on) meeting with Mitchell for
Tuesday 4/22 1:30-2:30. Please let me know whether these times are ok. I
hope to finalize the schedule Wednesday afternoon. If, when McCarthy returns
Monday, it turns out these times are bad, Anne Richardson will have Mitchell's
schedule and will try to juggle things around.
Mitchell's talk will be Monday 2:30-4:00. Unfortuantely, there is no chance
of changing its time if the time is bad for McCarthy.
Also, please make sure McCarthy gets this information!
Thanks for the help!
Haym
-------
∂15-Apr-86 2331 vijay@ernie.berkeley.edu AI Disc: Joseph Weizenbaum
Received: from [128.32.0.6] by SU-AI.ARPA with TCP; 15 Apr 86 23:31:29 PST
Received: by ernie.berkeley.edu (5.45/1.12)
id AA14126; Tue, 15 Apr 86 23:32:47 PST
Date: Tue, 15 Apr 86 23:32:47 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604160732.AA14126@ernie.berkeley.edu>
To: jmc@su-ai.arpa
Subject: AI Disc: Joseph Weizenbaum
About AI
One of the definitions of AI that floats around the computer environment is
that it is the intelligence exhibited by computers when they do things
which, were they done by people, would be considered evidence of the
performer's intelligence. One trouble with this definition is that it is
too broad. It would cover, for example, virtually the entire set of
applications and systems programs ever written. It would, after all,
require considerable intelligence on the part of human beings to solve
systems of differential equations, to invert matrixes, or to manage the
execution of jobs under severe constraints of time, priorities, etc.
Computers do all these things, yet, systems and applications programmers
don't think of themselves as working in AI. Moreover, the "AI community",
a social category on whose definition there is little, if any, consensus at
all, specifically excludes such computer professionals. That this
exclusion is arbitrary can be seen from the fact that literally no one
suggests that the computer systems that routinely land wide body aircraft,
such as Boeing 747's and DC-10's, a task that surely requires great
intelligence on the part of human pilots when they take the controls, no
one suggests that these systems are in the category AI. Ed Feigenbaum,
who, by the way, is emphatically thought not an AI worker by, for example,
the AI insiders at the MIT and Yale Universities' AI laboratories, excluded
MACSYMA, the spectacularly successful applied mathematics system developed
by Joel Moses at MIT concurrent with the developments of Stanford's
DENDRAL, from his list of existing "expert systems" published in his Fifth
Generation, on the ground (so he told me personally) that MACSYMA is not
AI!
Even though the above quasi definition of AI is unsatisfactory in many
repects, it is useful. I prefer it, at least in the present context, to
definitions that appeal to "the way humans do things" or to analogies to
the human brain and such. Under this definition, exercised with a little
bit of common sense and judgment, today's chess playing machines are in,
even though they don't choose (compute) their moves in the same way, or
even in a way resembling the way, chess masters choose their moves. Also
in are robots whose behavior is in part a consequence of their perception
of their environments by means of sensors such as video cameras, touch
sensitive fingers, and so on. Cruise missiles are in. So, of course, is
the set of computer systems developed under the heading "Cognitive
Science".
The reason I prefer to be that all inclusive is that, while I have no
quarrel at all with the idea that there is such a thing as artificial
intelligence, that there exist some, as seen from a performance point of
view, very impressive examples of it, I do not agree with the claim that
"the [computer's] ability to [think and learn and create] is going to
increase rapidly until - in the visible future - the range of problems they
can handle will be coextensive with THE RANGE TO WHICH THE HUMAN MIND HAS
BEEN APPLIED", as Herb Simon wrote almost 30 years ago - in 1958, to be
exact, that is, in the almost no longer visible past.
The artificial intelligencia and I have never disagreed about 1) that
artificial intelligence is possible, 2) that there is no way to put a limit
on the degree - that is, the amount or magnitude - of AI machines may in
future achieve. For all I know, machines may someday be "more intelligent"
than their human models or creators. In any case, I don't know how such an
assertion can be disproved - even apart from the necessary vagueness of its
language. This, by the way, appears to be a point on which Dreyfus and I
disagree. My point of departure from the artificial intelligencia is that I
insist a machine's intelligence, no matter how "great" - must always be
different from and, in certain very important respects, alien to human
intelligence, whereas they, the people claiming to represent AI, believe
that machines can be made to think as humans think, that they can be made
to understand human language as humans do, that, in other words they can
achieve, if that's the right word, an inner life no different in all but
trivial ways from the inner life of human beings.
My belief that machines cannot be made to understand human situations
essentially involved with such matters as respect, love, the psychological
identity (in Eric Ericson's sense, for example) of individuals, and so on,
leads me to the derivative belief that there are areas of human concern to
which AI machines OUGHT not to be applied even if, perhaps especially if,
such machines can be made to ask and answer questions, perhaps in natural
language, in those areas. The machine responses may well create a very
powerful impression, i.e., the illusion, that the machine "understands", it
must, however, reach its conclusions on bases which human beings ought not
to accept in such human problem situations. Machines ought not, for
example, practice clinical psychiatry - so-called talk therapies - nor act
as judges or jurors in human conflict resolution efforts. This conclusion
is not at all sensitive to the manner of achieving AI, that is, whether AI
is realized on Von Neuman machines, connection machines, machines designed
to duplicate the neural (or whatever) structure of the brain. Individual
human beings are the products of their biological constitution and of their
histories. No artifact can have had a history of experiences remotely
resembling that of a human being. Nor can the history of any individual
human being be expressed by a finite bit string (which could be given to a
computer in lieu of its having the human experience itself). I think these
points are obvious. Those who disagree with me on these fundamental points
are free to believe that here, if anywhere, is an outcropping of faith,
theology, belief or whatever, in other words of something that they believe
need not be taken seriously in a discussion of scientific matters.
That the computer is and can be a powerful metaphor, in terms of which much
of the functioning human can be discussed and, in a certain sense,
understood goes without saying in this ninth decade of the 20th century.
Consistent with the position I have here sketched, however, I think there
is more to the life of the mind (at least mine) than cognition. An
understanding of the human mind resting entirely on the computer metaphor
must, in my view, be necessarily grossly incomplete. That it may
nevertheless be useful, even very useful, I don't dispute.
Of those, however, who believe that the improvement in computer chess
playing over the last dozen years or so is a triumph of cognitive science
and not due mainly to the increasing raw power of the computers involved, I
would like to ask what new psychological principles have been discovered
and deployed to account for the strength of todays machine chess play.
-------
- Joseph Weizenbaum
∂16-Apr-86 1129 RA Meeting Tom Mitchell
Haym Hirsh scheduled your meeting with Mitchell for Monday 4/21 1:30-2:00,
and the search committee (which you are on) meeting with Mitchell for
Tuesday 4/22 1:30-2:30. Please let me know whether these times are ok.
If these times are bad, Anne Richardson will have Mitchell's
schedule and will try to juggle things around.
Mitchell's talk will be Monday 2:30-4:00. Unfortuantely, there is no chance
of changing its time if the time is bad for you.
Rutie
-----
∂16-Apr-86 1204 RA telex from Milan
This telex arrived today from The American Consulat General, US & Foreign
Services, Milan
Re telex no. 0552, April 11, 1986, visit of Italian businessmen to US research
centers.
We would be very grateful if you could informa us ASAP about your visit
and availability to meet with us during your stay in Milan. We understand
your are scheduled to speak at the Work Revisited Conference on April 19.
Thank you for your cooperation.
Carmen G. D'Aloisio
Commercial Officer
∂16-Apr-86 1419 LES DARPA $ spending
To: facil@SU-AI.ARPA
I either have or will request permission to buy the following equipment by
the end of today. This listing is organized by vendor, with individual
items grouped by project. [Project codes given in brackets are explained
at the end].
Relative to the last budget that you saw, Dave Cheriton's slice has
enlarged (as usual) by $3.5k, KSL's has gone down by $4.6k because of
better prices and much of the stuff requested by the Computer Facilities
Group is not included because they have not provided specifications and
prices -- this includes the VAX spares and logic analyzer. The two scopes
that were budgeted for a total of $12 are being dropped on the grounds
that the fancier scope requested by Cheriton can serve the needs of the
department.
Assuming that we still will spend $60k for the VAX spares and the logic
analyzer, I estimate that we can still afford to commit another $20k or so
for other equipment, leaving an additional $20k for frieght, installation,
consumables, etc. Recent requests have been received for an Ethernet
print server to go in the trailer with the TI Explorers (~$10k), a
Laserwriter to go with the Macintoshes at KSL ($5k) and spare parts for
Symbolics machines.
I propose that we make the final decision on fund allocations not later
than Friday, April 25, which is two days after the planned meeting with
Symbolics on maintenance issues, recognizing that there may or may not be
enough time after that to get approval and complete the purchases. We
could have a meeting, say at noon that day. Or if you are all tired of
this, I can flip a biased coin to decide.
Les Earnest
----------------------------
Alliant
Quantity Model & Description Unit Price Total
[QLISP]
1 FX/8 multiprocessor system (partial cost) 120,000 120,000
--------
120,000
Proteon
Quantity Model & Description Unit Price Total
[CF]
1 p4200-10 Pro-NET-10 Base Unit, 110V 7,545 7,545
1 p4203-10 Ethernet Base Unit 110V 6,800 6,800
1 p4213 Ethernet Multibus I/F & driver 2,950 2,950
1 p4299 ProNET-Linkway rack mount kit 215 215
1 p5600 IP Packet Forwarder 1,200 1,200
--------
18,710
Sun Microsystems
Quantity Model & Description Unit Price Total
MEMORY EXPANSION [CF]
10 2/50-OPT-12 2 MByte memory expansion 2,100 21,000
48 2/120-OPT-10 1 MByte memory expansion 1,050 50,400
--------
71,400
SPARE PARTS [CF]
1 3/50M-4 Workstation 5,530 5,530
FILESERVER [CF]
1 3/180S-8-R1 processor, 8 MB memory 8,130 8,130
1 3/180S-675-R1 6250 tape drive 11,830 11,830
1 370-1012 SMD disk controller 2,450 2,450
1 540-1251 VME to Multibus adapter 560 560
1 3/180S-960 76-inch full height rack 2,730 2,730
--------
25,700
FILESERVER [KSL]
1 3/180S-8-R1 processor, 8 MB memory 8,130 8,130
1 3/180S-650 60 MByte cartridge drive 2,800 2,800
1 370-1012 SMD disk controller 2,450 2,450
1 540-1251 VME to Multibus adapter 560 560
1 3/180S-960 76-inch full height rack 2,730 2,730
1 SUN SRC Unix source code 1,000 1,000
--------
17,670
WORKSTATIONS [KSL]
3 3/75M-4 Workstation 11,130 33,390
WORKSTATIONS [ZM]
2 3/75M-4 Workstation 11,130 22,260
COLOR WORKSTATIONS [ML]
2 3/160C-4 Color Workstation 24,430 48,860
FILESERVER UPGRADES [DC]
2 3/180S-8 server (8 MB) 18,130 36,260
2 540-1251 VME-to-Multibus adapter (SMD) 560 1,120
1 3/180S-960 76-inch full height rack 2,730 2,730
--------
40,110
MULTICAST AGENTS [DC]
2 2/50M-4 workstation 5,530 11,060
GATEWAY [DC]
2 3004 Sun-3 single board computer 9,485 18,970
1 540-1248 Backplane w/hardware 665 665
--------
19,635
NON-CAPITAL [DC]
1 530-1079-10 72" external command cable 200 200
1 530-1097-11 24" internal command cable 70 70
1 530-1080-10 72" external data cable 100 100
3 530-1064-11 24" internal data cable 35 105
2 340-1340-01 SMD panel 35 70
2 340-1124-01 Support brackets (set) 10 20
2 530-1079-01 24" external command cable 100 200
--------
765
--------
296,380
System Industries
Quantity Model & Description Unit Price Total
FILESERVER [CF]
2 Model 9751 Fujitsu disk drive, cable set 7,800 15,600
FILESERVER [KSL]
2 Model 9751 Fujitsu disk drive, cable set 7,800 15,600
--------
31,200
Tektronix
Quantity Model & Description Unit Price Total
[DC]
1 2467 Oscilloscope with options 09, 11 & 2A 11,629 11,629
--------
11,629
Miscellaneous
Quantity Model & Description Unit Price Total
[MISC]
1 Miscellaneous costs 101,897 101,897
--------
101,897
TOTALS
CF Computer Futility Group 136,940
QLISP McCarthy's Misconceptions 120,000
KSL Knowledge Syncophancy Lab 66,660
DC Cheriton's Disintegrated Systems 83,199
ML Linton's Multimedia Madness 48,860
ZM Manna's Moochers 22,260
MISC Miscellaneous 101,897
========
579,816
∂16-Apr-86 1507 binford@su-whitney.arpa DARPA $ spending
Received: from SU-WHITNEY.ARPA by SU-AI.ARPA with TCP; 16 Apr 86 15:07:04 PST
Received: by su-whitney.arpa with Sendmail; Wed, 16 Apr 86 15:07:48 pst
Date: Wed, 16 Apr 86 15:07:48 pst
From: Tom Binford <binford@su-whitney.ARPA>
To: LES@SU-AI.ARPA
Cc: facil@SU-AI.ARPA
In-Reply-To: Les Earnest's message of 16 Apr 86 1419 PST
Subject: DARPA $ spending
Les
I would like to put up for consideration a SUN 3/160 or
two SUN 3/50s for Cedar.
Tom
∂16-Apr-86 1624 CLOUTIER@SU-SIERRA.ARPA NAE Nomination Forms
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 16 Apr 86 16:24:50 PST
Date: Wed 16 Apr 86 16:25:54-PST
From: Mary Cloutier <CLOUTIER@SU-SIERRA.ARPA>
Subject: NAE Nomination Forms
To: JMC@SU-AI.ARPA
cc: cloutier@SU-SIERRA.ARPA
Dr. McCarty:
I requested and received additional blank NAE nomination
forms. If you would let me know how many you would like I will
have them hand-delivered to you.
Mary Cloutier
-------
∂16-Apr-86 1742 VAL Reminder: Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
A THEORY OF DIAGNOSIS FROM FIRST PRINCIPLES
Raymond Reiter
Department of Computer Science, University of Toronto
and
The Canadian Institute for Advanced Research
Thursday, April 17, 4pm
MJH 252
ABSTRACT
Suppose given a description of a system, together with an
observation of the system's behaviour which conflicts with the way
the system is meant to behave. The diagnostic problem is to determine
those components of the system which, when assumed to be functioning
abnormally, will explain the discrepancy between the observed and
correct system behaviour.
We propose a general theory for this problem.The theory requires
only that the system be described in a suitable logic. Moreover, there
are many such suitable logics, e.g., first order, temporal, dynamic,
etc. As a result, the theory accomodates diagnostic reasoning in a wide
variety of practical settings, including digital and analogue circuits,
medicine, and database updates. The theory leads to an algorithm for
computing all diagnoses, and to various results concerning principles
of measurement for discriminating between competing diagnoses. Finally,
the theory reveals close connections between diagnostic reasoning and
non-monotonic reasoning.
∂16-Apr-86 1808 lantz@su-gregorio.arpa my Sun's
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 16 Apr 86 17:55:11 PST
Received: by su-gregorio.arpa with Sendmail; Wed, 16 Apr 86 17:55:55 pst
Date: 16 Apr 1986 1755-PST (Wednesday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: facil@sail
Cc:
Subject: my Sun's
Vaughan suggests that the color Suns really need 8MB of memory and
should also have the optional graphics processor. Incremental prices
for these are:
Stanford price
4 MB memory $2800
3/160C-201 $4130
yielding a total additional cost of $13860. Is possible?
Keith
∂16-Apr-86 2229 vijay@ernie.berkeley.edu AI DISC: David Rumelhart
Received: from [128.32.0.6] by SU-AI.ARPA with TCP; 16 Apr 86 22:28:00 PST
Received: by ernie.berkeley.edu (5.45/1.12)
id AA08440; Wed, 16 Apr 86 22:29:34 PST
Date: Wed, 16 Apr 86 22:29:34 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604170629.AA08440@ernie.berkeley.edu>
To: jmc@su-ai.arpa
Subject: AI DISC: David Rumelhart
(Here is a brief statement.)
Since I suspect different discussants have different conceptions of what
Artificial Intelligence is, it will be worth while for me be as explicit as I
can about my view of the enterprise. AI, like many other fields, can be
separated into three distinct, but interrelated activities. I would
characterize these as (1) theoretical AI, (2) experimental AI and (3) applied
AI. I will sketch below my basic understanding of these three activities.
Theoretical AI
I take theoretical AI to be a branch of mathematics -- a branch devoted
to the development of mathematical systems inspired by considerations of human
cognitive behavior. The activities of the theorist generally involve the
development of formalisms which are thought to be appropriate for expressing
algorithms capable of human-like behavior. Of course, as in all mathematics,
such algorithms and formalisms are not true or false. They are merely useful
or unuseful for some particular task. In this sense AI is not a science any
more than mathematics is a science. There is no empirical domain against
which AI should be held accountable. Of course, psychologists (or AI
practitioners acting as psychologists) may well select some of the
representations and algorithms developed by theoretical AI and state
psychological theories in terms of those formalism. Such theories, like all
scientific theories are then subject to empirical verification. I will say more
about using AI formalisms for stating psychological theories below.
Experimental AI
AI differs from other branches of mathematics in that they do not
ordinarily justify their work on the basis of theorems they might
prove, rather the justification is normally done by the development of computer
programs of one kind or another which are designed to explore and evaluate their
ideas. In this sense, AI is empirical -- do the algorithms do what they were
designed to do? Are there better ones for the kinds of tasks practitioners
have in mind etc? For these reasons computer programs play a unique and
important role in informing and justifying work in AI. They, of course, also
play an important role in giving the AI practitioner a unique perspective on
cognition and therefore, indirectly, shape the kinds of work done by
theoretical AI.
Applied AI
Applied AI is the application of the techniques and formalisms
developed by theoreticians to real problems. There is a fine line here between
applied AI and software engineering in general. AI techniques may be employed
to do whatever tasks computers might be called upon to do. Expert systems are
the most common applications today, but robotics, image analysis, speech
recognition, computer aided instruction and other possibilities are areas where AI techniques can and have been developed.
Cognitive Science
If you are a Cognitive Scientist interested in stating theories
about how people function, it is natural to look to AI as a source for
hypotheses and formalisms within which to state our theories. This, I
think, is because the kinds of mathematics that were developed for the
purposes of expressing theories in the physical sciences have not seemed
to carry over very well to cognitive science. In any case, the
past fifteen years or so have seen an upswing in the use of AI concepts
in accounting for human cognitive processes. However, this does not
necessarily mean that the formalisms that AI has already developed are
the ones which, in the end, will be appropriate or useful. It does
seem clear enough that many of the ideas which dominate modern AI have
been useful in taking a first cut at describing human cognitive
processes. It has seemed to me, however, that many of the ideas from
AI have been a bit misleading and have led people to propose theories
which are probably dead ends. (Note, this does not reflect on
theoretical AI, only on its use for expressing psychological models.)
AI systems have proven to be notoriously brittle and rigid. It seems
to me that this is largely due to the fact that the dominant conception
of computation in AI is based on our understanding of how conventional
computers work. My colleagues and I have become increasingly convinced
that the symbol processing architecture that has served AI more or less
since its inception, probably will not be of as much use in the
description of human cognitive processes (especially those which take
place in fractions of a second) and that we should look elsewhere for
our inspiration for building accounts of cognition. Of course, AI's
tie to symbol processing as the basic paradigm is only historical.
Indeed AI practitioners over the years have considered alternative
computational systems and today a number of AI practitioners are actively
pursuing alternatives.
Parallel Distributed Processing
Parallel distributed processing or PDP for short is the name we give to
the particular brand of "connectionist" models that I see as providing a more
powerful and appropriate formalism within which to formulate models of
cognition. The basic idea is to attempt to give a very general
characterization of what we call "brain style" processing and then to develop
models within that framework. What we have found is that the algorithms most
appropriate for the brain-style architectures are very different than those
appropriate for symbol processing type systems. Moreover, we find that these
algorithms which are natural for PDP models seem better suited as models of
human information processing. They exhibit a number of useful properties
including: ability to rapidly solve a very general "best match" problem,
graceful degradation (rather than brittleness), ability to formulate simple,
powerful learning procedures which let them readily adapt to their environment,
ability to allow a very large number of mutual "soft constraints" to be quickly
and efficiently taken into account in finding the best match, the ability to
naturally represent similarity and automatically generalize to new situations,
naturally form prototypes and a number of other similar features which
seem important in accounting for human cognitive processing. Of
course, the development of these ideas is still in its infancy. It
will probably turn out that some of the features which we now think are
important will not, in the end, turn out to be and features of brains
which we are currently ignoring will turn out to be critical. Still,
we believe that the general strategy of looking directly at brains and
asking about the nature of brain-style processing will ultimately prove
to be a valuable strategy in our understanding of the nature of human
intelligence. Moreover, we suspect that developments in PDP style computation
will ultimately contribute to AI and, to a certain degree, liberate AI from the
historical grip of logic based and symbolic processing systems and encourage
the search for other more appropriate architectures.
D. E. Rumelhart
Institute for Cognitive Science
University of California, San Diego
∂17-Apr-86 0327 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Paper
Received: from CS.UCL.AC.UK by SU-AI.ARPA with TCP; 17 Apr 86 03:27:13 PST
Received: from aiva.edinburgh.ac.uk by 44d.Cs.Ucl.AC.UK via Janet with NIFTP
id a000149; 16 Apr 86 12:15 BST
From: Alan Bundy <bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk>
Date: Wed, 16 Apr 86 12:01:24 GMT
Message-Id: <257.8604161201@aiva.ed.ac.uk>
To: jmc <@cs.ucl.ac.uk,@cs.ucl.ac.uk:jmc@su-ai.arpa>
Subject: Paper
Cc: bundy%aiva.edinburgh.ac.uk@cs.ucl.ac.uk
John
I enjoyed your visit. Hurry back!
You mentioned a paper on circumscription and induction
schemata which I was to send you a message about to remind you to
send it. Please regard yourself as reminded.
Alan
∂17-Apr-86 1414 RA Mr. Baltes
Baltes from the College of Engineering and Mines at the University of
Arizona called (602) 621 1443. He will call back Monday.
∂17-Apr-86 1622 LES DARPA Equipment Finish (?)
To: facil@SU-AI.ARPA, "@XTRA.DIS[1,LES]"@SU-AI.ARPA
Having requested permission to buy the remaining equipment listed in my
last message, I received a phone call today from Robin Simpson (ONR Rep.)
saying that they couldn't possibly approve such purchases because the
contract was about to run out. I tried to explain that this was an
equipment contract and that there was no research to be performed but he
insisted that the auditors wouldn't accept last-minute purchases.
Whereas it is normally unacceptable to buy equipment at the end of an
R.&D. contract, this is clearly a case where the normal rules shouldn't
apply. Nevertheless, it appears that the easiest way around the problem
is to get another no-cost extension. Simpson sees no reason why it should
not be granted. Six months should be enough. I have written and
forwarded through channels a letter requesting the extension.
Assuming that the extension will be granted, I propose that we tie up the
loose ends of the DARPA equipment purchase in a meeting on Wednesday,
April 23 at 5:00pm (same day as the Symbolics meeting) in MJH 220.
∂17-Apr-86 1653 WIEDERHOLD@SU-SCORE.ARPA Industrial lecturers
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 17 Apr 86 16:50:24 PST
Date: Thu 17 Apr 86 16:51:11-PST
From: Gio Wiederhold <WIEDERHOLD@SU-SCORE.ARPA>
Subject: Industrial lecturers
To: jmc@SU-AI.ARPA
Message-ID: <12199666789.36.WIEDERHOLD@SU-SCORE.ARPA>
Are you still handling them?
I heard John Sowa talk on Conceptual structures, with many references to Peirce'
s concepts. Very inspiring to all present. He is at IBM yorktown and
if asked could probpably arrange to come.
He has a 1984 book out (addison Wesley) by that title.
I would like to vote again as well fro Reid Smith. his work at
Schlumberger is quite relevant to various projects here.
Gio
-------
∂17-Apr-86 1657 gnelson@decwrl.DEC.COM Industrial Lecturship
Received: from DECWRL.DEC.COM by SU-AI.ARPA with TCP; 17 Apr 86 16:56:56 PST
Received: from magic.ARPA (magic) by decwrl.DEC.COM (4.22.03/4.7.34)
id AA16924; Thu, 17 Apr 86 16:58:16 pst
Received: by magic.ARPA (4.22.04/4.7.34)
id AA01785; Thu, 17 Apr 86 16:56:09 pst
Date: Thu, 17 Apr 86 16:56:09 pst
From: gnelson@decwrl.DEC.COM (Greg Nelson)
Message-Id: <8604180056.AA01785@magic.ARPA>
To: jmc@sail
Subject: Industrial Lecturship
When can I expect to hear Stanford's decision? I am starting
to make plans for next year.
Greg
∂17-Apr-86 1758 LES Possible time change
To: facil@SU-AI.ARPA, "@XTRA.DIS[1,LES]"@SU-AI.ARPA
Tom R. has a conflict with the proposed late afternoon time and would
prefer to meet right after the Symbolics meeting, at 1:30pm on April 23.
Howzat for the rest of you folks?
∂18-Apr-86 0449 BONNIE%upenn.csnet@CSNET-RELAY.ARPA AI Fellowships
Received: from CSNET-RELAY.ARPA by SU-AI.ARPA with TCP; 18 Apr 86 04:42:25 PST
Received: from upenn by csnet-relay.csnet id cg15779; 18 Apr 86 7:29 EST
From: Bonnie Webber <Bonnie%upenn.csnet@CSNET-RELAY.ARPA>
Subject: AI Fellowships
To: jmc@su-ai.arpa
Date: Thu, 17 Apr 86 15:03 EST
John - I think you want to amend your statement that "Historically
conceptual scientific breakthroughs have most often come from well-prepared,
strongly motivated young men free to follow their own ideas." It insults
half our community.
B.
∂18-Apr-86 0815 RICHARDSON@SU-SCORE.ARPA AI Prospects and Visits
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 18 Apr 86 08:15:16 PST
Date: Fri 18 Apr 86 08:16:04-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Prospects and Visits
To: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
shah@SU-SIERRA.ARPA, grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA
Message-ID: <12199835161.14.RICHARDSON@SU-SCORE.ARPA>
As promised, here are the confirmed dates for the visits of our
prospective AI candidates:
April 21-22 Tom Mitchell
April 28 David Etherington
May 1-2 Yoav Shoham
May 5 John Batali
May 15 Natarajan Shankar
The schedule for Tom Mitchell's visit is already in place (thanks to
Haym Hirsh) and I will begin to work on scheduling for Etherington's
visit shortly.
-Anne
-------
∂18-Apr-86 1031 RA AI prospects
Shankar will come on May 15. Etherington will come on April 28 (instead of
Shankar). Shoham will give his talk on May 1. The interviews will take place
May 2.
∂18-Apr-86 1328 binford@su-whitney.arpa Possible time change
Received: from SU-WHITNEY.ARPA by SU-AI.ARPA with TCP; 18 Apr 86 13:28:30 PST
Received: by su-whitney.arpa with Sendmail; Fri, 18 Apr 86 13:29:10 pst
Date: Fri, 18 Apr 86 13:29:10 pst
From: Tom Binford <binford@su-whitney.ARPA>
To: LES@SU-AI.ARPA
Cc: facil@SU-AI.ARPA, @SU-AI.ARPA@su-whitney.arpa
In-Reply-To: Les Earnest's message of 17 Apr 86 1758 PST
Subject: Possible time change
That's a conflict with me Les. A meeting that I can't change.
∂18-Apr-86 1347 LES Meeting Confirmation
To: facil@SU-AI.ARPA, "@XTRA.DIS[1,LES]"@SU-AI.ARPA
It appears that, as usual, there is no generally acceptable time for the
Facilities Committee meeting. Having received a number of confirmations
for 4/23 at 1:30pm, I now confirm that time for the meeting. I regretfully
request that Tom B. and anyone else who can't make it then appoint a proxy.
∂18-Apr-86 1443 RICHARDSON@SU-SCORE.ARPA Etherington/AI Prospect
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 18 Apr 86 14:43:35 PST
Date: Fri 18 Apr 86 14:28:25-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Etherington/AI Prospect
To: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA, shah@SU-SIERRA.ARPA
Message-ID: <12199902943.23.RICHARDSON@SU-SCORE.ARPA>
Here is the schedule that I propose. Please let me know if this will work
for you.
April 28
--------
11:30 - 12:00 Grosof/Hirsh
1:15 - 1:45 Buchanan
2:00 - 2:30 McCarthy
2:30 - 3:30 Search Committee
3:30 - 4:00 Genesereth
4:00 - 5:00 talk
Thanks,
Anne
-------
∂18-Apr-86 1457 VARDI@IBM.COM
Received: from IBM.COM by SU-AI.ARPA with TCP; 18 Apr 86 14:57:05 PST
Date: 18 Apr 86 14:49:46 PST
From: VARDI@IBM.COM
To: jmc@su-ai, stan@sri-ai
Subject: SIGART
Is any of you familiar with the ACM Special Interest Group on Artificial
Intelligence? Do you know whether it is a SIG with a good reputation?
Moshe
∂18-Apr-86 1900 LES DARPA update
To: JMC, CLT, RPG
∂18-Apr-86 1551 LES Computer Buy and other issues
To: Squires@USC-ISI.ARPA
This note covers three information items.
PARALLEL PROCESSORS. As you know, we are planning to buy the Alliant and
Encore machines under the DARPA equipment contract (MDA903-83-C-0188). We
have just hit a snag that apparently will require another approval from
your office.
A couple of weeks ago we sent a request for permission to buy the
equipment to the local contract monitor (Robin Simpson of ONR). Yesterday
I heard from him that he can't approve an equipment buy so near the end of
the contract (May 31). I tried to explain that this was an equipment
contract and that there was no research to be performed but he insisted
that the auditors wouldn't accept last-minute purchases.
Whereas it is normally unacceptable to buy equipment at the end of an
R.&D. contract, this is clearly a case where the normal rules shouldn't
apply. Nevertheless, it appears that the easiest way around the problem
is to get another no-cost extension to the contract. Simpson sees no
reason why it should not be granted. Six months should be enough. I have
written a letter to him requesting the extension. I expect that it will
percolate to you eventually.
AI RESEARCH. John McCarthy's AI research contract (N00039-82-C-0250) is
scheduled to run out next January, so we ought to get a new proposal to
you soon. I am curious whether the follow-on work, if accepted, is likely
to be supported under an extension to the existing contract or whether it
would more likely be considered as an additional task under the blanket
research contract (N00039-84-C-0211). This question obviously doesn't
have to be resolved immediately.
OPERATING SYSTEM RESEARCH. We have some ideas for a new kind of human
interface to oerating systems that we call the "Editor-based Operating
System." We would like to do an experimental implementation of this idea
using Unix as a substrate, but recognize that a full implementation
will require more fundamental restructuring. We would like to bounce an
outline of this scheme off you. (When Saul Amarel came through earlier
this week, I asked who in IPTO should review such a proposal. He pointed
at you). Unless you beg off, I will send you something next week.
Cheers,
Les
∂18-Apr-86 1851 squires@ipto.ARPA Re: Computer Buy and other issues
Received: from IPTO.ARPA by SU-AI.ARPA with TCP; 18 Apr 86 18:51:01 PST
Received: by ipto.ARPA (4.12/4.7)
id AA18257; Fri, 18 Apr 86 21:51:57 est
Date: Fri 18 Apr 86 21:51:51-EST
From: Stephen Squires <SQUIRES@IPTO.ARPA>
Subject: Re: Computer Buy and other issues
To: LES@SU-AI.ARPA
Cc: SQUIRES@IPTO.ARPA
Message-Id: <VAX-MM(187)+TOPSLIB(118) 18-Apr-86 21:51:51.IPTO.ARPA>
In-Reply-To: Message from "Les Earnest <LES@SU-AI.ARPA>" of 18 Apr 86 1551 PST
EQUIPMENT
The local ONR office is taking a very narrow view. To keep them happy
ask for the no cost extension and ask them for an estimate when they
while completed the action. If you believe it will help you can have
them contact me and I will have them talk to our program managment
office that will confirm the broader contract context in which this
work is going on. I do not want your purchase action to be delayed
unnecessarily. Since the initial part of QLISP will be a task on
the tasking contract it should be in place within about two weeks
after the ARPA Order is signed. The ARPA Order has been prepared
and should be ready in about a week. Please advise me about the
ordering, delivery, and need timing so I can judge the extent of
the action that needs to be taken to fix this problem.
McCARTHY Projects
I have no objections to having the next contract be an extension
of the existing contract. I want to insure that there is a tasking
type contract in place with sufficient duration, technical, and
financial scope to insure that we can act promply as the budget
situation improves. Since tasking contracts seem to take extra
time to establish and to extend, this process needs to be started
as quickly as possible. The only reason to have a new contract in
this case is if there is a reason to change agents.
Editor OS
I will be happy to learn about your ideas in this area. There are
some interesting related issues that we can discuss. Do you know
of Bob Balzer's work in this area?
QLISP
I have described the general approach that the QLISP project will
be taking to several people within the office including the rationale
for using the Alliant. The approach is being very well received
because of the emphasis that it places on the quality of the
implementation on the ability gracefully migrate from workstations
to parallel servers. I expect that several AI groups will be very
interested in using this system sooner rather than later. Therefore,
I have some timing questions and suggestions on related issues:
1. It would be very desirable to have an early capability that is
simply Common Lisp on the workstation (Sun) with a message passing
package from the Sun to the server (Alliant) and within the server.
This would provide a way to have the effect of multiple Common Lisp
systems with a high performance interconnect through message
passing with shared memory. While this is not a research issue, it
would be a very useful tool. Gabriel told me some time ago that
this is a natural step along the way. What is your estimate of
when this could be available during the project? I beleive that
this would accelerate the transtion of various AI projects to
initial parallel systems which is very important to the community.
2. Given that the Sun workstations use 4.x BSD, and the fact that
this is also being used on the Alliants at Aargon, and the fact
the the MACH OS nucleus from CMU which supports 4.x BSD is
designed to provide the kind of transparency that is needed...
It seems that at some point in the project it would be useful
to transition to this variant. Future results of the DARPA
distributed systems community will be focusing on MACH and I would
like to begin to introduce these new capabilities into the AI
community through these kinds of projects.
-------
∂18-Apr-86 1908 vijay@ernie.berkeley.edu AI DISC: NOTE
Received: from [128.32.0.6] by SU-AI.ARPA with TCP; 18 Apr 86 19:07:51 PST
Received: by ernie.berkeley.edu (5.45/1.12)
id AA24647; Fri, 18 Apr 86 19:09:06 PST
Date: Fri, 18 Apr 86 19:09:06 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604190309.AA24647@ernie.berkeley.edu>
To: jmc@su-ai.arpa
Subject: AI DISC: NOTE
Just a reminder that we'd like you're comments about
Group 1 and the Middle, as soon as possible - as the target
date is today!
-VR RA JS DB
∂20-Apr-86 0000 JMC
Bookstore charge slips
∂20-Apr-86 1951 RPG Invitation
We were invited to submit the Qlisp paper, in a shorter form, to
the book collectionn, ``Parallel Computing and Computers for
Artificial Intelligence,'' edited by Janusz Kowalik of Boeing,
and published by Springer. The deadline for final papers is
Halloween. Are you interested? I can do most of the work.
-rpg-
∂21-Apr-86 0825 RICHARDSON@SU-SCORE.ARPA AI Search Committee
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 21 Apr 86 08:23:33 PST
Date: Mon 21 Apr 86 08:24:19-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Search Committee
To: jmc@SU-AI.ARPA
cc: ra@SU-AI.ARPA
Message-ID: <12200623092.14.RICHARDSON@SU-SCORE.ARPA>
Haym Hirsh has informed me that you have confirmed your appointment for
today with Tom Mitchell. Will you also be able to attend an AI search
committee meeting tomorrow (4/22) from 1:30 - 2:30?
-Anne
-------
∂21-Apr-86 1020 guibas@decwrl.DEC.COM re: Miklos Simonovits
Received: from DECWRL.DEC.COM by SU-AI.ARPA with TCP; 21 Apr 86 10:19:36 PST
Received: from magic.ARPA (magic) by decwrl.DEC.COM (4.22.03/4.7.34)
id AA18479; Mon, 21 Apr 86 10:21:09 pst
Received: by magic.ARPA (4.22.04/4.7.34)
id AA00800; Mon, 21 Apr 86 10:19:00 pst
From: guibas@decwrl.DEC.COM (Leo Guibas)
Message-Id: <8604211819.AA00800@magic.ARPA>
Date: 21 Apr 1986 1018-PST (Monday)
To: John McCarthy <JMC@SU-AI.ARPA>
Cc: guibas@decwrl.DEC.COM
Subject: re: Miklos Simonovits
In-Reply-To: Your message of 20 Apr 86 2252 PST.
<8604210654.AA10132@decwrl.DEC.COM>
good. then i plan to tell miklos that he can be here for one quarter
next year. i'd personally prefer spring, but winter may also be ok.
who deals with getting a formal invitation letter sent, deciding on salary,
etc. ?
thanks, l.
∂21-Apr-86 1118 gnelson@decwrl.DEC.COM re: Industrial Lecturship
Received: from DECWRL.DEC.COM by SU-AI.ARPA with TCP; 21 Apr 86 11:14:17 PST
Received: from magic.ARPA (magic) by decwrl.DEC.COM (4.22.03/4.7.34)
id AA18714; Mon, 21 Apr 86 10:36:25 pst
Received: by magic.ARPA (4.22.04/4.7.34)
id AA01189; Mon, 21 Apr 86 10:34:02 pst
From: gnelson@decwrl.DEC.COM (Greg Nelson)
Message-Id: <8604211834.AA01189@magic.ARPA>
Date: 21 Apr 1986 1033-PST (Monday)
To: John McCarthy <JMC@SU-AI.ARPA>
Cc: berg@SU-SCORE.ARPA, gnelson@decwrl.DEC.COM
Subject: re: Industrial Lecturship
In-Reply-To: Your message of 20 Apr 86 2310 PST.
<8604210713.AA10376@decwrl.DEC.COM>
Kathryn Berg asked me which quarter I preferred, and I said the fall.
I did not get the impression from her that everything (or even much of
anything) was settled.
If you have scheduled me for the fall, I'd like to know more about what
to expect. A ballpark estimate of the number of students would be
helpful. Will there be undergraduates? Can I hire a TA, and will the
department pay for photocopying lecture notes? Will I have office space
for holding office hours?
Greg
∂21-Apr-86 1338 RICHARDSON@SU-SCORE.ARPA Yoav Shoham
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 21 Apr 86 13:36:51 PST
Date: Mon 21 Apr 86 13:28:28-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Yoav Shoham
To: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
shah@SU-SIERRA.ARPA, grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA
Message-ID: <12200678462.29.RICHARDSON@SU-SCORE.ARPA>
Following is a proposed schedule for Yoav Shoham's visit to our department.
Please let me know if this will work for you.
Talk at 4:00 on May 1
May 2:
11:30 - 12:00 Hirsh/Grosof
1:15 - 1:45 Buchanan
2:00 - 2:30 McCarthy
3:30 - 4:30 Search Committee
4:30 - 5:00 Genesereth
-Anne
-------
∂21-Apr-86 1540 FEIGENBAUM@SUMEX-AIM.ARPA [Ellie Engelmore <EEngELmore@SUMEX-AIM.ARPA>: SJ Mercury/Book Review request]
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 21 Apr 86 15:39:20 PST
Date: Mon 21 Apr 86 15:41:05-PST
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: [Ellie Engelmore <EEngELmore@SUMEX-AIM.ARPA>: SJ Mercury/Book Review request]
To: jmc@SU-AI.ARPA
cc: eengelmore@SUMEX-AIM.ARPA
Message-ID: <12200702603.30.FEIGENBAUM@SUMEX-AIM.ARPA>
John, as an old Rozak antagonist, do you want to do a review of Rozak's
book? I have not seen the book. It might be a chance to put in print an
answer to this type of thinking (Dreyfus and Rozak, and Weizenbaum though
he does not have a new book).
Ed
---------------
Mail-From: EENGELMORE created at 21-Apr-86 12:04:05
Date: Mon 21 Apr 86 12:04:04-PST
From: Ellie Engelmore <EEngELmore@SUMEX-AIM.ARPA>
Subject: SJ Mercury/Book Review request
To: Feigenbaum@SUMEX-AIM.ARPA
cc: EEngELmore@SUMEX-AIM.ARPA
Stanford-Phone: (415)723-4878
Message-ID: <12200663098.66.EENGELMORE@SUMEX-AIM.ARPA>
Mark Johnson (408)920-5825
SJ Mercury-News
They are looking for a reviewer for Theodore Rozak's book:
CULT OF INFORMATION:
The Folklore of Computers and the True Art of Thinking
Peter Carey recommended you. Do you want to make another
recommendation?
-------
-------
∂21-Apr-86 1545 FEIGENBAUM@SUMEX-AIM.ARPA [Anne Richardson <RICHARDSON@SU-SCORE.ARPA>: Faculty Candidates]
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 21 Apr 86 15:43:24 PST
Date: Mon 21 Apr 86 15:45:12-PST
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: [Anne Richardson <RICHARDSON@SU-SCORE.ARPA>: Faculty Candidates]
To: nilsson@SU-SCORE.ARPA, buchanan@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA
Message-ID: <12200703355.30.FEIGENBAUM@SUMEX-AIM.ARPA>
DEAR NILS AND BRUCE (WHO I REALIZE ARE IN DAYTON RIGHT NOW) AND JOHN,
WHO ARE THESE PEOPLE? DO YOU KNOW THEM? IS THE FIELD OF AI SO BIG THAT I AM
UNAWARE OF WHO IS COMING UP IN IT?.........PUZZLED, ED
---------------
Return-Path: <RICHARDSON@SU-SCORE.ARPA>
Received: from SU-SCORE.ARPA by SUMEX-AIM.ARPA with TCP; Mon 21 Apr 86 13:30:18-PST
Date: Mon 21 Apr 86 11:16:48-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Faculty Candidates
To: faculty@SU-SCORE.ARPA
Message-ID: <12200654492.14.RICHARDSON@SU-SCORE.ARPA>
As you are aware, we have a number of searches going on within the department
for faculty candidates. The following AI candidates will be visiting us on
the indicated dates. Should you have a special interest in meeting with them,
please let me know.
David Etherington on April 28
Yoav Shoham on May 2 (talk on May 1)
John Batali on May 5
Natarajan Shankar on May 15
-Anne
-------
-------
∂21-Apr-86 1606 FEIGENBAUM@SUMEX-AIM.ARPA re: [Ellie Engelmore <EEngELmore@SUMEX-AIM.ARPA>: SJ Mercury/Book Review request]
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 21 Apr 86 16:02:57 PST
Date: Mon 21 Apr 86 16:04:38-PST
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: re: [Ellie Engelmore <EEngELmore@SUMEX-AIM.ARPA>: SJ Mercury/Book Review request]
To: JMC@SU-AI.ARPA
cc: eengelmore@SUMEX-AIM.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Mon 21 Apr 86 15:51:00-PST
Message-ID: <12200706892.30.FEIGENBAUM@SUMEX-AIM.ARPA>
Thanks, John. Ellie will close the loop with the SJ Mercury people re your
willingness to do this.....Ed
-------
∂21-Apr-86 2248 HST 30 years conference
what is the state of your thinking concerning the 30 years LISP conference?
do you like the idea of giving talks to LISP related subjects?
do you prefer to give talks on historic events?
did i select the right people?
∂22-Apr-86 0906 WINOGRAD@SU-CSLI.ARPA PhD Minor requirement
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 22 Apr 86 09:05:58 PST
Date: Tue 22 Apr 86 09:05:32-PST
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: PhD Minor requirement
To: phdcom@SU-AI.ARPA
cc: o.oakley@LOTS-B, WINOGRAD@SU-CSLI.ARPA
The following was proposed to me by some students from other departments
interested in the CS minor. The old version of the minor required a
Masters pass on the comp. Since we no longer have our own masters
students taking the comp, this is not appropriate, since it would
lead to extremely bimodal distribution on the exam.
Let me know if you approve the following, and also how things are
going on getting the comp syllabus together (our last major piece
of work). We should meet on it and put it out to the faculty within
the next couple of weeks. --t
----------
Date: Tue 22 Apr 86 02:00:30-PST
From: celia oakley <O.OAKLEY@LOTS-B.STANFORD.EDU>
Subject: the cs phd minor proposal
To: winograd@SU-SCORE.ARPA
Message-ID: <12200815367.113.O.OAKLEY@LOTS-B.STANFORD.EDU>
PROPOSED COMPUTER SCIENCE PHD MINOR REQUIREMENT
prepared by Stan Schneider and Celia Oakley
April 21, 1986
1.0 Introduction
To make the Computer Science PhD examination more comprehensive, many
changes have been proposed. Unfortunately, these proposed modifications
do not specify requirements for the PhD minor. The purpose of this
document is to propose a new set of requirements for the PhD Minor in
Computer Science.
1.1 Report Structure
Section 2.0 presents a summary of the PhD minor requirements for other
relevant departments. Section 3.0 discusses the relative merits of
various types of requirements, and presents proposals for the CS PhD
minor.
1.2 Summary of Recommendations
To alleviate the difficulties imposed by mixing PhD and minor students
taking the comprehensive examinations, we propose a coursework based
minor requirement. The recommended requirement of 15 units of graduate
work is similar to the requirements of other related departments.
Breadth, depth, and programming experience are provided by requiring
specific course combinations.
We propose the following requirement:
"15 units of CS coursework, including at least three of the
Master's core courses to provide breadth and one course numbered
above 300 to provide depth. One of the courses taken must include
a significant programming project to demonstrate programming
proficiency. A GPA of 3.0 or better must be maintained."
2.0 Summary of other departmental requirements
(Taken from "Courses and Degrees 1985-1986")
2.1 Electrical Engineering
a) MS depth requirement.
3 courses in a single area, selected from 1 of 27 areas.
b) 15 total units in EE -- must be approved by the PhD committee.
c) A GPA of at least 3.0 is required.
2.2 Mechanical Engineering
a) Student must obtain a minor advisor.
b) 15 units of approved graduate-level (200+) courses
or 9 units of graduate courses, and pass ME qualifying oral
exam in 2 of 8 areas.
c) The above courses must satisfy the depth or breadth
requirement of the ME master's degree.
2.3 Aeronautics and Astronautics
a) Student must obtain a minor advisor.
b) 15 units of approved graduate-level courses in AA.
2.4 Mathematics
a) Student must first demonstrate that several undergraduate-
level requirements have been met, either at previous
universities, or by taking specified 100 level courses.
b) 18 units of approved 200 level courses in Mathematics.
2.5 Economics
a) Student must pass the written PhD qualifying examinations in
3 of 5 fields of economics.
2.6 Physics
a) The student must complete three courses, comprising 9 units.
These are chosen from one of the three specified sets.
2.7 Psychology
a) The student is expected to have completed the equivalent
of an A.B. in Psychology.
b) 15 units must be at Stanford, including 2 courses above the
200 level.
2.8 Operations Research
a) 15 units of 200 or higher level courses.
b) An average grade of 3.0 or higher.
3.0 Proposal
The present requirement of satisfactory performance on the PhD
Comprehensive exams has the disadvantage of combining PhD candidates
with minor candidates. The exam must thus be written for two different
types of student, with differing backgrounds. Also, PhD candidate
performance evaluation is skewed by the number and quality of minor
candidates taking the exam. A coursework based minor requirement would
alleviate these difficulties, while still insuring well-qualified students.
Since the CS field is unusually broad, many diverse requirements are
indicated. Sufficient breadth and depth, along with significant
programming experience should be incorporated. This should be achieved
in a program of difficulty comparable to the above departments.
Consonant with this, we propose a selection of courses from the CS
Master's core requirements to provide breadth. An additional 300 level
course should be required to promote depth. Programming proficiency
could be demonstrated by requiring one of a selection of programming
intensive courses.
We propose the following requirement:
"15 units of CS coursework, including at least three of the
Master's core courses to provide breadth and one course numbered
above 300 to provide depth. One of the courses taken must include
a significant programming project to demonstrate programming
proficiency. A GPA of 3.0 or better must be maintained."
A possible list of programming intensive courses:
CS 223B, CS 242(?), CS 243, CS 245, CS 246, CS 248, CS 306(?),
CS 327A(?), CS 340, CS 342, CS 343, CS 346, CS 346, CS 368(?),
CS 446, EE 288(?) CS 108B(?).
(An example list, we are not familiar with all these.)
Possible additional requirements/modifications:
- Approved advanced courses may be substituted for the core requirement
above. (eg CS 346 for CS 246 for sufficiently prepared students.)
- Four core courses could be required. This has the disadvantage of
discouraging depth for students with specific interests.
- All 15 units must be numbered 200 or above.
- Two courses from one of the five CS Master's specialty areas must be
included.
- Minor programs must be approved by a CS minor faculty advisor.
-------
∂22-Apr-86 1011 KARP@SUMEX-AIM.ARPA Re: PhD Minor requirement
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 22 Apr 86 10:02:33 PST
Date: Tue 22 Apr 86 10:03:38-PST
From: Peter Karp <KARP@SUMEX-AIM.ARPA>
Subject: Re: PhD Minor requirement
To: WINOGRAD@SU-CSLI.ARPA
cc: phdcom@SU-AI.ARPA, o.oakley%LOTS-B@SUMEX-AIM.ARPA
In-Reply-To: Message from "Terry Winograd <WINOGRAD@SU-CSLI.ARPA>" of Tue 22 Apr 86 09:26:04-PST
Message-ID: <12200903316.53.KARP@SUMEX-AIM.ARPA>
I would certainly think all 15 units should be numbered 200 or above.
Peter
-------
∂22-Apr-86 1130 nttlab!NTT-20!Goto@su-shasta.arpa Common Business Communication Language
Received: from SU-SHASTA.ARPA by SU-AI.ARPA with TCP; 22 Apr 86 10:51:39 PST
Received: by su-shasta.arpa with TCP; Tue, 22 Apr 86 10:51:21 pst
Received: by ntt.junet (4.12/4.7JC-7) CHAOS with CHAOS-MAIL
id AA15590; Tue, 22 Apr 86 14:55:58 jst
Message-Id: <8604220555.AA15590@ntt.junet>
Date: 22 Apr 1986 1453
From: Shigeki Goto <nttlab!NTT-20!Goto@su-shasta.arpa>
Subject: Common Business Communication Language
To: JMC@SU-AI.ARPA
Cc: kawasima@nttyrl.ntt.junet, nttlab!NTT-20!goto@su-shasta.arpa
John,
One of my colleagues at NTT Laboratories got interested in your work
on the Common Business Language. Could you send your paper to him?
Thanks in advance.
-- Shigeki, sg@sail = "nttlab!goto"@Shasta
===== Forwarded message =====
From: Nobuo Kawashima
NTT Commun's and Inf. Proc. Labs.
Subject: Common Business Communication Language
Dear Prof. McCarthy:
I am writing you to ask a favor. When I attended at the International
Symposium on Prospects and problems of Interpreting Telephony, which
was held on Apr. 12, 1986 in Tokyo, Dr. Philip. R. Cohen of SRI
International mentioned your Common Business Communication Language
as a suggestion for the research on interpreting telephony.
I would appreciate it if you could send me a copy of your papers on
Common Business Communication Language. Following is the address of
my laboratory:
Nobuo Kawashima
Commun. Proc. Programs Sect.
NTT Commun's and Inf. Proc. Labs.
1-2356, Take
Yokosuka, Kanagawa
238-03 JAPAN
Thank you in advance.
---------
-------
∂22-Apr-86 1249 LES re: The next contest . . .
[In reply to message rcvd 20-Apr-86 22:51-PT.]
Not bad . . . in fact, it could be shortened to "Gango4," but what if
we get more processors?
Nonquantitative names in a similar vein would include "Gangbuster,"
"Gangway" or "Gangbang," but I guess that is a bit too crude.
How about "MOB?"
∂22-Apr-86 1306 VAL Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
There will be no seminar on April 24. Please notice that next week we meet at a
different place.
CHRONOLOGICAL IGNORANCE:
time, knowledge, nonmonotonicity and causation
Yoav Shoham
Yale University
Thursday, May 1, 4pm
Room 380X, Mathematics Building
We are concerned with the problem of reasoning about change within a formal system.
We identify two problems that arise from practical considerations of efficiency
and naturalness of expression: the persistence problem (otherwise known as the
frame problem, and a new, but no less evil, initiation problem. In this talk we
concentrate on the latter one.
We propose a new logic that allows efficient and natural reasoning about change
and which avoids the initiation problem. The logic, called the logic of
chronological ignorance, is a fusion of recent ideas on temporal logic,
modal logic of knowledge, and nonmonotonic logic.
We identify a special class of theories, called causal theories, and show these
have elegant model-theoretic properties which make reasoning about causal theories
very easy.
Finally, we contrast our logic with previous work on nonmonotonic logics in
computer science, and discuss its connection to the philosophical literature
on causation.
∂22-Apr-86 1454 pratt@su-navajo.arpa PhD Minor requirement
Received: from SU-NAVAJO.ARPA by SU-AI.ARPA with TCP; 22 Apr 86 14:38:40 PST
Received: by su-navajo.arpa with Sendmail; Tue, 22 Apr 86 14:39:16 pst
Date: Tue, 22 Apr 86 14:39:16 pst
From: Vaughan Pratt <pratt@su-navajo.arpa>
Subject: PhD Minor requirement
To: phdcom@su-ai.ARPA
I like the idea of not having to deal with Ph.D. minors in the
comprehensive exam.
-v
∂22-Apr-86 1607 ME net adr
∂22-Apr-86 1605 JMC
I have mislaid what you told me about what my ultimate net address will be.
ME - It will be "JMC@SAIL.STANFORD.EDU".
∂22-Apr-86 1616 ME Lathrop
To: CLT
CC: JMC
The IMPRINT job (the spooler) gives up after failing to communicate with
the printer for some some length of time (a minute?). Indeed, this
usually means the printer is off or out of paper (the Imagen won't talk at
that point). Basically, what this means is that you should always turn
the printer on before you try to print something. There's basically
nothing wrong, the messages just let you know that the printer isn't
ready.
sometimes it claims to have finished when only part
of the pages have been printed and it prints a title
page with the error message
Error in impress:
Page 1: 159 undefined glyphs on page
(or something similar)
This shouldn't happen in general. If it ever happens repeatably, you
should save a copy of the file you're printing. I suppose it might happen
if you're trying to use characters that don't exist in some font you're
using, but in that case, it should be repeatable. If it isn't repeatable,
I have to suspect some problem in the printer, since the protocol for the
TTY line to the printer provides for checksums, so there should never be
any uncorrected transmission errors.
∂22-Apr-86 2028 GENESERETH@SUMEX-AIM.ARPA Re: Yoav Shoham
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 22 Apr 86 20:27:54 PST
Date: Tue 22 Apr 86 20:20:34-PST
From: Michael Genesereth <GENESERETH@SUMEX-AIM.ARPA>
Subject: Re: Yoav Shoham
To: RICHARDSON@SU-SCORE.ARPA
cc: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, shah@SU-SIERRA.ARPA,
grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA
In-Reply-To: <12200678462.29.RICHARDSON@SU-SCORE.ARPA>
Message-ID: <12201015626.12.GENESERETH@SUMEX-AIM.ARPA>
Anne, Thats fine with me. I can see him at 4:30, only I won't be able
to make the search committtee inquisition session.
mrg
-------
∂22-Apr-86 2259 CLT books to by re ebos
Peterson, J. L. and Silbershatz, A.[1985]
Operating System Concepts
Addison-Wesley
Kerrnighan, B.W. and Pike, R.[1984]
The UNIX programming environment
Prentice-Hall
[best according to Peterson and Silberschatz OSC]
∂23-Apr-86 0803 RICHARDSON@SU-SCORE.ARPA AI Prospect David Etherington
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 23 Apr 86 08:02:53 PST
Date: Wed 23 Apr 86 07:41:05-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Prospect David Etherington
To: jmc@SU-AI.ARPA
cc: ra@SU-AI.ARPA
Message-ID: <12201139512.10.RICHARDSON@SU-SCORE.ARPA>
On April 28, will you be available to meet with Etherington from 2 - 2:30
and with the search committee from 2:30 - 3:30?
-Anne
-------
∂23-Apr-86 0803 RICHARDSON@SU-SCORE.ARPA AI Prospective/Shoham
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 23 Apr 86 08:03:22 PST
Date: Wed 23 Apr 86 07:47:06-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Prospective/Shoham
To: jmc@SU-AI.ARPA
Message-ID: <12201140607.10.RICHARDSON@SU-SCORE.ARPA>
Will you be able to meet with Shoham on May 2 from 2 - 2:30 and with the
search committee from 3:30 - 4:30?
-Anne
-------
∂23-Apr-86 0900 JMC
323-8161 x224, Marian about audiologist appt.
∂23-Apr-86 0912 CLT calendar item
thu 1-may 9:00 okner
∂23-Apr-86 1123 SJM Kugel's address
c/o Javelin [where he works]
1 Kendall Square
Cambridge, Mass. 02142
∂23-Apr-86 1209 CLT japan
JAPAN[1,CLT] has two letters relating to the collaboration.
The first is a request for a no cost extension and the
second is an answer to the request for information.
If they look ok to you I will as Rutie to turn them into
letters for you to sign.
Do you have any papers that you want mentioned?
∂23-Apr-86 2053 RESTIVO@SU-SCORE.ARPA Review of Sophist Dreyfus
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 23 Apr 86 20:41:55 PST
Date: Wed 23 Apr 86 20:42:38-PST
From: Chuck Restivo <Restivo@SU-SCORE.ARPA>
Subject: Review of Sophist Dreyfus
To: JMC@SU-AI.ARPA
Message-ID: <12201281788.36.RESTIVO@SU-SCORE.ARPA>
[cwr]
Professor McCarthy,
I am interested in completing a review of Hubert Dreyfus' new
book, and completed the following paper on his first book as
an exercise.
I would appreciate your comments and in particular, whether
you think this would be worthwhile.
chuck
The Intent of Dreyfus: A Review of What Computers Can't Do
=======================
Between tools and language there is certainly some degree
of evolutionary or biological relationship: they are rooted
in similar cognitive processes; imagery and structured
planning are required for both. .....To tie the origin of
language to the onset of tool manufacture, however, seems
unreasonable when we consider the obvious advantages of
some form of language in the basic cognitive and social
developments that may have predated tool industries. It is
equally valid to suggest that tool manufacture became
possible because the previous evolution of a primitive
spoken language laid the appropriate cognitive foundations.
(Richard E. Leakey and Roger Lewin, Origins, 1977, pp.
204-5)
In {What Computers Can't Do: The Limits of Artificial} Intelligence},
(rev. ed. 1979), Hubert Dreyfus takes up the question of the
relationship between a very modern tool, the digital computer, and an
extremely elaborate cognitive system, the human mind. Dreyfus
believes that it will not be possible to create a computer system
that will match wits with humans. Dreyfus challenges those who
maintain that artificial intelligence, the project of fashioning a
computer capable of behaving intelligently, can be grasped. "The
claim is made that this sort of machine---a Turing machine--- which
expresses the essence of a digital computer can, in principle, do
anything that human beings can do---that it has, in principle, only
those limitations shared by man." (id. p. 192) Dreyfus' critique
of artificial intelligence is many-sided. He contests the
biological, physiological, epistemological and ontological
assumptions that he finds at the root of this new discipline. He
proposes an alternative conception of human intelligence that is
premised in the apprehension of intentions in a context that makes
transparent human needs and their relation to the world. By way of a
response to Dreyfus' arguments, I wish to draw attention to some of
the limitations in his conception of the computer as a rule-based
mechanism. I shall describe briefly how a system of representation
might address the problems Dreyfus has identified concerning human
intentionalities. To present this system of intentional
representation, I shall utilize Hofstadter's notions of levels and
active symbols, and use as an example certain problems in the
representation of counterfactual statements. While this will not
answer fully Dreyfus' attack on artificial intelligence----only the
construction of a truly intelligent machine would do this, and I
believe that such a feat is decades, if not centuries or aeons
away---it will show one way in which our relationship to the modern
computer can be reconceived. As humans we must take hold of our
tools and set artificial intelligence upon a more humane course.
Dreyfus' Critique of Artificial Intelligence
After reviewing some of the most significant work in different
periods of artificial intelligence--- Cognitive Simulation, Semantic
Information Processing, Microworlds, and Representation--- Dreyfus
introduces part II of his book, "Assumptions Underlying Persistent
Optimism", by setting out his own analysis of the various
presuppositions figuring into the assertion that artificial
intelligence could rival human intelligence.
Thus the assumption that man functions like a
general-purpose symbol-manipulating device amounts to
A biological assumption that on some level of
operation---usually supposed to be that of the
neurons---the brain processes information in discrete
operations by way of some biological equivalent of on/off
switches.
A psychological assumption that the mind can be viewed
as a device operating on bits of information according to
formal rules. Thus, in psychology, the computer serves as a
model of the mind as conceived of by empiricists such Hume
(with the bits atomic impressions) and idealists such as
Kant (with the program providing the rules). Both
empiricists and idealists have prepared the ground for this
model of thinking as data processing----a third -person
process in which the involvement of the "processor" plays
no essential role.
An epistemological assumption that all knowledge can be
formalized, that is, that whatever can be understood can be
expressed in terms of logical relations, more exactly in
terms of Boolean functions, the logical calculus which
governs the way the bits are related according to rules.
Finally, since all information fed into digital
computers must be in bits, the computer model of the mind
presupposes that all relevant information about the world,
everything essential to the production of intelligent
behavior, must in principle be analyzable as a set of
situation-free determinate elements. This is the
ontological assumption that what there is, is a set of
facts each logically independent of all the others. (p.
156)
Dreyfus introduces the third section "Alternatives to the Traditional
Assumptions", with this conclusion: "The psychological,
epistemological, and ontological assumptions have this in common:
they assume that man must be a device which calculates according to
rules on data which take the form of atomic facts." As he continues,
this reduction of the human intellect to rules, the upshot of his
analysis of the errors lurking within artificial intelligence, was
produced by the convergence of two powerful cultural forces:
technological advance and philosophical tradition.
Such a view is the tidal wave produced by the confluence of
two powerful streams: first, the Platonic reduction of all
reasoning to explicit rules and the world to atomic facts
to which alone such rules could be applied without the
risks of interpretation; second, the invention of the
digital computer, a general-purpose information-processing
device, which calculates according to explicit rules and
takes in data in terms of atomic elements logically
independent of one another. (p. 231)
These passages illustrate what I think is central to Dreyfus' attack
on artificial intelligence: to see virtually all the significant
research in the discipline as being at one with a unitary
metaphysical tradition extending back to Plato, whose ramifications
are dealt with in terms of biology, physiology, epistemology and
ontology, and to reject this rigidly defined tradition and its
implications when forced by its proponents to accept the proposition
that the mind could be a thing. Dreyfus mocks artificial
intelligence in much the same way that Hegel derided phrenology in
his day. But I think it is wrong to hold that by assuming a
scientific attitude towards cognition we thereby degrade it . What
needs to be defined of course is in what such a scientific attitude
consists. In Dreyfus' words I sense an emotional outcry. Winograd
has identified this type of response to work in artificial
intelligence with great accuity.
There are many reasons why one can feel uncomfortable with
the tendency to adopt the same orientation towards people
(who are the prototype for autonomous beings), and towards
machines (or organizations). It isn't that in doing so
someone is right or wrong, accurate or inaccurate, but that
they are accepting (often unwittingly) attitudes and role
relations that can be dehumanizing and destructive of the
social structure. "Extracts from 'Notes on John Searle's
"Notes on Artificial Intelligence" Terry Winograd
(unpublished p.3)
To the extent that Dreyfus' critique of artificial intelligence is
rooted in the emotional abhorrenece of thinking of people as machines
and ultimately as things, no response is possible. This is a serious
problem and everyone should take notice of it. There are too many
instances in recent history of entire political orders being founded
on the destruction of the human spirit, particularly for the benefit
of the Race or God, to ignore or slight Dreyfus' challenges to the
artificial intelligence community on this score. On the other hand,
if we look beyond our immediate emotional reactions to the
proposition that the mind might be an object of inquiry, and seek to
educate ourselves, then it seems equally true that there is a great
deal to be learned from the project of artificial intelligence. One
might hope for an uncovering of the means whereby thoughts are
transformed into words and the other ways in which the mind makes
itself known. If by making this emotional response explicit one is
able to inquire more deeply into the current problems of artificial
intelligence then I think it is also possible to reject Dreyfus'
theoretical analysis of its supposed limitations. In part I wish to
suggest that mechanism has become a much more subtle and amorphous
issue than it was in Descartes' day. I also hope to show that Dreyfus
confuses Plato with Descartes and that recent exponents of artificial
intelligence have indeed proposed approaches to cognitive science
that are compatible with the alternative metaphysical stance that
Dreyfus assumes. But at this stage it is necessary to leave his
critique of artificial intelligence and proceed to examine his own
conception of the proper relationship between the mind and the
machine.
The structure of my argument will be to show that Dreyfus is
ultimately trapped in a self-contradiction: his alternative
metaphysical conception relies upon the previous tradition that he
criticizes and also makes possible a system of intentional
representation that might meet the standards of proficiency that he
expects from a truly humane artificial intellect. I think that
Dreyfus is wrong on both of these points because he has failed to
understand the notion of levels in modelling cognition and because he
has not reflected upon his own efforts to make the bases of
intelligence explicit. I shall focus on the question of whether
intentions can be understood as active symbols, but first it is time
to examine Dreyfus' own ideas in greater detail.
Dreyfus' Proposed Alternative
To demonstrate the fundamental difference between human and machine
intelligence, Dreyfus appeals to three inter-related themes: that the
body plays a significant role in processing information; that human
beings can perceive order without having recourse to formal rules;
and that situations are uniquely defined in relationship to
previously given human needs. "These considerations are supported by
a general theory of human experience as being-already-in-a-situation
in which the facts are always already interpreted." (p. 290) Thus
Dreyfus conjures up the image of a huge gulf separating computer
power from human reason (to borrow Weizenbaum's phrase).
The first element of Dreyfus' metaphysic is the body. He notes that
the body coordinates different types of sensory information in a
unique manner.
These bodily skills enable us not only to recognize objects
in each single sense modality, but by virtue of the felt
equivalence of our exploratory skills we can see and touch
the same object. A computer to do the same thing would
have to be programmed to make a specific list of the
characteristics of a visually analyzed object and compare
that list to an explicit list of traits recorded by moving
tactical receptors over the same object. This means that
there would have to be an internal model of each object in
each sense modality, and that the recognition of an object
seen and felt must pass through the analysis of that object
in terms of common features. (p. 249)
Here Dreyfus' is apparently claiming that the body integrates
experience in a manner that a machine cannot simulate. Unless a
computer could combine information obtained through different sensory
channels, it would lack the human sensitivity that intelligence
seemingly requires. Although this is an important objection, I do
not think that it is insurmountable. It might prove difficult to
build sensing devices that would rival human eyes, ears and flesh,
but given such devices, I do not see why the information obtained
from them could not be grouped in appropriate forms. If Dreyfus is
simply saying that when we see something we coordinate this
experience with our tactile sensations, then this is not a serious
objection. Reading further, however, one discovers that there is a
more profound sense in which the body figures into reasoning. "The
important thing about skills is that, although science requires that
the skilled performance be described according to rules, these rules
need in no way be involved in producing the performance." (p. 253)
Dreyfus suggests that our immediate access to sensory experience
provides a mode for information processing that a machine could not
replicate. In mastering a skill, such as riding a bicycle, it is not
necessary for humans to recognize, make explicit, and recollect the
explicit rules whereby the skill is carried out. Dreyfus maintains
that rules only become necessary when we communicate this sensory
experience to others. But if one is simply interested in mastering a
skill, rather than learning how to teach it, one can rely entirely on
the unconscious integration of experience: we see the road, we feel
the resistance in the handlebars, we maintain a sense of balance, and
to traverse the course from here to there on the bicycle we need only
remember, in some intuitive sense, how those sights and sounds and
tactile impressions were previously integrated. This is the gist of
Dreyfus' argument.
Dreyfus' description of the role of the body in sustaining human
intelligence merges into his statements regarding our ability to
perceive order without being aware of formal rules. He notes that
human beings perceive patterns well. "A gestalt defines what counts
as the elements it organizes; a plan or a rule simply organizes
independently defined elements." (p. 245) Dreyfus argues that humans
experience an already ordered world.
Human beings are somehow already situated in such a way
that what they need in order to cope with things is
distributed around them where they need it, not packed away
like a trunk full of objects, or even carefully indexed in
a filing cabinet. This system of relations which makes it
possible to discover objects when they are needed is our
home or our world. (p. 260)
He also urges that this orderliness of our cosmos would not be
accessible to other intellects.
Martian might have to proceed in a very unfamiliar
context if he were on earth, but if he shared no human
purposes his task of sorting out the relevant from the
irrelevant, essential from the inessential, would be as
hopeless as that of the computer. (p. 265)
At this stage, Dreyfus is able to synthesize his remarks on the
orderliness of the human universe with his earlier remarks about the
integration of experience. Although human purposes order our world,
there is no reason to suppose that they could ever be reduced to
formal rules that a machine might model and thereby comprehend.
Our context-guided activity in terms of which we constantly
modify the relevance and significance of particular objects
and facts is quite regular, but the regularity need not and
cannot be completely rule governed. As in the case of
ambiguity tolerance, our activity is simply as rule
governed as is necessary for the task at hand---the task
itself, of course, being no more precise than the rules.
Wittgenstein, like Heidegger, sees the regulation of
traffic as paradigmatic:
The regulation of traffic in The streets permits and
forbids certain actions on the part of drivers and
pedestrians; but it does not attempt to guide the
totality of their movements by prescription. And it
would be senseless to talk of an 'ideal' ordering of
traffic which would do that; in the first place we
should have no idea what to imagine as this ideal. If
someone wants to make traffic regulations stricter on
some point or other, that does not mean that he wants
to approximate to such an ideal.
This contextual regularity, never completely rule governed,
but always as orderly as necessary, is so pervasive that
it is easily overlooked. Once, however, it has been
focused on as the background of problem solving , language
use, and other intelligent behavior, it no longer seems
necessary to suppose that all ordered behavior is rule
governed. The rule-model only seems inevitable if one
abstracts himself from the human situation as philosophers
have been trying to do for two thousand years, and as
computer experts must, given the context-free character of
information processing in digital machines. (p. 271)
Although I believe Dreyfus is correct in speaking of non-rule
governed orderly human behavior, he does not provide a very elaborate
account of the bases of its orderliness at this stage in his
discussion. He merely indicates that the orderliness is that which is
necessary for the "task at hand". Yet what is this task? In the
next section , Dreyfus brings us back to the theme of human needs.
This provides the context for human activity that computers allegedly
will never be able to replicate.
We have seen, however, that our concrete bodily needs
directly or indirectly give us our sense of the task at
hand, in terms of which our experience is structured as
significant or insignificant. These needs have a very
special structure, which, ...does resemble artistic
creation. When we experience a need we do not at first
know what it is we need. We must search to discover what
allays our restlessness or discomfort...... Rather, in
discovering what they need they make more specific a
general need which was there all along but was not
determinate...When a man falls in love he loves a
particular woman, but it is not that particular woman he
needed before he fell in love. However, after he is in
love, that is after he has found that this
particular relationship is gratifying, the need becomes
specific as the need for that particular woman, and the man
has made a creative discovery about himself.(pp. 276-77)
It is this notion of creation after the fact that seems central to
Dreyfus' discussion of human intelligence. Our needs provide the raw
material for interpretation of experience. They presumably supply us
with a background of standards by which we can judge the incoming
stream of sensations and make sense of them. Our needs create
certain ill-defined expectations that become sharper and more
structured as our experience is elaborated.
These considerations are supported by a general theory of
human experience as being-already-in-a-situation in which
the facts are always already interpreted. This theory also
suggests that the ultimate situation in which human beings
find themselves depends on their purposes, which are in
turn a function of their body and their needs,and that
these needs are not fixed once and for all but are
interpreted and made determinate by acculturation and thus
by changes in human self-interpretation. Thus in the last
analysis we can understand why there are not facts with
built-in significance and no fixed human forms of life which
one could ever hope to program.....Computers can only deal
with facts, but man---the source of facts---is not a fact
or set of facts, but a being who creates himself and the
world of facts in the process of living in the world. This
human world with its recognizable objects is organized by
human beings using their embodied capacities to satisfy
their embodied needs. There is no reason to supposed that
a world organized in terms of these fundamental human
capacities should be accessible by any other means. (pp.
290-291)
Since computers cannot compete with human intellects, Dreyfus
concludes that we should not think of them as autonomous creatures.
We must think in the short run of cooperation between men
and digital computers, and only in the long run of
nondigital automata which, if they were in a situation,
would exhibit the forms of "information processing"
essential in dealing with our nonformal world. (p. 304)
Dreyfus' conclusion is sensible at this stage: it would be unwise to
treat computers as autonomous beings so long as we have such a
shallow conception of what they represent and so long as they lack an
adequate conception of themselves. In this last respect they perhaps
differ from other sentient beings such as dolphins or the proverbial
Martians. While it is perhaps necessary now to regard machines as
entities subservient to our needs, this would not be the case if they
attained the marks of intelligence Dreyfus claims they never will
have. Thus we must understand computers as tools, but we should also
be aware that as they change, our conception of what a tool is will
also change. As the line between human and artificial intelligence
becomes blurred, we should be sensitive to its reciprocal effects: it
brings the danger of reducing humans to a shadow of themselves by
making it more possible to treat others as things; but it should also
hold the promise of making us more aware of our affinities with other
entities in the universe. If the project of artificial intelligence
were ever carried out to completion, then we might emerge as
pantheists: to view the mind as arising from matter would overcome
the separation between the species and its surroundings. It is now
appropriate to consider whether Dreyfus' arguments are truly so
compelling as to deny the possibility of artificial intelligence. His
is a strong claim but in its strength there are signs of weakness. To
formulate a response to Dreyfus' critique of artificial intelligence I
shall survey some of his other remarks concerning language and then
introduce some notions developed by Hofstadter. I shall endeavor to
relate the notion of active symbols that may form a basis for human
intelligence to the ideas of levels and intentions. From this I shall
suggest some pre-requisites for consciousness in computers and
sketch what I take to be the limits of objective discourse. This
will serve as a general introduction to the problem of intentional
representation that I shall address in the penultimate section.
Symbols, Intentions and Levels
...we return to the question of isomorphism between
brains...we have some intuitive sense that, although people
differ from us in important ways, they are still "the same"
as we are in some deep and important ways.... we are all
forced, by external realities, to construct certain class
symbols and triggering paths in the same way. These core
symbols are like the large cities, to which everyone can make
reference without ambiguity. .....The fact is that a large
proportion of every human's network of symbols is universal.
....We take what is common to all of us so much for granted that
it is hard to see how much we have in common with other
people. ....What we notice about another person immediately is
not the standard overlap, because that is taken for granted as
soon as we recognize the humanity of the other person;
rather we look beyond the standard overlap and generally find
some major differences, as well as some unexpected, additional
overlap. (D. Hofstadter, Godel, Escher, Bach, pp. 375-6)
Buried deep in Hofstadter's notion of active symbols, or more simply,
symbols, lies most of what I have to say in response to Dreyfus. The
notion of a symbol that is both a carrier and a creator of meaning is
somewhat alien and thus it must be confronted with a charitable
spirit. Although it is not clear how symbols are physically
realized, how they come into being, how they organize themselves, and
how they ultimately determine their own firing patterns, if we
postulate the existence of active symbols, it will allow us to
obtain a much firmer grasp on the meaning of intelligence and the
possibility of its realization in computers. To engage in this
discourse, I ask that you imagine the mind as a storehouse of symbols
linked to one another by signalling threads. When a portion of one
symbol is activated, it responds by activating some of the other
symbols to which it is linked. If the activation rates are sustained
for long periods of time, eventually a group of closely related
symbols may be chunked up into something more like a single entity.
Symbols therefore have potentially several levels of detail that can
be flushed out if one chooses to focus on a particular individual. To
carry this analogy to its conclusion, one can liken thought to a
series of moves between different symbols via their respective
activating threads. One can traverse from virtually any symbol in
the mind to any other. The firmest and most likely paths constitute
our knowledge; unlikely paths represent our beliefs; and nearly
impossible journeys instantiate our personal visions of the absurd or
illogical. One may also assume a rough correspondence between symbols
and the elements of language, but the precise relationship between
mental symbols and linguistic entities remains somewhat obscure.
Seizing upon this notion of symbols, it may be possible to show the
limitations in Dreyfus' critique of artificial intelligence. First,
however, it is necessary to return to his discussion of language.
To have a complete theory of what speakers are able to do,
one must not only have grammatical and semantic rules but
further rules which would enable a person or a machine to
recognize the context in which the rules must be applied.
Thus there must be rules for recognizing the situation, the
intentions of the speakers, and so forth. But if the
theory then requires further rules in order to explain how
these rules are applied, as the pure intellectualist
viewpoint would suggest , we are in an infinite regress.
Since we do manage to use language, this regress cannot be
a problem for human beings. If AI is to be possible, it
must also not be a problem for machines. (pp. 203-4)
Dreyfus holds up the problem of context as being the ultimate
stumbling block in natural language processing. As he notes, human
beings "bottom-out" in the regress of contexts at the level of the
body. This provides a locus for the human needs sustaining all
cognitive acts of interpretation. Dreyfus maintains that a computer
could not operate in a similar manner. When pressed to the extreme,
however, he sees this as a limitation on human self-knowledge.
There remains only one possible "solution." The computer
programmer can make up a hierarchy of contexts and general
rules for how to organize them for the computer. He does
this by appealling to his general sense of what is
generally relevant and significant for a human being. In
some situations, however, any fact may become important.
To formalize this so that the computer could exhibit human
flexibility, the programmer would have to be able to make
explicit all that he normally takes for granted in being a
human being. However, once he tries to treat his own
situation as if he were a computer looking at it from the
outside, the computer programmer is himself faced with an
infinity of meaningless facts whose relevance and
significance could only be determined in a broader context.
(p. 289)
The way to overcome the seemingly insoluble problems of context is to
imagine the possibility of creating symbols that would represent
intentions across contextual boundaries. If our symbolic apparatus
were sufficiently powerful to approximate and model what Dreyfus
referred to as the "task at hand", this information could be passed
along in the representation of atomic facts in a knowledge-base.
Another way of perceiving this point is to consider the similarities
between what Dreyfus thinks of as the "task at hand"----something
that may be regarded more generally as the human condition---and what
Hofstadter recognized as the most common elements of human
experience. Both authors postulate the existence of a certain set of
core concepts and intentions that can be transported from one mind to
another. Where they seem to differ is on the possibility of these
symbolic linkages being formalized. However, if one agrees, with
Hofstadter, that symbols can be shared between different human
beings, then there is little point in arguing that they cannot be
passed along to machines. Symbols in different humans will bottom out
in different neural structures. What human beings share, therefore,
is not intelligence on the lowest level of representation----the
neural level---but form and content at some higher level of
understanding. The question is not whether electromechanical relays,
transistors and resistance densities in silicon chips can model the
activity of neurons, but whether something corresponding to symbols
can be recreated in software.
As I have presented it, the possibility of machine intelligence
reduces to the problem of symbolic representation and this
presupposes a notion of different levels in cognition. In some crude
sense this notion of different levels corresponds to what Dreyfus
speaks of as a hierarchy of contexts. One of the reasons, perhaps,
why this approach to cognition has been ignored is that the levels
make more complex any analysis of the mind. As Dreyfus noted, any
fact may be relevant to the human understanding of some problem. This
corresponds to the idea that the hierarchy of levels is potentially
infinite, extending across the most intimate and the most remote
aspects of human experience. Intentions are needed at many different
places in the theory:
-atomic level would be in the representation of sentences
-molecular level might involve something like scripts
-higher levels would have to do with personal and group actions
-higher levels still might relate to economic development, broadly
conceived
-highest levels might have to do with physical laws
These are but a few of the levels that would have to be dealt with in
an intelligent system. If we think of the machine's intelligence
operating on symbols, then the lowest levels are those where
immediate context----relative to human purposes----is most important.
At the lowest machine levels we deal purely with conventions. At
higher levels, purposes and intentions become less clearly defined.
One must examine larger and larger areas of context to determine
meaning. In this respect, something like physical laws might stand
at the ultimate stages of context. At the present time I am unable
to offer a definitive statement of what these individual levels might
be and how they would relate to one another. Instead I would like to
present some examples of levels and discuss how they cut across one
another. [I'll also venture that the representation of statements at
each level should have the same form....cf. the identity of local and
global models in fractals.]
In computer science and logic, levels can be related directly to the
notion of symbols. Consider the following schema:
ATOM STRING MATRIX (SCREEN) N=FOLD (FILE)
Atoms are elementary symbols; strings are sets of atoms, matrices
are sets of strings (in two dimensions), and files are arbitrary
n-tuples of the earlier levels. Although this is not a standard
analysis of symbolic levels, I think it is sufficiently detailed to
approximate the different sets of contexts that a computer would have
to consider in interpreting sets of symbols. Notice how all the
lowest orders of complexity along with the very highest orders are
compressed into two levels, ATOMS and N-FOLD's respectively. This
forces one to abstract and generalize. In treating the levels
symbolically, we should recognize that they could expand at either
extreme. One can also imagine an expansion of the levels between two
other levels, as between MATRICES and N-FOLD's. Now, imagine if the
information represented by strings of atomic symbols were represented
differently, e.g. in something like frames. To a certain extent, the
same leveling structure could be worked into the representation of
data in this domain.
ATOM STRING MATRIX (SCREEN) N=FOLD (FILE)
"HERE AND NOW" SEQUENCE LIST SYSTEM UNIVERSE
In this representation, the basis set consists of all the "HERE AND
NOW"'s that the system has experienced, either directly through some
sort of sensory apparatus or indirectly via its access to other data.
Individual experiences are collated into sequences that span periods
of time and space. These are built up into what I shall refer to as
list structures (recall that they are sets of sets of elements). At
the highest level stands the SYSTEM UNIVERSE, i.e. all the symbols to
which the system has access presented in an arbitrarily high
dimension. We can also apply the notion of levels to human contexts.
ATOM STRING MATRIX (SCREEN) N=FOLD (FILE)
"HERE AND NOW" SEQUENCE LIST SYSTEM UNIVERSE
ACTION/GOAL PERSON GROUP SPECIES
If we think of human activities in "HERE AND NOW"'s as being defined
by ACTION/GOAL's, then a person can be seen as a sequence of such
ACTION/GOAL's. A person accordingly becomes a set of values in
continual realization. A set of persons would comprise a group such
as a family. Groups would then have their own goals which would in
turn have an impact upon the ACTION/GOAL's of their constituents.
Finally, the SPECIES would constitute the group of all GROUP's.
Although these are very rough sketches of the effects of levels, each
relates to a simple form.
ATOM STRING MATRIX (SCREEN) N=FOLD (FILE)
"HERE AND NOW" SEQUENCE LIST SYSTEM UNIVERSE
ACTION/GOAL PERSON GROUP SPECIES
SYMBOL SET SET OF SETS SET OF ALL SETS
This may be one of the simplest and most abstract means of treating
levels. It is important to note that the meaning of symbolic
representations at one level will be affected by the determinations
made at both the superior and the inferior levels. It is not
immediately clear how these different notions of levels relate to one
another, but they do provide the basis for a response to Dreyfus.
Rule-bound activities can be seen as affecting the relationship
between symbols on one level. As the context shifts to higher
levels, however, rules for operation at the lower level may be
over-ridden. Similarly, changes at the lower level can also affect
changes in the rules at the higher levels. Without presenting a
mathematical demonstration, I am postulating that the existence of
levels is sufficient to contradict the notion of rule-bounded
behavior which Dreyfus has imposed upon all mechanical reasoning. The
next step is to postulate the existence of control structures, which
turn out to be intentions, that can operate across these
representational levels. Intentions can be seen as creating a
non-rule bound transition between levels. The important point to gain
from this discussion of levels is that they soften the "rule-like"
character of computers, permitting them to reason in a much more
"fluid" manner, and that although humans and computers bottom out at
different places in their representations of concepts they should both
be able to represent the active symbols that appear to be the true
bases of intelligence.
As important as symbols and levels are for an understanding of
intelligence, there remains the further notion of intentions that is
needed to complete the picture I am sketching of the mind and its
imitation in artificial intelligence. One point that seems crucial is
that intentions cut across the different levels of understanding.
This can be seen in Dreyfus' own critique of artificial intelligence.
In Dreyfus' case, his intention is to render implausible the claims
of artificial intelligence. In considering the relationship between
human and machine intelligence, he has himself adopted a leveling
structure very reminiscent of computer organization. The limitations
of intelligence are expressed on four levels: first the hardware
level (the biological assumption); then the software level (the
physiological assumption); next the level of information access (the
epistemological level); and lastly the level of possible information
(the metaphysical level). Once the basic assumption of rule-governed
behavior is dismissed, Dreyfus' arguments can be dismantled at each
level of analysis. The biological assumption is not important if, as
I have argued, intelligence is realized at a higher symbolic level.
The physiological assumption can be maintained if one recognizes that
symbols operating at different levels will not function in the
limited, rule-governed mode that Dreyfus assumes. The
epistemological and metaphysical assumptions are also transformed by
an appreciation of levels and intentions: the world, and what we can
know it, does not consist entirely of facts, but of facts realized at
various levels of understanding together with intentions that cut
across these levels, supplying a context that elaborates the meaning
of individual rules. This is admittedly a rather summary dismissal of
Dreyfus' claims, but once once appreciates the power of different
levels in representing our knowledge of the world, the entire thrust
of Dreyfus' critique appears misguided.
If intentions and levels are as central to human reasoning as Dreyfus
suggests, and if they can be understood as relating to one another in
the manner I have indicated, then it should not be surprising to find
that they also appear in one of the most important tools of human
reasoning, language. I have presented intentions as operations
between different levels of context. One might say that they act
upon functions at the different levels, adjusting them to specific
contexts. We use intentions to cross these symbolic discontinuities.
The same thing occurs in language, where we try to use symbolic
entities to pass information between different people. The problem of
hermeneutics is to reconcile the different intentions of the author
and recipient (reader) of a statement with the form of the signal
(message) itself. Intentions are abstract entities that can give
meaning to words in radically different contexts. The extent to
which these intentions are shared between different speakers
(writers) defines the limits of objective discourse.
[So we try to make a grand analogy.....wetware in the brain is like
wetware in language. Maybe it's not so much a question of the
symbols being active, as of the interpreters being active. Talk here
about the different levels in a message that Doug presents: (1)
knowing that it's a message (2) breaking the code (3) finding the
speaker's ultimate intention.....example of the necessity of (3)
would be the evaluation that the CIA chief makes of an intercepted
Russian signal....is it a dummy or is it real.....another example
might be in some Zen commentaries.]
The limitations of an intentional representational system will
correspond to the limitations on the expressive power of the set of
possible intentions and the limitations on the levels of context to
which these intentions are applicable. One thing that seems
absolutely necessary for artificial intelligence to mimic the human
mind is to find some way of representing intentions relating to the
self. For a machine to have a sense of context for its actions, it
must be able to represent itself and be able to organize its actions
as intentions relating to this model of the self. The system of
levels and intentions must therefore be recursive: these systems
bottom out in self-definitions. [Query:what does this mean at the
machine level? An intention to create, or retain, symbols?]
The final point I wish to emphasize in this rather cursory treatment
of levels, symbols and intentions is that there are limits to the
objective discourse that we can imagine machines being able to
comprehend. I wish to assert also, however, that these limits are
co-extensive with our own ability to communicate with one another. In
this respect, the limits of artificial intelligence that Dreyfus
describes are simply the limits of reason, where reason is taken as
that that can be unveiled through logos. Once these limitations are
established, one can assume a different attitude towards the
machines. One can accept the possibility of a mechanistic account of
reasoning and begin to understand its peculiarities, particularly
with respect to the problems of self-reference. An example of such
an undertaking is ↑&Godel, Escher, Bach\&. The aim of these studies
should be to illuminate the points where consistency and
completeness must yield to the other characteristics embodied in
art, beauty and simplicity.
As I emphasized above, the issue is not whether a person
"really is" an autonomous agent or a deterministic
machine. In choosing to apply a mechanistic explanation
(be it from AI or elsewhere), we are making it so. As with
the use of mental terms for machines, I have ambivalent
feelings about making use of this kind of explanation. The
ultimate social and moral consequences of treating human
beings primarily as mechanisms are repugnant. But this
extreme does not follow inexorably from the use of
mechanisms as one domain of interpretation for human
action. It would be futile to argue against all uses of
the mechanistic mode of understanding for people. Not only
would it be counter to a large part of our tradition, but it
would impoverish our ability to understand, by eliminating
one potentially relevant way of thinking. What is
important is to be aware of the ways in which we make use
of different kinds of explanation, and of the effects it
can have. The trend in our modern society is to emphasize
the mechanistic domain of explanation, and it is important
to take responsibility for the ways in which our own uses
of language (particularly as public interpreters of
cognition) affect that direction. (8) (Winograd, supra, p.
8)
At the present time, the limits of discourse are largely defined by
the traditional metaphysic that Dreyfus attacked. To a large degree
this is the only manner in which we are able to communicate our
knowledge of our selves to one another. Parts of this tradition are
found in the notion of rule-bound action that Dreyfus attacks, but
other parts are found simply in the presupposition that we can make
experience intelligible to our selves and communicate it to others. In
this respect, Dreyfus, by presenting his alternative metaphysic,
joins in the tradition of rationality. By making the analysis of
intentions and human contexts part of the discussion of intelligence,
and by doing so in a manner that challenges the objective validity of
the previous tradition, Dreyfus is in point of fact extending the
limits of objective discourse. He is attempting to enrich our
collective stock of active symbols and to relate those symbols more
closely to the pragmatic tasks at hand in our human world. In all of
this he is making explicit the presuppositions for machine
intelligence. I am suggesting that Dreyfus' critique of artificial
intelligence ultimately strangles itself in a self-contradiction: by
speaking of what cannot be spoken by machines he may ultimately
induce them to speak.
These are not, however, necessarily the only limits to objective
discourse. These limits are further defined by our reservoirs of
knowledge concerning the physical universe, and our ability to relate
this knowledge to the satisfaction of human needs. A complete system
of intentional knowledge would therefore extend from the
circumstances of the individual to the conditions of the species and
would be transversed by our scientific knowledge.
Artificial intelligence is best understood as a project that is in a
continual state of realization. Obviously it has a long way to go.
What is surprising is that the study must become recursive. The
project will be to create a person, but this may entail our
discovering much more about the real world and our own subjective
constitutions.. In its boldest forms, artificial intelligence
recreates the grand project of an universal science, but to
appreciate its achievements, we must also acknowledge how far it has
to go, how unenlightened our own self-conceptions are. My question
for the machine in a Turing test would be to interpret the Delphic
oracle's remarks as they were related to Socrates' cousin. When the
machine can interpret the question----who is the wisest person-- as a
project, then it will have learned to behave intelligently. The mark
of its increasing intelligence will be its ability to interpret all
experience in a constructive manner, including such assertions as 2+
2=5. All of this meant to suggest that Dreyfus is caught in a
contradiction. His presentation of an alternative metaphysic
presupposes the possibility of a rational discourse that is much more
part of traditional metaphysics than the simple notion of rule-based
reasoning. By speaking of intentions in the way that he does, and by
calling for phenomenological investigations of human reasoning in
humane contexts, Dreyfus adumbrates the studies that will be
necessary to construct an artificial intellect. This makes possible
a system of intentional representation that I shall next discuss. I
shall present this as an alternative to traditional systems of logic
that rely upon truth-values to represent judgments. This will leave
us with two major problems: how to relate intentions to the active
mental symbols I have postulated previously, and how to give values to
the intentions spanning all levels of intelligence. In particular
one needs to show how intentions would bottom out in the machine
level.
A System of Intentional Representation
The error that underlies much of Dreyfus' thinking is his diagnosis
of the traditional objectivity of judgments. Dreyfus sees in the
history of Western metaphysics a series of attempts to render
explicit all that is or can be known.
The story of artificial intelligence might well begin
around 450 B.C. when (according to Plato) Socrates demands
of Euthyphro , a fellow Athenian who, in the name of piety,
is about to turn in his own father for murder: "I want to
know what is characteristic of piety which makes all
actions pious... that I may have it to turn to and to use
as a standard whereby to judge your actions and those of
other men." Socrates is asking Euthyphro for what modern
computer theorists would call an "effective procedure," "a
set of rules which tells us, from moment to moment,
precisely how to behave." (p. 67)
Contrary to Dreyfus, I think the dialogues illustrate not a search
for effective procedures that will produce given actions in all
imaginable contexts, but a search for the intentions that actually
guide human conduct. Socrates' strategy is to use reason to
illuminate the soul: this is the essence of his activity as a
philosophical midwife, and I am suggesting that it is the same
activity that needs to be carried out to construct an intelligent
machine.
This reality, then, that gives their truth to the objects
of knowledge and the power of knowing to the knower, you
must say is the idea of good, and you must conceive it as
being the cause of knowledge, and of truth in so far as
known. Yet as they both are, knowledge and truth, in
supposing it to be something fairer still than these you
will think rightly of it. But as for knowledge and truth,
even as in our illustration it is right to deem light and
vision sunlike, but never to think that they are the sun,
so here it is right to consider these two their
counterparts, as being like the good or boniform , but to
think that either of them is the good is not right. Still
higher honor belongs to the possession and habit of the
good.
An inconceivable beauty you speak of, he said, if it is the
source of knowledge and truth, and yet itself surpasses
them in beauty. For you surely cannot mean that it is
pleasure.
Hush, said I , but examine the similitude of it still further
in this way. (Republic, 509,a, tr. Shorey)
Socrates, in trying to explain the nature of the good, has laid bare
the source of Glaucon's interest in the subject, and he chastises him
for thinking that the good could be mere pleasure. What Socrates is
doing is fashioning an active symbol that can take hold in the minds
of others. He is able to do this by relating it to other symbols,
such as pleasure, that already have meaning for the members of his
audience. Knowledge therefore is always linked to the
pre-theoretical understanding of a person, and this in turn relates
to the person's intentions. Something similar is present in
Winograd's model of discourse.
The figure illustrates three independent sets of
structures: the text itself (in the center) ; the cognitive
structures in the speaker relevant to its production (on
the left); and the cognitive structures in the hearer
relevant to its comprehension (on the right). The model
includes long-term structures that existed prior to the
specific discourse event. In addition, since this is a
snapshot taken at some point in the midst of discourse,
there are a number of current "instantiated schemas" that
represent the knowledge created and acquired during the
discourse. One is the language user's analysis of the
events of the conversation. This analysis includes multiple
levels; for example, it may include an analysis of the
specific semantic content, the words used, the tone and so
on. Both speaker and listener have models of the events of
the conversation. In addition, each participant has a
model of the other person, including his or her knowledge,
current goals, and processing state. This component
captures the cooperative aspect of language that is such an
important feature of communication. The speaker's model of
the listener makes it possible to tailor the communication;
whereas the listener's model of the speaker makes it
possible to interpret the communication in a
context-dependent way. These models are developed during
the conversation, and are used for designing and
comprehending utterances. Some of them, in turn may be
stored to serve as knowledge for use in producing and
comprehending future discourses. (T. Winograd, A Framework
for Understanding Discourse, p. 68)
Rather than representing knowledge in a series of objectively true
statements, perhaps knowledge should be related to individual
statements of intention. This is the shift from an objective to a
purposeful or relativistic theory of knowledge. The significance of
such a change can be best understood by considering problems in the
representation of counterfactual statements. After discussing some
of the options available to one choosing a representation of
counterfactuals, I shall conclude this review of Dreyfus' work by
summarizing the different theses at issue and by pointing out some of
the problems that still need to be addressed.
!Wanting and Wishing
Fate sometimes intercedes to take vengeance upon the unsuspecting by
offering what one has asked for but does not necessarily want.
Consider, for example, this excerpt from, John Brunner's The Traveler
in Black (1971).
Dame Rosa...trusted the promise the one in black had made, and
looked forward with impatience to the earliest moment she could
closet herself with her books and apparatus and rehearse with
improvements the most relevant of her formulae. Her family
had in the past been counted among the most lascivious of Ys, and
excessive indulgence by its womenfolk in the pleasures of the
bed had often threatened to overpopulate the resources of their
not inconsiderable estate. Accordingly there was a cellar where
surplus children had for generations been discretely disposed of,
not by crude and brutal means but by consigning the problem of
their nourishment to the fates. She entered this cellar by a bronze
door which she locked with a heavy key, and passed between rows of
wooden stalls in each of which a set of rat-gnawed bones lay on
foul straw, gyves about one ankle. She had chosen this place after
much thought. Surely, she reasoned, the point of departure to
eternity of so many spirits must be imbued with a peculiar
potency.... She had, the books stated, to keep her eyes shut
until she had completed the recital of a cantrip that lasted
eight whole pages in miniscule script. There were two pages to go
when she heard the first rustling and clicketing behind her. There
was one page to go when the first touch came on her fleshy thigh.
Desperately wanting to know what marvels her work had brought
about, she raced through the last page, and on the concluding word
came the first bite. Thirty starving children mad with hunger,
their teeth as keen as any rat's, left gnaw-marks on her bones too.
The rather grisly story of Dame Rosa, who like the other citizens of
Ys is granted the power to call her ancestors back from the grave, is
an extreme example of a rather commonplace occurrence. This story has
much in common with the Monkey's Paw: a person states a desire for
something. That desire is satisfied, through fate or some other
agency. The desire thus satisfied, however, yields a state of
affairs that is not what the person wanted. With suitable artistic
talent, this brief schema could be turned into a tragedy, romance or
comedy, depending on the relationship between the anticipated and
actual outcomes of the act of wanting or wishing. The problem I am
am interested in examining is how statements of this general type
should be represented and what significance, if any, these representa-
tional issues may have for a more general theory of judgment.
Very simple examples of statements expressing a wish or a want
include the ordinary utterances of material desire. "I wish I had a
Mercedes." "I want to go to Mexico." "If only I had my own
computer." These token expressions signify the set of statements
indicating some form of wanting or wishing. The simplest statements
involve only a subject, verb and object, as in the statement, "I want
money!" One might be tempted to represent the meaning of such
statements using a simple, first-order relation between the subject,
"I" and the object "money". One could represent the statement in
this manner:
(1) R(I, money)
Here "R" signifies the relationship of wanting or wishing holding
between the referents of the terms "I" and "money." This formulation
however, must sometimes give way, and for the moment, I shall ignore
it to consider another variation. This second formalism might be
applied more straightforwardly to statements of the form
"I want a Porsche." One can imagine the statement "I want money!"
being spoken to a bank teller by an impatient gunman, but "I want a
Porsche" is a casual remark that might be made to a close friend over
lunch. as the context changes, the statement slides off from an
ostensive expression of immediate desire to a more abstract
formulation of counterfactual intention. The object of desire is
less of a thing and more of a state of affairs, i.e. of having a
Porsche. At an even more abstract level one might mumble "I want to
be a philosopher." Or one could go so far as to say, "I wish Nixon
had remained in office." Now the object has expanded into a clause
that is almost an entire sentence. Accordingly a new
representational scheme might be offered.
(2) R(I,p)
This formalism suggests that wanting or wishing concerns a
relationship between the subject of the sentences and another
sentence. An alternative at this stage, might be to treat the
phrase"I want..." as a modal operator that stands outside the entire
expression "p." This might lead to the following type of formalism:
(3)W{I} (p)
In this case, there might be different operators for each of the
different possible subjects of such a sentence. The mention of modal
operators, however, suggests a fourth alternative that I would like
to consider in greater detail. The relationship between these types
of sentences might be described as one involving the subject and a
possible world, viz. the possible world (or worlds) described by the
sentence within a sentence. Thus informed, one might offer:
(4) R(I, pw) [(4') R'(I,p,pw)]
The reference of "pw" is presumably described by the phrase following
"I want...." in the sentences used above as examples. In many cases,
the reference of "pw" will be described by phrases ("...Nixon had
remained in office")("...to be a philosopher"), or even by
articulated nouns and single words ("...a Porsche.") ("...money.").
On this analysis, however, incomplete sentences would be expanded so
that they adequately described possible worlds. "I want to be a
philosopher " could be recast as "I want (a state of affairs in which
I am a philosopher)."). "I wish that Nixon had remained in office."
might re-emerge as "I want (a state of affairs in which Nixon is the
President).". Although this particular parsing is unwieldy and would
probably never find its way into the patterns of natural language,
for the logician, it may suitably describe the characteristic flavor
of a wish. What remains unclear is whether this type of formalism
describes a specific state of affairs, i.e. one identifiable possible
world among many, or whether it describes any state of affairs in
which the relevant criterion is satisfied.
The choice between (2) and (4) is remarkably subtle. In the case of
(2) the representation expresses a relationship between a person and
a sentence or proposition. In the case of (4) the relationship
exists between a person and a possible world. The former is a
descriptive relationship that intuitively raises the problem of
interpretation, whereas the latter is a characteristically
referential relationship that seemingly sidesteps the issues of
hermeneutics. Another way of making the same point might be to say
that (2) is more of an epistemologically adequate representation (in
McCarthy's sense), because one can assume that the proposition "p"
has some meaning for the wisher, whereas (4) may ultimately be a more
satisfactory metaphysical representation because one can certainly
postulate possible worlds whose descriptions are impossible,
inexpressible, ambiguous or simply meaningless in the object language
of the believer. (One rather far-fetched example of this phenomenon
might be case of an extremely pious monk who asks for a just
world. The description of the world may be impossible for the monk to
imagine e.g. "A world from which God is absent." (If we take the
Grand Inquisitor as an interlocutor.) One might be able to describe
the possible state of affairs here contemplated in a language that
could be interpreted by third persons, but not by the monk. Thus the
representational scheme of (4) might be preferred over that of (2).)
Several other issues are raised by the choice of representations for
wishes and desires. These issues concern our assumptions about the
knowledge of the subject and our assumptions about judgments. If one
chooses to represent statements of wanting or wishing as relation-
ships that obtain between the subject of the sentence and a possible
world, this may lead to improper assumptions regarding the subject's
knowledge of the possible world or worlds referred to by the
sentence. Recall the example of Dame Rosa. By reciting her magical
oath, she requested a possible world in which her ancestors'
infanticidees would return to life. What she did not imagine is that
a possible world satisfying the description she had given would also
be one in which she too would be gnawed to death. A representational
treatment of wanting may also rely upon an unsophisticated treatment
of judgments. The relationship "R(I,pw)" fixes truth-claims
concerning the desire of "I" for the characteristic "C" that "pw"
satisfies. For example, if I desire a world in which pigs have
wings, "C" would require that in such a world those ordinarily
humble, terrestrial creatures would fly about merrily like newly
feathered angels. Thus if the relationship holds, one could
conclude, using Frege's notation,
/---C(pw)
On the descriptive account, one makes the more circumspect judgment,
/---R'(I,C(pw))
I believe that this judgment should be adjudged more circumspect
because the relationship asserted concerns the knowledge of the
reference of "I" rather than the characteristics of a possible world.
I am therefore assuming that we will have an easier time coming to
know what someone thinks than we will have trying to find out what
things might be like if they were otherwise.. If one adopts a
descriptive formalism, one can question the fate that was dealt to
Dame Rose. Her advocate might claim that she had merely stated a
wish involving herself and the description of some possible world.
Her description did not exclude her being gnawed to death, but it did
not include it either. She might maintain that her wanting and
wishing was either indeterminate or somehow constrained by a higher
order set of goals and purposes, including, for example, the desire
to remain alive. The meaning of Dame Rosa's statement might then be
represented in the following manner:
(5) Intention (I,R (I,p))
Now, the traveler may still object that this is not the complete
meaning of the statement because it is communicated (implicitly) to
others who have their own purposes fulfilled by the chosen language.
Thus, one might prefer:
(6) Interpretation (The Traveler in Black, Intention (I, R(I,p)))
This formulation would still allow fate to have some of its force in
situations such as those of Dame Rosa, where one could show that the
interpretation was in keeping with the original speaker's higher
order intentions. A more general form of this type of judgment is:
(7) Interpretation (X, Intention (Y, Information (Z)))
The referents of "X" and "Y" may be loosely defined as "agents"
(persons and perhaps also machines or other entities) while "Z" takes
as its referent a text or other object capable of transmitting
information. An argument proposing the use of (7) as a
representational scheme might be formulated by criticizing the
traditional use of truth-values to indicate judgments. The upshot of
all this might be a richer appreciation of Dame Rosa's situation and
of the problems involved in representing wants and wishes. If the
lust for power makes one a victim of such lust, then we should say
that Dame Rosa got what she asked for and what she wanted. Her
problem may have been that she lacked ~knowledge, self-knowledge in
particular. Her confusion was not about the nature of the possible
world she desired, but about the nature of her desires.
The problem of Dame Rosa suggests that we require a representational
apparatus that will make intentions transparent at all levels of
information. Ideally, intentional markers would each statement at
each contextual level. How this might be carried out is an
outstanding problem in the theory of intentional representation.
!
Concluding Theses
To conclude this review of What Computers Can't Do, it is helpful to
summarize some of the different theses involved in the claim that
human intelligence differs from machine intelligence. In its simplest
guise, Dreyfus' claim amounts to this:
(1); Human intelligence is not rule bound; artificial intelligence
is therefore impossible.
As I have argued, the weak version of this thesis, b.b2;Ased upon a
naive conception of rules, is undermined by a theory of levels in
both human and machine reasoning. A stronger version of this simple
thesis, one which discusses both rules and levels, is not offered by
Dreyfus, and I shall therefore ignore it.
A second thesis concerns the relationship between intelligence and
context.
(2) Human intelligence uses context, and purposes; therefore
artificial intelligence must also do so.
While (2) seems quite convincing, particularly as it is described in
part III of Dreyfus' book, this is not an argument against the
possibility of artificial intelligence; it merely establishes that a
scheme of represen tation must deal with the problems of levels, and
this holds the promise of being an active area of research.
(3) Intentions and "wetware" belong intrinsically to the body.
Similarly, one can accept (3) and still maintain that artificial
intelligence is possible. It is necessary to appreciate that
intelligence operates at and above the symbolic level. Therefore to
the extent that we are able to create symbols that represent and
mimic intentions, we may be able to produce intelligent machines.
(4) Reason is not the same as calculation according to rules.
(4) can also be accepted without it being a challenge to artificial
intelligence. As I have argued, rules operate within the context of
intentions that transcend the levels at which rules operate. Reason
requires that we transverse several levels in our thinking. But once
again, we can conclude that Dreyfus has not shown that it would be
impossible for machines to perform such feats.
(5) Reason requires language.
Thinking may be understood as the passage between several levels of
understanding (or through certain sequences of symbols) in accordance
with some intention. This occurs also in language where we cross the
symbolic discontinuities between individuals. Ultimately it is in
our abilities to manipulate symbols through language that we will
find the tools to carry out the project of artificial intelligence.
In so doing, we will gain a firmer grasp upon the problems of human
intentionalities.
-------
∂24-Apr-86 0032 Carnese@SRI-KL.ARPA Re: May 14 CPSR meeting
Received: from SRI-KL.ARPA by SU-AI.ARPA with TCP; 24 Apr 86 00:32:19 PST
Date: Wed, 23 Apr 1986 18:57 PST
From: Carnese@SRI-KL
To: jmc at sail
Subject: Re: May 14 CPSR meeting
cc: suchman at xerox
John,
The meeting will be at 7:30 in the hall of the First Presbyterian Church,
1140 Cowper, Palo Alto.
Unfortunately, I will have to be out of town that week. Lucy will make
sure that you have an overhead projector.
-- Dan
∂24-Apr-86 0841 RA receipt
Just wanted to remind you to look for the receipt for the hotel in Frankfurt.
∂24-Apr-86 0936 SJG re: pointer needed
[In reply to message rcvd 20-Apr-86 22:49-PT.]
Thanks (do you remember what for?).
Matt
∂24-Apr-86 0944 RPG F-111
From what I understand, no US fighter/bomber contains more avionics
than the F-111. There are two crew members, who sit next to each other because
they want an avionics bubble behind the cockpit. I don't think the
laser-directed missles and bombs are mounted on anything but these
guys outside the US.
The thing was originally thought to be able to take off from a carrier,
but it kept falling into the sea. The Navy, then, bought none and the
Air Force bought 500 - 1000. Maybe the LLNL crowd knows more than I do.
-rpg-
∂24-Apr-86 0959 SJG re: pointer needed
[In reply to message rcvd 24-Apr-86 09:43-PT.]
"Where did the phrase `qualification problem' come from", I asked.
I have a draft of a paper about planning and counterfactuals.
Want a copy?
Matt
∂24-Apr-86 1129 CHEADLE@SU-SCORE.ARPA Gray Tuesday
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 24 Apr 86 11:29:39 PST
Date: Thu 24 Apr 86 11:10:00-PST
From: Victoria Cheadle <CHEADLE@SU-SCORE.ARPA>
Subject: Gray Tuesday
To: PACK@SU-SUSHI.ARPA
cc: jmc@SU-AI.ARPA
Office: Margaret Jacks 258, 723-1519
Message-ID: <12201439688.43.CHEADLE@SU-SCORE.ARPA>
The Computer Science Department faculty met on March 20, 1986 to
evaluate the progress of each of our PhD students. This was the first
of two meetings held this year. Our next evaluation meeting, Black
Friday, is tentatively scheduled to be held on Tuesday, June 3th.
The purpose of the meetings is to ensure that all PhD students are
making ``reasonable progress'' towards completion of the degree. The
first meeting is intended to detect potential problems, to alert
advisors and other appropriate faculty to those problems, and to
initiate corrective action if possible. Students are expected to
correct problems by the second meeting.
During the meeting records of each student are checked for accuracy
and completeness. If you have any questions about your records, see
Victoria Cheadle (the Administrator of the Ph.D. Program). Each
student's progress is compared to the timetable which defines
``reasonable progress'' toward the PhD. (Copies of the ``Requirements
in the PhD Program'' are available from Victoria Cheadle.) Subjective
opinions are expressed on how each student is doing. Comments from
the student's advisor and from other appropriate faculty members are
solicited.
I would like to remind you that continuous registration is a
requirement of the PhD program. You must be registered every quarter
except summer, unless you have officially been granted leave of
absence.
Just to make sure our records are up to date, let me show you the
basic information we have on file about you:
Kaelbling, Leslie P. Advisor: McCarthy PhD Entry: 1-84
Comp Writ: Pass Comp Prog: Not Taken Quarter: 4.5
Qual: Not Taken Honors Coop: SRI
Teach (%): 50 Cand Begin: Not Filed
COMP 3-82 MS PASS
COMP 2-84 MS PASS WRITTEN
COMP 3-84 PHD PASS WRITTEN
Sincerely,
Nils Nilsson
Chairman
-------
∂24-Apr-86 1152 CHEADLE@SU-SCORE.ARPA Gray Tuesday
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 24 Apr 86 11:52:34 PST
Date: Thu 24 Apr 86 11:31:53-PST
From: Victoria Cheadle <CHEADLE@SU-SCORE.ARPA>
Subject: Gray Tuesday
To: JJW@SU-AI.ARPA
cc: JMC@SU-AI.ARPA, RPG@SU-AI.ARPA, Ullman@SU-AIMVAX.ARPA
Office: Margaret Jacks 258, 723-1519
Message-ID: <12201443672.29.CHEADLE@SU-SCORE.ARPA>
The Computer Science Department faculty met on March 20, 1986 to
evaluate the progress of each of our PhD students. This was the first
of two meetings held this year. Our next evaluation meeting, Black
Friday, is tentatively scheduled to be held on Tuesday, June 3th.
The purpose of the meetings is to ensure that all PhD students are
making ``reasonable progress'' towards completion of the degree. The
first meeting is intended to detect potential problems, to alert
advisors and other appropriate faculty to those problems, and to
initiate corrective action if possible. Students are expected to
correct problems by the second meeting.
During the meeting records of each student are checked for accuracy
and completeness. If you have any questions about your records, see
Victoria Cheadle (the Administrator of the Ph.D. Program). Each
student's progress is compared to the timetable which defines
``reasonable progress'' toward the PhD. (Copies of the ``Requirements
in the PhD Program'' are available from Victoria Cheadle.) Subjective
opinions are expressed on how each student is doing. Comments from
the student's advisor and from other appropriate faculty members are
solicited.
I would like to remind you that continuous registration is a
requirement of the PhD program. You must be registered every quarter
except summer, unless you have officially been granted leave of
absence.
Just to make sure our records are up to date, let me show you the
basic information we have on file about you:
Weening, Joseph S. Advisor: McCarthy PhD Entry: 1-80
Comp Writ: Pass Comp Prog: Pass Quarter: 18
Qual: MTC, Pass Qual Qtr.: 1-82
Teach (%): 50 Cand Begin: 1-82 Cand End: 1-87
G81: 3-83 Seminar: 1-85
Dissertation: Parallel Execution of LISP Programs
Readers: Gabriel Ullman
COMP 2-80 PHD PASS WRITTEN
COMP 1-82 PHD PASS PROG
EVAL 2/01/1982 Programming project.
EVAL 6/01/1982 Needs programming project by summer.
EVAL 2/01/1983 Needs G81 by 11/83
QUAL 1-81 FAIL MTC
QUAL 1-82 PASS MTC
Although your advisor did not attend the meeting, other members of your
Reading Committee expressed concern over you not producing more tangible
work. I'd like to strongly encourage you to submit a draft of your
thesis to your full committee by Black Friday.
Sincerely,
Nils Nilsson
Chairman
-------
∂24-Apr-86 1200 RICHARDSON@SU-SCORE.ARPA AI Candidate - Batali
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 24 Apr 86 12:00:13 PST
Date: Thu 24 Apr 86 11:59:10-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Candidate - Batali
To: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
shah@SU-SIERRA.ARPA, grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA
Message-ID: <12201448639.10.RICHARDSON@SU-SCORE.ARPA>
Following is the schedule I propose for Batali's visit on May 5. Please
let me know if this will work for you.
10:30 - 11:30 Nilsson
11:30 - 12:00 Grosof/Hirsh
1:15 - 1:45 Buchanan
2:00 - 2:30 McCarthy
2:30 - 3:30 Search Committee
3:30 - 4:00 Genesereth
4:00 - 5:00 Talk
-Anne
-------
∂24-Apr-86 1346 BRONSTEIN@SU-SCORE.ARPA Re: We should give diplomacy a chance in Libya.
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 24 Apr 86 13:46:51 PST
Date: Thu 24 Apr 86 13:38:07-PST
From: Alexandre Bronstein <BRONSTEIN@SU-SCORE.ARPA>
Subject: Re: We should give diplomacy a chance in Libya.
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 24 Apr 86 07:01:00-PST
Message-ID: <12201466651.18.BRONSTEIN@SU-SCORE.ARPA>
I was wondering where you were all this time!
Glad to have you back.
By the way, your message is Z most sensible one I've seen on the subject.
I'm not even sure anybody in Washington has planned anything that far ahead
(I'm serious). Have you sent a few copies of it to the White House? (or
Shultz)
You never know, someone might read it, and talk about it to the right person...
Alex
-------
∂24-Apr-86 1645 NILSSON@SU-SCORE.ARPA Re: Batali
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 24 Apr 86 16:42:31 PST
Date: Thu 24 Apr 86 16:42:48-PST
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Re: Batali
To: JMC@SU-AI.ARPA
cc: NILSSON@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 24 Apr 86 13:44:00-PST
Message-ID: <12201500273.37.NILSSON@SU-SCORE.ARPA>
I haven't heard of any other Batali papers, but I'll try to find
out if there are some and get copies.
-------
∂24-Apr-86 1720 EENGELMORE@SUMEX-AIM.ARPA Rozak Book Review
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 24 Apr 86 17:19:56 PST
Date: Thu 24 Apr 86 17:21:32-PST
From: Ellie Engelmore <EEngELmore@SUMEX-AIM.ARPA>
Subject: Rozak Book Review
To: JMC@SU-AI.ARPA
cc: EEngELmore@SUMEX-AIM.ARPA
Stanford-Phone: (415)723-4878
Message-ID: <12201507324.15.EENGELMORE@SUMEX-AIM.ARPA>
John, I don't know whether the loop has been closed. I have not been
able to reach Mark Johnson (408)920-5825 at the San Jose Mercury.
However, I talked to his answering machine and told it that he
should call you.
Let me know if there is anything else you want me to do, loopwise.
ellie
-------
∂24-Apr-86 1811 Y.YDUJ@LOTS-B.STANFORD.EDU re: fun (from SAIL's BBOARD)
Received: from LOTS-B by SU-AI.ARPA with TCP; 24 Apr 86 18:11:47 PST
Date: Thu 24 Apr 86 18:11:07-PST
From: Judy Anderson <y.yduJ@LOTS-B.STANFORD.EDU>
Subject: re: fun (from SAIL's BBOARD)
To: JMC@SU-AI.ARPA
cc: SU-BBOARDS@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 24 Apr 86 17:42:00-PST
Message-ID: <12201516350.117.Y.YDUJ@LOTS-B.STANFORD.EDU>
No, doesn't "hypo" mean little, and "hyper" mean lots? Therefore,
hypocracy is little governmetn. Big government would be hypercracy...
Judy.
-------
∂24-Apr-86 2030 JMC
Shoham,productivity essay
∂25-Apr-86 0800 JMC
ibm about ebos
∂25-Apr-86 0953 RA taking the afternoon off
I need to take the afternoon off today.
∂25-Apr-86 1146 RA leaving
I am leaving now.
Have a nice weekend.
∂25-Apr-86 1156 NILSSON@SU-SCORE.ARPA Etherington
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 25 Apr 86 11:56:18 PST
Date: Fri 25 Apr 86 11:52:09-PST
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Etherington
To: jmc@SU-AI.ARPA, buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
grosof@SU-SCORE.ARPA, hirsh@SUMEX-AIM.ARPA, shah@SU-SIERRA.ARPA
cc: nilsson@SU-SCORE.ARPA
Message-ID: <12201709504.41.NILSSON@SU-SCORE.ARPA>
David Etherington will be visiting Monday, Apr 28. According to the
schedule Anne has devised for him, all of the search committee (except
Haresh) will be seeing him individually. Perhaps we don't need a massed
search committee meeting with him? Would anyone, instead, like to join
me in taking him out to dinner Monday night? I won't suggest spouses
this time. -Nils
-------
∂25-Apr-86 1527 avg@su-aimvax.arpa Orals committee?
Received: from SU-AIMVAX.ARPA by SU-AI.ARPA with TCP; 25 Apr 86 15:25:54 PST
Received: by su-aimvax.arpa with Sendmail; Fri, 25 Apr 86 15:28:05 pst
Date: Fri, 25 Apr 86 15:28:05 pst
From: Allen VanGelder <avg@diablo>
Subject: Orals committee?
To: jmc@sail
Are you interested in being on my orals committee? My thesis title is
Logic Programming for Parallel Implementation. One chapter will be
on Negation as Failure. I am trying for June 6, the Friday of dead week.
∂25-Apr-86 1634 RICHARDSON@SU-SCORE.ARPA Etherington
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 25 Apr 86 16:33:56 PST
Date: Fri 25 Apr 86 16:14:23-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Etherington
To: AISEARCH: ;
Message-ID: <12201757242.16.RICHARDSON@SU-SCORE.ARPA>
A/ The Search Committee will NOT be meeting from 2:30 - 3:30 on April 28
with Etherington.
B/ Nils Nilsson would like to take Etherington to lunch on April 28 from
11:30 - 12:30. Those wishing to accompany them, please let me know by
9:30 so that I can make the reservation.
-Anne
-------
∂25-Apr-86 1639 RICHARDSON@SU-SCORE.ARPA Etherington
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 25 Apr 86 16:39:13 PST
Date: Fri 25 Apr 86 16:35:21-PST
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Etherington
To: AISEARCH: ;
cc: val@SU-AI.ARPA, Binford@SU-WHITNEY.ARPA, nilsson@SU-SCORE.ARPA,
rosenbloom@SUMEX-AIM.ARPA
Message-ID: <12201761058.16.RICHARDSON@SU-SCORE.ARPA>
Here is the schedule we have for Etherington on Monday, April 28.
TIME VISITING WHERE
10:30-11:00 Paul Rosenbloom Jordan 436
11:00-11:30 Vladimir Lifschitz MJH 362
11:30-12:30 lunch with Nilsson meet in MJH 216
1:15-1:45 Bruce Buchanan Welch A1112
2:00-2:30 John McCarthy MJH 356
2:30-3:00 Tom Binford MJH 220
3:00-3:30 Benjamin Grosof MJH 220
3:30-4:00 Michael Genesereth MJH 234
4:00-5:00 Talk Jordan 050
5:00-5:30 Haresh Shah* (this one is definitely not yet confirmed)
-------
∂25-Apr-86 2016 vijay@ernie.berkeley.edu
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 25 Apr 86 20:16:43 PST
Received: by ernie.berkeley.edu (5.45/1.12)
id AA17472; Fri, 25 Apr 86 20:17:27 PST
Date: Fri, 25 Apr 86 20:17:27 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604260417.AA17472@ernie.berkeley.edu>
To: jmc@su-ai.arpa
∂25-Apr-86 2017 vijay@ernie.berkeley.edu AI DISC.
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 25 Apr 86 20:17:33 PST
Received: by ernie.berkeley.edu (5.45/1.12)
id AA17500; Fri, 25 Apr 86 20:18:09 PST
Date: Fri, 25 Apr 86 20:18:09 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604260418.AA17500@ernie.berkeley.edu>
To: jmc@su-ai.arpa
Subject: AI DISC.
Hello,
Unfortunately, ucbernie has been down most of this week and
as a result you're mail might not have been received by
the machine.
Thus, please resend any mail that you sent earlier this week
or last weekend. Remember, you're responses don't need
to be long at all... but we DO need you're responses immediately!
Thanks,
- VR
P.S. Please Acknowledge Receipt of this mail.
∂25-Apr-86 2016 vijay@ernie.berkeley.edu
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 25 Apr 86 20:15:44 PST
Received: by ernie.berkeley.edu (5.45/1.12)
id AA17451; Fri, 25 Apr 86 20:16:22 PST
Date: Fri, 25 Apr 86 20:16:22 PST
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8604260416.AA17451@ernie.berkeley.edu>
To: jmc@su-ai.arpa
∂26-Apr-86 1406 @SU-SCORE.ARPA:BUCHANAN@SUMEX-AIM.ARPA FYI -- maybe 55 but it will only take us 10 years!
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 26 Apr 86 14:03:42 PST
Received: from SUMEX-AIM.ARPA by SU-SCORE.ARPA with TCP; Sat 26 Apr 86 13:59:55-PST
Date: Sat 26 Apr 86 14:00:34-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: FYI -- maybe 55 but it will only take us 10 years!
To: ksl-exec@SUMEX-AIM.ARPA
cc: rosenbloom@SUMEX-AIM.ARPA, nilsson@SU-SCORE.ARPA, mccarthy@SU-SCORE.ARPA
Message-ID: <12201995025.17.BUCHANAN@SUMEX-AIM.ARPA>
Mail-From: JARDETZKY created at 23-Apr-86 14:27:36
Date: Wed 23 Apr 86 14:27:35-PST
From: Oleg Jardetzky <JARDETZKY@SUMEX-AIM.ARPA>
Subject: [jardetzk@cory.berkeley (Paul Jardetzky):]
To: PROTEAN@SUMEX-AIM.ARPA
Message-ID: <12201213512.24.JARDETZKY@SUMEX-AIM.ARPA>
Dear colleagues,
I could not resist forwarding to you my son's (Paul Jardetzky,UCB senior)
unsolicited opinion of our research.(His plans are to go into CS,but he
just came back from Hawaii).
---------------
Return-Path: <jardetzk%cory@BERKELEY.EDU>
Received: from ucbvax.berkeley.edu by SUMEX-AIM.ARPA with TCP; Wed 23 Apr 86 13:53:15-PST
Received: by ucbvax.berkeley.edu (5.48/1.11)
id AA13452; Wed, 23 Apr 86 11:30:52 PST
Received: by cory.berkeley.edu (5.45/5.13)
id AA25799; Wed, 23 Apr 86 11:29:52 PST
Date: Wed, 23 Apr 86 11:29:52 PST
From: jardetzk@cory.berkeley (Paul Jardetzky)
Message-Id: <8604231929.AA25799@cory.berkeley.edu>
To: jardetzk@sumex-aim.stanford.edu
How many AI people does it take to change a lightbulb?
At least 55:
The problem space group (5):
One to define the goal state.
One to define the operators.
One to describe the universal problem solver.
One to hack the production system.
One to indicate about how it is a model of human lightbulb
changing behavior.
The logical formalism group (16):
One to figure out how to describe lightbulb changing in
first order logic.
One to figure out how to describe lightbulb changing in
second order logic.
One to show the adequecy of FOL.
One to show the inadequecy of FOL.
One to show show that lightbulb logic is non-monotonic.
One to show that it isn't non-monotonic.
One to show how non-monotonic logic is incorporated in FOL.
One to determine the bindings for the variables.
One to show the completeness of the solution.
One to show the consistency of the solution.
One to show that the two just above are incoherent.
One to hack a theorm prover for lightbulb resolution.
One to suggest a parallel theory of lightbulb logic theorm
proving.
One to show that the parallel theory isn't complete.
...ad infinitum (or absurdum as you will)...
One to indicate how it is a description of human lightbulb
changing behavior.
One to call the electrician.
The robotics group (10):
One to build a vision system to recognize the dead bulb.
One to build a vision system to locate a new bulb.
One to figure out how to grasp the lightbulb without breaking it.
One to figure out how to make a universal joint that will permit
the hand to rotate 360+ degrees.
One to figure out how to make the universal joint go the other way.
One to figure out the arm solutions that will get the arm to the
socket.
One to organize the construction teams.
One to hack the planning system.
One to get Westinghouse to sponsor the research.
One to indicate about how the robot mimics human motor behavior
in lightbulb changing.
The knowledge engineering group (6):
One to study electricians' changing lightbulbs.
One to arrange for the purchase of the lisp machines.
One to assure the customer that this is a hard problem and
that great accomplishments in theory will come from his support
of this effort. (The same one can arrange for the fleecing.)
One to study related research.
One to indicate about how it is a description of human lightbulb
changing behavior.
One to call the lisp hackers.
The Lisp hackers (13):
One to bring up the chaos net.
One to adjust the microcode to properly reflect the group's
political beliefs.
One to fix the compiler.
One to make incompatible changes to the primitives.
One to provide the Coke.
One to rehack the Lisp editor/debugger.
One to rehack the window package.
Another to fix the compiler.
One to convert code to the non-upward compatible Lisp dialect.
Another to rehack the window package properly.
One to flame on BUG-LISPM.
Another to fix the microcode.
One to write the fifteen lines of code required to change the
lightbulb.
The Psychological group (5):
One to build an apparatus which will time lightbulb
changing performance.
One to gather and run subjects.
One to mathematically model the behavior.
One to call the expert systems group.
One to adjust the resulting system so that it drops the
right number of bulbs.
-------
-------
∂27-Apr-86 1252 CLT
\def\N{\mathop{N}}
∂27-Apr-86 1836 shoham@YALE.ARPA Re: Chronological ignorance
Received: from YALE-BULLDOG.ARPA by SU-AI.ARPA with TCP; 27 Apr 86 18:35:11 PDT
Received: by Yale-Bulldog.YALE.ARPA; 27 Apr 86 20:19:54 EDT (Sun)
Date: 27 Apr 86 20:19:54 EDT (Sun)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8604280019.AA21593@Yale-Bulldog.YALE.ARPA>
Subject: Re: Chronological ignorance
To: Vladimir Lifschitz <VAL@SU-AI.ARPA>
Cc: jmc@su-ai.ARPA, nilsson@score.ARPA, thomason@c.cs.cmu.edu,
ginsberg@sumex-aim.ARPA, bmoore@sri-ai.ARPA
In-Reply-To: Vladimir Lifschitz <VAL@SU-AI.ARPA>, 27 Apr 86 1547 PDT
Yoav,
I have a question about Definition 9 on p. 8 of your paper. Assume that there
is a time point t1 which is less than t0. ∀Then every immediate knowledge
sentence whose f.t.p. is t0 is equivalent to a sentence whose f.t.p. is less
than t0, because we can conjunctively append a tautology which mentions t1 to
the base sentence x. This seems to imply that conditions 1 and 3 cannot be
satisfied together. Do I overlook something?
Vladimir
Of course not; it's I who overlooked something.
There are two answers to your question.
In the paper you have (and a shorter version submitted to AAAI) I gave a simplified
version of chron. ign., and in the process introduced some bugs, I'm afraid.
The actual definition of chr. minimality of Kripke structures is more complex, and in it
you rely not only on the ftp of propositions, but on their "time signature", which is
the tuple of ftp's of all subformulas. It actually gets a bit complicated.
But let me resolve the issue by a second answer. For the purposes of the paper,
(and, I'm beginning to think, for quite a variety of purposes), you
can alter the definition of minimality. When you compare models you look not at any
immediate knowledge sentences, but only at atomic ones: those of the form
K(TRUE(t,p)), where p is a primitive proposition (in the FO case this of course will
mean propositions of the form K(TRUE(t,r(....))) ). I believe this is sufficient
to take care of what correctly disturbed you. For example, you can't compare
models by the satisfaction of K(TRUE(1,true))&K(TRUE(3,p)); you must
consider each conjunct independently.
Please let me know if you still find bugs. Those aside, do you understand my
motivation? Do you agree with my approach to constructing nonmonotonic logics?
I expect you'll have more than one objection.
Regards,
Yoav.
-------
∂28-Apr-86 1151 VARDI@IBM.COM
Received: from IBM.COM by SU-AI.ARPA with TCP; 28 Apr 86 11:51:21 PDT
Date: 28 Apr 86 11:37:10 PDT
From: VARDI@IBM.COM
To: stan@sri-ai, jmc@su-ai
Subject: SIGART
Thanks for your response. Here is why I was asking.
The ACM Symp. on Principles of Database Systems (PODS) has been recently
branching in the direction of knowledge bases. In the past PODS was
sponsored by SIGACT (automata and computability theory) and SIGMOD
(management of data). In view of PODS' direction, we (the PODS executive
committee) are considering seeking also SIGART sponsorship, but we wanted
to know that we are not making a mistake.
Moshe
∂28-Apr-86 1206 lantz@su-gregorio.arpa Encore
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 28 Apr 86 12:06:14 PDT
Received: by su-gregorio.arpa with Sendmail; Mon, 28 Apr 86 12:06:52 pdt
Date: 28 Apr 1986 1206-PDT (Monday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: facil@sail
Cc:
Subject: Encore
My quasi-reliable rumor mongers have it that Hambricht and Quist is
trying very hard to pull the plug on Encore -- as in have them merge
with whoever. At last word they were down to there last $17M (as if
lots of companies have that much to begin with, but Encore apparently
started out with $120M!) and H&Q was beginning to think they were all
form and little content.
Just stirring the froth, Keith
∂28-Apr-86 1207 RA lunch
I am going out for lunch; will be back around 1:30.
∂28-Apr-86 1416 VAL re: Chronological ignorance
To: shoham@YALE.ARPA
CC: JMC@SU-AI.ARPA, nilsson@SU-SCORE.ARPA, thomason@C.CS.CMU.EDU,
ginsberg@SUMEX-AIM.ARPA, bmoore@SRI-AI.ARPA
[In reply to message from shoham@YALE.ARPA sent 27 Apr 86 20:19:54 EDT.]
Thanks for your comments. I'd like to see the details of your first answer.
In the second answer, you propose to look only at atomic knowledge sentences
when comparing models. It seems that then it wouldn't hurt if only such knowledge
sentences were allowed in the language. In other words, we can apparently avoid
treating K as a modal operator, and instead use two predicate constants, TRUE and
K-TRUE, so that we would be back in the language of classical logic. If this can
be done then perhaps we wouldn't need Kripke models, or maybe it wouldn't be even
necessary to distinguish between TRUE and K-TRUE, i.e., we'll be working with
something very close to the system of Section 2.1. Is that possible? If yes then
I think it would be an important simplification.
Vladimir
∂28-Apr-86 1613 VAL re: Chronological ignorance
To: shoham@YALE.ARPA
CC: JMC@SU-AI.ARPA, nilsson@SU-SCORE.ARPA, thomason@C.CS.CMU.EDU,
ginsberg@SUMEX-AIM.ARPA, bmoore@SRI-AI.ARPA
[In reply to message from shoham@YALE.ARPA sent 27 Apr 86 20:19:54 EDT.]
One more question. Let's make your shooting story slightly more complex and
assume that the gun is loaded either at t=1 or at t=2.
Can you handle that, either in TETL (but perhaps with axioms of a more
general form than it toy causal theories), or in a generalization of it? The
difficulty I see is that chronological minimization can give you the undesirable
consequence that loading actually occured at t=1. If the answer is yes, can you
do that on the basis of your simple second answer to my original question?
Vladimir
∂28-Apr-86 2000 SJM
p.m. work on produc[1,sjm]
∂29-Apr-86 0933 CLT calendar item
sat 3-may 18:00 dinner chez hurd
∂29-Apr-86 0957 RICHARDSON@SU-SCORE.ARPA Yoav Shoham
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 29 Apr 86 09:57:08 PDT
Date: Tue 29 Apr 86 09:46:14-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Yoav Shoham
To: jmc@SU-AI.ARPA, val@SU-AI.ARPA
Message-ID: <12202724234.10.RICHARDSON@SU-SCORE.ARPA>
Would either or both of you be interested in taking Yoav Shoham to lunch
at the Faculty Club on Friday, May 2?
-Anne
-------
∂29-Apr-86 0958 RA library
If it's ok with you, I will put all the library material
which goes into the new filing cabinet in newlib[1,jmc]
∂29-Apr-86 1057 SJM computer failure
I am forced to conclude that an explicable (but totally minor)
meltdown of the computer prevented your receiving my reminder at 8:00
yesterday about PRODUC[1,sjm].
Susie
∂29-Apr-86 1107 KOLKOWITZ@SU-SCORE.ARPA
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 29 Apr 86 11:07:06 PDT
Date: Tue 29 Apr 86 11:07:20-PDT
From: Dan Kolkowitz <KOLKOWITZ@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Tue 29 Apr 86 10:03:00-PDT
Message-ID: <12202738998.38.KOLKOWITZ@SU-SCORE.ARPA>
I started to work on it on Friday. I'll pick up again today. My proposed
fix is to take our sendmail software and port it to the RT. Then we can
start using the regular sendmail config file that is used around Stanford,
a great convenience. I brought it over. Now I have to continue with the port.
The alternative is to debug their sendmail stuff which, I believe, is a more
unpleasant proposition.
Dan
-------
∂29-Apr-86 1200 JMC
lunch with Street
∂29-Apr-86 1204 LLW@S1-A.ARPA Pournelle's Affair
Received: from S1-A.ARPA by SU-AI.ARPA with TCP; 29 Apr 86 12:04:27 PDT
Date: 29 Apr 86 1154 PDT
From: Lowell Wood <LLW@S1-A.ARPA>
Subject: Pournelle's Affair
To: jmc@SU-AI.ARPA
CC: LLW@S1-A.ARPA
∂29-Apr-86 1148 JMC@SU-AI.ARPA
Received: from SU-AI.ARPA by S1-A.ARPA with TCP; 29 Apr 86 11:48:40 PDT
Date: 29 Apr 86 1151 PDT
From: John McCarthy <JMC@SU-AI.ARPA>
To: llw@S1-A.ARPA, rah@S1-A.ARPA
If you are going to Pournelle's next, I'd like to go along.
[John: I'm planning to go, and would be delighted to have your company.
However, I have to be in DC on the 9th for Senate SDI testimony, so
detailed plans on how/when/whether to get to LA for the weekend are still
a trifle uncertain. I'll try to get them into shape this weekend, and
let you know. Lowell]
∂29-Apr-86 1215 GSMITH@SUMEX-AIM.ARPA AI Qualifying Exams, 5/20 and 5/23
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 29 Apr 86 12:15:05 PDT
Date: Tue 29 Apr 86 12:16:23-PDT
From: Grace Smith <GSMITH@SUMEX-AIM.ARPA>
Subject: AI Qualifying Exams, 5/20 and 5/23
To: TOB@SU-AI.ARPA, EAF@SUMEX-AIM.ARPA, Genesereth@SUMEX-AIM.ARPA,
JMC@SU-AI.ARPA, Nilsson@SUMEX-AIM.ARPA, Rosenbloom@SUMEX-AIM.ARPA,
Winograd@SUMEX-AIM.ARPA
cc: BGB@SUMEX-AIM.ARPA, Cheadle@SU-SCORE.ARPA, Waleson@SUMEX-AIM.ARPA,
Gsmith@SUMEX-AIM.ARPA
Message-ID: <12202751568.29.GSMITH@SUMEX-AIM.ARPA>
The following students have signed up for the AI quals:
Mike Dixon
Glenn Healey
Andy Golding
John Woodfill
Liam Peyton
Leslie Kaelbling
John Mohammed
Eric Schoen
Arif Merchant
Kim McCall
Rich Washington
Amy Unruh
Please tell me if you feel there is a conflict in your personally being on the
examining team for any of these students. My preference is that advisors
should not examine advisees.
As we have discussed previously, Nils, Ed, Paul, Terry and I are scheduled to
administer the quals on Tuesday morning, May 20th, and Tom, Mike, John, Nils and I
are scheduled for Friday morning, May 23rd. We will divide into three teams
and administer six exams each morning. The exams will begin at 8:30; the
period from noon until 1:00 will be spent evaluating the results.
-------
∂29-Apr-86 1312 CLT SAIL lines
To: BJORK@SU-SCORE.ARPA
CC: JMC@SU-AI.ARPA
Our terminals at home now work again.
Thanks
∂29-Apr-86 1336 RA John Kwapisz
John Kwapisz from Science and Engineering Committee on SDI would like you to
call him back (202) 547 5607.
∂29-Apr-86 1407 VAL Shoham's paper
Here are some comments I sent to Shoham regarding his paper.
... First, I'd like to see more clearly what we gain by using
modalities instead of classical logic. Second, I suspect that chronological
minimization is only good for some limited class of problems, and accepting it as
a part of logic can make some things provable that are not intuitively true. My
example with loading a gun at an instant that isn't completely specified was
intended to be an example of that...
For instance, I received your manuscript either last Thursday or last Friday, I
don't remember exactly. If you can derive from this by pure logic that this
happened on Friday, I'll say that your logic is too strong. I'd like to see how
you handle this and a few other problems of this kind.
∂29-Apr-86 1424 TAJNAI@SU-SCORE.ARPA Dr. Peled of IBM
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 29 Apr 86 14:24:03 PDT
Date: Tue 29 Apr 86 14:22:19-PDT
From: Carolyn Tajnai <TAJNAI@SU-SCORE.ARPA>
Subject: Dr. Peled of IBM
To: jmc@SU-AI.ARPA
Message-ID: <12202774494.12.TAJNAI@SU-SCORE.ARPA>
Dr. Abe Peled of IBM Yorktown Heights will be here on Tuesday,
May 20. I am setting up appointments for him:
1:30
2:00
2:30
3:00
in Nils' conference room. You mentioned that you were trying to get
money from IBM. Would you like a half hour appointment to tell him
about your project(s)?
Please advise,
Carolyn
-------
∂29-Apr-86 1519 RA Sally Tuttle
Sally Tuttle (5-3005) Stanford TV called re your lecture on May 6.
∂29-Apr-86 2000 SJM
pm a good time to work on PRODUC[1,sjm]
∂30-Apr-86 0114 HST lisp standardization and conference
i was told that you are member of the ANSI steering committee for the standardization
zation of LISP. is that true? did you already something for them?
what about the 30-years conference? are you not interested?
∂30-Apr-86 0925 RICHARDSON@SU-SCORE.ARPA Yoav Shoham
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 30 Apr 86 09:25:30 PDT
Date: Wed 30 Apr 86 09:08:22-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Yoav Shoham
To: nilsson@SU-SCORE.ARPA, buchanan@SUMEX-AIM.ARPA, rosenbloom@SUMEX-AIM.ARPA,
grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA, winograd@SU-CSLI.ARPA,
jmc@SU-AI.ARPA, shah@SU-SIERRA.ARPA, genesereth@SUMEX-AIM.ARPA
Message-ID: <12202979486.13.RICHARDSON@SU-SCORE.ARPA>
Following is the schedule for Yoav Shoham's visit:
5/1 2:15 - 3:15 Nils Nilsson MJH 216
3:15 - 3:45 Bruce Buchanan MJH 220
4:15 - 5:00 Talk Bldg. 380, Room 380 X
5:00 - 6:00 Students MJH 252
6:30 Dinner TBA
5/2 11:00 - 11:30 Paul Rosenbloom Jordan 436
11:30 - 12:00 Grosof/Hirsh MJH 220
12:00 - 1:00 lunch TBA
1:15 - 1:45 Terry Winograd MJH 238
2:00 - 2:30 John McCarthy MJH 356
3:30 - 4:00 Haresh Shah MJH 220
4:30 - 5:00 Michael Genesereth MJH 234
-------
∂30-Apr-86 1034 NILSSON@SU-SCORE.ARPA Re: AI Qualifying Exams, 5/20 and 5/23
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 30 Apr 86 10:34:44 PDT
Date: Wed 30 Apr 86 10:25:20-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Re: AI Qualifying Exams, 5/20 and 5/23
To: GSMITH@SUMEX-AIM.ARPA
cc: TOB@SU-AI.ARPA, EAF@SUMEX-AIM.ARPA, Genesereth@SUMEX-AIM.ARPA,
JMC@SU-AI.ARPA, Rosenbloom@SUMEX-AIM.ARPA, Winograd@SUMEX-AIM.ARPA,
BGB@SUMEX-AIM.ARPA, Cheadle@SU-SCORE.ARPA, Waleson@SUMEX-AIM.ARPA,
NILSSON@SU-SCORE.ARPA
In-Reply-To: <12202751568.29.GSMITH@SUMEX-AIM.ARPA>
Message-ID: <12202993497.32.NILSSON@SU-SCORE.ARPA>
I'd like not to be on the sub-group which examines Leslie Kaelbling.
(I'm not her adviser, but I think I might be on her committee, and I
worked with her when I was at SRI.) -Nils
-------
∂30-Apr-86 1131 LES Asking for the RT (again)
To: JMC
CC: CLT
In fact, we asked for the RT for EBOS in the first place and submitted the
text of the EBOS proposal to justify the loan. It sounds to me as though
someone there is confused.
∂30-Apr-86 1151 VAL Shoham
Anne Richardson asks whether any or both of us would like to take Shoham to lunch
on Friday. If you want to take him to lunch and want me to accompany you, please
let me know.
∂30-Apr-86 1341 RPG Rumor
To: JMC, LES
Do you know anything about this rumor that Yale pressed criminal
charges against Schank?
-rpg-
∂30-Apr-86 1500 JMC
phone admittees
∂30-Apr-86 1526 RA coming in late
I will come in late tomorrow, between 11:00 and 11:30.
∂30-Apr-86 1509 SJM produc
Here's my mission: I think we ought to finish this essay this week. I
am feeling the need to step up the pace, which probably involves making
your life hell, or at least purgatory. That's how books get written, so
they tell me. And if it's in print, it must be true. Anyway, what do
you think?
Susie
∂30-Apr-86 1538 CLT to try kcl
/usr/kcl/okcl/unixport/kcl
∂30-Apr-86 1542 kolk@su-carmel.arpa ibmrtpc1
Received: from SU-CARMEL.ARPA by SU-AI.ARPA with TCP; 30 Apr 86 15:42:08 PDT
Received: by su-carmel.arpa with Sendmail; Wed, 30 Apr 86 15:42:48 pdt
Date: Wed, 30 Apr 86 15:42:48 pdt
From: Dan Kolkowitz <kolk@su-carmel.arpa>
Subject: ibmrtpc1
To: jmc@sail
Is there a contact for technical questions/problems for this machine. There
are some examples of strange behavior that I'd like to ask someone about if
I could.
Thanks,
Dan
∂30-Apr-86 1646 VAL Reminder: Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
Please notice that this time we meet at a new place.
CHRONOLOGICAL IGNORANCE:
time, knowledge, nonmonotonicity and causation
Yoav Shoham
Yale University
Thursday, May 1, 4pm
Room 380X, Mathematics Building
We are concerned with the problem of reasoning about change within a formal system.
We identify two problems that arise from practical considerations of efficiency
and naturalness of expression: the persistence problem (otherwise known as the
frame problem, and a new, but no less evil, initiation problem. In this talk we
concentrate on the latter one.
We propose a new logic that allows efficient and natural reasoning about change
and which avoids the initiation problem. The logic, called the logic of
chronological ignorance, is a fusion of recent ideas on temporal logic,
modal logic of knowledge, and nonmonotonic logic.
We identify a special class of theories, called causal theories, and show these
have elegant model-theoretic properties which make reasoning about causal theories
very easy.
Finally, we contrast our logic with previous work on nonmonotonic logics in
computer science, and discuss its connection to the philosophical literature
on causation.
∂30-Apr-86 1824 vijay@ernie.berkeley.edu Re: are you there?
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 30 Apr 86 18:24:51 PDT
Received: by ernie.berkeley.edu (5.50/1.12)
id AA08458; Wed, 30 Apr 86 18:25:50 PDT
Date: Wed, 30 Apr 86 18:25:50 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605010125.AA08458@ernie.berkeley.edu>
To: JMC@su-ai.arpa
Subject: Re: are you there?
Hi,
This is the first opportunity I've had to login for several straight
minutes. Unfortunately, ucbernie has been down due to work on
an ethernet connection. Things should run smoothly from now on.
Right now, I'm processing all the mail received. I've received
your responses to all papers sent. They are about to be forwarded
to everyone, and you should be getting some responses to those
soon.
Thanks,
Vijay
∂30-Apr-86 1857 vijay@ernie.berkeley.edu AI DISC: Hofstadter's reply
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 30 Apr 86 18:57:11 PDT
Received: by ernie.berkeley.edu (5.50/1.12)
id AA09572; Wed, 30 Apr 86 18:57:41 PDT
Date: Wed, 30 Apr 86 18:57:41 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605010157.AA09572@ernie.berkeley.edu>
To: jmc@SU-AI.ARPA
Subject: AI DISC: Hofstadter's reply
==============================================================================
Reply by D. Hofstadter to the Dreyfus statement
I find myself largely in agreement with the final paragraph of the
statement by Hubert and Stuart Dreyfus in which they suggest that we will
get much closer to human mentality when we shift away from the idea of
"a local symbolic representation of high-order macrostructural features
describing the world and turn instead to some sort of microstructural
distributed, holistic representation that is directly amenable to association,
generalization, and completion." However, I feel that this kind of view
ignores the importance of building a bridge between micro- and macro-levels.
It is fine to have a simulated neural net that as a whole does certain
things in a more fluid way than old-AI symbolic representations could,
but if it remains a mystery how that fluidity emerges, then we have
hardly gained anything over simply knowing that brains do it, "somehow".
Totally distributed neural activity is in some sense continuous,
flowing hither and thither in its amazingly parallel way throughout
a vast network. But when we think and utter things, we feel something
much more discrete is happening. We summarize vast patterns of neural
firings in small discrete chunks called "words" and "idioms". Although
the rules of thought themselves are almost undoubtedly not capturable at
the level of those discrete chunks, the most important part of the study
of thinking will have to be to relate the swarms of micro-activity to
the selection of chunks of that sort.
How does a country make a national policy decision? The old AI approach
might be likened to modeling the head of state while ignoring the masses of
people down below. This is insufficient because the huge currents of public
opinion are what allow a head of state to be elected and to make decisions.
Ignoring the subgovernmental level of a country is a poor idea if one wishes
to understand where that country's decisions come from.
By contrast, the new connectionist approach might be likened to modeling
the vast masses while ignoring the governmental structure up above. This is
insufficient because the power is concentrated in the government, and more
particularly in the upper echelons of government. This hierarchy acts as
a kind of "discretizing funnel" through which the collective level of the
masses channels its collective desires. Much of the public may be of mixed
mind or confused, with just impressionistic and continuously changing feelings
filling their brains. However, as pressures from these vast distributed
feelings surge upwards through the layers of governments towards the ultimate
seat of decision-making, the process becomes more and more localized and
discrete. Eventually, at the top level, a single very clear discrete decision
has to emerge.
Obviously, such high-level decisions cause great consequences for the
country as a whole, which have tremendous repercussions, over a period of
time, on the vast masses. This then influences how the waves of feelings
will flow on the "micro-level", which of course then cause new macro-decisions
to emerge, and so on. Thus we have a complex interrelation of low-level and
high-level activity.
Moreover, since not just one but many micro- and macro-decisions are
going on in parallel, the decision-making process of a country cannot be
thought of as a simple temporal alternation between activity on low and high
levels of a nation, but must rather be thought of as a constantly on-going
mishmosh of high-level activity and low-level activity and intermediate-level
activity, with causal effects constantly propagating upwards and downwards
through the various levels of the hierarchy. Out of all this, there emerges
what seems to the outside to be a relatively understandable, modelable
"personality" -- a collection of traits typical of that country but not
of any other country.
Between the two extremes I earlier described for modeling a country,
there is a continuum of possibilities. For instance, rather than modeling
just the head of state, one can add on to it a model of the cabinet. Then
one could add on a model of the legislature, or some part of it. And then
one could try to chunk the populace as a whole into units larger than
individual people. Thus by gradually moving to a finer and finer grain
of chunks, one could simulate the country more and more accurately. What
one loses with a sufficiently fine-grained model is the explanatory power
of a small system. What one gains is perfect accuracy. By contrast, what
one loses with a coarse-grained model is accuracy, and in compensation one
gains clarity.
If one believes, however, in something such as "Only people exist, and
governments are a fiction", I don't see how one will come out with a good
model of how nations decide. And I think that may be the position of the
Dreyfus brothers. Conversely, if one believes that "Only governments exist
and not people", or worse, "Only heads of state exist and not people", then
one will go equally wrong. The problem of AI is how to mediate between
what is a very clear macro-level on which things appear to be discrete, and
what is a very clear micro-level on which things appear to be continuous.
(I do not mean to imply that what a neuron does alone seems to be continuous,
but rather that when huge swarms of neurons are observed from afar, their
individuality vanishes and one sees a pattern of flow, just as grains of
sand in an hourglass seem to form a perfect fluid.)
Much of the most beautiful physics and mathematics comes from delineating
the subtle and elusive properties of the boundaries between continuous and
discrete phenomena, and I think that as AI goes further, we will see an
increasingly beautiful explanation of how discrete-seeming cognitive
macrobehavior -- chunks called "concepts" and "ideas" and "images" and
so on -- emerges out of continuous subcognitive microbehavior. People who
concentrate only on the substrate are missing the big picture -- and there
really IS a big picture there, which is why old-fashioned AI is so beguiling,
and has some truth to it. Therefore, the ultimate challenge to AI is to
reconcile concepts and other macro-regularities of symbolic thought with
the awesome non-symbolic substrate out of which they derive their wondrous
fluidity.
=============================================================================
Reply by D. Hofstadter to the Winograd statement
This statement has more to do with science and society than with the
nature of AI. While I fully agree with Winograd on the vital importance of
integrating one's scientific work with one's social responsibilities, and
salute him for stressing that in this forum, I do not share his view that
social systems are being reorganized "to fit the properties of particular
computer implementations". It could be that I simply have not recognized
this phenomenon, but if so, it would help to have some clear examples of it.
Winograd seems to be concerned with the explosion of AI as a technology
("expert systems", "knowledge engineering", and such). I agree that that
aspect of AI is largely amoral and gives the public a very distorted view of
the nature of computers and perhaps of people themselves. However, I feel
that the more limited venture of trying to understand human mentality by
using psychology, philosophy, and computer models -- cognitive science, in
short -- is much less threatening to humanity.
Some people see scientific progress as constantly creating new types
of dangers to humanity (indeed, to life in general). I tend to see science
itself as more neutral, and to see human nature as the culprit. Scientific
discoveries have fantastic potential in either direction, bad or good, and
so I would not curtail science in any way. I would rather try to educate
people to the complexities of using scientific discoveries, and the manifold
repercussions that may propagate down the decades and centuries. Winograd's
statement is generally along those lines, but I feel it does not have much
to do with the nature of scientific AI.
=========================================================================
Reply by D. Hofstadter to the Searle statement
Searle wishes to know where he is wrong. Here is the answer: His
propositions 1 and 3 ["Programs are purely formal (i.e., syntactical)" and
"Minds have mental contents (i.e., semantic contents)"] are right. His
proposition 2 ["Syntax is neither equivalent to nor sufficient by itself
for semantics"] is wrong. Brains provide the counterexample. Since his main
conclusions (1, 4, and 5) rest on proposition 2, the rest of his statement
is irrelevant.
==========================================================================
Reply by D. Hofstadter to the Weizenbaum statement
I tend to agree with Weizenbaum's opinion that AI systems will remain
incredibly different from humans for the foreseeable future. I don't know,
however, why he undermines his credibility by making the statement "Nor can
the history of any individual human being be expressed by a finite bit string".
This amounts to saying "A given human mind (soul, etc.) is infinite". If he
were simply to say that it is very very large, I would have no problem with
it, but what he said sounds too religious for me.
I feel that Weizenbaum's and my gut reactions to certain aspects of
AI come from a very similar part of our "souls", and that this part rebels
against the simplistic, cold, overly rational, mechanical vision of human
emotional experience that many AI people proclaim and that AI as a field
gives the impression of espousing. This is very sad. People are unbelievably
much more complex than today's machines, and we should revel in this fact.
Often, when I am asked whether we will have achieved "real AI" in 50 years,
I reply, "Oh, no, I am much more optimistic than that -- I think it will take
hundreds of years, if we ever get there". And I am very serious about calling
that view "optimistic". To me, our minds' and souls' subtlety and sublimity
seem wonderful; it would be a tragedy to have our minds' mysteries packageable
on a chip worth $2. But even with my sympathy for "where he's coming from",
I just cannot agree with Weizenbaum on his seeming claim of the infinitude
of the human soul.
Like Winograd, Weizenbaum has taken this forum as an opportunity to
comment on social points more than on scientific ones. As with Winograd,
I am happy to see such points made, and made in forceful ways, but I personally
feel that AI as a scientific activity is quite benign, and therefore feel that
this forum ought to concentrate on AI as science rather than on AI as social
activity.
===========================================================================
Reply by D. Hofstadter to the Rumelhart statement
I am very much in sympathy with most of Rumelhart's statement, and
particularly with his concluding section on "Parallel Distributed Processing".
I simply will restate my conviction that the new "bottom-up" approach that
is represented by "connectionist" or "PDP" approaches must ultimately somehow
join up with the "top-down" approach that stresses epistemological issues.
I would emphasize that top-down approaches need not take logic as the
cornerstone of mind, even though traditionally they have done so. A top-down
approach simply begins with observed macro-phenomena and seeks to make models
of them that are as accurate as possible. Those phenomena could be irrational
phenomena as well as highly rational ones. The only question is whether all
top-down approaches will ultimately wind up needing a kind of neural substrate
-- or at least a statistical substrate -- to endow them with the flexibility
that human mentality so clearly exhibits. I think that is very likely, but
it would be interesting to see if one can't get partly there by having a
"slightly statistical" substrate, and one that might look very different
indeed from neural networks. Some of my own research is in fact exploring
such pathways to fluid mentality.
As a last comment on Rumelhart's statement, I would say that I was
a bit bewildered by his labeling theoretical AI "a branch of mathematics".
Mathematicians have occasionally shown some interest in the mind, but real
minds are far too irrational and messy to interest mathematicians. That is
why their models of mind have tended to be so relentlessly logic-based (think
of John McCarthy's view of mind, for instance). I think physicists come much
closer to the spirit of mind-exploration than mathematicians do, but even so,
I would never call theoretical AI "a branch of physics". Mathematics is what
mathematicians do, and I have yet to meet a mathematician (as distinguished
from a former mathematician) doing AI.
============================================================================
∂30-Apr-86 1941 vijay@ernie.berkeley.edu AI DISC: Charniak's Response
Received: from ERNIE.BERKELEY.EDU by SU-AI.ARPA with TCP; 30 Apr 86 19:40:53 PDT
Received: by ernie.berkeley.edu (5.50/1.12)
id AA09403; Wed, 30 Apr 86 18:53:31 PDT
Date: Wed, 30 Apr 86 18:53:31 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605010153.AA09403@ernie.berkeley.edu>
To: jmc@SU-AI.ARPA
Subject: AI DISC: Charniak's Response
Progress is Our Most Important Product
Eugene Charniak
Department of Computer Science
Brown University
Should AI researchers continue to do research in AI, or
should they stop and turn their attention to some other
field, such as the social implications of computers, neuro-
physiology or basket weaving? This is, I take it, the ques-
tion before us.
Actually we can refine the issue somewhat. None of the
participants here are going change their views on this ques-
tion because of new ideas they pick up during the discus-
sion. Therefore we are really arguing for the minds of the
less committed, and in particular the young, since without
them any research effort is doomed. My personal belief is
that the allegiance of the young can only be had if they
perceive the field as important, and as one in which pro-
gress has been made and will continue to be made. There can
be no doubt as to the importance of AI, but what about pro-
gress? Are their new ideas which are interesting enough
that a new Ph.D. student should consider spending the new
few years of his or her life exploring them? Should new
faculty members bet their tenure cases on them? For that
matter, are the ideas interesting enough so that those of us
with tenure will continue to explore them when we know that
all we absolutely have to do is show up for class? Progress
is Dreyfus's concern, and he is right to concentrate on it,
although he is wrong in his assessment of the field.
In looking for progress the best place to start is AI
work on vision - in particular low-level vision. This is
not my specialty but anyone with an eye for good research
cannot help but be impressed. I can remember an official of
ARPA telling me (this was about 1971) that while AI language
work seemed in good shape, vision seemed hopeless. I argued
with him, but he did have a point. It seemed impossible to
get good lines, even from scenes with all complexity
removed, and adding complexity was too terrible to contem-
plate. What neither of us knew was that AI vision work was
about to experience a revolution. So, for example, where
previously vision was viewed as a two-step process (get the
lines, identify the objects), we now recognize an extra step
in the way from light intensities to internal representation
- a step variously called the "intrinsic image", the "2 1/2
D sketch", etc. The basic theory is that it is possible to
extract good guesses as to surface orientation, depth, light
intensity, and surface reflectance simply on the basis of
the most general sort of information about how light and
physical objects work. Furthermore we now have the
beginnings of theories for all of those "complicating fac-
tors": texture, motion, stereo vision, the motion of objects
due to eye movement (optical flow), shadows, etc. Even
better, these factors typically help the viewer estimate the
intrinsic image components. They do not just complicate
things. Finally, within these theories we find numerous
links to psychophysics and, increasingly, neurophysiology.
It is instructive to note that NONE of the anti-AI position
papers so much as allude to the work in low-level vision.
It is hard to dismiss, so it seems to be ignored.
If low-level vision is the most obvious success story
it is by no means the only one. Another fascinating area of
work is that on non-monotonic logic. Again it is useful to
put the situation in perspective. The earliest work here
seems to be the THNOT statement in Planner. Planner used
the idea of negation as failure (via THNOT) and it quickly
gained currency because it offered solutions to several
problems, most obviously the frame problem. However, it
should be remembered that at this point (1970 or so) THNOT
was almost a programming trick. It has only been in the
last few years that the work of Doyle, McCarthy, McDermott,
Reiter, etc. has put the topic on a firm mathematical foot-
ing. We are now gaining some ideas of exactly what can and
cannot do with the notion. In particular, we now know that
the primary problem seems to be with multiple models (or
multiple fix-points, depending on the formulation you use).
What is consistent in one model may not be in another, and
in such cases the logics either allow for no conclusion to
be drawn or for contradictory conclusions. Now the original
examples of multiple models were artificial -- that is, peo-
ple could not reach a conclusion either. But a recent paper
by Hanks and McDermott presents a natural example for which
none of the most common logics can reach the right conclu-
sion. There are now at least two theories which claim to
fix this problem and I am waiting to see what will happen
next.
Another area of interest is connectionism. Like Wino-
grad, and unlike Hofstadter and Rumelhart, I include connec-
tionism within AI. Hofstadter's problem is that he has been
taking the connectionists too much at their word. Occasion-
ally connectionists talk as if with their network just "do
the right thing" and that it is impossible to explain their
workings in traditional AI terms. In fact, this is rarely
the case. Firstly, this only applies to distributed connec-
tionist, since those of a more localist orientation (the
"Rochester school") have obvious AI connections. But even
for distributed connectionists (Boltzman machines, back pro-
pagation, etc.) this is not an accurate characterization of
what happens. In the examples I have seen it is typically
easy for an AI type such as myself to understand the work.
One common transformation is that individuals (in the logic)
map into the set of facts known about them, and then each
fact maps into a set of connectionist nodes. Once one
recognizes the transformation it is not that hard to map
between the connectionist model and more traditional AI
terms.
In saying this I do not mean to say that connectionists
are just rehashing more typical AI. They are exploring an
architecture in which some things which are difficult in AI
are easy (relaxing to a best fit), while other things which
are comparatively easy in AI become incredibly hard (logical
deduction). But it is still AI.
There are many other areas of AI that look exciting
now. The recent work in "explanation-based" learning at
last seems to offer some hope of showing how some things can
be learned with only one example. The work in plan-
recognition now unifies story comprehension, speech acts,
and visual plan detection. Work in planning now views the
problem as a special case of reasoning about time.
Mathematical probability is now being dusted off and applied
to many AI problems, mostly expert system type applications,
but not exclusively.
And these are just the well-known examples. Equally
important are the ideas which have not received wide circu-
lation, but are interesting to a few. One of particular
importance to me is my current work on a marker-passing
model of language comprehension. This model has all sorts
of nice properties: marker-passing paths can be interpreted
as conditional proofs (so "believing" a path corresponds to
believe the conditions), word-sense disambiguation happens
because one of the conditions specifies the sense of a word,
noun-phrase reference occurs because one condition asserts
equality between the new object and an old one, etc. Each
of us have such ideas, and as long as some of them pan out
the field will continue.
This is why I have concentrated on the progress issue
to the exclusion of the other ones which have been brought
up. For example, I must confess not to really care that
much about the Chinese room problem. It is not that I do
not have an opinion on it. I like Searle's reformulation
because it makes clear where the fallacy is - proposition
three "Minds have mental contents (i.e. semantic contents)".
The first part is fine, but the "i.e. semantic contents" is
contentious. Searle has confused the everyday notion of
semantics (for which the equality between semantic and men-
tal is true) and the technical notion (as used in the seman-
tics of logic) where it is by no means obvious. Fodor in
his "Methodological Solipsism" paper quite explicitly argues
that mental events cannot be semantic in the sense that
Searle is using the term and to my mind he is clearly right.
Actually, I cannot even imagine how mental entities could be
semantic in the technical sense, but I suppose that it must
be an empirical issue, which raises the question of whether
Searle intends his argument to be empirical or not.
But as I said, arguing about Chinese Rooms seems point-
less to me. If Searle's argument is empirical then it is
sorely lacking in substantiating detail, and from historical
precedent we can be sure that the ultimate decision will be
made by scientists voting with their feet. (Note how pro-
gress comes in.) If Searle's argument is not empirical then
I suppose it must be about the definition of "understand-
ing". In this case no matter what we in AI accomplish
Searle can say "but it does not REALLY understand". Need-
less to say, if we ever do get a machine to handle human-
like conversation (much less make accurate predictions of
psychological and neurophysiological results) nobody will
pay much attention to this particular definition of "under-
stand" (progress again!).
My conclusion is simple - the final arbiter of good
science is the ideas created and by this measure we are
doing OK. Basket-weaving doesn't have a chance.
∂30-Apr-86 2000 JMC
send Sarah her taxes file
∂30-Apr-86 2000 JMC
jobs also
∂30-Apr-86 2039 RWW kcl
I have forwarded your message to Fausto. He will get back to you.
Richard
∂30-Apr-86 2045 RWW fausto@sumex
∂30-Apr-86 2047 RWW mad
I spoke briefly to Nafeh today. He seems to be having an upheaval
Any data?
Richard
∂30-Apr-86 2225 UMA@IBM.COM
Received: from IBM.COM by SU-AI.ARPA with TCP; 30 Apr 86 22:25:45 PDT
Date: 30 Apr 86 11:20:31 SET
From: UMA@IBM.COM
To: jmc@su-ai.arpa
Hi John,
this is my last day here at the lab and within electronic-mail's
reach. If there is someone I could visit on my Russia trip so let me
know. My address until the end of July is:
Ursula Maydell
Saumerstr. 6
CH-8800 Thalwil
Switzerland
Phone: (01) 720-1727
My Russia trip is from June 15 to July 2, but I will leave for northern
Europe on June 4th already.
Hope all is well with you and the new family!! All the best.
Ursula
∂01-May-86 0102 HST lisp 30 yeArs
there was no relation between the two things.
the idea was to organize the lisp conference in october/september 1988 in cambridge.
in some earlier mail i sent you a plan proposal and a list of autthors
i think we should make advertisements now - to have time for the authors
to prepare. we should organize an exhibition then where we
present pictures of the involved people, old documents, listings etc..
(there was a 30 years fortran conference and they did such things.
you propsed to take weinreb as local organisator in cambridge.
please find the plan message and think about the problem.
∂01-May-86 0900 JMC
John Cocke
∂01-May-86 1208 HST 30years lisp
john,you misunderstood me.you don't have to organize it.
but you should stand behinnd me.therefore you should agree with my plans.
i want to act in your interest.therefore: please look in that old message.
i proposed themes and speakers.
∂01-May-86 1419 TAJNAI@SU-SCORE.ARPA Inference Corporation
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 1 May 86 14:19:41 PDT
Date: Thu 1 May 86 14:06:10-PDT
From: Carolyn Tajnai <TAJNAI@SU-SCORE.ARPA>
Subject: Inference Corporation
To: jmc@SU-AI.ARPA
Message-ID: <12203295842.22.TAJNAI@SU-SCORE.ARPA>
John, when do you plan to do your faculty liaison visit to
Inference?
Carolyn
-------
∂01-May-86 1426 TAJNAI@SU-SCORE.ARPA re: Inference Corporation
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 1 May 86 14:26:13 PDT
Date: Thu 1 May 86 14:26:24-PDT
From: Carolyn Tajnai <TAJNAI@SU-SCORE.ARPA>
Subject: re: Inference Corporation
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 1 May 86 14:23:00-PDT
Message-ID: <12203299527.22.TAJNAI@SU-SCORE.ARPA>
They are entitled to one day a year consulting, seminar, or equivalent
as a Forum benefit. Since you go anyway, they may want someone else
to go. If you go specifically as a Forum rep, then $2K will be
transferred into your unrestricted account.
Carolyn
-------
∂01-May-86 1434 GIUNCHIGLIA@SUMEX-AIM.ARPA
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 1 May 86 14:34:05 PDT
Date: Thu 1 May 86 14:35:39-PDT
From: Fausto Giunchiglia <GIUNCHIGLIA@SUMEX-AIM.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Wed 30 Apr 86 21:17:00-PDT
Message-ID: <12203301211.56.GIUNCHIGLIA@SUMEX-AIM.ARPA>
I gave one copy of the kcl manual to your secretary
-- Fausto
-------
∂01-May-86 1630 RA leaving
It's Thursday and I am leaving at 4:30.
See you tomorrow.
∂01-May-86 1642 LES
To: JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, JJW@SU-AI.ARPA, RPG@SU-AI.ARPA,
Bosack@SU-SCORE.ARPA, tom@SU-SCORE.ARPA
Alliant shipped
Steve Ross of Alliant (213 534-8943) just called to tell us that the FX/8
was shipped on April 29 even though we have not received the Fed's
approval yet. Sounds as though they were eager to get another delivery in
April.
It will probably arrive here around Tuesday, May 6. Ross plans to be
in this area on Thursday 5/8, so if the power is ready by then he can
install it. If not, he will have to make another trip.
I am trying to find out about high speed data lines from here to Lucid (as
an alternative to microwave) and am getting the runaround from Pac Bell &
Stanford Telecommunications so far.
∂01-May-86 1731 VAL Etherington and Shoham
Maybe I want to take back what I said when you asked me to compare their work.
Recently I noticed something I didn't see before: Yoav, unfortunately, is
sometimes very sloppy, and this should be taken into account. Two very serious
revisions of the basic definitions in his last draft were necessary to make his
theorems true: one discussed in my correspondence with him this week, the other
noticed by Moshe during his talk. Compare this to David's clean handling of the
non-trivial model theory he needed in his thesis. Nils has asked me to write
letters about their work, so I'll go back to their papers and think more about it.
∂01-May-86 1801 NILSSON@SU-SCORE.ARPA
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 1 May 86 18:01:50 PDT
Date: Thu 1 May 86 18:02:23-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
cc: NILSSON@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 1 May 86 15:27:00-PDT
Message-ID: <12203338844.11.NILSSON@SU-SCORE.ARPA>
I'd rather have a free day than an extra $2k in my unrestricted. But
I'll go if no one else does and if it can be worked in and if it's
necessary.
-------
∂01-May-86 1827 NILSSON@SU-SCORE.ARPA cbcl
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 1 May 86 18:26:55 PDT
Date: Thu 1 May 86 18:13:35-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: cbcl
To: jmc@SU-AI.ARPA
cc: nilsson@SU-SCORE.ARPA
Message-ID: <12203340882.11.NILSSON@SU-SCORE.ARPA>
Would you like to discuss your cbcl ideas with some of my students
during the time 1:30-3:15 Tuesday, May 6? (That's this very
next Tuesday.)
-------
∂01-May-86 1953 JACOBS@su-sushi.arpa re: the undersecretary of idiocy (from SAIL's BBOARD)
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 1 May 86 19:53:09 PDT
Date: Thu 1 May 86 19:53:09-PDT
From: Joseph D. Jacobs <JACOBS@su-sushi.arpa>
Subject: re: the undersecretary of idiocy (from SAIL's BBOARD)
To: JMC@su-ai.arpa
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 1 May 86 18:58:00-PDT
Message-ID: <12203359010.14.JACOBS@su-sushi.arpa>
John:
I agree entirely with your statement that the money is SDI
money. It is theirs to decide what to do with. It is not money that
they are "taking away from people". Saying something to that extent
was part of my intent in quoting the sentence in the first place ( I
took it from someone's earlier bboard message ).
Joseph
-------
∂01-May-86 2127 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 1 May 1986
Previous Balance 6.30
Monthly Interest at 1.5% 0.09
Current Charges 6.00 (bicycle lockers)
-------
TOTAL AMOUNT DUE 12.39
Please deliver payments to Debbie Woodward, room 040, Jacks Hall.
Make checks payable to: STANFORD UNIVERSITY.
To ensure proper crediting, please include your Pony account name on your check.
Note: Payment recordation takes up to three weeks after delivery of a payment
(but not beyond the next billing date).
Bills are payable upon presentation. Interest of 1.5% per month will be
charged on balances remaining unpaid 25 days after bill date above.
Your last Pony payment was recorded on 3/16/86.
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.
∂02-May-86 0742 TOM@SU-SCORE.ARPA
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 2 May 86 07:40:34 PDT
Date: Fri 2 May 86 07:31:21-PDT
From: Thomas Dienstbier <TOM@SU-SCORE.ARPA>
To: LES@SU-AI.ARPA
cc: JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, JJW@SU-AI.ARPA, RPG@SU-AI.ARPA,
Bosack@SU-SCORE.ARPA, TOM@SU-SCORE.ARPA
In-Reply-To: Message from "Les Earnest <LES@SU-AI.ARPA>" of Thu 1 May 86 16:42:00-PDT
Message-ID: <12203486111.10.TOM@SU-SCORE.ARPA>
I doubt that we will have power installed before the end of next week.
I will have a better feel if and when they start working on it, which
they haven't as yet. But, I do at this time have made space the Alliant.
I am waiting to receive some lititure on General Datacomm datasets
which are suppose to be quite good and fast. 144KB over phone lines.
Stanford telecommuncations still have not figured out how to install
telephones so I don't see much help there as far as high speed data is
concerned.
tom
-------
∂02-May-86 0800 JMC
dog support
∂02-May-86 0840 RICHARDSON@SU-SCORE.ARPA John Batali
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 2 May 86 08:38:41 PDT
Date: Fri 2 May 86 08:38:38-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: John Batali
To: rosenbloom@SUMEX-AIM.ARPA, nilsson@SU-SCORE.ARPA, winograd@SU-CSLI.ARPA,
Binford@SU-WHITNEY.ARPA, buchanan@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA,
shah@SU-SIERRA.ARPA, grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA,
genesereth@SUMEX-AIM.ARPA
Message-ID: <12203498361.12.RICHARDSON@SU-SCORE.ARPA>
Here is the agenda for John Batali's visit on Monday, May 5:
10:00 - 10:30 Paul Rosenbloom Jordan 436
10:30 - 11:00 Nils Nilsson MJH 216
11:00 - 11:30 Terry Winograd MJH 238
11:30 - 12:00 students MJH 220
12:00 - 1:00 lunch with Binford/Nilsson Faculty Club
1:15 - 1:45 Bruce Buchanan Welch A1112
2:00 - 2:30 John McCarthy MJH 356
2:30 - 3:00 Haresh Shah MJH 220
3:00 - 3:30 Hirsh/Grosof MJH 220
3:30 - 4:00 Michael Genesereth MJH 234
4:15 - 5:00 Talk Jordan 050
6:30 dinner TBA
-------
∂02-May-86 0847 TAJNAI@SU-SCORE.ARPA re: Dr. Peled of IBM
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 2 May 86 08:47:46 PDT
Date: Fri 2 May 86 08:48:19-PDT
From: Carolyn Tajnai <TAJNAI@SU-SCORE.ARPA>
Subject: re: Dr. Peled of IBM
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Tue 29 Apr 86 15:45:00-PDT
Message-ID: <12203500123.11.TAJNAI@SU-SCORE.ARPA>
Jeff Ullman is in the same situation not knowing his exact schedule
for May 20. I'll put you both down for 1:30 in Nils' conference room.
Surely one of you can make it, or you can each take 15 minutes.
Carolyn
-------
∂02-May-86 0915 NILSSON@SU-SCORE.ARPA re: cbcl
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 2 May 86 09:15:06 PDT
Date: Fri 2 May 86 09:06:15-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: re: cbcl
To: JMC@SU-AI.ARPA
cc: NILSSON@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 1 May 86 18:50:00-PDT
Message-ID: <12203503388.30.NILSSON@SU-SCORE.ARPA>
there hasn't been a prepared lecture yet in our group!
-------
∂02-May-86 1206 MOGUL@SU-SCORE.ARPA re: the undersecretary of idiocy (from SAIL's BBOARD)
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 2 May 86 12:01:07 PDT
Date: Fri 2 May 86 11:38:56-PDT
From: Jeffrey Mogul <MOGUL@SU-SCORE.ARPA>
Subject: re: the undersecretary of idiocy (from SAIL's BBOARD)
To: JMC@SU-AI.ARPA, SU-BBOARDS@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 1 May 86 18:58:00-PDT
Message-ID: <12203531184.13.MOGUL@SU-SCORE.ARPA>
From JMC:
"You can say what you want, but we can take your money away if
we don't like it." - Joseph D. Jacobs.
This can be interpreted inaccurately. The money referred to isn't
money people already have; it's new SDI money as I understand it.
First off, the quote you attribute to Joseph actually comes from
someone else; Joseph was quoting him, not necessarily agreeing
with him.
Second, my understanding of the article I posted was that Hicks was
not referring only to SDI research money, but all DoD research
money (SCI, ONR, ALV, and all your favorite non-SDI acronyms).
If he was only talking about SDI money, I would be less upset (since
in the current political climate, I see taking SDI money as implicit
approval of the program. If SDI were truly a research project,
with the possibility of rejecting the hypothesis that "SDI will work",
then it would be wrong to exclude researchers who might support
rejection.)
Third, while it is true that Hicks did not speak of cancelling
current grants, most researchers count on renewals or new grants
if their current work and new proposals have scientific merit and
meet the goals of the funding agency. Hicks is saying "I won't
fire you today, but in a year or two you'll be begging over at
NSF."
Fourth, I don't want to argue that US scientists have constitutional
rights to DoD money. Hicks may be entirely within the law (unless I've
overlooked something) but his policy is rotten and corrupt. The
DoD could, within the law, scrap all its weapons and issue swords
to the infantry, but it would be lousy policy, no? So is this one.
-------
∂02-May-86 1134 SJM carolyn
Hazel is looking for Carolyn because Timothy is crying a lot. Do you know
where she is? Ask her to call home. (Hazel called around 11:20)
Susie
∂02-May-86 1531 MOGUL@SU-SCORE.ARPA re: Purge of dissident scientists? (from SAIL's BBOARD)
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 2 May 86 15:31:44 PDT
Date: Fri 2 May 86 15:29:11-PDT
From: Jeffrey Mogul <MOGUL@SU-SCORE.ARPA>
Subject: re: Purge of dissident scientists? (from SAIL's BBOARD)
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Fri 2 May 86 14:48:00-PDT
Message-ID: <12203573098.47.MOGUL@SU-SCORE.ARPA>
Please return to the original message I sent on this topic, or
the April 25 (26?) issue of Science, if you're not sure what Hicks
said. If you can interpret his remarks as applying only to SDI
funding, then I suppose my complaint is groundless -- but if he
intends to cut off, say, David Parnas, who is clearly not anti-defense,
then his policy is one of retribution for political differences and
I think it's sad. You're probably right that he'll back down; but
he'll back down because responsible scientists make loud noise, not
because he'll have a flash of insight.
I also hope you're not so naive as to believe that the only issue
is whether DoD will get a "system for shooting down missles"; if
SDI were being sold to the public (and the price is likely to be huge)
with clear specs, then we could argue about the wisdom of meeting
those specs as well as the feasibility. But you and both know that
the administration is selling SDI as an alternative to arms control,
as a means of population protection. The feasibility of meeting this
utopian spec is most certainly in question, and the danger of closing
off alternatives is most grave -- especially if we bankrupt the nation
and then SDI doesn't work.
-Jeff
-------
∂03-May-86 1121 HALPERN@IBM.COM
Received: from IBM.COM by SU-AI.ARPA with TCP; 3 May 86 11:21:18 PDT
Date: 3 May 86 10:33:54 PDT
From: HALPERN@IBM.COM
To: jmc@su-ai
Subject: Comments on your paper; conjecture settled
John, I think I can settle your conjectures on the propositional
case of Normality Logic, although I'm not sure you'll be pleased
with the answers. First of all, there are always minimal models for
satisfiable formulas; lots of them, in fact. Indeed, either a
formula is satisfiable in a model with no anamolies (and there
are infinitely many distinct such models) or it is not satisfiable
in a model with no anamolies. In this case, if the formula is
satisfiable at all, then ALL models are minimal. Any two models
for the formula are incomparable in terms of their anamolies
(in this sense, normality logic does bear some relationship to
the work on "all I know is".) The proofs of these claims are
pretty straightforward.
First of all, suppose a formula pi can be satisfied in a model
with no anomalies. It is quite easy to check whether
such a model exists: you simply consider all the subformulas
of the formula and see if they can be assigned truth values
in a consistent way with no anomalies. Suppose they can.
In this case, suppose p1, ..., pk are the subformulas for
which N(pj) holds. We can then extend to all formulas by
saying N(q) holds for a formula q iff q is a logical consequence
of p1&...&pk. We can also set all primitive propositions not
subformulas of pi to be true. This gives us a model of pi where
there are no anomalies. But we can easily modify this
model to get lots of others. In some sense, this is the
model where the LEAST number of things are normal, but I can
easily make more things normal.
Now suppose pi has no models where there are no anomalies.
Suppose I1 and I2 are distinct models for pi. Then I claim
that I1 and I2 are incomparable in their anomalies. It thus follows
that all models are minimal. An argument like that above actually
shows there are infinitely many such models. To prove the claim,
suppose that phi is in anomalies(I1) but not anomalies(I2).
Thus N(phi) is true in I1, but phi is false. There are two
possibilities for I2: either (1) both N(phi) and phi are true,
or (2) N(phi) is false. Since there are anomalies in I2, there
must be some formula psi such that N(psi) is true in I2, but
psi is false. In case (1), consider the formula
(psi or ~phi). (On my terminal ? is a negation symbol; I'm not
sure how it comes out on yours.) Since N(psi) is true in I2,
and (psi or ~phi) is a logical consequence of psi, then
N(psi or ~phi) must be true in I2. (This follows from your
conditions 2 and 3 on interpretations. To see this, note
that psi => (psi or ~phi) is a tautology, so that
N(psi => (psi or ~phi)) is true in I2. Putting that together
with the assumption that N(psi) is true in I2, we get the
desired result using condition 3.) But since by assumption
psi is false and phi is true in I2, (psi or ?phi) is in
abnormal(I2). Since (psi or ~phi) is true in I1 (since
phi is false in I1), (psi or ~not phi) cannot be
in abnormal(I1). Thus I1 and I2 are incomparable.
In case (2), consider the formula (psi or N(phi)). A similar
argument to that used for case 1 shows that N(psi or N(phi))
is true in I2, so that the formula (psi or N(phi)) is in
abnormal(I2) but not in abnormal(I1).
I'm afraid that this argument shows that minimal models as
you've defined them are not very interesting. You might
want to consider what happens if you look for models that
have the least number of purely propositional abnormalities,
or, as you've suggested, look at more complex orderings
among abnormalities.
One thing that bothers me in your discussion about the birds
is the use of NN. Your explanation of why you need it seems
awfully ad hoc. I think we all have some intuitions of what
"normally" really means, and it would be nice if we could get
a modal logic that captured them. (That's what I meant when
I asked you about semantic conditions on normality.) Instead
of using the NN as you've done in your example, you might consider
using a whole class of N's, perhaps subscripted by the rationals.
If a>b, then Na is a stronger normality condition than Nb.
Thus, in the example with the birds, you can capture the fact that
condition (3) is meant to supersede condition (1) by using a larger
subscript on N. (To be honest, I haven't thought this through very
carefully, but it seemed like an interesting possibility. I
suggested rationals rather than integers simply because that way
there's always room to put in new condition "in between" pre-existing
ones.)
To change the subject somewhat, I also picked up a copy of Hopcroft's
report yesterday and read it reasonably carefully. I have lots of
comments for him. While I think the idea of doing such a report is
very important, I don't think the key ideas in a number of areas are
brought out very well. As it stands, it reads to me a bit like
computer science is the science of finding good algorithms for very
important problem x (where x= linear integer programming, solving
linear equations, primality testing, planesweeping, etc.) It's not
that these aren't important problems! But it seems that to give
integer linear programming more space than the whole area of AI
somewhat distorts the picture. Am I right in assuming
that Hopcroft is looking for comments?
Do keep me informed on the latest developments in normality logic.
-- Joe
∂04-May-86 2146 PMF Weekend Plans Shift
To: JMC@SU-AI.ARPA, llw@S1-A.ARPA
John, I found out late Friday that I have to be in Washington all day next
Friday, and thus can't get back here much before 2100, making a very late
drive to LA from Livermore (after a long day in DC and in the air). I've
therefore decided to fly directly from Washington to LA, and will be
arriving there about 2000. I'd be delighted to provide local-area ground
transport for you over the weekend and back to the airport on Sunday late
PM, if this would be of interest to you (and my arrival time on Friday
evening reasonably convenient for you). I'll be leaving the Lab about 1145
tomorrow AM, but would be happy to discuss arrangements with you anytime
before then (or via Gloria, after then). Sorry for the change in travel
plans. Lowell 415-4229058.
∂04-May-86 2227 LES Facilities Committee Meetings of 4/23/86, etc.
To: facil@SU-AI.ARPA, "@XTRA.DIS[1,LES]"@SU-AI.ARPA
This note reports on two meetings held on April 23 and some subsequent
developments. A meeting with representatives of Symbolics to discuss
hardware and software maintenance issues raised by Tom Rindfleisch in an
earlier letter to them. The Facilities Committee also met to prioritize
the "final" set of things to be bought with DARPA facilities funds.
Subsequent analysis reveals that we can apparently buy everything that
has been requested.
The meeting with Symbolics scheduled for noon actually began about 12:45pm
because of an apparent communication failure. Representing Symbolics
management were J. Jay Weber, Director of Customer Service, and Lawrence
J. Rostetter, National Manager of Software Support. After we beat up on
them over their outrageous pricing and poor service, they agreed to try to
work out a less expensive and more responsive program for maintaining our
fleet of Symbolics machines. A first step was to be a survey of the
existing machines to determine what is needed to bring them up to current
hardware and software standards. [The survey has since been completed.
We await their proposal.]
In the subsequent Facilities Committee meeting, it was the consensus of
the committee was that the remaining available DARPA facility funds
(about $90k) should be spent according to the following priority:
1. VAX spares and diagnostics (about $50k)
2. 4 Meg. LCF memory boards for Sun 2/120s instead of 2 one-meg. boards ($5k).
3. DEC softaware and hardware upgrades for NAXOS ($5.3k).
4. Sun 3/160M-8 for Binford ($19.6k).
5. Imagen 8/300 Ethernet Printer, to displace one for TI trailer ($10.4k).
6. Extra Sun memory and floating point processor for Lantz ($13.8k).
7. Apple Laserwriter for KSL (about $6k).
8. Apple Laserwriter for Cheriton's lab (about $6k).
9. More Ethernet transceivers and cables ($300 each).
The Chairman was authorized to use this priority list as a guide in making
further decisions about how to spend the available funds.
[The balance of this note reports developments subsequent to the meeting.]
I was happy to hear from our local ONR Rep. Rob Simpson on 5/2 that our
request for a six-month contract extension of the DARPA facilities
contract has been approved, as has the purchase of the Alliant FX/8.
I also learned from Alliant that they had shipped our FX/8 on 4/28 even
though the approval had not been received. (They apparently wanted to
boost their April sales figures.)
Dave Cheriton has decided that he would be better off with one Sun 3/75
instead of the two 3/50's he was slated to get. Since the difference in
cost is small ($70 more), I find this acceptable. I note, however, that
Dave never seems to think of anything that would permit him to get by on
less money.
Choosing from among the best quotations for VAX spares that were finally
received by May 1 gave a total cost of $62k, which is $12k over budget.
On the other hand, it appears that buying a Sun 3/50 for spare parts is
not sensible, since we are now buying no other 3/50s and the parts in
the 3/50 are not interchangable with 3/75s or other workstations. It
appears, then, that this item should be dropped, saving $5.5k and leaving
the spares budget just $6.5k higher than expected.
As it turns out, the proposed purchase of 24 four-meg. memories from LCF
for the existing Sun 2/120s instead of 1 meg. memories from Sun was
ill-considered on two counts:
(1) the actual cost increase would have been $13.2k rather than $5k as
someone asserted during the meeting and
(2) these replacements would have resulted in 48 megabytes of perfectly
good memory being discarded.
The reason that the existing 2/120 memory boards could not be used is that
2/120 systems can address at most 4 megabytes. A more sensible approach
is to buy half as much 4-meg memory (12 boards) and use the 1-meg. boards
these displace to bring the rest of the 2/120s up to the 4-meg. maximum.
In other words, half the systems will have 4-meg. boards and the other
half will have 4 1-meg. boards, which just fit in the available card
slots. The net effect of this change will be a saving of $18.6k.
In still another development, Forest Baskett came by to inquire whether
CSNet might be permitted to place an ARPAnet gateway at Stanford that
would connect to a local branch of their network that includes a number of
industrial research groups. He said that he has a Microvax-based gateway
that can do the job.
If the DEC Microvax gateway were placed at Stanford, it should be
technically straightforward to use it also as a gateway to the Stanford
Ethernet system. Such a gateway would therefore provide still another
developmental and functional alternative or backup to the gateways
allegedly available from Cisco Systems and Proteon and would share with
the Cisco unit the beneficial property of being "free." I think that we
should consider the DEC alternative if it bears up under a closer look.
Bottom Line
I am happy to say that because of the economies discussed above,
principally the substantial savings on add-in memories for existing
Sun 2/120s, we apparently can buy everything that has been requested.
If we manage to convince ourselves that we do not need the Proteon
gateway, then we will be able to buy something else useful. The
following working budget retains the Proteon gateway and includes a dozen
Ethernet transceivers for good measure. I hope to resolve the gateway
question within the next two weeks. Opinions are invited.
Les Earnest
Facilities Committee Chair
----------------------------------------------------------------------------
DARPA Equipment Budget
Alliant
Quantity Model & Description Unit Price Total
[QLISP]
1 FX/8 multiprocessor system (partial cost) 120,000 120,000
--------
120,000
Apple
Quantity Model & Description Unit Price Total
[KSL]
1 Laserwriter 6,000 6,000
[CDS]
1 Laserwriter 6,000 6,000
--------
12,000
Computer Clearing House
Quantity Model & Description Unit Price Total
VAX MAINTENANCE [CF]
1 DEC DMF32 MUX (used; 8 ch. terminal board) 800 800
1 DEC M7475 TU80 interface (used) 1,075 1,075
--------
1,875
DEC
Quantity Model & Description Unit Price Total
VAX MAINTENANCE [CF]
1 TU80-AA 1600bpi tape drive 5,625 5,625
1 RA60-aa rack mount disk drive 7,875 7,875
1 TU78-AF 6250bpi slave disk drive 11,475 11,475
1 SV-LV55H-EK VSll RD53TK50 SWLIC RQDX3V 120 13,000 13,000
(Microvax 11)
1 ZD201-CW VAX 11750 COMPLT. MT/TU/58 DIAG. 7,000 7,000
1 MD-VAX-00 VAX MAINT. DOC./DIAG. LIST. 3,000 3,000
--------
47,975
[NAXOS]
1 QD058-HM VAX ADA EDU UPD 16MT9(MEDIA) 2,600 2,600
1 DEUNA-M UNIBUS ETHERNET CONTROLLER 1,888 1,888
--------
4,488
--------
52,463
Imagen
Quantity Model & Description Unit Price Total
ETHERNET PRINTER [EP]
1 8/300 Ethernet Print Server, 2 meg. memory 10,356 10,356
--------
10,356
Insyde
Quantity Model & Description Unit Price Total
[All parts manufactured by DEC.]
VAX MAINTENANCE [CF]
1 L0002 DPM 1,400 1,400
1 L0003 MIC 1,400 1,400
1 L0004 UBI 1,400 1,400
1 L0007 RH750 massbus interface 1,600 1,600
1 L0011 16K mem. contrl. 400 400
1 L0016 256K mem. contrl. 1,400 1,400
3 M8728-AA 1/4 meg. 64K mem. board 594 1,782
1 M7762 RL11 Contrl. 750 750
1 RK6/7-TA RK06 tester head. alaign. 2,485 2,485
--------
12,617
LCF
Quantity Model & Description Unit Price Total
MEMORY EXPANSION [CF]
12 LCF 4 meg. boards for Sun 2/120 2,650 31,800
--------
31,800
Proteon
Quantity Model & Description Unit Price Total
[CF]
1 p4200-10 Pro-NET-10 Base Unit, 110V 7,545 7,545
1 p4203-10 Ethernet Base Unit 110V 6,800 6,800
1 p4213 Ethernet Multibus I/F & driver 2,950 2,950
1 p4299 ProNET-Linkway rack mount kit 215 215
1 p5600 IP Packet Forwarder 1,200 1,200
--------
18,710
Sun Microsystems
Quantity Model & Description Unit Price Total
MEMORY EXPANSION [CF]
10 2/50-OPT-12 2 MByte memory expansion 2,100 21,000
--------
21,000
FILESERVER [CF]
1 3/180S-8-R1 processor, 8 MB memory 8,130 8,130
1 3/180S-675-R1 6250 tape drive 11,830 11,830
1 370-1012 SMD disk controller 2,450 2,450
1 540-1251 VME to Multibus adapter 560 560
1 3/180S-960 76-inch full height rack 2,730 2,730
--------
25,700
FILESERVER [KSL]
1 3/180S-8-R1 processor, 8 MB memory 8,130 8,130
1 3/180S-650 60 MByte cartridge drive 2,800 2,800
1 370-1012 SMD disk controller 2,450 2,450
1 540-1251 VME to Multibus adapter 560 560
1 3/180S-960 76-inch full height rack 2,730 2,730
1 SUN SRC Unix source code 1,000 1,000
--------
17,670
WORKSTATIONS [KSL]
3 3/75M-4 Workstation 11,130 33,390
WORKSTATIONS [MM]
2 3/75M-4 Workstation 11,130 22,260
COLOR WORKSTATIONS [LMM]
2 3/160C-8 Color Workstation 27,230 54,460
2 3/160C-201 Graphics Processor 4,130 8,260
--------
62,720
FILESERVER UPGRADES [CDS]
2 3/180S-8 server (8 MB) 18,130 36,260
2 540-1251 VME-to-Multibus adapter (SMD) 560 1,120
1 3/180S-960 76-inch full height rack 2,730 2,730
--------
40,110
MULTICAST AGENTS [CDS]
1 3/75M-4 workstation 11,130 11,130
GATEWAY [CDS]
2 3004 Sun-3 single board computer 9,485 18,970
1 540-1248 Backplane w/hardware 665 665
--------
19,635
NON-CAPITAL [CDS]
1 530-1079-10 72" external command cable 200 200
1 530-1097-11 24" internal command cable 70 70
1 530-1080-10 72" external data cable 100 100
3 530-1064-11 24" internal data cable 35 105
2 340-1340-01 SMD panel 35 70
2 340-1124-01 Support brackets (set) 10 20
2 530-1079-01 24" external command cable 100 200
--------
765
--------
254,380
System Industries
Quantity Model & Description Unit Price Total
FILESERVER [CF]
2 Model 9751 Fujitsu disk drive, cable set 7,800 15,600
FILESERVER [KSL]
2 Model 9751 Fujitsu disk drive, cable set 7,800 15,600
--------
31,200
Tektronix
Quantity Model & Description Unit Price Total
[CDS]
1 2467 Oscilloscope with options 09, 11 & 2A 11,629 11,629
--------
11,629
Miscellaneous
Quantity Model & Description Unit Price Total
[CF]
12 Ethernet transceivers and cables 300 3,600
[MISC]
1 Miscellaneous costs 19,186 19,186
--------
22,786
TOTALS BY GROUP New Change from
Total preceding budget
[CDS] Cheriton's Disintegrated Systems 89,269 +6,070
[CF] Computer Futility Group 178,877 +41,937
[EP] Conflict-of-interest Printer 10,356 +10,356
[KSL] Knowledge Syncophancy Lab 72,660 +6,000
[LMM] Linton's Multimedia Madness 62,720 +13,860
[MM] Manna's Moochers 22,260 0
[NAXOS] Gio's Noxious Waste 4,488 +4,488
[QLISP] McCarthy's Misconceptions 120,000 0
[MISC] Miscellaneous Slush Fund 19,186 -82,711
======== ========
579,816 0
∂05-May-86 0817 EPPLEY@SU-SCORE.ARPA Rutie
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 5 May 86 08:17:01 PDT
Date: Mon 5 May 86 08:17:31-PDT
From: LaDonna Eppley <EPPLEY@SU-SCORE.ARPA>
Subject: Rutie
To: JMC@SU-AI.ARPA
cc: Bscott@SU-SCORE.ARPA
Message-ID: <12204280948.18.EPPLEY@SU-SCORE.ARPA>
Rutie called that she would not be able to come in today. If you need
anything, please let me know.
LaDonna
-------
∂05-May-86 0900 JMC
ring and hat size now.
∂05-May-86 0906 RICHARDSON@SU-SCORE.ARPA Jean-Claude Latombe
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 5 May 86 09:06:46 PDT
Date: Mon 5 May 86 09:05:53-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Jean-Claude Latombe
To: cannon@SU-SIERRA.ARPA, jmc@SU-AI.ARPA, binford@SU-WHITNEY.ARPA,
reynolds@SU-SCORE.ARPA, bscott@SU-SCORE.ARPA, grosof@SU-SUSHI.ARPA,
rv@SU-AI.ARPA
Message-ID: <12204289755.13.RICHARDSON@SU-SCORE.ARPA>
Jean-Claude Latombe, a robotics candidate, will be visiting CSD on May 8.
I propose that the search committee meet with him from 2:30 - 3:30 that
day. Please let me know if this will work for you.
-Anne
-------
∂05-May-86 0930 CLT dinner with edwina and oliver
edwina says oliver might be here thursday evening (May 8)
and if so could we come then. Its ok with me. What about you
∂05-May-86 0949 CONTRERAS@SU-SCORE.ARPA message
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 5 May 86 09:49:41 PDT
Date: Mon 5 May 86 09:15:10-PDT
From: Tina Contreras <CONTRERAS@SU-SCORE.ARPA>
Subject: message
To: JMC@SU-AI.ARPA
Message-ID: <12204291443.21.CONTRERAS@SU-SCORE.ARPA>
Janet of Dr. Lipset's office phoned, regarding a Luncheon at the Faculty
club he would like you to attend. 3-4741.
Tina
-------
∂05-May-86 0959 CLT dinner with edwina and oliver
well if Timothy would rather chew his toe than have dinner
we'll let him
∂05-May-86 1016 BIL@SU-SIERRA.ARPA Have you ever made a mistake?...
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 5 May 86 10:16:46 PDT
Date: Mon 5 May 86 10:17:55-PDT
From: Bil Lewis <BIL@SU-SIERRA.ARPA>
Subject: Have you ever made a mistake?...
To: JMC@SU-AI.ARPA
John,
This Sat. I am going to be taking part in a debate on the
subject of SDI (I'll be defending, a challanging position!). As I
fully expect the opposition to quote lots of folks about how
impossible the technological requirements are to meet, I'd like
to have a good rebuttal.
It seemed to me that one effective tact would be to point out
how even the very best scientists can be completely wrong. Rutherford's
comments about the impracticality of splitting the atom pop to mind as
a good choice. Unfortunately, he isn't going to be the most recognizable
figure for many of the judges. You, however, may be.
So, I was wondering... Do you recall having make any notable
predictions which completely missed the mark and would also be
easily comprehensible for a semi-technical audience? And of course
something you wouldn't mind being quoted for.
my thanks, -Bil
-------
∂05-May-86 1023 VAL re: Comments on your paper; conjecture settled
[In reply to message sent 3 May 86 10:33:54 PDT.]
Maybe you want to make interpretations partial functions: a model of π assigns
truth values to parts of π, but not necessarily to other formulas. In this case
we should probably replace conditions 2, 3 and 4 in the definition of
interpretation by something like this: if T is assigned to Nπ1,...,Nπk and π is
a logical consequence of π1,...,πk then only T can be assigned to Nπ. That may
be a way to eliminate Joe's disjunctions.
∂05-May-86 1347 RICHARDSON@SU-SCORE.ARPA Robotics Search Committee
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 5 May 86 13:47:44 PDT
Date: Mon 5 May 86 13:41:22-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Robotics Search Committee
To: cannon@SU-SIERRA.ARPA, reynolds@SU-SCORE.ARPA, binford@SU-WHITNEY.ARPA,
jmc@SU-AI.ARPA, grosof@SU-SUSHI.ARPA, rv@SU-AI.ARPA
Message-ID: <12204339904.17.RICHARDSON@SU-SCORE.ARPA>
We would like to have a follow-up robotics search committee meeting as soon
as possible after Latombe's visit. Will you be available to meet on Monday,
May 12 from 4:00 - 5:00?
-Anne
-------
∂05-May-86 1502 CLT
i did phone cate
∂05-May-86 1531 LES Request for IBM RT PCs for Editor-based Operating System
To: Eustis@SU-SIERRA.ARPA
CC: Nilsson@SU-SCORE.ARPA, JMC@SU-AI.ARPA
We appreciate having been given the use of an IBM RT in support of the
Editor-based Operating System (EBOS) research that we outlined in a
message to you on November 1, 1985. I understand that the unit we have is
considered a "loaner" and that its assignment to this project has not yet
been formally approved. We wish to reconfirm our interest in having the
continued use of this machine. In fact, we could make very good use of a
second machine of similar characteristics if one could be made available
within the next two or three months.
So far we have connected the first unit to our Ethernet and some people in
the department have managed to get a Common Lisp system working, which is
one of the things we needed in support of EBOS. It appears that we will
shortly have a site license that will make this software, from Kyoto
University, available to other RT users at Stanford at no cost. It will
also likely become a commercial product outside Stanford.
We would appreciate it if you and your committee would favorably consider
our request for the use of two RTs. If you would like to have additional
technical descriptions of our plans, we will be happy to make them
available.
Les Earnest
∂05-May-86 1625 VAL Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
ON THE RELATION BETWEEN DEFAULT THEORIES AND AUTOEPISTEMIC LOGIC
Kurt Konolige
SRI International and CSLI
Thursday, May 8, 4pm
MJH 252
Default theories are a formal means of reasoning about defaults: what
normally is the case, in the absence of contradicting information.
Autoepistemic theories, on the other hand, are meant to describe the
consequences of reasoning about ignorance: what must be true if a
certain fact is not known. Although the motivation and formal
character of these systems are different, a closer analysis shows that
they bear a common trait, which is the indexical nature of certain
elements in the theory. In this talk I will show how default theories
can be reanalyzed as a restricted type of indexical theory. The
benefits of this analysis are that it gives a clear (and clearly
intuitive) semantics to default theories, and combines the expressive
power of default and autoepistemic logics in a single framework.
∂05-May-86 1706 MOGUL@SU-SCORE.ARPA re: Purge of dissident scientists? (from SAIL's BBOARD)
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 5 May 86 17:06:22 PDT
Date: Mon 5 May 86 17:06:25-PDT
From: Jeffrey Mogul <MOGUL@SU-SCORE.ARPA>
Subject: re: Purge of dissident scientists? (from SAIL's BBOARD)
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Fri 2 May 86 17:45:00-PDT
Message-ID: <12204377231.17.MOGUL@SU-SCORE.ARPA>
Perhaps I phrases that wrong. A more pejorative statement would be
"The administration is selling SDI to the public as a way of avoiding
serious arms control, because for whatever reason the administration is
unwilling or unable to negotiate with the Soviets."
I think the country has arms control options, but the administration
doesn't want to control arms. Certainly the recent reversal of long-
standing US policy on a comprehensive test ban indicates that the
administration is unwilling to enter into even a completely verifiable
treaty, so it isn't simply distrust of the USSR.
My own guess is that many factions are to blame; some think we can
outspend the USSR, some don't trust them to observe treaties, some
are building a first-strike system (that's what is most dangerous
about SDI) and some just want to enrich the weapons industry. There
may even be people with sincere motives, but I don't think it matters.
-------
∂05-May-86 1803 pjd@RIACS.ARPA Scientific Contributions of CS
Received: from RIACS-HYDRA.ARPA by SU-AI.ARPA with TCP; 5 May 86 17:58:03 PDT
Date: Mon, 5 May 86 17:52:13 pdt
From: Peter Denning <pjd@RIACS.ARPA>
Message-Id: <8605060052.AA17441@hydra.ARPA>
Received: by hydra.ARPA (4.12/4.01)
Mon, 5 May 86 17:52:13 pdt
To: jeh@cornell
Subject: Scientific Contributions of CS
Cc: avd%brown@csnet-relay, gries@cornell, lampson@decwrl, mccarthy@su-ai
John,
I just received your draft document and have skimmed it.
It gives me a considerable amount of heartburn. Let me
explain why.
In late 1984, American Scientist invited me to write
a regular column for them on the Science of Computing. The
purpose of that column was to expose scientists in many disciplines
to the richness of computer science. I gave considerable
thought -- long and hard -- to the question of how to
present computer science in such a way as to achieve this
goal. I wanted to identify the intellectual achievements
of the the field. I wanted to avoid stating as fundamental
questions things that will change when the technology
changes. I wanted to break away from the naive Scientific
American view of computer science, which is that it is either
theory or microelectronics. I wanted to achieve a balanced
perspective of the field. One of the results of this
process was my first article, ``What is computer science?''
published in January 1985.
An effect of this was that ACM asked me to chair a committee
to identify the core of computer science and recommend a new
approach to introducing students to it. I agreed to do this
and organized the ACM task force on the core of computer science.
The task force has recently revised the AmSci article and
is ready to incorporate it into the core definition.
David Gries has a copy.
The document you are drafting portrays computer science in
what I call the naive Scientific American style: it makes the
field seem to an outsider as mostly theory, with a handful
of obeisances to hardware. Where are the intellectual
accomplishments in programming methodology? Architecture?
Networks? Operating Systems? Parallel computing? Graphics?
In your draft, there are a few nods to distributed computing,
but nothing so solid as to complexity theory, cryptography, algorithm
theory, artificial intelligence, numerical analysis, primality
testing, and the like. The draft document gives a very
imbalanced view of the field. This may reflect the interests
of the committee.
It seems to me that those of us who are interested in explicating
computer science to outsiders, in a way that makes clear the
basis on which we believe it to be a peer of older sciences,
ought to work together and give one story. I think it would be
bad for the field if the NSF commissioned a subcommittee that
issued different recommendations from another committee whose
report is likely to be endorsed by ACM and IEEE. I think the fact
that two incongruent reports surfaced at about the same time
would confirm the perception among physicists, chemists, and
others, that we don't have our act together, undoing any
of the good work either report sought to achieve.
I repeat that the draft document does a disservice to the field
by giving an excessive amount of emphasis to the intellectual
achievements on the theory side. There are many intellectual
achievements in other areas as well. Your report would confirm
the impression that the most solid aspects of computer science
are those deriving from mathematics, which undermines the
feeling within the field that the mathematics and engineering
are inseparable, undermines the effort to get wider use
of experimental methods, and undermines the effort to build
up computing laboratories around the country.
Let's have a conversation on the phone about this and see
if we can find a way that our two efforts will lead in the
same direction and help the field.
Peter
415-694-6139
∂05-May-86 2014 LES
To: JMC@SU-AI.ARPA, greep@SU-AI.ARPA, CLT@SU-AI.ARPA
Draft EBOS proposal
[Here is a modified concatenation of texts written by Greep and JMC.
I have added or modified a few words here and there but tried to stick
to the basic story. I also added some words at the end.
This is still a rather sparse proposal, I think, and would be improved by
saying more about how we plan to proceed. Please let me know of any
disagreements. Meanwhile I will put together still another budget.
We likely will want to say things a bit differently in versions aimed
at IBM and DARPA. This one is aimed at IBM, though it doesn't use IBM
terminology. -Les]
Motivation
Computer operating systems that allow direct user interaction have now
been commercially available for over 20 years, yet the user interfaces on
most general-purpose computer systems are still not markedly different
from the earliest ones. While there is some benefit to long-term
stability, the experience gained in two decades of using these systems
points out a number of design limitations. Furthermore, early systems
were designed for configurations with small memories and hardcopy
terminals, thus restricting the possibilities for creating more reasonable
user interfaces.
Many CRT terminals built in the last several years have the capability of
changing arbitrary parts of the screen, thus making it possible to have a
user interface that allows considerably greater flexibility than the
simple prompt/command model currently enployed. In addition, the trend is
towards workstations that provide graphics facilities. The dramatic
decline in memory prices over the last decade means that even a
workstation intended for a single user has enough memory so that the
system designers no longer need to spend the bulk of their effort in
keeping the code down to a minimum.
Unfortunately, few systems for workstations make good use of their
capabilities. A system running on many general-purpose workstations is
Unix, which was designed for computers with very small memories and
hardcopy terminals. Some vendors have multi-window systems that are
useful but do not provide any major increase in functionality. On
Digital's latest version of their system, for example, each window acts
exactly like a terminal on a conventional Unix system, so the only
advantage is that multiple terminals can be simulated. Sun Microsysems
has a somewhat more advanced system, in which some editing functions can
be performed on the data in the window. There is also a version of the
Emacs editor that provides this capability in Unix. However, these
systems are still basically designed around the teleprinter model of
interaction.
Although individual utility programs that are not teleprinter-oriented,
such as text editors, have been available for many years, they constitute a
distinct minority and there is no standard model of interaction on which
they are based. A typical program thus has no provisions for correcting
or re-entering any data, other than the primitive backspace and
line-delete characters.
Another feature lacking or not cleanly provided in existing systems is the
ability to combine several programs. Experience with pipes in Unix has
shown that having a number of utility programs and a simple way to
interconnect them greatly simplifies many everyday tasks. The pipe
mechanism in Unix is restricted in practice to a single input stream and a
single output stream per process. This limitation is not inherent in
Unix, but there is no standard convention for the use of streams other
than standard input, standard output, and standard error, and the shell
does not even have a good way of dealing with standard error.
Most systems either do not provide programmability at the command level
or do so in an idiosyncratic way. This means that a new language must be
learned and that facilities which are available in conventional programming
languages, e.g. debuggers, compilers, and syntax-directed editors, are
generally nonexistent for command level programs.
A further problem with existing systems is that all character handling
procedures are based on a limited character set, usually ASCII. This is
marginally adequate for normal English prose but does not meet the needs of
technical writing (e.g. math and logic symbols) or even of prose in most
other languages. On a system based around a bitmap display, there is no
reason to impose this limitation. The ability to incorporate a variety of
symbols and general graphical data would be very useful both in
constructing programs that provide graphics functions and in composing
textual materials that use special symbols.
We propose to design and implement a new kind of system that will rectify
the shortcomings outlined in the preceding paragraphs.
Proposed Development
The main premise of the project is that the proper environment for the
interactive user of a computer is a display editor. This provides the
ability to re-edit both previous input and new output in order to generate
new input and also to file the output as desired. Various other
improvements in editors and operating systems are planned. Thus the goals
of the project include the following.
1. The shell and interactive programs will be normally
operated out of a new advanced display editor. Programs for which
this is not suitable will be able to control the interaction themselves.
2. We will explore the possibility of writing the editor and our
other programs in Common Lisp. The full facilities of Common Lisp would
then be available for macros for controlling the editor itself, the
operating system and user programs. Our experience has shown us that if
the editor provides good enough interactive facilities, then many user
programs can rely on them thus allowing simpler programming of interactive
programs.
3. The editor and the operating system will be kept fully
programmable. Thus anything a person can do interactively, he will
be able to write programs to do. This means that all output seeable
by a user must also be readable by programs. Interactive programs
that use the display in a non-standard way will be able to violate
this condition. System status information will be maintained in editable
files.
4. The editor and file system will accomodate arbitrary
character sets. This advance is easy to make in the editor itself and in
bit map displays and modern printers, but keyboards present a problem for
which we intend to provide and explore a variety of solutions. The editor
will be constructed in such a way that it will be possible to add graphical
constructs such as line drawings and textured areas, though these features
will likely not be included in the initial implementation.
5. Our planned solution to the arbitrary character set problem
will allow both for standard keyboards and for keyboards adapted to
special tasks, e.g. the use of mathematics, APL or foreign languages.
Thus a special character can appear on the screen either because the user
has pressed a key on a special keyboard or because he has pressed a
suitable sequence of keys on an ordinary keyboard. The same sequence of
bytes will inhabit the file in either case.
6. The system will permit processes with multiple inputs and
outputs to be linked together to accomplish a given task. The user
will still be able to monitor and control the process network.
Initial Implementation
We plan to put together an experimental system based on the ideas outlined
above using an IBM RT PC running a Unix kernal. We are currently undecided
on whether this system should be based on the Berkeley 4.2 or AIX versions
of Unix, but it appears that 4.2 would be easier to use. We expect to do
most of our programming in Common Lisp.
A new shell, including the proposed display editor, and a number of
utility programs must be written. The emphasis will be on getting the
interactive qualities of the system right in the first implementation
rather than on optimizing performance. We think that a system showing the
basic ideas can be assembled in about one year within an effort of about
two person-years.
If the initial implementation is successful, it likely will be desirable
to do a deeper reimplementation, including a complete new operating
system, that emphasizes processing efficiency. Toward this end, we plan
to limit and structure the interfaces between the programs that we write
and the Unix environment so that it will be possible to embed our programs
in the new environment without having to completely rewrite them.
∂05-May-86 2246 pjd@RIACS.ARPA Further Comments
Received: from RIACS-HYDRA.ARPA by SU-AI.ARPA with TCP; 5 May 86 22:46:17 PDT
Date: Mon, 5 May 86 22:39:11 pdt
From: Peter Denning <pjd@RIACS.ARPA>
Message-Id: <8605060539.AA19575@hydra.ARPA>
Received: by hydra.ARPA (4.12/4.01)
Mon, 5 May 86 22:39:11 pdt
To: jeh@cornell
Subject: Further Comments
Cc: avd%brown@csnet-relay, gries@cornell, lampson@decwrl, mccarthy@su-ai
John,
I've read the draft report on the contributions of computer
science and would like to offer some further comments. I
regret that some of my comments may stike you as quite strongly
worded; I feel quite strongly on these topics. I made a change
of career from Purdue to RIACS, from academia to research lab,
because I felt strongly about working together with scientists
from other disciplines on problems where traditional science
and computer science can make equal contributions. In the
course of this pursuit, I have worked side by side for three
years with computational physicists, chemists, fluid dynamicists,
human factors experts, flight simulation experts, biologists,
and others in the audience for whom your report is intended.
I think I have developed a pretty clear perception of what is
on their minds -- where are their perceptions of computer
science, what are their skepticisms, etc. I would like very
much to see traditional scientists come to understand and
perceive computer science as a peer science, not as a technology
that can be handled by any competent lab technician in their
disciplines.
What are the perceptions of computer science as seen by these
other scientists? My strongest impression is that they feel
computer scientists are too pontifical: when they try to
interact with us, they get lectures on what is the best
programming language, algorithm technology, computational
model, or computer architecture. They have a perception
that we computer scientists want to tell them about all our algorithms and
programming theories, but we don't really want to listen to their
problems and enter a true partnership to solve them. This is an
unfortunate perception, but it is real. In my opinion, we should
avoid describing computer science in terms that will play into this
perception.
Other scientists also have a perception that computer people are of
exactly two types: highly mathematical ones with little interest
in practical problems in science, and hackers. Many experimental
physicists or chemists view the mathematical types with the same
jaundice as the pure-theory types in their own disciplines. They
view the hackers as technological wizards, whom they equate with
their own lab technician wizards, whom they believe are extremely
valuable, but whom they do not regard as scientists. They perceive
that the computer scientists who are experimentalists -- i.e.,
have some appreciation for the theory and know how to set up and
run a genuine experiment -- are few in number. You may not like
these perceptions, but they are real. We have to deal with them.
Unfortunately, in my opinion, your draft report does nothing to
dispel these illusions. It plays right into the weakest perceptions
other scientists have of computer science. Examples:
1. Virtually all the discussion extols the virtues of the
mathematical contributions of computer science. There is
little about the intellectual achievements in microelectronics,
architecture, software engineering, etc. -- even though other
scientist percieve that there have been significant achievements
in these areas.
2. Virtually all the persons named in the draft are mathematicians,
and the most glowing descriptions are reserved for the Turing
Award winners who happen to be members of the committee. This
will not go unnoticed by your intended audience.
3. Most of the problems cited in the report as fundamental sound
esoteric to experimental scientists. Do you expect anyone to
take you seriously when you characterize ``Bob knows that Alice
knows that Bob loves her,'' as exemplifying a fundamental
question in computer science? Or when you say that ``gridlock''
has actually been understood longer by computer scientists
than by New York police? Or that moviegoers who must deal
with unsynchronized clocks and uncertain message delivery times
face similar problems with computers on a network? Statements
like these make our discipline sound trivial to outsiders.
4. There is no serious attempt in the draft report to
identify fundamental intellectual in the more experimental
areas of computer science -- e.g., programming languages,
programming methodology, architecture, database and
information retrieval systems, human interface,
parallel systems, distributed systems, or operating systems.
Aren't Algol, Fortran, Pascal, and LISP intellectual
achievements? What about assertion-oriented programming?
The stored-program computer? Fast hardware units for
arithmetic? Relational databases? Best-match document
systems? Core graphics systems? E&S graphics workstations?
Windows and mice? CSP? Systolic processing? Dataflow
computation? Client-server LANs? Network protocols?
UNIX? Level structured operating systems? If you think these
items do not include fundamental intellectual accomplishments
that can be fully appreciated by other scientists, you need
to rethink your whole perception of who these other scientists
are. I realize that's a strong statement, but I mean it.
5. Some of the favorite current theory areas are identified
as fundamental even though most of the world outside the
group of immediate researchers and funding agents would not
agree. I am thinking of cryptography, primality testing,
semantics, temporay logic, and some (but not all)
the results quoted from distributed computing.
I repeat: your draft report and the report being put together
by the ACM task force are incongruent and it will not serve
the field to issue two reports that are at odds with one
another. Let's talk as soon as possible to find a way of
coordinating our efforts.
Peter
415-694-6139
∂06-May-86 0740 RICHARDSON@SU-SCORE.ARPA Principia Group
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 6 May 86 07:40:00 PDT
Date: Tue 6 May 86 07:40:25-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Principia Group
To: jmc@SU-AI.ARPA
cc: ra@SU-AI.ARPA
Message-ID: <12204536339.14.RICHARDSON@SU-SCORE.ARPA>
This is to reconfirm that you will be meeting today, (May 6), with Nils
Nilsson and his Principia Group in MJH 252 from 1:30 - 3:15.
-------
∂06-May-86 0956 RA John Kwapisz
Please call John Kwapitz from Science and Engineering on SDI; his. tel.
(202) 547 5607.
Also, Franklin from Dina Bolla returned your call. He said that he had tried
to call you yesterday but couldn't reach you.
∂06-May-86 1011 RA Re: hat and ring sizes and trip
[Reply to message recvd: 05 May 86 14:51 Pacific Time]
Re hat size I called someone who exports hats to Europe and he claims that
7 1/2 American = 60 European. As to the ring I was told that a ring size
is an international measurement.
Rutie
-----
∂06-May-86 1028 BURY@su-sushi.arpa CS 306 grade from Fall, remote
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 6 May 86 10:27:11 PDT
Date: Tue 6 May 86 10:26:49-PDT
From: Robert Bury <BURY@su-sushi.arpa>
Subject: CS 306 grade from Fall, remote
To: jmc@su-ai.arpa, givan@su-sushi.arpa
Message-ID: <12204566631.24.BURY@su-sushi.arpa>
Professor McCarthy, Bob Givan:
Has any progress been made on grading my assignments and final exam
from last fall's CS 306, which I took as a remote student?
I would appreciate your cooperation in completing this course.
Robert Bury
-------
∂06-May-86 1040 NILSSON@SU-SCORE.ARPA Re: Batali
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 6 May 86 10:37:27 PDT
Date: Tue 6 May 86 10:36:39-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Re: Batali
To: JMC@SU-AI.ARPA
cc: NILSSON@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Tue 6 May 86 00:43:00-PDT
Message-ID: <12204568420.25.NILSSON@SU-SCORE.ARPA>
The MacQueen matter is just an instance of a more general problem
that arises if the theory search committee recommends an immediate
filling of the Andy Yao slot. Although it is true that if we don't
act now we may lose a theory person we might like to have come, it
is also true that we lose someone (like MacQueen) whom we might
like to have come later. (Because I don't expect that we can
get approval for another theory person for a couple of years.) Does
the theory search committee want to reconsider their most recent
recommendation? -Nils
-------
∂06-May-86 1040 NILSSON@SU-SCORE.ARPA
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 6 May 86 10:37:38 PDT
Date: Tue 6 May 86 10:37:05-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
To: jmc@SU-SCORE.ARPA
Message-ID: <12204568500.25.NILSSON@SU-SCORE.ARPA>
I sent you the wrong msg. Pse ignore.
-------
∂06-May-86 1119 RICHARDSON@SU-SCORE.ARPA re: Robotics Search Committee
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 6 May 86 11:19:02 PDT
Date: Tue 6 May 86 11:19:28-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: re: Robotics Search Committee
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Mon 5 May 86 13:49:00-PDT
Message-ID: <12204576215.14.RICHARDSON@SU-SCORE.ARPA>
John,
What about the meeting on May 8? Will you be able to attend that?
-Anne
-------
∂06-May-86 1159 RICHARDSON@SU-SCORE.ARPA Robotics Search Committee/May 14
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 6 May 86 11:57:53 PDT
Date: Tue 6 May 86 11:57:56-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Robotics Search Committee/May 14
To: cannon@SU-SIERRA.ARPA, reynolds@SU-SCORE.ARPA, binford@SU-WHITNEY.ARPA,
jmc@SU-AI.ARPA, grosof@SU-SUSHI.ARPA, rv@SU-AI.ARPA
Message-ID: <12204583218.14.RICHARDSON@SU-SCORE.ARPA>
May 12 does not appear to be a good day for the follow-up Robotics Search
Committee. What about May 13 from 4:30-5:30?
-Anne
-------
∂06-May-86 1325 RA John Kwapisz
Kwapitsz from Science and Engineering on SDI called again; very anxious
to talk to you. His tel. (202) 547 5607.
∂06-May-86 1338 LES Facilities Budget Revision
To: facil@SU-AI.ARPA, "@XTRA.DIS[1,LES]"@SU-AI.ARPA
The budget that I distributed Sunday night saying that we could buy
everything on the list seemed too good to be true. It was.
While biking to work yesterday morning I remembered that I had deferred
entering Tom Binford's requested Sun workstation because I did not have a
complete specification.
Having gotten the specification and entered it, I find that we can buy
only down to Priority Level 6 (Lantz's hardware) and must drop the two
Laserwriters and the Ethernet transceivers. Even at that, we will have
only $15k for miscellaneous shipping charges and cabling, which is a bit
tight.
I also learned that Laserwriters are now available for only $3.8k.
If we decide to drop the Proteon gateway from our buying list, then,
we should be able to get both Laserwriters and a few extra transceivers.
Sorry for the muddle.
∂06-May-86 1402 binford@su-whitney.arpa Jean-Claude Latombe
Received: from SU-WHITNEY.ARPA by SU-AI.ARPA with TCP; 6 May 86 14:02:33 PDT
Received: by su-whitney.arpa with Sendmail; Tue, 6 May 86 13:59:16 pdt
Date: Tue, 6 May 86 13:59:16 pdt
From: Tom Binford <binford@su-whitney.ARPA>
To: RICHARDSON@SU-SCORE.ARPA
Cc: cannon@SU-SIERRA.ARPA, jmc@SU-AI.ARPA, reynolds@SU-SCORE.ARPA,
bscott@SU-SCORE.ARPA, grosof@SU-SUSHI.ARPA, rv@SU-AI.ARPA
In-Reply-To: Anne Richardson's message of Mon 5 May 86 09:05:53-PDT
Subject: Jean-Claude Latombe
I have it on my calendar now.
∂06-May-86 1405 binford@su-whitney.arpa Robotics Search Committee
Received: from SU-WHITNEY.ARPA by SU-AI.ARPA with TCP; 6 May 86 14:05:35 PDT
Received: by su-whitney.arpa with Sendmail; Tue, 6 May 86 14:02:18 pdt
Date: Tue, 6 May 86 14:02:18 pdt
From: Tom Binford <binford@su-whitney.ARPA>
To: RICHARDSON@SU-SCORE.ARPA
Cc: cannon@SU-SIERRA.ARPA, reynolds@SU-SCORE.ARPA, jmc@SU-AI.ARPA,
grosof@SU-SUSHI.ARPA, rv@SU-AI.ARPA
In-Reply-To: Anne Richardson's message of Mon 5 May 86 13:41:22-PDT
Subject: Robotics Search Committee
That coincides with robotics seminar, but
I will come.
∂06-May-86 1639 RICHARDSON@SU-SCORE.ARPA Robotics Search Committee/May 21
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 6 May 86 16:27:39 PDT
Date: Tue 6 May 86 16:24:13-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Robotics Search Committee/May 21
To: jmc@SU-AI.ARPA
Message-ID: <12204631695.11.RICHARDSON@SU-SCORE.ARPA>
One more proposal for the robotics search committee follow-up meeting...
May 21 (Wednesday) from 4:00 - 5:00. Will this work for you?
-Anne
-------
∂06-May-86 1831 binford@su-whitney.arpa Robotics Search Committee/May 14
Received: from SU-WHITNEY.ARPA by SU-AI.ARPA with TCP; 6 May 86 18:31:19 PDT
Received: by su-whitney.arpa with Sendmail; Tue, 6 May 86 18:27:58 pdt
Date: Tue, 6 May 86 18:27:58 pdt
From: Tom Binford <binford@su-whitney.ARPA>
To: RICHARDSON@SU-SCORE.ARPA
Cc: cannon@SU-SIERRA.ARPA, reynolds@SU-SCORE.ARPA, jmc@SU-AI.ARPA,
grosof@SU-SUSHI.ARPA, rv@SU-AI.ARPA
In-Reply-To: Anne Richardson's message of Tue 6 May 86 11:57:56-PDT
Subject: Robotics Search Committee/May 14
OK. I have CS colloquium this quarter, but I can start things
and come to the meeting.
∂06-May-86 2000 SJM
productivity and construction cry out for help
∂07-May-86 0656 JJW RA support for summer
To: JMC
CC: LES
Will you be able to support me as a research assistant this summer?
∂07-May-86 0955 CLT rt unix upgrade
To: "@RT.DIS[1,CLT]"@SU-AI.ARPA
The new version of Unix for the RT will be installed
this Friday afternoon. The process will destroy the
current world entirely. I will back up the system
sometime after 11pm thursday. If you do anything after
that time it is likely to disappear, so you should
arrange to save it elsewhere temporarily.
The directory will be restructured when we restore the
world - mainly to facilitate dumping and keeping track
of our modifications. More on this later.
∂07-May-86 1114 NILSSON@SU-SCORE.ARPA [Tom Garvey <Garvey@SRI-AI.ARPA>: Danny Cohen visit]
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 7 May 86 11:14:05 PDT
Date: Wed 7 May 86 10:23:25-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: [Tom Garvey <Garvey@SRI-AI.ARPA>: Danny Cohen visit]
To: jmc@SU-AI.ARPA
Message-ID: <12204828156.42.NILSSON@SU-SCORE.ARPA>
fyi
---------------
Return-Path: <GARVEY@SRI-AI.ARPA>
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Wed 7 May 86 10:12:43-PDT
Date: Wed 7 May 86 10:09:59-PDT
From: Tom Garvey <Garvey@SRI-AI.ARPA>
Subject: Danny Cohen visit
To: AIC-Staff: ;
Folks,
For those of you who are interested, Danny Cohen
(Dr.-software-can-handle-the-SDI-battle-management-function-reliably)
will visit SRI on Monday, May 19. He will review the Eastport Study
and field questions (aka, the-panel-that-Parnas-quit (and reviled))
from 1115 to 1215 in Bldg. 408, Room 95. I have been asdvised that
SRI visitors are welcome, so this is your big chance to find out just
what Star Wars has in common with the phone system.
Cheers,
Tom
-------
-------
∂07-May-86 1150 GIVAN@su-sushi.arpa summer work
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 7 May 86 11:50:25 PDT
Date: Wed 7 May 86 11:50:06-PDT
From: Bob Givan <GIVAN@su-sushi.arpa>
Subject: summer work
To: jmc@su-ai.arpa
Message-ID: <12204843935.15.GIVAN@su-sushi.arpa>
Hello Prof. McCarthy,
I called last night between 8:30 and 9 pm but could not reach you...
I hope to be able to find you in today, as i must make my job decisions
very shortly, and need to briefly discuss such with you.
I will come by at 1:40pm to try and meet...if this is a bad time could you
send me mail regarding a better one? (any time later today, preferably after
4 pm)
Thanks,
Bob Givan.
-------
∂07-May-86 1230 RA John Coke
John Coke called; he said you have his number.
∂07-May-86 1233 CLT dinner on Thursday.
Oliver is free on Thursday. Shall I tell Edwina we can come.
(She say's Oliver's 60th is Sat)
∂07-May-86 1235 CLT calendar item
mon 7-jul - wed 9-jul CLT @ Program Transformation Workshop -- Harvard
∂07-May-86 1332 LES Draft EBOS proposal
To: JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, greep@SU-AI.ARPA
[Carolyn pointed out that I used JMC's earlier EBOS text rather than
the stuff he wrote last. (Don't understand why JMC didn't complain.)
This edition is essentially unchanged up to the section on "Proposed
Development" but is based on the newer text thereafter.
I just received a call from Eustis responding to my request for a second
RT. He said that he wants a current description of the proposed research
project. I propose to send him this stuff, incorporating any further
ideas that emerge soon. Comments? Suggestions? -Les]
Motivation
Computer operating systems that allow direct user interaction have now
been commercially available for over 20 years, yet the user interfaces on
most general-purpose computer systems are still not markedly different
from the earliest ones. While there is some benefit to long-term
stability, the experience gained in two decades of using these systems
points up several design limitations. Furthermore, early systems were
designed for configurations with small memories and hardcopy terminals,
thus restricting the possibilities for creating more reasonable user
interfaces.
Many CRT terminals built in the last several years have the capability of
changing arbitrary parts of the screen, thus making it possible to have a
user interface that allows considerably greater flexibility than the
simple prompt/command model currently enployed. In addition, the trend is
towards workstations that provide graphics facilities. The dramatic
decline in memory prices over the last decade means that even a
workstation intended for a single user has enough memory so that the
system designers no longer need to spend the bulk of their effort in
keeping the code down to a minimum.
Unfortunately, few systems for workstations make good use of their
capabilities. A system running on many general-purpose workstations is
Unix, which was designed for computers with very small memories and
hardcopy terminals. Some vendors have multi-window systems that are
useful but do not provide any major increase in functionality. On
Digital's latest version of their system, for example, each window acts
exactly like a terminal on a conventional Unix system, so the only
advantage is that multiple terminals can be simulated. Sun Microsysems
has a somewhat more advanced system, in which some editing functions can
be performed on the data in the window. There is also a version of the
Emacs editor that provides this capability in Unix. However, these
systems are still basically designed around the teleprinter model of
interaction.
Although individual utility programs that are not teleprinter-oriented,
such as text editors, have been available for many years, they constitute a
distinct minority and there is no standard model of interaction on which
they are based. A typical program thus has no provisions for correcting
or re-entering any data, other than the primitive backspace and
line-delete characters.
Another feature lacking or not cleanly provided in existing systems is the
ability to combine several programs. Experience with pipes in Unix has
shown that having a number of utility programs and a simple way to
interconnect them greatly simplifies many everyday tasks. The pipe
mechanism in Unix is restricted in practice to a single input stream and a
single output stream per process. This limitation is not inherent in
Unix, but there is no standard convention for the use of streams other
than standard input, standard output, and standard error, and the shell
does not even have a good way of dealing with standard error.
Most systems either do not provide programmability at the command level
or do so in an idiosyncratic way. This means that a new language must be
learned and that facilities which are available in conventional programming
languages, e.g. debuggers, compilers, and syntax-directed editors, are
generally nonexistent for command level programs.
A further problem with existing systems is that all character handling
procedures are based on a limited character set, usually ASCII. This is
marginally adequate for normal English prose but does not meet the needs of
technical writing (e.g. math and logic symbols) or even of prose in most
other languages. On a system based around a bitmap display, there is no
reason to impose this limitation. The ability to incorporate a variety of
symbols and general graphical data would be very useful both in
constructing programs that provide graphics functions and in composing
textual materials that use special symbols.
Proposed Development
Present operating systems are unnecessarily inconvenient in a variety of
ways, as discussed above. They were lacking in systematic design to begin
with and have gotten worse as ad hoc features with ad hoc syntaxes have
been added. We propose to design and implement a new kind of system that
will rectify the principal shortcomings outlined above.
The main premise of the project is that the proper environment for the
interactive user of a computer is a display editor. This provides the
ability to re-edit both previous input and new output in order to generate
new input and also to file the output as desired. Various other
improvements in editors and operating systems are planned. Thus the goals
and approach that we will use include the following.
1. The shell and interactive programs will be normally
operated out of a new advanced display editor. Programs for which this is
not suitable will be able to control the interaction themselves. The
editor and the operating system will be kept fully programmable. Thus
anything a person can do interactively, he will be able to write programs
to do. This means that all output seeable by a user must also be readable
by programs. Interactive programs that use the display in a non-standard
way will be able to violate this condition. System status information
will be maintained in editable files.
2. The command language will be a sugared programming language ---
in fact, sugared Lisp. The user will have the full capability of Lisp in
compounding the basic operations into command programs, i.e. macros
(called execs in IBMese). In our examples, we sugar the Lisp by using
conventional functional notation, but "command programs" will have Lisp
internal notation format. Whether they will be represented as list
structure isn't decided. File operations, compiling, etc. will be treated
as functions, e.g. one will have statements like
foo.bin := fortran(foo.for).
The function fortran then represents the effect of the Fortran compiler
and can be composed with other functions. The Unix pipes provide for
composition of functions but in a non-standard syntax and make the curious
assumption that all functions are unary. We will consider carrying this
functional idea to an extreme. Namely, typing a single text character to
a text editor may be regarded as executing an assignment statement of the
form
file := insert(file,char,place).
Operations on files are also obvious candidates to be regarded assignment
statements and will be compoundable in such statements as
foo.bin,foo.error := fortran(append(foo1.for,precompile(foo2.alg)).
The requirement of programmability means that whatever compound operations
the user can do himself, he can conveniently make command programs do for
him. Of course, it must be convenient to include conditionals in command
programs, and this is where most macro and exec systems fall down.
Conditionals are an afterthought and are done awkwardly. Keeping the
command language programmable is to have a higher priority than making the
system controllable by the mouse. Thus any control operations involving a
mouse are sugared versions of the basic language.
3. The command language, the editor language and any others
provided as part of the system will obey what may be called the "Chomsky
principle": the syntax never requires a constant at any point. Chomsky
claimed in his "Reflections on Language" that human languages have this
property; some linguists dispute this. Anyway we want it. It will always
accept a term designating an entity of the required type. Thus whereever
a file name is acceptable, a file-valued expression, e.g. a file-valued
conditional expression, is also acceptable. Many programming languages
have this property, but apparently no operating systems.
4. The sugaring will be both syntactic and semantic. Syntactically,
one will simply type the character while in a context in which typing the
character inserts it in the file. Other operations will be sugared in
appropriate ways, e.g. one can write
(1) fortran(foo)
for the above assignment statement.
5. By semantic sugar we mean that in appropriate contexts, the
operations that are the most obvious interpretations of the base forms
are not carried out, but something easier is done. For example, (1)
above doesn't make a new copy of the file, but merely inserts a character
in the core image of the file.
6. By editor-based, we mean that the user is always interacting
with the system via a powerful editor, e.g. something like EMACS. All
input is available in the editor for re-editing and re-submission, and all
output is available for editing. This facility has been useful in the E
editor on WAITS but requires substantial expansion. Our experience has
shown us that if the editor provides good enough interactive facilities,
then many user programs can rely on them thus allowing simpler programming
of interactive programs.
7. It is time to allow the user arbitrary character sets. This
presents no difficulty for the internal operation of editors, and only
requires that they admit and recognize that some characters are
represented by multiple bytes. Similarly bit map displays (IBMese APA
displays) allow the display of arbitrary characters, but keyboards present
a problem for which we intend to provide and explore a variety of
solutions. The editor will be constructed in such a way that it will be
possible to add graphical constructs such as line drawings and textured
areas, though these features will likely not be included in the initial
implementation.
Our planned solution to the arbitrary character set problem will allow
both for standard keyboards and for keyboards adapted to special tasks,
e.g. the use of mathematics, APL or foreign languages. Thus a special
character can appear on the screen either because the user has pressed a
key on a special keyboard or because he has pressed a suitable sequence of
keys on an ordinary keyboard. The same sequence of bytes will inhabit the
file in either case.
8. EBOS will take advantage of large memories; we are quite
prepared to make a system that won't run with less than 4 megabytes if
that will make the system run faster and more smoothly. The basic
technique is to keep large amounts of directory information in main memory
while the system is running. If a user's command refers to certain files,
our goal is that the only disk operations executed are those that read and
write those specific files.
9. The system will be designed to be efficient with very large
numbers of files and with very large files. If a reader wishes to treat a
book as a single file, that should work efficiently.
(It is our opinion that paging has basically failed in its goal of making
small memories look like large ones. It is useful mainly as a means of
address translation, and introduces inefficiencies even when used for
that. The fastest computers, i.e. the Crays, don't use it. However, it
looks like EBOS will be implemented on computers that have paging, so
we'll have to minimize the effects of its inefficiencies. The same
remarks apply to a lesser degree to caches. In the long run it will
probably turn out that memory hierarchies are more efficiently handled by
compiled software than by built-in hardware features.)
10. The system will permit processes with multiple inputs and
outputs to be linked together to accomplish a given task. The user
will still be able to monitor and control the process network.
Initial Implementation
We plan to put together an experimental system based on the ideas outlined
above using the IBM RT PC running Unix 4.2. A new shell, including the
proposed display editor, and a number of utility programs must be written.
The emphasis will be on getting the interactive qualities of the system
right in the first implementation rather than on optimizing performance.
We think that a system showing the basic ideas can be assembled in about
one year within an effort of about two person-years.
If the initial implementation is successful, it likely will be desirable
to do a deeper reimplementation, including a complete new operating
system, that emphasizes processing efficiency. Toward this end, we plan
to limit and structure the interfaces between the programs that we write
and the Unix environment so that it will be possible to embed our programs
in the new environment without having to completely rewrite them.
∂07-May-86 1413 RA Doug Spitz
Doug Spitz (212) 512 1401 called, re Dr. John Gilmore, who gave your name as
reference.
∂07-May-86 1443 RICHARDSON@SU-SCORE.ARPA Follow-up Robotics Meeting/May 21
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 7 May 86 14:42:53 PDT
Date: Wed 7 May 86 14:38:26-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Follow-up Robotics Meeting/May 21
To: cannon@SU-SIERRA.ARPA, reynolds@SU-SCORE.ARPA, jmc@SU-AI.ARPA,
binford@SU-WHITNEY.ARPA, grosof@SU-SUSHI.ARPA, rv@SU-AI.ARPA
cc: bscott@SU-SCORE.ARPA
Message-ID: <12204874582.33.RICHARDSON@SU-SCORE.ARPA>
There will be a Robotics Search Committee meeting on May 21 from 4:00 - 5:00
in MJH 220. (This is in addition to the meeting tomorrow with Latombe.)
-Anne
-------
∂07-May-86 1533 CLT Remarks on Draft EBOS proposal
To: LES@SU-AI.ARPA, JMC@SU-AI.ARPA, greep@SU-AI.ARPA
Perhaps before the final paragraph of the motivation section
one could say something like
Most operating systems don't provide adequate data structures
to represent information, for communication between programs
or for interaction with users.
One of the winning features of Unix is the uniform treatment
of strings as a means of representing and communicating information.
However strings are often clumsy and inconvenient to use.
In a Lisp-like environment there is an even richer collection of data
structures making it easier to represent many things and to
share information by structure sharing.
re 2. surely there is more power to Lisp than just macros
presumably the full capability of function definition using recursion
and conditional expressions is available too as well as the
choice of interpreting or compiling programs, macros and commands.
In the previous version you had a more detailed paragraph about
interaction with multiple processes through multiple windows in a very
flexible manner. I liked that paragraph and wanted to add the requirement
that `arbitrary' communication among processes be possible. For example
information can be sent from a window belonging to one process to another
process or put in another window.
∂07-May-86 1810 LES re: Remarks on Draft EBOS proposal
To: CLT@SU-AI.ARPA
CC: JMC@SU-AI.ARPA, greep@SU-AI.ARPA
[In reply to message rcvd 07-May-86 15:33-PT.]
1. The idea of sharing data structures between processes is something to
consider, but it is not clear to me that it is practical in a Unix
environment. Rather than specifying this as a goal, I suggest that we
keep it in mind as an "interesting idea if practical." The primary
purpose of the initial implementation is to get the external properties of
the system right; we can afford to compromise a bit on the innerds for
now.
2. Following your suggestion, I expanded the "full capabilities of Lisp"
statement in 2. to include recursion and conditional expressions.
3. I can't find the paragraph you mention involving "interaction with
multiple processes through multiple windows." In fact, in the previous
version, the neither of the words "window" or "windows" appears in the
section on "Proposed Development."
∂07-May-86 1934 LES Datalink alternatives
To: RPG@SU-AI.ARPA, JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, JJW@SU-AI.ARPA
CC: Bosack@SU-SCORE.ARPA, Tom@SU-SCORE.ARPA
The known alternative ways of linking our local Ethernet to one at Lucid
(3 miles away) are microwave, Pac Bell T-1 service, and leased-wire with
modems. The properties of these alternatives are apparently as follows.
1. A home-brew microwave link has allegedly been developed and could be
installed, though there is some uncertainty about the length of time it
would take to get a license. (It is allegedly feasible to operate it in
bootleg mode in the meantime with modest risk.) The construction and
installation cost is believed to be under $30k and the maintenance cost
should be negligible. The lead time would be "a few weeks." It would
provide a data rate of about 3 Mb. [Source: Len Bosack]
2. T-1 data service would use existing Stanford fiber-optic links across
campus and PacBell fibers from there through the central office to Lucid.
It would have an installation cost of about $4.8k plus the cost of an
Ethernet gateway at each end and would cost $1177 per month for the
service. I haven't found out the installation lead time yet. It would
provide a data rate of 1.544 Mb. [Source: Leslie Sanberg, Pac Bell,
542-6806].
3. Some newly-available modems allegedly can provide about 100 kb data
rate over copper wire at a cost of $1.6 k plus two gateways and a
negligible monthly charge for the wires (~ $40). Len and Tom D. have
acquired a pair and will test them in the next couple of weeks. If it
works, the installation time for this should be something like two weeks.
[Source: Tom D.]
If the 100 kb data rate will suffice and if the new modems actually work,
alternative 3 looks most attractive. Failing that, the break-even period
between the microwave and T-1 service is apparently less than 2 years, so
the microwave is attractive for the long haul. It also seems to involve
higher technical risks. Comments are invited.
∂08-May-86 0810 RICHARDSON@SU-SCORE.ARPA AI Faculty Candidate/Shankar
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 8 May 86 08:09:33 PDT
Date: Thu 8 May 86 08:02:28-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Faculty Candidate/Shankar
To: buchanan@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA, shah@SU-SIERRA.ARPA,
hirsh@SUMEX-AIM.ARPA, grosof@SU-SUSHI.ARPA
Message-ID: <12205064642.13.RICHARDSON@SU-SCORE.ARPA>
Natarajan Shankar, an AI faculty candidate, will be visiting CSD on
Thursday, May 15. Following is the schedule that I propose - please let
me know if it will work for you.
11:00 - 11:30 Nilsson
1:15 - 1:45 Buchanan
2:00 - 2:30 McCarthy
2:30 - 3:00 Shah
3:00 - 3:30 Hirsh/Grosof
3:30 - 4:00 Genesereth
5:15 Talk
Thanks,
-Anne
-------
∂08-May-86 0810 RICHARDSON@SU-SCORE.ARPA AI Search Committee
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 8 May 86 08:10:50 PDT
Date: Thu 8 May 86 08:07:23-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Search Committee
To: buchanan@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA, shah@SU-SIERRA.ARPA,
grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA
Message-ID: <12205065535.13.RICHARDSON@SU-SCORE.ARPA>
We would like to have an AI search committee meeting to discuss the candidates
on Monday, May 19 from 3:00 - 4:00. Please let me know if this will work for
you.
Thanks,
Anne
-------
∂08-May-86 0938 ullman@su-aimvax.arpa MS candidate Hirani
Received: from SU-AIMVAX.ARPA by SU-AI.ARPA with TCP; 8 May 86 09:38:42 PDT
Received: by su-aimvax.arpa with Sendmail; Thu, 8 May 86 09:41:16 pdt
Date: Thu, 8 May 86 09:41:16 pdt
From: Jeff Ullman <ullman@diablo>
Subject: MS candidate Hirani
To: jmc@sail, mayr@diablo, oliger@navajo, pratt@navajo
Sounds like he was unaware that he should apply for the
Ph. D. program.
Is he good enough that we want to encourage him to come?
If so, I can offer a TA-ship in 108 for the fall, and there
must be some other courses that could be used to support
him the other quarters.
---jeff
∂08-May-86 0954 VAL Reminder: Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
ON THE RELATION BETWEEN DEFAULT THEORIES AND AUTOEPISTEMIC LOGIC
Kurt Konolige
SRI International and CSLI
Thursday, May 8, 4pm
MJH 252
Default theories are a formal means of reasoning about defaults: what
normally is the case, in the absence of contradicting information.
Autoepistemic theories, on the other hand, are meant to describe the
consequences of reasoning about ignorance: what must be true if a
certain fact is not known. Although the motivation and formal
character of these systems are different, a closer analysis shows that
they bear a common trait, which is the indexical nature of certain
elements in the theory. In this talk I will show how default theories
can be reanalyzed as a restricted type of indexical theory. The
benefits of this analysis are that it gives a clear (and clearly
intuitive) semantics to default theories, and combines the expressive
power of default and autoepistemic logics in a single framework.
∂08-May-86 1038 SJM good news v bad news
I have finished THE GOOD NEWS IS THE BAD NEWS IS WRONG by Wattenberg and
filed my impressions under gdnews[1,sjm]. That is the good news. Bad
news is you must still decide if you want to read book yourself. My
condolences.
Susie
e
∂08-May-86 1054 RICHARDSON@SU-SCORE.ARPA AI Search/May 21
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 8 May 86 10:53:57 PDT
Date: Thu 8 May 86 10:32:07-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Search/May 21
To: buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA,
shah@SU-SIERRA.ARPA, grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA
Message-ID: <12205091883.15.RICHARDSON@SU-SCORE.ARPA>
Plan B:
How about Wednesday, May 21 from 2:15 - 3:45 for the AI Search Committee
meeting?
-Anne
-------
∂08-May-86 1054 NILSSON@SU-SCORE.ARPA re: [Tom Garvey <Garvey@SRI-AI.ARPA>: Danny Cohen visit]
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 8 May 86 10:53:49 PDT
Date: Thu 8 May 86 10:32:06-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: re: [Tom Garvey <Garvey@SRI-AI.ARPA>: Danny Cohen visit]
To: JMC@SU-AI.ARPA
cc: NILSSON@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Wed 7 May 86 17:18:00-PDT
Message-ID: <12205091882.38.NILSSON@SU-SCORE.ARPA>
I don't know about non-sri visitors. You might check with Tom.
-------
∂08-May-86 1113 LES Alliant name
To: Bosack@SU-SCORE.ARPA
CC: Tom@SU-SCORE.ARPA, JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, JJW@SU-AI.ARPA,
RPG@SU-AI.ARPA
We wish to call the new Alliant system Gang-of-Four with nickname Go4.
∂08-May-86 1150 CHEADLE@SU-SCORE.ARPA Conferral of MSAI diplomas
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 8 May 86 11:50:33 PDT
Date: Thu 8 May 86 11:35:02-PDT
From: Victoria Cheadle <CHEADLE@SU-SCORE.ARPA>
Subject: Conferral of MSAI diplomas
To: feigenbaum@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA,
rosenbloom@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA, winograd@SU-CSLI.ARPA,
tob@SU-AI.ARPA
Office: Margaret Jacks 258, 723-1519
Message-ID: <12205103337.9.CHEADLE@SU-SCORE.ARPA>
I'm looking for a volunteer to give out the MSAI diplomas at
the departmental Commencement ceremonies June 15. Nils will
be handing out the Ph.D. diplomas, Joe Oliger will do the
CSMS. Please let me know if you would be willing to do this.
The ceremony will probably start around 12:30 and finish by
2:00 p.m.
Thanks,
Victoria
-------
∂08-May-86 1224 knapp@gvax.cs.cornell.edu Report on scientific computing
Received: from CU-ARPA.CS.CORNELL.EDU by SU-AI.ARPA with TCP; 8 May 86 12:24:02 PDT
Received: by cu-arpa.cs.cornell.edu (5.31/4.30)
id AA01999; Thu, 8 May 86 15:24:37 EDT
Date: Thu, 8 May 86 15:25:45 edt
From: knapp@gvax.cs.cornell.edu (Susan Knapp)
Message-Id: <8605081925.AA09265@gvax.cs.cornell.edu>
Received: by gvax.cs.cornell.edu (4.12/4.30)
id AA09265; Thu, 8 May 86 15:25:45 edt
To: jmc@su-ai.arpa
Subject: Report on scientific computing
Here are three lists. The first consists of areas where we
have coverage, the second where we have promises, and the third where
we have received no response at all. Please add any topic that you
think necessary and suggest some names of people I can contact. We
don't need to be exhaustive but we do need to be balanced.
Coverage adequate:
Theory
Algorithms
Linear Programming
Cryptography
AI
Numerical Analysis
Distributed Computing
Coverage promised:
Symbolic Computing Bob Caviness
Vision Joe Traub
Object Reconstruction Azriel Rosenfeld
(Tomography)
Coverage needed:
Semantics
Languages
Architecture Charles Seitz (suggested)
VLSI Lynn Conway (suggested)
Parallel Computing
Graphics
Operating Systems
Data Bases
Natural Language / Speech Understanding
Programming Methodology
∂08-May-86 1226 alliant!jat@EDDIE.MIT.EDU alliant system for q-lisp
Received: from MIT-EDDIE.ARPA by SU-AI.ARPA with TCP; 8 May 86 12:26:27 PDT
Received: by EDDIE (5.31/4.7) id AA06702; Wed, 7 May 86 17:53:02 EDT
Received: by alliant.ARPA (4.12/4.7)
id AA02221; Wed, 7 May 86 14:29:31 edt
Date: Wed, 7 May 86 14:29:31 edt
From: alliant!jat@EDDIE.MIT.EDU (Jack Test)
Message-Id: <8605071829.AA02221@alliant.ARPA>
To: mit-eddie!jmc@su-ai.ARPA
Subject: alliant system for q-lisp
Cc: mit-eddie!clt@su-ai.ARPA, mit-eddie!les@su-ai.ARPA
Everybody here at Alliant is enthusiatic about working with your
group at Stanford on the Q-Lisp project. It is my understanding
that we have already shipped you a 1-CE system which is to be
upgraded to 4-CE's in the near future.
For your information, the version of the operating system to be
initially installed with your system will support preliminary
porting efforts for Common Lisp. The only thing it will not
support is dynamic loading of text. This capability is present
in the latest version of the operating system that is currently
under final testing here at Alliant. This version supports
execution out of data space and should be available for use at
Stanford later this month or early next month. (I had hoped it
would be available prior to shipping your system but it has fallen
behind schedule by a couple of weeks. I will see to it that your
system gets it as soon as possible).
When the new operating system is available, I plan on coming out
to Stanford to touch base with your group at Stanford and with
the people at Lucid. In the meantime, I hope to stay in touch
with you via mail.
-Jack
∂08-May-86 1344 kolk@su-navajo.arpa Alliant name and address
Received: from SU-NAVAJO.ARPA by SU-AI.ARPA with TCP; 8 May 86 13:44:23 PDT
Received: by su-navajo.arpa with Sendmail; Thu, 8 May 86 13:44:14 pdt
Date: Thu, 8 May 86 13:44:14 pdt
From: Dan Kolkowitz <kolk@su-navajo.arpa>
Subject: Alliant name and address
To: clt@sail, jjw@sail, jmc@sail, les@sail, rpg@sail, tom@score
The names are registered and the ip address is 36.8.0.118 (the names are
Gang-of-Four and Go4).
∂08-May-86 1410 RINDFLEISCH@SUMEX-AIM.ARPA Request for Help
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 8 May 86 14:09:55 PDT
Date: Thu 8 May 86 14:10:44-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Request for Help
To: KSL-Exec@SUMEX-AIM.ARPA, Nilsson@SU-SCORE.ARPA, JMC@SU-AI.ARPA,
LES@SU-AI.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA
Message-ID: <12205131681.21.RINDFLEISCH@SUMEX-AIM.ARPA>
I've gotten a request (with amplifying information appended below) to help some
people working on a distributed chess system from the Univ of Alberta get set
up and compete in Germany this June. The "help" involves kludging network
access (TYMNET, ARPANET, etc.) so they can run on 20 SUN's which SMI is
donating access to. Does anyone know if the Alberta program is credible? Is
it something of interest to our parallel computation efforts? And bottom line,
is there any reason we should say yes to this request?
Tom R.
---------------
Return-Path: <gingell%opus@SUN.COM>
Received: from sun.com by SUMEX-AIM.ARPA with TCP; Thu 8 May 86 10:06:31-PDT
Received: from snail.sun.com (snail-ptp) by sun.com (3.2-/SMI-3.0)
id AA09392; Thu, 8 May 86 09:20:26 PDT
Received: from opus.sun.uucp by snail.sun.com (3.2/SMI-3.0)
id AA10401; Thu, 8 May 86 09:23:48 PDT
Received: by opus.sun.uucp (3.2/SMI-3.0DEV3)
id AA22303; Thu, 8 May 86 09:23:57 PDT
Date: Thu, 8 May 86 09:23:57 PDT
From: gingell%opus@SUN.COM (Rob Gingell)
Message-Id: <8605081623.AA22303@opus.sun.uucp>
To: Rindfleisch@SUMEX-AIM.ARPA
Subject: Re: Tymnet
Thanks for your message -- I appreciate the issues you've raised and
understand the need to satisfy the bureaucracy.
The University of Alberta is a contestant in the World Computer Chess
championships this year. The contest is held every 3 years, this year
the site is in Germany.
Alberta's program runs on a LAN of processors, in this case they will
be using a LAN of 20 Sun-3's which we're configuring for them here at
Sun. Alberta will be accessing the network here before the contest to
finish tuning their program against the configuration, and during the
contest for actual competition. The contest organizers will pay for
some communications costs, but not for anything like a long-distance
call to California from Germany. It is possible that Alberta could
use Tymnet back to their facilities, and then dial down to California,
but that's still relatively expensive. I volunteered to ask around
various facilities in the Bay Area to see if we could cut this down
even more -- hence my communication with you.
The people who would be using your facilities would all be people at
Alberta -- no Sun employee is involved here. Actually if it appears
like this might be something you could accomodate I'll get out of the
loop and just let you all and Alberta work on it.
Sun's interest in all this is to support a worthwhile academic exercise,
one which uses our machines in an interesting way. We certainly will get
some publicity out of this, though we're not getting any funds from Alberta
or anyone else for supporting this. We can accomodate this relatively
unusual request because of the brief usage it makes of our machines --
machines which are destined for other uses once the contest is over.
I realize it may be difficult for you to accomodate this kind of access,
and if it is, please feel free to decline -- and I appreciate your entertaining
it far enough to ask for more information.
- Rob
-------
∂08-May-86 1623 RA leaving now
It's Thursday and I am leaving early. See you tomorrow.
∂08-May-86 1750 LES Next meeting
To: facil@SU-AI.ARPA, Nilsson@SU-SCORE.ARPA
Having received no grumbles, I am now emitting the next-to-last (?)
requests for approval to spend the DARPA funds as previously reported.
The next meeting of the Facilities Committee will be on Wednesday, May 28
at noon in MJH 220. The main topic will be departmental policy on
"ownership" and administration of equipment, including terminals. I will
circulate a proposed policy in advance.
∂08-May-86 2206 GOLUB@SU-SCORE.ARPA re: Letter to USSR
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 8 May 86 22:06:15 PDT
Date: Thu 8 May 86 22:02:04-PDT
From: Gene H. Golub <GOLUB@SU-SCORE.ARPA>
Subject: re: Letter to USSR
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 8 May 86 18:30:00-PDT
Phone: 415/723-3124
Message-ID: <12205217485.14.GOLUB@SU-SCORE.ARPA>
It's a visa form I need to send so I guess I cann't send a telex.
Also, the Russian sent me a telex via some company in New Jersey and
it took a week to get via US mail to me. It's a pity communication is
so badly constricted.
GENE
-------
∂09-May-86 0708 HST scott
do you remember to have taken part of scott's lambda calculus seminar in stanford 1963?
rd 1963?
∂09-May-86 1050 RA transparencies
Do you still need the transparencies? If you are done, I will return them
to Rosemary Napier.
∂09-May-86 1243 jeh@gvax.cs.cornell.edu NSF report on computing
Received: from CU-ARPA.CS.CORNELL.EDU by SU-AI.ARPA with TCP; 9 May 86 12:43:29 PDT
Received: by cu-arpa.cs.cornell.edu (5.31/4.30)
id AA07262; Fri, 9 May 86 15:44:14 EDT
Date: Fri, 9 May 86 15:45:34 edt
From: jeh@gvax.cs.cornell.edu (John E. Hopcroft)
Message-Id: <8605091945.AA20033@gvax.cs.cornell.edu>
Received: by gvax.cs.cornell.edu (4.12/4.30)
id AA20033; Fri, 9 May 86 15:45:34 edt
To: mccarthy@su-ai.arpa
Subject: NSF report on computing
.nf
TO: Subcommittee of NSF Advisory Committee for Computer Research
FROM: John Hopcroft
RE: Report on Scientific Computing
DATE: May 9, 1986
Having discussed the report with each of you let me summarize the
discussions. There is a need to get material on a wide variety of topics
outside of theory. The report as it stands has material on maybe 20%
of computer science. Furthermore, the material covered must be expanded
if it is to be intelligible to the average scientist. The implication
is that the report would grow to a size that is unmanageable.
In discussion with Rabin, the idea of restructuring the report came
up. Select five major areas: Foundations, Languages, Archecture, AI and
Applications. The first four areas would get 8-10 pages each. Within the area
there would be a 1-2 page overview and then a selection of two topics for
in-depth development of 3-4 pages each. The last area, Applications,
would be hard to treat this way and thus we would try eight topics at 2 pages
per topic.
Please let me have your comments. If you agree with the approach, we will
need names of people for each area and each subtopic. Suggestions would
be most welcome. The subtopics listed are only tentative and suggestions
here would be helpful.
1. Foundations 8-10 pages
2. Languages 8-10 pages
specification and correctness
object oriented languages
3. Architecture 8-10 pages
parallel machines
distributed systems
4. AI 8-10 pages
computer vision
temporal logic
5. Applications 16 pages, 2 pages/topic
graphics
VLSI design tools
scientific computing - linear algebra, finite element methods,
linear programming, relation to archecture
CAD/CAM
robotics
data bases - query languages, relational data bases
expert systems
man-machine interfaces
∂09-May-86 1435 RA doctor's appointment
I have a doctor's appointment at 2:50; I will be back by 4:00.
∂09-May-86 1630 VAL pointwise circ'n paper
To: JMC@SU-AI.ARPA, SJG@SU-AI.ARPA, nilsson@SU-SCORE.ARPA,
grosof@SU-SCORE.ARPA
You will be receiving copies of a new draft. There are notational changes
(which make it incompatible with the earlier versions), and new material is added
in Section 8 on the model-theoretic meaning of pointwise circumscription.
I should send a shorter version to the publisher in three weeks, so that
comments and criticisms at this time would be very much appreciated.
Vladimir
∂09-May-86 1650 ME bike locker unlocked with bike in it
I noticed your bike locker unlocked just now and it had
a bike in it. I closed the door to it, but without locking
it, to keep the door from swinging out and blocking the path.
You should keep the locker closed and locked, even when it
has no bike in it.
∂11-May-86 1050 jeh@gvax.cs.cornell.edu re: NSF report on computing
Received: from CU-ARPA.CS.CORNELL.EDU by SU-AI.ARPA with TCP; 11 May 86 10:49:55 PDT
Received: by cu-arpa.cs.cornell.edu (5.31/4.30)
id AA04779; Sun, 11 May 86 13:51:00 EDT
Date: Sun, 11 May 86 13:52:31 edt
From: jeh@gvax.cs.cornell.edu (John E. Hopcroft)
Message-Id: <8605111752.AA04730@gvax.cs.cornell.edu>
Received: by gvax.cs.cornell.edu (4.12/4.30)
id AA04730; Sun, 11 May 86 13:52:31 edt
To: JMC@su-ai.arpa
Subject: re: NSF report on computing
I got your message. Give me a call when you get back and we can discuss how you want AI handled. 607 255-7416
∂11-May-86 1231 NILSSON@SU-SCORE.ARPA Jobs (Steve)
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 11 May 86 12:30:58 PDT
Date: Sun 11 May 86 12:29:14-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Jobs (Steve)
To: jmc@SU-AI.ARPA
cc: nilsson@SU-SCORE.ARPA
Message-ID: <12205899635.12.NILSSON@SU-SCORE.ARPA>
Steve Jobs wants to talk to you and me sometime this week about what
he is doing in his new company (Next, Inc.). Will you have some time
available? If yes, you might suggest two or three possibilities for
Anne and me to propose to Steve.
-------
∂11-May-86 1351 NILSSON@SU-SCORE.ARPA AI search
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 11 May 86 13:51:08 PDT
Date: Sun 11 May 86 13:51:29-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: AI search
To: jmc@SU-AI.ARPA, shah@SU-SIERRA.ARPA, buchanan@SUMEX-AIM.ARPA,
genesereth@SUMEX-AIM.ARPA, grosof@SU-SUSHI.ARPA, hirsh@SU-SUSHI.ARPA
cc: nilsson@SU-SCORE.ARPA
Message-ID: <12205914611.12.NILSSON@SU-SCORE.ARPA>
At our last mtg people wanted me to call some people about
recommendations for our AI search. Here are summaries of the
phone calls I have made so far:
5/8/86 Hector Levesque: Wants to stay in Toronto.
5/7/86 James Allen: He cannot move now, but we agreed to keep
him in mind for a possible natural language position in a few years.
5/5/86 Don Perlis: He is coming up for tenure in a year. He likes
Maryland and doesn't want to move now. His wife teaches industrial
psychology.
5/5/86 Ray Reiter: He knows of no additional candidates. He doesn't
want to give up his 5-year Canadian Institute fellowship.
4/30/86 Allen Newell: He recommends Carl Elbeling (systems, not AI).
He could recommend no one this year. Next year or two: David Speier,
Steve Minton.
4/30/86 Jerry Sussman: He doesn't want to move. Watch David Chapman
coming out in a year or two; doing work on planning. Also Dave
McAllister doing work on theorem proving.
4/30/86 Bob Beyer: Besides Shankar, he says we should keep our eyes
on Warren Hunt and Mike Gordon, Boyer-Moore students doing work on
hardware verification using the Boyer-Moore theorem prover. Also,
in theorem proving, he likes Chou Shang-ching who has some new approaches
based on the work of Wu in China. (Maybe we can talk Boyer himself
into coming in a couple of years.)
4/25/86 Woody Bledsoe: Thought of no one in AI. Watch Hassan
Ait-kaci in systems (now at MCC).
I think we all feel that we will have among the candidates we are
seeing someone whom we can proudly recommend and who will be a
big plus for the department. I am scheduling us to meet at 2:15
on Wednesday, May 21. I don't see any reason why we can't come out
of that meeting with a decision.
Letters are beginning to come in on the candidates. We should have
an adequate number in time for our meeting. I am phoning stragglers.
(Jerry Sussman says he spends 20% of his time writing letters and wants
MIT and Stanford to sign a mutual moratorium on asking each other for
letters.)
-Nils
-------
∂11-May-86 2035 rms@PREP.AI.MIT.EDU I'm visiting
Received: from MIT-PREP.ARPA by SU-AI.ARPA with TCP; 11 May 86 20:35:11 PDT
Received: by PREP.AI.MIT.EDU; Sun, 11 May 86 23:36:47 edt
Date: Sun, 11 May 86 23:36:47 edt
From: rms@PREP.AI.MIT.EDU (Richard M. Stallman)
Message-Id: <8605120336.AA01430@prep>
To: jmc@su-ai.arpa
Subject: I'm visiting
I am coming to California for two weeks starting May 13.
Would anyone like a talk? Or is there a need for courses
in GNU Emacs use/customization, or any other such work?
∂12-May-86 0832 RICHARDSON@SU-SCORE.ARPA Shankar
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 12 May 86 08:32:44 PDT
Date: Mon 12 May 86 08:33:01-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Shankar
To: jmc@SU-AI.ARPA
cc: ra@SU-AI.ARPA
Message-ID: <12206118777.13.RICHARDSON@SU-SCORE.ARPA>
Will you be available to meet with Shankar on Thursday, May 15 from
2:00 - 2:30?
-Anne
-------
∂12-May-86 1229 LES DARPA Order received!
To: JMC, RPG, CLT, JJW
John Pucci called to say that he had just received the DARPA Order covering
Qlisp and expects to get a draft task description to us this week. (I hope
that it will resemble the one I wrote for him.) He thinks that we can make
the planned June 1 start date.
∂12-May-86 1444 CLT rt unix upgrade - revisited
To: "@RT.DIS[1,CLT]"@SU-AI.ARPA
Due to unsuccessful dump, we didn't bring up the
new Unix on the RT Friday. Dumping now seems to be
under control, so we will try again on Tuesday afternoon.
An incremental dump will be done sometime after 11pm tonight.
But no guarantees are made for changes made between then
and downtime.
The vanilla system will hopefully be up by the end of
Tuesday, but it may take a day or so to restore the rest
of the world. Have patience.
∂13-May-86 1404 RA Conference on Computers and Mathematics, July 29, 1986
Mike Beaver from Kaiser Electronics and Aerospace Corp. needs information
about the conference. Do you have anything re the conference I can mail him?
His tel. (408) 946 3000, ext. 461.
Thanks,
∂13-May-86 1407 RA Olivia Stuart, Time Mag.
Olivia Stuart from Time is writing an article on intelligent machines for
Time and would like to come in in order to take your picture. Her tel. (415)
434 5241. She will call again later this afternoon.
∂13-May-86 1432 @SUMEX-AIM.ARPA:MRC@PANDA PDP-6 CONS instruction
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 13 May 86 14:32:03 PDT
Received: from PANDA by SUMEX-AIM.ARPA with Cafard; Tue 13 May 86 11:55:28-PDT
Date: Tue 13 May 86 11:49:14-PDT
From: Mark Crispin <MRC%PANDA@SUMEX-AIM.ARPA>
Subject: PDP-6 CONS instruction
To: JMC@SU-AI.ARPA, Les@SU-AI.ARPA, ME@SU-AI.ARPA, JJW@SU-AI.ARPA
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12206416643.6.MRC@PANDA>
Gentlemen -
I am trying to find out what the CONS instruction on the Stanford
PDP-6 did. I'd like to know its opcode, what it took in AC and E, and
what it did with the operands. I think it would be fun to try to
microcode it on my 2020. If you remember how it worked or have any
documents lying around that may describe it, I'd like the information.
I've already asked Kotok but he's long since forgotten. I also looked
in FACIL.TED[H,DOC] and it just says "read the prints" (do we still
have the PDP-6 prints??).
-------
∂13-May-86 1511 VAL Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
NON-MONOTONICITY IN PROBABILISTIC LOGIC
Benjamin Grosof
Computer Science Department, Stanford University
Thursday, May 15, 4pm
MJH 252
I will discuss how to formalize the notion of non-monotonicity in
probabilistic reasoning, using the framework of Probabilistic Logic
(cf. Nils Nilsson). I will give some motivating examples of types of
non-monotonic probabilistic reasoning that seem to be found in
practice. There seems to be a relationship to default inheritance,
i.e. prioritized defaults of the type used in classic example of
whether birds and ostriches fly. Next, I introduce the idea of
maximizing conditional independence, which can be thought of as
maximizing irrelevance. This can be described more simply in terms of
non-monotonic reasoning on Graphoids (cf. Judea Pearl).
I conjecture that an important type of non-monotonicity in probabilistic
reasoning may be concisely expressed in terms of conditional
independence and Graphoids. Finally, I pose as an open question how
to formulate in the above terms the non-monotonic behavior of
maximizing entropy, a widely-used technique in probabilistic
reasoning.
∂13-May-86 1530 lantz@su-gregorio.arpa FYI
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 13 May 86 15:21:10 PDT
Received: by su-gregorio.arpa with Sendmail; Tue, 13 May 86 15:20:27 pdt
Date: 13 May 1986 1520-PDT (Tuesday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: phd-program@score, phdcom@sail, csl-everyone@sierra
Cc:
Subject: FYI
The proposals of the Systems PhD Committee were passed unanimously at
today's CSL faculty meeting. Beyond the fact that this means I've
survived one more committee, it also means that a substantial subset of
EE students (and faculty) have agreed to be subject to the same set of
FORMAL requirements as CS students (and faculty). I'd say that
represents another significant step towards (re)making the Computer
Systems Laboratory a truly joint laboratory. Let's keep it up!
Keith
∂13-May-86 1533 RA Judith Shoolery
Judith Shoolery, editor to Dr. Teller, would like to talk to you. Her
tel. 3-0425; (415) 857 1697 (home)
∂13-May-86 1548 GIVAN@su-sushi.arpa summer work
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 13 May 86 15:48:20 PDT
Date: Tue 13 May 86 15:47:41-PDT
From: Bob Givan <GIVAN@su-sushi.arpa>
Subject: summer work
To: jmc@su-ai.arpa
Message-ID: <12206460051.56.GIVAN@su-sushi.arpa>
Prof. McCarthy,
Hello, and I hope all went well in the east. (I am of course biased!)
If we need to get together, then possibly we should set a time?
Otherwise, you could make a salary offer via E-mail, and we could
get together as convenient. But I must resolve the salary issue
soon as I have two companies waiting on me.
I would like to meet, if possible.
Times which are OK, in order of preference:
1) Before 5 this afternoon (specify time!)
2) Between 7 and 8 this evening
3) Before 9:45am tomorrow morning
4) 1:15 tomorrow
99) 4:15 tomorrow
If none of these times will do, some mail message re: salary would
be appreciated.
Thanx,
Bob Givan
-------
∂13-May-86 1550 VAL MCC
They invited me to come again. I can go there any time during the week of June 2.
∂13-May-86 1803 JMC
Shoolery 857-1697
∂13-May-86 1806 NILSSON@SU-SCORE.ARPA Shankar
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 13 May 86 18:06:07 PDT
Date: Tue 13 May 86 18:06:16-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Shankar
To: buchanan@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA, shah@SU-SIERRA.ARPA,
grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA
cc: nilsson@SU-SCORE.ARPA
Message-ID: <12206485279.28.NILSSON@SU-SCORE.ARPA>
Don't forget we have Nataranja Shankar visiting this Thurs and Friday.
Due to scheduling overload, we are having Shankar give his talk at
5:15 on Thursday. Hope the late hour doesn't bother Shankar (or any
of us) too much. Anyone who wants to accompany Shankar and me for
lunch Thurs at the fac. club and/or Thurs eve for dinner, pls let
Anne R. know. I'll be busy Fri lunch, so anyone who wants to host
him for lunch Fri, pls speak up. -Nils
-------
∂13-May-86 1807 CLT shopping list
veal stew meat - 2.5 pounds
2 ears corn
2 zucchini
4 carrots
celery
green beans
cauliflower
chicken soup stock - 6 cans
buttermilk
milk
seltzer
hamburger
asparagus
∂13-May-86 1810 JJW Re: PDP-6 CONS instruction
To: MRC%PANDA@SUMEX-AIM.ARPA
CC: JMC@SU-AI.ARPA, LES@SU-AI.ARPA, ME@SU-AI.ARPA
The DECsystem-10/20 Hardware Manual (SAIL Operating Note 75A, 1977,
which looks like it was put together by SAIL and LOTS people) says
Opcode 257 is CONS. No one uses the CONS instruction anymore.
If you want to know how it works, read the prints. Bit 18 in
CONI APR, is the CONS flag, which interrupts on the APR channel.
Bit 20 in CONO APR, clears the CONS flag.
It looks like 257 was taken over by MAP on the KI and later machines,
so you'll need a new opcode at least.
SAILON-75A says it supersedes SAILON-26 and SAILON-71, so maybe one
of those manuals has a more complete description.
Joe
∂13-May-86 1819 CLT phone msg
call tsahi gozani (if i read hazel's writing correctly)
424-8274
∂13-May-86 2000 JMC
Bukovsky
∂13-May-86 2017 @SUMEX-AIM.ARPA:MRC@PANDA Re: PDP-6 CONS instruction
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 13 May 86 20:17:10 PDT
Received: from PANDA by SUMEX-AIM.ARPA with Cafard; Tue 13 May 86 20:17:42-PDT
Date: Tue 13 May 86 19:18:11-PDT
From: Mark Crispin <MRC%PANDA@SUMEX-AIM.ARPA>
Subject: Re: PDP-6 CONS instruction
To: JJW@SU-AI.ARPA
cc: JMC@SU-AI.ARPA, LES@SU-AI.ARPA, ME@SU-AI.ARPA
In-Reply-To: Message from "Joe Weening <JJW@SU-AI.ARPA>" of Tue 13 May 86 18:10:00-PDT
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12206498370.8.MRC@PANDA>
Oh. I thought CONS was 247, which is a free opcode. MIT uses 247 for
CIRC, which is a ROTC except that the bit goes into AC+1 the other way.
In other words CIRC AC,-↑D36 will put the bits of AC in reverse order
in AC+1 and the bits of AC+1 in reverse order into AC. Actually,
CIRC AC,↑D36 would do the same thing...
-------
∂14-May-86 0001 LES PDP-6 CONS instruction
To: mrc%panda@SUMEX-AIM.ARPA
CC: JJW@SU-AI.ARPA, JMC@SU-AI.ARPA, ME@SU-AI.ARPA
I have looked through the part of my files where I would expect to find
the CONS description and didn't. It might be lurking somewhere else but
I don't have time for an exhaustive search.
The CONS instruction was, of course, a bad idea. It cost a non-trivial
amount (several grand as I recall), made a negligible improvement in Lisp
system performance and made any software that used it not work on any
other PDP-6/10. While it serves as a basis for interestingly
ideosyncratic anecdotes, I recommend that you let that idea die in peace.
∂14-May-86 0621 FAHLMAN@C.CS.CMU.EDU copyright holder
Received: from C.CS.CMU.EDU by SU-AI.ARPA with TCP; 14 May 86 06:21:26 PDT
Received: ID <FAHLMAN@C.CS.CMU.EDU>; Wed 14 May 86 09:22:33-EDT
Date: Wed, 14 May 1986 09:22 EDT
Message-ID: <FAHLMAN.12206619304.BABYL@C.CS.CMU.EDU>
Sender: FAHLMAN@C.CS.CMU.EDU
From: "Scott E. Fahlman" <Fahlman@C.CS.CMU.EDU>
To: John McCarthy <JMC@SU-AI.ARPA>
Cc: cl-steering@SU-AI.ARPA
Subject: copyright holder
In-reply-to: Msg of 14 May 1986 01:20-EDT from John McCarthy <JMC at SU-AI.ARPA>
If copyrighting this document under "Common Lisp Steering Committee" (or
"Technical Committee" or both, whatever we decide is best) is legal it
could save us all a lot of hassle. We'd want to do something more
formal before we turn it over to someone for approval or publication,
but this could allow us to get started. The biggest remaining question
is whether Lucid and/or Digital Press will agree to turn over rights to
create a derivitive work to such a group.
-- Scott
∂14-May-86 0922 RICHARDSON@SU-SCORE.ARPA Shankar
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 14 May 86 09:22:25 PDT
Date: Wed 14 May 86 09:18:44-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Shankar
To: jmc@SU-AI.ARPA
cc: ra@SU-AI.ARPA
Message-ID: <12206651389.9.RICHARDSON@SU-SCORE.ARPA>
Will you be available to meet with AI candidate Shankar on Thursday, May 15
at 2:00?
-Anne
-------
∂14-May-86 1000 JMC
Givan
∂14-May-86 1004 RICHARDSON@SU-SCORE.ARPA AI Search Committee/May 21
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 14 May 86 10:04:49 PDT
Date: Wed 14 May 86 10:01:22-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Search Committee/May 21
To: jmc@SU-AI.ARPA
cc: ra@SU-AI.ARPA
Message-ID: <12206659149.9.RICHARDSON@SU-SCORE.ARPA>
Are you available for an AI Search Committee meeting on May 21 from
2:15 - 3:45?
-Anne
-------
∂14-May-86 1001 RA John Hopcroft
Hopcroft called and would like you to call him back at (607) 255 7416 re
AI section of the NSF report you are writing together.
∂14-May-86 1005 RA Susan Bishop, TV Ontario
Bishop is doing a TV series on AI and Expert Machines and would like to talk
to you. Her tel. (416) 484 2600. She will call back later.
∂14-May-86 1157 RA lunch
I am going out for lunch.
∂14-May-86 1350 RA Meeting with Shankar
Could you meet Shankar tomorrow, Thursday, 2:00-2:30. Please let me know.
Thanks,
∂14-May-86 1351 RA AI Search Committee
Is Wednesday, May 21, 2:15-3:45 ok with you for AI Search Committee? Please
let me know.
Thanks,
∂14-May-86 1452 NILSSON@SU-SCORE.ARPA Hooray!
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 14 May 86 14:51:50 PDT
Date: Wed 14 May 86 14:51:19-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Hooray!
To: pcohen@SRI-AI.ARPA
cc: stan@SRI-AI.ARPA, konolige@SRI-AI.ARPA, georgeff@SRI-AI.ARPA,
rperrault@SRI-AI.ARPA, genesereth@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA,
rosenschein@SU-SUSHI.ARPA
Message-ID: <12206711933.16.NILSSON@SU-SCORE.ARPA>
Alan Meyerowitz of ONR phoned me this morning to say that our proposal
"Intelligent Communicating Agents" is one of "four or five" (out of
over 40 submitted) selected in the "AI Track" of the URI Block Grant
Program to be site-visited. A site-visit panel has been assembled and
they would like to visit us on Monday June 2 (all day).
The panel consists of Alan Meyerowitz, Randy Shumaker (NRL), Michael
Shafto (ONR Cognitive Psychol Branch), Y.T. Chien (NSF), Vince
Sigilito (AFOSR), Susan Chipman (ONR Cog. Psy Branch), and
possibly Cmdr. Alan Sears (DARPA).
The panel would like to hear an overview of the proposal (I'll give that),
in-depth presentations on some of its parts (we need to talk about and
organize those) plus any supporting talks that we think important. They
want a chance to ask questions of all of us, inspect our facilities, etc.
Probably we should schedule a "field trip" to SRI to see Flakey, etc.
I'll get in touch with our proposed industrial and navy liasion people
to see if they can participate.
-Nils
-------
∂14-May-86 1507 KTRACY@su-sushi.arpa Your book and my VTSS160 paper
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 14 May 86 15:07:09 PDT
Date: Wed 14 May 86 15:06:24-PDT
From: Kim W. Tracy <KTRACY@su-sushi.arpa>
Subject: Your book and my VTSS160 paper
To: jmc@su-ai.arpa
Message-ID: <12206714681.52.KTRACY@su-sushi.arpa>
Is your book still to be finished by this fall? If so, would
you mind if I received an early version of it from you, so that I
could comment on it?
Also, I would like to know what comments you had on my paper about
the possibilities of computer teaching and to get it back.
--Kim Tracy
-------
∂14-May-86 1727 WINOGRAD@SU-CSLI.ARPA Meeting about syllabus and proposal for it
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 14 May 86 17:27:30 PDT
Date: Wed 14 May 86 17:26:01-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Meeting about syllabus and proposal for it
To: phdcom@SU-AI.ARPA
cc: jlh@SU-SONOMA.ARPA, WINOGRAD@SU-CSLI.ARPA
The following is a rough draft for a new syllabus, based on input from
Leo, John Hennessey, and others. We need to fix it up, and decide what
format to propose for the exam to go with it (e.g., do people take all
of the applications areas, or choose?). We should meet sometime soon
and get this approved at one of the faculty meetings between now and the
end of quarter. I propose that we discuss this at either 2:15 or 3:15
next Tuesday (May 20). Let me know if one of those is bad for you, and
I will set it definitely by Friday. The names in brackets indicate who
suggested a given item. --t
-------------
{\bf LANGUAGES AND SOFTWARE}
Harold Abelson and Gerald Sussman, {\sl Structure and Interpretation of
Computer Programs}, MIT, 1985 [Guibas]
Alfred V. Aho, Sethi, R., and Jeffrey D. Ullman, {\sl Compilers --
Principles, Techniques, and Tools}, Addison-Wesley, 1986. All except
sections 9.11-9.12, 10.9-10.13. [Hennessy]
M. Ben-Ari, Principles of Concurrent Programming, Prentice-Hall
International, 1982. Material on concurrent programming. [Hennessy]
Terrence W. Pratt, {\sl Programming Languages: Design and
Implementation}, Second edition, Prentice-Hall, 1984. [Hennessy]
{\bf HARDWARE AND OPERATING SYSTEMS}
Bell, C.G., Mudge, J.C., and McNamara, J.E., Computer Engineering -- A
DEC View of Hardware Systems Design, Digital Press, 1978. [Hennessy]
Kogge, P., The Architecture of Pipelined Computers, McGraw-Hill, 1981.
[Hennessy]
M. Morris Mano, {\sl Computer System Architecture}, Second Edition,
Prentice-Hall, 1982. Basic logic design, data representation, and
computer organization. Material taught in CS 108 and CS 112. [Hennessy]
James L. Peterson and Abraham Silberschatz, {\sl Operating System
Concepts}, Addison-Wesley, Second Edition, 1985. [Hennessy]
{\bf THEORY OF COMPUTATION}
[This section has gotten very long, since it combines material from 2
1/2 sections of the old exam (AA, MTC and the theory part of AI). It's
proportion has gone down from 5/12 to 1/4, so we need to trim]
Alfred V. Aho, John E. Hopcroft, and Jeffrey D. Ullman, {\sl Data
Structures and Algorithms}, Addison-Wesley, 1983. [Guibas]
Herbert B. Enderton, {\sl A Mathematical Introduction to Logic},
Academic Press, 1972, Chapters 1--2.
Michael R. Garey and David S. Johnson, {\sl Computers and
Intractability}, Freeman, 1979, Chapters 1--3, 7. [Guibas]
John E. Hopcroft and Jeffrey D. Ullman, {\sl Introduction to Automata
Theory, Languages, and Computation}, Addison-Wesley, 1979, Chapters
1--3, 4.1--4.6, 5--7, 8.1--8.5. [Guibas]
Donald E. Knuth, {\sl The Art of Computer Programming}, Volume 1, Second
Edition, Addison-Wesley, 1973, Section 1.2 (except for Subsections
1.2.9, 1.2.10, 1.2.11.2, and 1.2.11.3). [We should only leave out
1.2.11.1 - Guibas]
Zohar Manna, {\sl Introduction to Mathematical Theory of Computation},
McGraw-Hill, 1973, Chapters 1--3.
John McCarthy and Carolyn Talcott, {\sl LISP: Programming and Proving},
(available from McCarthy's secretary) 1980, Chapters 1--3. [replace
this with a real book if it doesn't come out soon - Guibas]
Nils Nilsson, {\sl Principles of Artificial Intelligence}, Kaufman,
1980, Chapters 4--6.
Sedgewick, Robert, {\sl Algorithms}, Addison-Wesley, 1983. [Guibas]
{\bf APPLICATIONS AND TECHNIQUES}
{\bf Artificial intelligence}
Elaine Rich, {\sl Artificial Intelligence}, McGraw-Hill, 1983.
{\bf DATABASES}
Jeffrey Ullman, {/sl Principles of Data Base Systems,}
Computer Science Press, 1982 [Jeff specified the chapters, but
I've lost his message --t]
Gio Weiderhold, {/sl Database Design}, 2nd Edition, McGraw
Hill, 1983. [Chapters?]
[we should pick one or the other -- Winograd]
{\bf GRAPHICS}
Newman and Sproull, Principles of Interactive Computer
Graphics,
Chapters 1--5, and 15--18. [Guibas]
{\bf NUMERICAL ANALYSIS}
Kendall E. Atkinson, {\sl An Introduction to Numerical
Analysis}, Wiley, 1978, Chapters 1--3, 5, 7, 8 (except
Sections 2.8, 2.10, 5.4). [material to be added later from
Scientific Computing course, when it is developed -- Oliger]
-------
∂14-May-86 2000 JMC
Roszak review
∂15-May-86 0000 JMC
III option
∂15-May-86 0131 HST lisp30
fitch is not the logician but an english lispler.
he has done a lot of work with developing an english lisp
implementation. moreover he did work in symbolic integration.
clinger is a scheme-guy (wrote the revised revised revised report on scheme
(perhaps one revised more or less)
white is the goy who programmed the first dendral for feigenbaum.
the other white is "jonl" from mit now at lucid.
do you have suggestions for the p.c.? perhaps weinreb, steele and bobrow?
∂15-May-86 0929 GIVAN@su-sushi.arpa
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 15 May 86 09:29:14 PDT
Date: Thu 15 May 86 09:28:33-PDT
From: Bob Givan <GIVAN@su-sushi.arpa>
To: jmc@su-ai.arpa
Message-ID: <12206915320.14.GIVAN@su-sushi.arpa>
i just want to remind you that i was expecting to hear from you regarding
next summer.
i hope we can work something out soon,
Bob Givan
-------
∂15-May-86 1050 perlis@mimsy.umd.edu circumscriptive aspects
Received: from MIMSY.UMD.EDU by SU-AI.ARPA with TCP; 15 May 86 10:50:25 PDT
Received: by mimsy.umd.edu (5.9/4.7) id AA15298; Thu, 15 May 86 13:53:08 EDT
Date: Thu, 15 May 86 13:53:08 EDT
From: Don Perlis <perlis@mimsy.umd.edu>
Message-Id: <8605151753.AA15298@mimsy.umd.edu>
To: jmc@su-ai.arpa
Subject: circumscriptive aspects
Cc: perlis@mimsy.umd.edu
John,
I have been puzzled for some time by your comment in at least one
version of your paper on formula circumscription, to the effect that aspects
are something of an embarrassment. It has always seemed to me that they
have a very natural interpretation. Namely, the aspect appropriate to a
given class (e.g., birds) with respect to some other class or property
(e.g., flying) is simply the relationship between the class and the property
(e.g., the typicality of birds being able to fly, or more generally, the
Q-ness of P). Formally this can be recorded simply as the pair
<`Bird',`Flying'>. Thus ab(aspect1,x) can be rewritten rather as
ab(x,`Bird',`Flying')
i.e., x is an abnormal member of the class of birds with regard to flying.
In fact, this way of writing aspects suggests a predicate such as
`Normally', which might be used as follows:
Normally(`P',`Q') -> Circum(`P & -Q')
That is, if P's are normally Q's, then we want to minimize the
extension of P & -Q. This can be written as a single wff if Circum is taken
to be either your second-order axiom but with the circumscribed wff replaced
by a variable, or a first-order axiom using sets as in the recent draft I
sent you.
Then aspects are implicit in the use of the Normally predicate, that
is, in the pair of arguments <`P',`Q'>. In general, any pair of properties
P and Q have an associated aspect or relation, which may be include
typicality or normalcy of Q given P in some cases.
Have I failed to grasp some other aspect of aspects (!) that you
were referring to?
Don
∂15-May-86 1059 GSMITH@SUMEX-AIM.ARPA AI Qual
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 15 May 86 10:59:20 PDT
Date: Thu 15 May 86 11:00:36-PDT
From: Grace Smith <GSMITH@SUMEX-AIM.ARPA>
Subject: AI Qual
To: TOB@SU-AI.ARPA, EAF@SUMEX-AIM.ARPA, GENESERETH@SUMEX-AIM.ARPA,
JMC@SU-AI.ARPA, NILSSON@SUMEX-AIM.ARPA, ROSENBLOOM@SUMEX-AIM.ARPA,
WINOGRAD@SU-CSLI.ARPA
cc: BGB@SUMEX-AIM.ARPA, Gsmith@SUMEX-AIM.ARPA, Cheadle@SU-SCORE.ARPA
Message-ID: <12206932078.27.GSMITH@SUMEX-AIM.ARPA>
As a final rundown on the Qual, here are the examining committees and
students for each day.
Tuesday, May 20:
Winograd Feigenbaum Nilsson
Rosenbloom Rosenschein Buchanan
Green Gabriel Wilkins
Students: Washington, Woodfill, Golding, McCall, Dixon, Healey
Friday, May 23:
Binford Genesereth Rosenbloom
McCarthy Buchanan Tenenbaum
Mittal Green Waldinger
Students: Merchant, Unruh, Schoen, Peyton, Kaelbling, Mohammed
Please plan to arrive a few minutes before 8:30; the room assignments will be posted on the door of MJH252.
The period from noon till 1:00 will be spent evaluating results and we
will provide sandwiches and drinks.
G.
-------
∂15-May-86 1124 RA airline tickets
I put your ticket for tomorrow and a Federal Express delivery in your in box
∂15-May-86 1147 VAL re: circumscriptive aspects
[In reply to message sent Thu, 15 May 86 13:53:08 EDT.]
This interpretation of aspects seems rather limited. For instance, it doesn't
seem to apply to their use in the sections on is-a hierarchies and on the
blocks world in your "Applications" paper, where aspects have many arguments
(more than any predicate symbols involved), and where the assertions
don't have the form "P's are normally Q's". We can write them in this form
by making P identically true and taking Q to be an appropriate lambda-expression,
but this seems rather artificial.
Vladimir
∂15-May-86 1328 GIVAN@su-sushi.arpa Re: reply to message
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 15 May 86 13:28:41 PDT
Date: Thu 15 May 86 13:27:51-PDT
From: Bob Givan <GIVAN@su-sushi.arpa>
Subject: Re: reply to message
To: JMC@su-ai.arpa
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Thu 15 May 86 09:45:00-PDT
Message-ID: <12206958882.53.GIVAN@su-sushi.arpa>
15$/hr. is fine with me. I would still like to talk more about what I'll
be doing, whenever that's convenient.
I can start the Thursday immediately following finals (6/12 ?).
[I will be available straight thru to fall qtr, with the exception of one
week or so in Sept for a visit East]
we had mentioned 3 different problems i might work on:
1. EBOS
2. QLISP
3. Formalizing common sense.
I am interested in them all, but in the reverse order (3 2 1).
If things go well this summer it is probable that I will have room in my
schedule to continue as an RA part time, if that will work.
Thanx much
Bob Givan.
-------
∂15-May-86 1359 RA Chuck Williams
Chuck Williams returned your call.
∂15-May-86 1414 VAL Reminder: Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
NON-MONOTONICITY IN PROBABILISTIC LOGIC
Benjamin Grosof
Computer Science Department, Stanford University
Thursday, May 15, 4pm
MJH 252
I will discuss how to formalize the notion of non-monotonicity in
probabilistic reasoning, using the framework of Probabilistic Logic
(cf. Nils Nilsson). I will give some motivating examples of types of
non-monotonic probabilistic reasoning that seem to be found in
practice. There seems to be a relationship to default inheritance,
i.e. prioritized defaults of the type used in classic example of
whether birds and ostriches fly. Next, I introduce the idea of
maximizing conditional independence, which can be thought of as
maximizing irrelevance. This can be described more simply in terms of
non-monotonic reasoning on Graphoids (cf. Judea Pearl).
I conjecture that an important type of non-monotonicity in probabilistic
reasoning may be concisely expressed in terms of conditional
independence and Graphoids. Finally, I pose as an open question how
to formulate in the above terms the non-monotonic behavior of
maximizing entropy, a widely-used technique in probabilistic
reasoning.
∂15-May-86 1621 RA leaving early
It is Thursday and I am leaving early. Have a nice trip.
∂15-May-86 1730 CLT
we have a reservation at la terrasse for 7:15
∂15-May-86 1732 CLT
fine
∂16-May-86 1200 jeh@gvax.cs.cornell.edu NSF report
Received: from CU-ARPA.CS.CORNELL.EDU by SU-AI.ARPA with TCP; 16 May 86 11:59:18 PDT
Received: by cu-arpa.cs.cornell.edu (5.31/4.30)
id AA06684; Fri, 16 May 86 14:59:05 EDT
Date: Fri, 16 May 86 15:00:00 edt
From: jeh@gvax.cs.cornell.edu (John E. Hopcroft)
Message-Id: <8605161900.AA05639@gvax.cs.cornell.edu>
Received: by gvax.cs.cornell.edu (4.12/4.30)
id AA05639; Fri, 16 May 86 15:00:00 edt
To: mccarthy@su-ai.arpa
Subject: NSF report
.rs
.nf
.pp
I have been in touch with most of you by telephone during the past
two weeks. The following is a consensus derived from those discussions.
.lp
(1) There is complete agreement to starting over along the lines of
the outline developed by Michael Rabin (attached).
.lp
(2) Everyone seems to agree to the five sections: Foundations, Languages,
Computer Systems, AI, and Applications, although it was felt that the actual
titles might be improved.
.lp
(3) Each of the first four sections would consist of a 1-2 page overview,
followed by a 3-4 page in-depth discussion of two selected topics.
.lp
(4) For foundations I am willing to write the 1-2 page overview and Rabin
is willing to write 3-4 pages on randomized algorithms. The preference
for the second topic apears to be parallel algorithms. I have left
NP-completeness and Algorithms on the outline in case someone wants to
argue strongly for one of them.
*ACTION ITEM: we need an ordered list of writers for parallel algorithms
if selected for second topic.
.lp
(5) For languages Scott has volunteered to do the 1-2 page overview. The
two preferred subtopics seem to be specification and correctness, and logic
programming. Again I have left three other possibilities on the outline
in case someone wishes to argue for one of them
*ACTION ITEM: we need an ordered list of potential authors for the
two subareas.
.lp
(6) For computer systems the consensus on the two subareas seems to be
parallel machines and distributed systems.
*ACTION ITEM: we need an ordered list for overview and both subareas.
.lp
(7) For AI, John McCarthy is willing to write the 1-2 page overview.
There is disagreement on subareas. I propose combining logic in AI
and Technology for developing production systems into one topic and
using computer vision for the second.
*ACTION ITEM: please advise on above and give lists of potential
authors for subareas.
.lp
(8) Should we rename "Applications" by "Technologies." The consensus
seems to be to select only 4-5 topics to avoid fragmentation.
Schwartz will write on robotics and Chuck Seitz will do VLSI.
Please propose a subset of topics assuming we include robotics and VLSI.
The set should be balanced.
.pp
From your responses I will proceed to contact authors in areas where I
sense we agree and get back where there is serious conflict.
.bp
.rs
.nf
1. Foundations: Hopcroft.
Randomized algorithms: Rabin.
Parallel algorithms: Cook, Valiant, Borodin.
NP-completeness: Cook, Karp.
Algorithms: Knuth.
2. Languages: Scott
Specification and correctness: Liskov, Hoare, Owicki, Boyer, Reynolds, Dijkstra.
Logic Programming: Fernando Pereira(SRI).
Object oriented languages: Small talk - Guy Steele (Thinking machine),
Larry Tessler (Apple)
Compiler optimization: Fran Allen.
Denotational Semantics:(can be included in specification and corectness)
abstraction: Liskov, Guttag, Burstall
concurrency: Lamport, Owicki, Chandy
3. Computer Systems: Hennessey, F. Corbato, Jerry Saltzer, Marurice Wilkes,
Peter Weinberger, Forest Baskett.
Parallel machines: Burton Smith, Charles Leiserson.
Gorden Bell, Dave Patterson(Berkeley)
Distributed systems: Barbara Liskov, Alfred Spector, Ric Rashid,
Nancy Lynch, Leslie Lamport, Roger Needham, Mike Schroeder
Special purpose architectures: Kung, Dennis, Arvind.
4. AI: John McCarthy 8-10 pages
Logic in AI: Nils Nilsson, Drew McDermott, Vladimir Lifshutz.
Technology for developing production systems: Alan Newell, Paul Rosenbloom, Bruce Buchanan(Stanford)
(the above two could be combined into one.)
Computer vision: Poggio(MIT), Shimon Ullman(MIT), Rosenfeld, Pat Winston.
5. Applications or technology 16 pages, 2 pages/topic 8 topics or 3-4 pages/topic 4 topics
VLSI design tools: Chuck Seitz, Lipton, Ousterhout.
robotics: Schwartz.
graphics: Van Dam, Sproull, Riseman (Utah), Requicha (Rochester).
scientific computing - linear algebra, finite element methods,
linear programming, relation to archecture: Gilbert Sprang(Applied Math MIT),
Gene Golub, Dennis, Gear, Bland.
CAD/CAM: Farin (Utah), Voelker, Requicha, Bajcsy
data bases - query languages, relational data bases: Jim Gray, Stonebraker, Eugene Wong, Ullman
expert systems: Bruce Buchanan (sp?) Stanford, Feigenbaum, John McDermott
man-machine interfaces: Jim Morris(CMU), Stuart Card(PARC).
cognitive psychology: Feldman, Newell.
symbolic computing: Caviness, Moses.
tomography:
Natural language processing: Waltz (Thinking machines), Grishman (NYU), Shank(Yale)
∂16-May-86 1258 WINOGRAD@SU-CSLI.ARPA Meeting Tuesday at 2:15
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 16 May 86 12:58:49 PDT
Date: Fri 16 May 86 12:56:37-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Meeting Tuesday at 2:15
To: phdcom@SU-AI.ARPA
cc: WINOGRAD@SU-CSLI.ARPA
We will meet at 2:15. Victoria, can you see what room is available? Thanks
--t
-------
∂16-May-86 1435 TAJNAI@SU-SCORE.ARPA Dr. Peled's visit
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 16 May 86 14:35:39 PDT
Date: Fri 16 May 86 12:19:12-PDT
From: Carolyn Tajnai <TAJNAI@SU-SCORE.ARPA>
Subject: Dr. Peled's visit
To: ullman@SU-SCORE.ARPA, JMC@SU-AI.ARPA, lantz@SU-SCORE.ARPA,
ungar@SU-SONOMA.ARPA, M@SU-SIERRA.ARPA
cc: Nilsson@SU-SCORE.ARPA
Message-ID: <12207208530.19.TAJNAI@SU-SCORE.ARPA>
I'm trying to rework the schedule for Dr. Peled. The following is
proposed:
Ullman 1:30 to 1:55
McCarthy 1:55 to 2:20
Lantz 2:20 to 2:45
Ungar 2:45 to 3:10
Flynn 3:10 to 3:35
Please let me know if this works for you.
Carolyn
-------
∂16-May-86 1625 SJM construction
I called Roger Halle, author of the NY Times article on construction. See
my file HALLE[1,sjm].
Susie
∂16-May-86 1645 VAL pointwise circ'n paper
I wrote a new introduction and an additional section on applications.
\noindent{\bf 1. Introduction}
\bigskip
{\sl Circumscription} (McCarthy 1980, 1986)
is {\sl logical minimization}, that is, the minimization of
predicates subject to restrictions expressed by predicate formulas.
The interpretation of a predicate symbol in a model can be described in two
ways. One is to represent a $k$-ary predicate by a
subset of $U↑k$, where $U$ is the universe of the model. This approach
identifies a predicate with its {\sl extension}.
The other possibility is to represent a predicate by a
Boolean-valued function on $U↑k$. These two approaches are, of course,
mathematically equivalent; but the intuitions behind them are somewhat
different, and they suggest different views on what ``minimizing a
predicate'' might mean.
If a predicate is a set then predicates are ordered by {\sl set inclusion}, and
it is natural to understand the minimality of a predicate as minimality
relative to this order. A {\sl smaller} predicate is a {\sl stronger}
predicate. A
predicate satisfying a given condition is minimal if it cannot be made stronger
without violating the condition. This understanding of minimality leads to
the usual definition of circumscription.
Let us accept now the view of predicates
as Boolean-valued functions, or, in other words,
as {\sl families of truth values}. Each predicate is a family of elements of the
ordered set $\{true,false\}$. Understanding ``smaller'' as ``stronger''
still makes sense; but another approach becomes also possible. We can
think of making a predicate smaller {\sl at a point} $\xi ε U↑k$ as changing its
value at that point from $true$ to $false$. As far as the values at other
points are concerned, we can require, in the simplest case, that they remain
the same; or we can allow them to change in an arbitrary way; or some of them
can be required to remain fixed, and the others alowed to vary.
This leads us to a new definition of circumscription, which, intuitively,
expresses the minimality of a predicate ``at every point''.
It can be interpreted as an ``infinite conjunction'' of ``local''
minimality conditions; each of these conditions expresses the
impossibility of changing the value of a predicate from $true$ to
$false$ at one point. (Formally, this ``infinite conjunction'' will be
represented by a universal quantifier). This is
what we call ``pointwise circumscription''.
We argue that the pointwise approach to circumscription is in some ways
conceptually simpler than the traditional global approach
and, at the same time, leads to generalizations with the additional
flexibility needed in applications to the theory of
commonsense reasoning. Its power is illustrated by applying it to two
problems, one posed in (Hanks and McDermott 1985) and the other in (McCarthy
1986). The pointwise approach is also used here for developing new methods for
computing global circumscription.
\noindent{\bf 10. General Remarks on Applications of Circumscription}
\bigskip
The extant work on the use
of circumscription for formalizing commonsense knowledge has been
justly criticized for the absence of clearly defined principles regarding
the choice of a circumscription policy. For new applications, we often have
to experiment with different sets of varied predicates and with different
assignments of priorities, and look for the circumscription
which allows us to derive
all the desirable consequences, while producing no contradictions or other
objectionable results. This situation is clearly unsatisfactory if
circumscription is to be used for the creation of a general-purpose
knowledge base that can be easily or even automatically extended.
Pointwise circumscription, with its richer repertoire of policies,
is liable to this criticism to an even larger degree.
A solution to this important open problem may possibly take the following form.
We may be able to describe a few sufficiently general
classes of circumscriptive theories, along with
their intended interpretations in the commonsense world, which would use
fixed (or at least clearly delineated) circumscription policies, so that
their properties would be well understood. For example, we may be able to prove
that a theory from such a ``manageable'' class cannot have two
essentially different models (under assumptions which make
multiple models unintuitive); a result of this kind would
show that the form of circumscription employed in this class of theories is
sufficiently strong. We may be able to prove that all these theories are
consistent, so that the circumscription is not too strong. Such
metamathematical results, in combination with the experience of using
theories of these types for formalizing commonsense knowledge,
may possibly give us reasonable confidence that no unpleasant surprises are to be
expected as long as we stay in these manageable parts of the world of
circumscription.
At present, we do not know any results of this kind. But some ``toy examples''
investigated in the literature on applications of circumscription
suggest a number of ideas that can serve as a point of departure for
this work, and we will mention some of them. In the formalizations
of taxonomic hierarchies with exceptions, the circumscription policy should
give a higher priority to more specific information. In reasoning about
actions, the laws of motion should be given a higher priority than the law
of inertia. (These examples are treated in (McCarthy 1986) without the use of
priorities, but this requires
additional ``cancellation'' axioms). In some forms of
temporal reasoning, preference should be given to minimization at earlier
instances of time (Hanks and McDermott 1986).
The study of ``manageable'' circumscriptive theories should be also related to
work on computing circumscription. Ideally, we would like to find classes of
theories which are both semantically manageable and computationally tractable.
∂16-May-86 1746 RA Olivia Stuart, Time
Stuart called re taking your picture for an article she is writing for Time.
Her tel. 434 5241. She'll call back Monday.
∂16-May-86 1748 RA AI Search Committee
Nils is anxious to know whether Wed., May 21, 2:15 - 3:45 is convenient
for you for AI Search Committee meeting.
∂16-May-86 1808 NILSSON@SU-SCORE.ARPA [Brian Fromme <FROMME@SRI-WARBUCKS.ARPA>: [Tom Garvey <GARVEY@SRI-IU.ARPA>: [Pony Express Mailer]
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 16 May 86 18:08:21 PDT
Date: Fri 16 May 86 17:37:07-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: [Brian Fromme <FROMME@SRI-WARBUCKS.ARPA>: [Tom Garvey <GARVEY@SRI-IU.ARPA>: [Pony Express Mailer]
To: jmc@SU-AI.ARPA
Message-ID: <12207266404.53.NILSSON@SU-SCORE.ARPA>
fyi
---------------
1) 15-May Brian Fromme [Tom Garvey <GARVEY@SRI-IU.ARPA>: [Pony Expre
2) 15-May Tom Garvey Danny Cohen's talk on 5/19/86
Message 1 -- ************************
Return-Path: <@SRI-AI.ARPA:FROMME@SRI-WARBUCKS.ARPA>
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Thu 15 May 86 14:06:58-PDT
Received: from SRI-WARBUCKS.ARPA by SRI-AI.ARPA with TCP; Thu 15 May 86 14:06:03-PDT
Date: Thu 15 May 86 13:44:45-PDT
From: Brian Fromme <FROMME@SRI-WARBUCKS.ARPA>
Subject: [Tom Garvey <GARVEY@SRI-IU.ARPA>: [Pony Express Mailer
<Mailer@SRI-]
To: aic-staff
Phone: (415) 859-3448
Date: Thu 15 May 86 11:37:50-PDT
From: Tom Garvey <GARVEY@SRI-IU.ARPA>
Subject: Danny Cohen's Visit
To: aic-staff
Cc: garvey
Message-ID: <VAX-MM(194)+TOPSLIB(120)+PONY(0) 15-May-86 11:37:50.SRI-IU.ARPA>
Folks,
A reminder and a change of venue for the Danny Cohen talk
about the Eastport Study (the infamous SDI software study). It will
take place from 11:15 to 12:15 in the Building G Conference Room.
Come hear how software will help save your miserable, SDI-bashing rear
end in spite of yourself.
Cheers,
Tom
PS -- If the trouble I've had sending this msg is any indication of
the state-of-the-art of software development, then Danny will have his
hands full!
-------
Message 2 -- ************************
Return-Path: <@SRI-AI.ARPA:GARVEY@SRI-WARBUCKS.ARPA>
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Thu 15 May 86 14:34:30-PDT
Received: from SRI-WARBUCKS.ARPA by SRI-AI.ARPA with TCP; Thu 15 May 86 14:34:04-PDT
Date: Thu 15 May 86 14:22:10-PDT
From: Tom Garvey <GARVEY@SRI-WARBUCKS.ARPA>
Subject: Danny Cohen's talk on 5/19/86
To: aic-staff
Message-ID: <VAX-MM(194)+TOPSLIB(120)+PONY(0) 15-May-86 14:22:10.SRI-WARBUCKS.ARPA>
Sorry to leave off the date.
Tom
-------
-------
∂17-May-86 1131 SJG should I sent this to BBoards about the uncertainty workshop?
sjg - I went to this workshop last year and was extremely disappointed.
The reason is that I emphatically did *not* find it a reasonable forum
in which to exchange ideas about uncertainty in AI. The speakers and
participants were extremely close-minded about their approaches; I heard
the following exchange several times:
X: I believe p.
Y: I believe -p.
X: I believe p. And anyone who believes -p is a jerk.
I left the workshop extraordinarily frustrated.
People who find these comments surprising would be well-advised to
have a look at Cheeseman's printed comments at the beginning of
last year's proceedings. He remarks that there is widespread
misunderstanding in AI of standard probabilistic methods, and that
he would name names, but that "libel laws forbid it".
Let me try and moderate this statement in a couple of ways. First,
I think that the problem the workshop professes to address - reasoning
under uncertainty - is an extremely important and interesting one.
The non-monotonic reasoning workshop in New Paltz in 1984 was wonderful;
I'm sorry a continuation hasn't been arranged. But I think that the
organization of *this* workshop makes it unlikely that it leads to
real progress.
Secondly, I will be the first to admit that I have been expounding
my own views on uncertainty in AI for a couple of years. I have
defended these views tenaciously but it is time, I think, to admit
that I was wrong. People who want to know the right answer should
go and ask McCarthy, who seems to have all of this much better
sorted out than I do.
Matt Ginsberg
P.S. Please don't tell McCarthy I said that.
∂17-May-86 1220 NILSSON@SU-SCORE.ARPA hopcroft report
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 17 May 86 12:20:10 PDT
Date: Sat 17 May 86 12:20:38-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: hopcroft report
To: jmc@SU-AI.ARPA
cc: nilsson@SU-SCORE.ARPA
Message-ID: <12207470934.12.NILSSON@SU-SCORE.ARPA>
I have a copy of the Hopcroft report and I think the AI section needs
to be beefed up considerably. And I don't have time to do it.
-------
∂17-May-86 1259 LES Alliant Schedule
To: JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, JJW@SU-AI.ARPA, RPG@SU-AI.ARPA
CC: Bosack@SU-SCORE.ARPA, Tom@SU-SCORE.ARPA
According to current plan, the power installation was presumably completed
on Friday, Steve Ross of Alliant will do the hardware installation on
Tuesday morning and someone else from Allaint will install the operating
system on Wednesday.
I eagerly await the outcome of tests on the new, cheap 100kb modems.
Pehaps we should optimistically assume that they will work and start the
paperwork going on a phone line from here to Lucid.
∂17-May-86 1808 CLT shopping list
strained fruits and veg 2 each of
pears
peaches
applesauce
applesauce+apricots
sweet potatos
carrots
squash
simlac - green label - no iron
banannas - 6
oranges-- eating
∂17-May-86 1818 ROMERO@SRI-AI.ARPA Symbolics Programmer
Received: from SRI-AI.ARPA by SU-AI.ARPA with TCP; 17 May 86 18:17:55 PDT
Date: Sat 17 May 86 18:19:07-PDT
From: Rob Romero <ROMERO@SRI-AI.ARPA>
Subject: Symbolics Programmer
To: jmc@SU-AI.ARPA
Home Address: Kappa Sigma 206 (Box 6537), Stanford
Work Address: SRI AI Center #EK248, 333 Ravenswood, Menlo Park
Home Phones: (415) 327-6521, 853-8891 Work: 859-5667, 859-2887
Prof. McCarthy,
I was directed to you in response to a query about the possibility
of a position (RAship?) somewhere in the Computer Science Dept. that
might require working knowledge with Symbolics Lisp Machines (SSRG,
perhaps?). I am a coterminal student in EES, but have taken extensive
portions of the CS MS curriculum (CS122, 242, 246, 248, 257A, 261,
222, 223, 224 ...), and I am looking for support for this summer and
next year. I would appreciate if you could point me in the right
direction. My resume follows. Thank you.
sincerely
Robert M. Romero
------------------------------------------------------------------------------
ROBERT MAGNUS ROMERO
P.O. Box 6537
535 Campus Drive
Stanford, CA 94305
(415) 327-6521
OBJECTIVE
To obtain a challenging summer position emphasizing applications of
Artificial Intelligence, including Expert Systems, and Lisp Machines.
EDUCATION
Stanford University, Stanford, California.
9/82-6/86 Honors Bachelor of Arts: Economics
9/82-6/86 Bachelor of Science: Electrical Engineering
4/86-6/87 Master of Science: Engineering-Economic Systems
Additional coursework in Operations Research and Mathematics.
EXPERIENCE
6/85-present
Student Associate, SRI Artificial Intelligence Center, Menlo Park,
California. Developed a graphical facility to interactively assert
predicates about the physical structure of a domain into a knowledge
base. This is part of a larger project at SRI, the Procedural Reasoning
System project, which will be used to accomplish real-time reasoning in
a fault-diagnosis expert system for the Space Shuttle.
6/84-9/84
Sofware Engineer, Texas Instruments Data Systems Group, Houston,
Texas. Designed and developed a highly interactive digital signal
processing facility based on a Lisp Machine. This facility was to be
used by geologists and petroleum engineers to filter and analyze seismic
traces. This was part of the Appaloosa project, a large Geophysical
Interpretation Expert System.
10/84-3/85
Consultant, Center for Study of Language and Information (CSLI),
Stanford, California. Assisted users with questions about Xerox Lisp
machine software and about the Tops-20 Operating System.
10/83-6/84
Operator, Low Overhead Timesharing System (LOTS), Stanford.
Executed the filesystem backups for Stanford's Undergraduate Computer
Center. Temporarily shut down two DEC-20's on Saturday mornings, saving
the four RP06's on magnetic tape.
1/83-6/84
Consultant, Stanford LOTS, Stanford, California. Assisted users
with both system-related questions and with programming problems in
Pascal, Lisp, Fortran, and assembly language.
1/83-6/84
Tutor, Learning Assistance Center, Stanford University, California.
Tutored students in courses in Computer Science, Mathematics, Physics,
Electrical Engineering, and Economics.
9/83-12/83
Grader, Electrical Engineering Department, Stanford University,
California. Corrected the homework assignments and programs for a
course in MC68000 Assembly language.
ADDITIONAL INFORMATION
* Chairman, Stanford I.E.E.E. Branch, 1984-1985
* Programming experience in ZetaLisp, Pascal, Fortran Coursework in
Prolog, Ada, and Modula-2.
* Fluency in Spanish and Portuguese. Two years of French courses.
* Member, Kappa Sigma Fraternity.
* Leisure activities include skiing, scuba diving, and tennis.
-------
∂18-May-86 1111 WEISS@su-sushi.arpa re: Turning off call waiting
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 18 May 86 11:11:35 PDT
Date: Sun 18 May 86 11:12:04-PDT
From: Gary M. Weiss <WEISS@su-sushi.arpa>
Subject: re: Turning off call waiting
To: JMC@su-ai.arpa
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Sun 18 May 86 10:52:00-PDT
Message-ID: <12207720596.22.WEISS@su-sushi.arpa>
This works from a Stanford phone. I assume it works for outside phones also.
-------
∂18-May-86 1703 IAM
To: LES, JMC, CLT
LICS
As you know I am giving a paper
on my recent work at the first symposium
of logic in computer science, to be held at M.I.T
in June. I am have also been invited to participate
in a program transformation workshop taking place at Harvard in July.
I was wondering if there is any travel money available
for me to help with this galavanting about?
Ian
∂18-May-86 2020 FRANK@SU-CSLI.ARPA re: Turning off call-waiting (from SAIL's BBOARD)
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 18 May 86 20:19:57 PDT
Date: Sun 18 May 86 20:17:07-PDT
From: Frank Chen <Frank@SU-CSLI.ARPA>
Subject: re: Turning off call-waiting (from SAIL's BBOARD)
To: JMC@SU-AI.ARPA
cc: SU-BBOARDS@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Sun 18 May 86 15:01:00-PDT
Also-known-as: IZZY087%UCLAVM.BITNET@WISCVM
Telephone: (213) 824-5706
I think I saw it in one of those "newsletters" that Pacific Bell includes
with your phone bill sometime last summer.
Frank
-------
∂18-May-86 2032 SIEGMAN@SU-SIERRA.ARPA Call Forwarding
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 18 May 86 20:32:27 PDT
Date: Sun 18 May 86 20:34:15-PDT
From: Tony Siegman <SIEGMAN@SU-SIERRA.ARPA>
Subject: Call Forwarding
To: jmc@SU-AI.ARPA
Message-ID: <12207822939.16.SIEGMAN@SU-SIERRA.ARPA>
There's a BBOARD TELECOM on Lots, on which there have been a series of
msgs in the past year re the fact that a "Call Waiting" signal seems to
break a modem connection if you are doing a terminal session on a line
when the "Call Waiting" signal comes in, and why this happens, and how
you can disable the Call Waiting feature for the duration of one call
by typing "star-something or other". I'm pretty sure the right codes
were in a small folder we got from Pac Tel when we ordered this service.
-------
∂19-May-86 0849 RA Re: reference
[Reply to message recvd: 15 May 86 17:25 Pacific Time]
The publication year for Roszak's book is 1969. It was published by Anchor Books.
∂19-May-86 0907 RA Expense Report (LA and Washington)
You did not give me a receipt for your LA car rental. Do you have it?
Thanks,
∂19-May-86 0918 CHEADLE@SU-SCORE.ARPA Ph.D. Committee meeting
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 19 May 86 09:18:49 PDT
Date: Mon 19 May 86 09:15:26-PDT
From: Victoria Cheadle <CHEADLE@SU-SCORE.ARPA>
Subject: Ph.D. Committee meeting
To: phdcom@SU-AI.ARPA
Office: Margaret Jacks 258, 723-1519
Message-ID: <12207961507.31.CHEADLE@SU-SCORE.ARPA>
will take place in Jacks 352, at 2:15 pm, Tuesday (tomorrow).
Victoria
-------
∂19-May-86 1041 CRISPIN@SUMEX-AIM.ARPA re: Turning off call-waiting (from SAIL's BBOARD)
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 19 May 86 10:40:56 PDT
Date: Mon 19 May 86 10:40:54-PDT
From: Mark Crispin <Crispin@SUMEX-AIM.ARPA>
Subject: re: Turning off call-waiting (from SAIL's BBOARD)
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Sun 18 May 86 15:01:00-PDT
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12207977066.21.CRISPIN@SUMEX-AIM.ARPA>
In general, the source is your friendly local phone hacker. There
is a "TELECOM@RUTGERS" mailing list where a lot of this folklore
is discussed.
-------
∂19-May-86 1056 RA TEX course
I would like to take the TEX course offered at SU this summer. The total
cost (for beginning and intermediate TEX) is $790. Would you be willing
to pay half of it from your unrestricted funds? Zohar will pay the other
half.
Thanks,
∂19-May-86 1224 RA Inference Board Meeting
The meeting, Friday May 23, is at 11:00 instead of 12:00. If it's a problem,
please call Thelma (213) 417 7997 (or tell me and I will call her). If
she doesn't hear from you, it means that the time change is no problem for
you.
∂19-May-86 1301 RA meeting
I am at a staff meeting; will be back in about an hour.
∂19-May-86 1322 VAL Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
AUTOEPISTEMIC LOGIC, STRATIFIED PROGRAMS AND CIRCUMSCRIPTION
Michael Gelfond and Halina Przymusinska
University of Texas at El Paso
Thursday, May 22, 4pm
MJH 252
In Moore's autoepistemic logic, a set of beliefs of a rational agent is described
by a "stable expansion" of his set of premises T. If this expansion is unique then
it can be viewed as the set of theorems which follow from T in autoepistemic logic.
Marek gave a simple syntactic condition on T which guarantees the existence of
a unique stable expansion. We will propose another sufficient condition, which
is suggested by the definition of "stratified" programs in logic programming.
The declarative semantics of such programs can be defined using fixed points
of non-monotonic operators (Apt, Blair and Walker; Van Gelder) or by means of
circumscription (Lifschitz). We show how this semantics can be interpreted in
terms of autoepistemic logic.
∂19-May-86 1348 LES EBOS Kickoff
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
It is now time to decide "What do we do first?"
It appears that eveyone can make it at 11:00am on Tuesday, 5/19.
JMC's office.
∂19-May-86 1448 SJM Dorothy Parker
Sunday night it came to me:
People who DO things exceed my endurance
My God! for a man who solicits insurance!
Is there a file clerk in my mind who has been ransacking the files for
36 hours looking for that penultimate line? Or did it just bob to the
top and get spotted like a license plate number? Or did I make it up?
Susie
∂19-May-86 1516 RICHARDSON@SU-SCORE.ARPA AI Search Committee/May 21
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 19 May 86 15:16:42 PDT
Date: Mon 19 May 86 15:15:21-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: AI Search Committee/May 21
To: buchanan@SUMEX-AIM.ARPA, genesereth@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA,
shah@SU-SIERRA.ARPA, grosof@SU-SUSHI.ARPA, hirsh@SUMEX-AIM.ARPA
Message-ID: <12208027029.25.RICHARDSON@SU-SCORE.ARPA>
There will be an AI Search Committee meeting on Wednesday, May 21 from
2:15 to 3:45 in MJH 220.
-------
∂19-May-86 1538 RA Re: TEX course
[Reply to message recvd: 19 May 86 15:28 Pacific Time]
I checked with Betty before I asked you and she said it cannot be charged
to a grant or a contract. I have no short term plan to depart.
Thanks.
∂19-May-86 1658 RA Hertz receipt
The receipt for your car rental in LA is here; I will use it for
the Expense Report.
∂19-May-86 1717 RA John Walecka
John Walecka from Brentwood Associates would like you to call him re
Dick Gabriel. His tel. (213) 826 6581.
∂20-May-86 0923 CONTRERAS@SU-SCORE.ARPA message
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 20 May 86 09:23:41 PDT
Date: Tue 20 May 86 09:22:46-PDT
From: Tina Contreras <CONTRERAS@SU-SCORE.ARPA>
Subject: message
To: JMC@SU-AI.ARPA
Message-ID: <12208224987.17.CONTRERAS@SU-SCORE.ARPA>
Lori of ACM in New York City phoned, regarding the paper you are submitting
about the touring awards, please call. 212 869-7440 Ext 311.
Tina
-------
∂20-May-86 1107 RA les
Les left the house about 10 minutes ago
∂20-May-86 1113 RA Olivia Stuart
Stuart called re your picture. She talked to Caddes but Caddes has only
black and white pictures and Stuart needs a color photo.
Please call her 434 5241.
∂20-May-86 1637 RA Steve Jobs
The meeting with Steve Jobs is Friday, May 23, 2:30 at NEXT which is at
3475 Deer Creek Road, Palo Alto.
To get there you go on Page Mill past Foothill Expressway west and when
you get to Deer
Creek you turn left and NEXT is part of the first complex on the left at
the farthest building from the road. It looks as if the place is still
under work, but don't let it fool you, it is all finished inside. It's
on the second floor.
∂20-May-86 1728 @SRI-AI.ARPA:LANSKY@SRI-WARBUCKS.ARPA YELLOW FORMS PLEASE!
Received: from SRI-AI.ARPA by SU-AI.ARPA with TCP; 20 May 86 17:28:01 PDT
Received: from SRI-WARBUCKS.ARPA by SRI-AI.ARPA with TCP; Tue 20 May 86 17:11:44-PDT
Date: Tue 20 May 86 16:51:58-PDT
From: Amy Lansky <LANSKY@SRI-WARBUCKS.ARPA>
Subject: YELLOW FORMS PLEASE!
To: genesereth@SUMEX-AIM.ARPA, georgeff@SRI-WARBUCKS.ARPA,
val@SRI-AI.ARPA, jmc@SRI-AI.ARPA, mcdermott-drew@YALE.ARPA,
morgenst@NYU-CSD2.ARPA, yoram@XX.LCS.MIT.EDU, pednault@SRI-AI.ARPA,
rosenschein@SUMEX-AIM.ARPA, shoham@YALE.ARPA, reid@AI.AI.MIT.EDU,
reid%oz@XX.LCS.MIT.EDU, suchman@XEROX.COM
Message-ID: <VAX-MM(194)+TOPSLIB(120)+PONY(0) 20-May-86 16:51:58.SRI-WARBUCKS.ARPA>
All of you have still not sent in your yellow forms finalizing your
plans at Timberline. Please send these in ASAP. For those who have
lost these forms, here is a recreation of the contents. PLEASE
BE DEFINITE IN YOUR ANSWERS... no "maybe I'll be staying on..." answers,
please!! (We are trying to finalize room reservations).
------------------------------------------------------------
PLEASE RETURN THIS FORM ASAP TO:
PLANNING + ACTION WORKSHOP
MARGARET OLENDER
SRI INTERNATIONAL, AI CENTER
333 RAVENSWOOD AVE.
MENLO PARK, CA 94025
Name:
Please circle which nights you want your room:
June July
29 30 1 2 3 4 5
How many people in your party?
Indicate form of your party:
(ex. couple, couple with 2 children, 2 singles, etc.)
For singles, roommate preference (optional):
Dietary Restrictions :
Stipend needs:
(Please break down into airfare, housing, transportation.
An attempt will be made to at least cover airfare costs.)
---------------------------------------------------------------
Here is the contents of the rest of that mailing....
Dear Workshop Attendees:
Now that the summer is rapidly approaching, we would like to
finalize our plans for the workshop. This letter is to get more
definite information from all of you about your stay at Timberline.
Please do not respond with indefinite answers -- we need to make
concrete room and food arrangements at this time.
Enclosed you will find a tentative workshop schedule as well as
more information about housing, payment, and stipends.
You will also find a yellow sheet with questions. PLEASE FILL THIS
OUT AS ACCURATELY AS POSSIBLE and return to the address given.
By the beginning of June we will be sending you a complete schedule
of talks as well as transportation details. Please feel free
to call us if you have further questions.
Amy Lansky Mike Georgeff
(415) 859-4376 (415) 859-4769
LANSKY@SRI-AI GEORGEFF@SRI-AI
Timberline Toll Free Number: (800) 547-1406
!
Arrival
-------
Attendees should fly into the Portland, Oregon airport. Timberline is
approximately a one hour drive from the airport. You may elect to
rent a car or arrange limousine service to and from Timberline. You
must make these arrangements on your own. The limousine service is
AES LIMOUSINE (telephone: (503)255-0325) and is $22.50 per person,
each way (there must be at least two passengers in the limo).
Directions to Timberline will be sent to you closer to the time of the
workshop.
Rooms
-----
We have reserved in advance 30 rooms for the nights of June 29,30 and
July 1. For those of you who want to stay on after the workshop, 10
rooms have been reserved for the nights of July 2,3,4,5. The cost of
the rooms is $61.60/per person (double occupancy), which includes 3
meals per day (starting on the night of the 29th). For those of you
bringing children (i.e. more than 2 people per room), each child is
$8.50 per day, which does NOT include meals. You will therefore have
to pay extra for meals for your children. However, children under 6
eat free.
The rooms are of a variety of types. Some have only one double or one
queen bed, some have two single beds, and the rest have a queen and a
single OR a double and a single. On the yellow form please note the
size and type of your party and exactly which nights you plan to
spend. If you are coming as a couple, a double or queen bed may be
assigned. Families with children will have priority in the rooms with
two beds. For those bringing 2 children, an extra cot can be brought
into your room. For those coming singly, double occupancy will be the
rule (obviously, only in those rooms with 2 beds!). If you have a
roommate preference, please indicate on the yellow form.
The 10 rooms for July 2-5 will be assigned first-come-first-serve.
If we need more rooms, we will try to make arrangements for those nights.
For those of you wanting to stay extra nights (besides June 29--July 5),
please make separate arrangements DIRECTLY WITH TIMBERLINE. Their toll
free number is (800) 547-1406. (One room has already been reserved
for Tate and Drummond for June 27-28). When you call, you should
indicate that you are part of the SRI workshop. (They may have you
talk to the sales office rather than the regular reservations desk).
Food
----
On the evening of your arrival (June 29), dinner will be served in
the public dining room (the Cascade Room). It will be open seating
and a limited menu. After that, breakfast, lunch, and dinner (up through
lunch on July 2) will be served in a private dining room. Only a fixed
menu will be available. If you have special food restrictions,
please advise on the yellow form.
Payment
-------
Each of you will be responsible for paying your full bill upon checking
out of Timberline. Those of you receiving stipends will receive a check
separately, which you can apply to your costs as you desire.
Stipends
--------
We have only a limited source of support for this workshop.
Therefore, we will be awarding stipends only for those who need them.
The approximate award will be $900 for those coming from England or
Australia, $400 for East Coasters or other non-West Coast origins, and
$200 for West Coasters. Our aim is to at least cover airfare costs.
Obviously only a limited number of these awards can be made, so please
indicate on the form the severity of your need!
!
TENTATIVE WORKSHOP SCHEDULE
---------------------------
Sunday June 29:
People arrive in the afternoon and evening
Dinner in the Cascade Dining Room -- public dining
Monday June 30:
Breakfast: 8:00-9:00
Session I: 9:00-10:30 (2 papers)
Coffee Break: 10:30-10:45
Session II: 10:45-12:15 (2 papers)
Lunch: 12:15-1:30
Session III: 1:30-3:45 (3 papers)
Discussion 4:00-5:00
Bar: 6:30-7:30
Dinner: 7:30-9
Tuesday July 1:
Breakfast: 8:00-9:00
Session IV: 9:00-10:30 (2 papers)
Coffee Break: 10:30-10:45
Session V: 10:45-12:15 (2 papers)
Lunch: 12:15-1:30
Session VI: 1:30-3:00 (2 papers)
Bar: 6:30-7:30
Dinner: 7:30-8:30
Discussion: 8:30-10
Wednesday July 2:
Breakfast: 8:30-9:30
Session VII: 9:30-11:00 (2 papers)
Coffee Break: 11:00-11:15
Session VIII: 11:15-12:45 (2 papers)
Lunch: 12:45-2
-------
∂20-May-86 1758 LANSKY@SRI-WARBUCKS.ARPA [Pony Express Mailer <Mailer@SRI-WARBUCKS.ARPA>:
Received: from SRI-WARBUCKS.ARPA by SU-AI.ARPA with TCP; 20 May 86 17:35:49 PDT
Date: Tue 20 May 86 17:23:50-PDT
From: Amy Lansky <LANSKY@SRI-WARBUCKS.ARPA>
Subject: [Pony Express Mailer <Mailer@SRI-WARBUCKS.ARPA>:
Undeliverable message]
To: val@SU-AI.ARPA, jmc@SU-AI.ARPA
Message-ID: <VAX-MM(194)+TOPSLIB(120)+PONY(0) 20-May-86 17:23:50.SRI-WARBUCKS.ARPA>
In-Reply-To: Message from "Anna Karlin <KARLIN@su-sushi.arpa>" of Tue
20 May 86 10:35SU-AI.ARPA
Oops, had the wrong address... Here's another try:
------------------
---------------
Return-Path: <>
Date: Tue 20 May 86 17:13:09-PDT
From: Pony Express Mailer <Mailer@SRI-WARBUCKS.ARPA>
Subject: Undeliverable message
To: LANSKY@SRI-WARBUCKS.ARPA
Reply-To: <Postmaster@SRI-WARBUCKS.ARPA>
Pony Express was unable to deliver mail to the following
recipients for the reasons listed.
val @ SRI-AI.ARPA : 550 No such local mailbox as "val", recipient rejected
Undelivered message follows:
----------------------------------------
Date: Tue 20 May 86 16:51:58-PDT
From: Amy Lansky <LANSKY@SRI-WARBUCKS.ARPA>
Subject: YELLOW FORMS PLEASE!
To: genesereth@SUMEX-AIM.ARPA, georgeff@SRI-WARBUCKS.ARPA,
val@SRI-AI.ARPA, jmc@SRI-AI.ARPA, mcdermott-drew@YALE.ARPA,
morgenst@NYU-CSD2.ARPA, yoram@XX.LCS.MIT.EDU, pednault@SRI-AI.ARPA,
rosenschein@SUMEX-AIM.ARPA, shoham@YALE.ARPA, reid@AI.AI.MIT.EDU,
reid%oz@XX.LCS.MIT.EDU, suchman@XEROX.COM
Message-ID: <VAX-MM(194)+TOPSLIB(120)+PONY(0) 20-May-86 16:51:58.SRI-WARBUCKS.ARPA>
All of you have still not sent in your yellow forms finalizing your
plans at Timberline. Please send these in ASAP. For those who have
lost these forms, here is a recreation of the contents. PLEASE
BE DEFINITE IN YOUR ANSWERS... no "maybe I'll be staying on..." answers,
please!! (We are trying to finalize room reservations).
------------------------------------------------------------
PLEASE RETURN THIS FORM ASAP TO:
PLANNING + ACTION WORKSHOP
MARGARET OLENDER
SRI INTERNATIONAL, AI CENTER
333 RAVENSWOOD AVE.
MENLO PARK, CA 94025
Name:
Please circle which nights you want your room:
June July
29 30 1 2 3 4 5
How many people in your party?
Indicate form of your party:
(ex. couple, couple with 2 children, 2 singles, etc.)
For singles, roommate preference (optional):
Dietary Restrictions :
Stipend needs:
(Please break down into airfare, housing, transportation.
An attempt will be made to at least cover airfare costs.)
---------------------------------------------------------------
Here is the contents of the rest of that mailing....
Dear Workshop Attendees:
Now that the summer is rapidly approaching, we would like to
finalize our plans for the workshop. This letter is to get more
definite information from all of you about your stay at Timberline.
Please do not respond with indefinite answers -- we need to make
concrete room and food arrangements at this time.
Enclosed you will find a tentative workshop schedule as well as
more information about housing, payment, and stipends.
You will also find a yellow sheet with questions. PLEASE FILL THIS
OUT AS ACCURATELY AS POSSIBLE and return to the address given.
By the beginning of June we will be sending you a complete schedule
of talks as well as transportation details. Please feel free
to call us if you have further questions.
Amy Lansky Mike Georgeff
(415) 859-4376 (415) 859-4769
LANSKY@SRI-AI GEORGEFF@SRI-AI
Timberline Toll Free Number: (800) 547-1406
!
Arrival
-------
Attendees should fly into the Portland, Oregon airport. Timberline is
approximately a one hour drive from the airport. You may elect to
rent a car or arrange limousine service to and from Timberline. You
must make these arrangements on your own. The limousine service is
AES LIMOUSINE (telephone: (503)255-0325) and is $22.50 per person,
each way (there must be at least two passengers in the limo).
Directions to Timberline will be sent to you closer to the time of the
workshop.
Rooms
-----
We have reserved in advance 30 rooms for the nights of June 29,30 and
July 1. For those of you who want to stay on after the workshop, 10
rooms have been reserved for the nights of July 2,3,4,5. The cost of
the rooms is $61.60/per person (double occupancy), which includes 3
meals per day (starting on the night of the 29th). For those of you
bringing children (i.e. more than 2 people per room), each child is
$8.50 per day, which does NOT include meals. You will therefore have
to pay extra for meals for your children. However, children under 6
eat free.
The rooms are of a variety of types. Some have only one double or one
queen bed, some have two single beds, and the rest have a queen and a
single OR a double and a single. On the yellow form please note the
size and type of your party and exactly which nights you plan to
spend. If you are coming as a couple, a double or queen bed may be
assigned. Families with children will have priority in the rooms with
two beds. For those bringing 2 children, an extra cot can be brought
into your room. For those coming singly, double occupancy will be the
rule (obviously, only in those rooms with 2 beds!). If you have a
roommate preference, please indicate on the yellow form.
The 10 rooms for July 2-5 will be assigned first-come-first-serve.
If we need more rooms, we will try to make arrangements for those nights.
For those of you wanting to stay extra nights (besides June 29--July 5),
please make separate arrangements DIRECTLY WITH TIMBERLINE. Their toll
free number is (800) 547-1406. (One room has already been reserved
for Tate and Drummond for June 27-28). When you call, you should
indicate that you are part of the SRI workshop. (They may have you
talk to the sales office rather than the regular reservations desk).
Food
----
On the evening of your arrival (June 29), dinner will be served in
the public dining room (the Cascade Room). It will be open seating
and a limited menu. After that, breakfast, lunch, and dinner (up through
lunch on July 2) will be served in a private dining room. Only a fixed
menu will be available. If you have special food restrictions,
please advise on the yellow form.
Payment
-------
Each of you will be responsible for paying your full bill upon checking
out of Timberline. Those of you receiving stipends will receive a check
separately, which you can apply to your costs as you desire.
Stipends
--------
We have only a limited source of support for this workshop.
Therefore, we will be awarding stipends only for those who need them.
The approximate award will be $900 for those coming from England or
Australia, $400 for East Coasters or other non-West Coast origins, and
$200 for West Coasters. Our aim is to at least cover airfare costs.
Obviously only a limited number of these awards can be made, so please
indicate on the form the severity of your need!
!
TENTATIVE WORKSHOP SCHEDULE
---------------------------
Sunday June 29:
People arrive in the afternoon and evening
Dinner in the Cascade Dining Room -- public dining
Monday June 30:
Breakfast: 8:00-9:00
Session I: 9:00-10:30 (2 papers)
Coffee Break: 10:30-10:45
Session II: 10:45-12:15 (2 papers)
Lunch: 12:15-1:30
Session III: 1:30-3:45 (3 papers)
Discussion 4:00-5:00
Bar: 6:30-7:30
Dinner: 7:30-9
Tuesday July 1:
Breakfast: 8:00-9:00
Session IV: 9:00-10:30 (2 papers)
Coffee Break: 10:30-10:45
Session V: 10:45-12:15 (2 papers)
Lunch: 12:15-1:30
Session VI: 1:30-3:00 (2 papers)
Bar: 6:30-7:30
Dinner: 7:30-8:30
Discussion: 8:30-10
Wednesday July 2:
Breakfast: 8:30-9:30
Session VII: 9:30-11:00 (2 papers)
Coffee Break: 11:00-11:15
Session VIII: 11:15-12:45 (2 papers)
Lunch: 12:45-2
-------
-------
-------
∂21-May-86 0506 FEIGENBAUM@SUMEX-AIM.ARPA AI Search/Opinion
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 21 May 86 05:05:57 PDT
Date: Wed 21 May 86 05:06:25-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: AI Search/Opinion
To: nilsson@SU-SCORE.ARPA, buchanan@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA,
genesereth@SUMEX-AIM.ARPA
Message-ID: <12208440464.14.FEIGENBAUM@SUMEX-AIM.ARPA>
I just woke up, panicked that I had forgotten last evening to
write to you the gist of my conversation with Nils yesterday
afternoon about the "AI search". I know the decision is coming up
within a few hours, so I want to pass on thoughts (but briefly, and
get back to sleep).
I read through all the AI Search folders yesterday. This year, the
candidates are awesome. I wish we could have them all, and Nils and I
spent some time discussing how to get more than one.
I was specially impressed by the file on Mitchell. After Gibbons talked
to me, Nils, and all of us about his desire to use one of his "famous
five" slots for an AI person with an engineering orientation to inspire
the diffusion of AI to the various disciplines of the Engineering
School, I thought long and hard about who I thought would satisfy this
important need and would at the same time be prominent in AI and
an inspiration to students (much needed). The best candidate I could
think of was Mitchell. Even though he was my recomendee,
I was still awed by the accumulation of
letters: e.g. Newell's sensational letter, or Bledsoe's letter.
Mitchell not only has the "engineering orientation" (for example,
he won the best-paper award at the 1984 VLSI-CAD conference), but he
offered/promised to me that if appointed he was willing to play the
role that Gibbons indicated (it would not be a burden but a pleasure,
within bounds). However, the important thing is that he combines this
with world-class performance in the mainline of traditional AI science.
He won the Computers and Thought Award for the version space approach
to machine learning; his work on LEX, and the learning apprentice approaches
is excellent; and overall he is at least tied for first in the
best-in-the-world ratings in machine learning research, in my view. He
is also much talked about in the AI community as an educator. Finally,
he is a wonderful person that one would really like to have
as a collaborator and colleague.
Mitchell told me that he would accept a Stanford job in preference
to the CMU job that is being proferred.
I would suggest a "mixed strategy":
a) while Mitchell is available (which is now, and won't be again for
many years if ever if he goes to CMU), we move to get him as a tenured
appointment. This gives us another high-quality AI person with
permanence, i.e. builds some permanent structure onto the AI faculty
and adds some structure to our long-term relation to engineering. In
this way, we will not be blown by the winds of six years from now concerning
"what shall the department do with this slot?", such as we might be if
a junior person appointed now didn't quite make it. In Mitchell we're
getting a known quantity of excellence and we have him permanently.
b) since Mitchell can be thought of partly as a "systems" person as
well as an AI person (as Vaughan Pratt is thought of partly as a
theorist and partly as a systems person), we leverage the situation
a bit and try to eek out a second appointment for one of the
remarkable new-PhDs in the AI Search. Nils and I discussed a couple
of alternative ways of pursuing this.
In short: re Mitchell, this is a tremendous opportunity. Let's grab
it. Let's also grab for one of the others.
-------
∂21-May-86 0900 JMC
formal wear
∂21-May-86 1009 JJW Alliant
To: JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, LES@SU-AI.ARPA, RPG@SU-AI.ARPA,
greep@Camelot, Bosack@SU-SCORE.ARPA, Tom@SU-SCORE.ARPA
The Alliant folks installing the machine will be giving a presentation
about how to use it today at 1:30 p.m. I've reserved MJH 352 for this.
If people are unable to attend, we'll probably schedule a meeting later
this week, or next week, to review the same information.
Please forward this message to anyone I've forgotten.
∂21-May-86 1021 UNIETIS@su-sushi.arpa cs306
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 21 May 86 10:21:06 PDT
Date: Wed 21 May 86 10:19:29-PDT
From: David R. Unietis <UNIETIS@su-sushi.arpa>
Subject: cs306
To: jmc@su-ai.arpa
Message-ID: <12208497455.21.UNIETIS@su-sushi.arpa>
I was a student in your CS306 class last fall, and I have not yet received
a grade for the class (my final was in within the 8-day time period, but
after the deadline to get a grade turned in that quarter). I am an honors
co-op student, and I need to get a grade for this class in order to
graduate.
Thanks,
David Unietis (unietis@sushi)
-------
∂21-May-86 1021 RICHARDSON@SU-SCORE.ARPA Robotics Search Committee
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 21 May 86 10:21:27 PDT
Date: Wed 21 May 86 10:16:49-PDT
From: Anne Richardson <RICHARDSON@SU-SCORE.ARPA>
Subject: Robotics Search Committee
To: cannon@SU-SIERRA.ARPA, reynolds@SU-SCORE.ARPA, binford@SU-WHITNEY.ARPA,
grosof@SU-SUSHI.ARPA, vistnes@SU-AI.ARPA, jmc@SU-AI.ARPA,
bscott@SU-SCORE.ARPA
Message-ID: <12208496972.13.RICHARDSON@SU-SCORE.ARPA>
There will NOT be a robotics search committee meeting today (May 21) as
originally planned.
-------
∂21-May-86 1146 RA
* grade
A * grade means "no grade reported". It is usually there when things
weren't turned in on time. The * is sort of like the grade (or NC) is
on hold for the time being.
∂21-May-86 1443 VAL Mints' address
The address of the Institute of Cybernetics in Tallinn is: Akadeemia Tee 21.
∂21-May-86 1509 WINOGRAD@SU-CSLI.ARPA PhD Program revision
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 21 May 86 14:54:26 PDT
Date: Wed 21 May 86 14:48:33-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: PhD Program revision
To: Phdcom@SU-AI.ARPA, phd-program@SU-SCORE.ARPA, faculty@SU-SCORE.ARPA,
jlh@SU-SONOMA.ARPA
cc: Winograd@SU-CSLI.ARPA
At the last full faculty meeting, it was voted to approve the plans for
the revision of the PhD program, subject to further discussion on the
exact makeup of the comprehensive exam. At one of the upcoming faculty
meetings this Spring, we will need to take final action. The committee
has collected syllabus information and met to discuss the comprehensive.
This message includes our draft proposal. Victoria Cheadle
(CHEADLE@SCORE) can give you a complete updated version of the proposal,
incorporating these. Also since the previous meeting, an additional
issue has come up that was not in the original proposal: the
requirements for students from other departments to get a PhD Minor in
Computer Science. A proposal for that appears at the end of this
message.
------
Summary of Comprehensive Exam:
Instead of the current six one-hour exams, we will have three 3-hour
exams. These can be passed separately (e.g., in different quarters or
years). Comps will still be given in Winter and Spring. A single
committee will be responsible for putting together the full exam
(although they are free to divvy up the work by sections). The three
areas are labelled "Systems [covering the old Hardware and Software
exams]", "Theory [MTC and AA]", and "Specialized areas [AI, NA, and new
material on Graphics and Databases]". The following is a draft of a
potential syllabus for the three, along with the names of the people who
will be cleaning up the details. If you have comments, please let them
know. We want to get general approval at the faculty meeting for the
overall structure. The comp committee will of course be free to modify
the readings each year.
In deciding what should go into the "Specialized areas" exam, we took
into account a number of considerations. An "ideal" topic would have
all of the following:
1) It is considered a significant topic (one that every student
should be familiar with) by computer scientists in general.
2) It involves specialized methods that are not part of the regular
systems or theory curriculum.
3) There is substantial faculty interest and expertise in the topic
at Stanford.
4) Courses are taught here that cover the material.
5) There is a good introductory-graduate text that incorporates the
relevant material.
Not every topic fits all of these, and we did not expect to find a hard
and fast rule for deciding what should be in. Presumably it will change
over time as the field (and our department) evolves. The exact
proportions of the different topics will be subject to the decision of
the committee making up the exam. That is, the fact that there are four
topics need not mean that each will count for exactly 25% of the score.
A student must pass the entire "specialized areas" part with an
acceptable overall score (not based on a minumum in each area). We felt
that it would raise too much confusion to try to keep track of results
of scores on these subtopics in a context of changes over the years.
------
SYSTEMS [to be reviewed by Lantz and Hennessey]
Harold Abelson and Gerald Sussman, Structure and Interpretation of
Computer Programs, MIT, 1985
Alfred V. Aho, Sethi, R., and Jeffrey D. Ullman, Compilers --
Principles, Techniques, and Tools, Addison-Wesley, 1986. All except
sections 9.11-9.12, 10.9-10.13.
Bell, C.G., Mudge, J.C., and McNamara, J.E., Computer Engineering -- A
DEC View of Hardware Systems Design, Digital Press, 1978.
M. Ben-Ari, Principles of Concurrent Programming, Prentice-Hall
International, 1982. Material on concurrent programming.
Kogge, P., The Architecture of Pipelined Computers, McGraw-Hill, 1981.
M. Morris Mano, Computer System Architecture, Second Edition,
Prentice-Hall, 1982. Basic logic design, data representation, and
computer organization. Material taught in CS 108 and CS 112.
James L. Peterson and Abraham Silberschatz, Operating System Concepts,
Addison-Wesley, Second Edition, 1985.
Andrew Tannenbaum, Structure Computer Organization, Chapters ?
Terrence W. Pratt, Programming Languages: Design and Implementation,
Second edition, Prentice-Hall, 1984.
------
THEORY [to be reviewed by Guibas and Pratt]
Alfred V. Aho, John E. Hopcroft, and Jeffrey D. Ullman, Data Structures
and Algorithms, Addison-Wesley, 1983.
Herbert B. Enderton, A Mathematical Introduction to Logic, Academic
Press, 1972, Chapters 1--2.
Michael R. Garey and David S. Johnson, Computers and Intractability,
Freeman, 1979, Chapters 1--3, 7.
John E. Hopcroft and Jeffrey D. Ullman, Introduction to Automata Theory,
Languages, and Computation, Addison-Wesley, 1979, Chapters 1--3,
4.1--4.6, 5--7, 8.1--8.5.
Donald E. Knuth, The Art of Computer Programming, Volume 1, Second
Edition, Addison-Wesley, 1973, Section 1.2 (except for Subsection
1.2.11.3)
Zohar Manna, Introduction to Mathematical Theory of Computation,
McGraw-Hill, 1973, Chapters 1--3.
Nils Nilsson, Principles of Artificial Intelligence, Kaufman, 1980,
Chapters 4--6.
Sedgewick, Robert, Algorithms, Addison-Wesley, 1983.
--------
SPECIALIZED AREAS [To be reviewed by Winograd and Rosenbloom]
ARTIFICIAL INTELLIGENCE
Elaine Rich, Artificial Intelligence, McGraw-Hill, 1983.
DATA BASES
Jeffrey Ullman, Principles of Data Base Systems, Computer Science
Press, 1982
GRAPHICS
Newman and Sproull, Principles of Interactive Computer Graphics,
Chapters 1--5, and 15--18.
NUMERICAL ANALYSIS
Kendall E. Atkinson, An Introduction to Numerical
Analysis, Wiley, 1978, Chapters 1--3, 5, 7, 8 (except
Sections 2.8, 2.10, 5.4).
--------------
Proposal for PhD Minor
The old version of the minor required a Masters pass on the comp. Since
we no longer have our own masters students taking the comp, this is not
appropriate, since it would lead to extremely bimodal distribution on
the exam. Some students did a careful examination of requirements from
other departments, and came up with the following. More details are on
{SCORE}<winograd>phd-minor.txt
"15 units of CS coursework, numbered 200 or above, including at
least three of the Master's core courses (to provide breadth) and
one course numbered above 300 (to provide depth). One of the
courses taken must include a significant programming project to
demonstrate programming proficiency. A GPA of 3.0 or better must
be maintained."
-------
∂21-May-86 1504 VAL mcc visit
Boyer said that my visits need not be tightly coupled with yours and that he'd like
me to come at the beginning of June. So let's go on different dates this time.
∂21-May-86 1517 RA Jack Harper
Jack Harper from Daedalux Machines (313) 988 3383 called. He will call
again tomorrow.
∂21-May-86 1520 RA Academic costume
I called the bookstore and they can Federal Express it for you from LA by
next Wednesday. I need the following:
your height
your weight
jacket size
shirt sleeve length
Your hat size is 7.5 am I right?
They need this in order to phone in the order.
Thanks,
∂21-May-86 1558 KARP@SUMEX-AIM.ARPA Review of the comp reading list
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 21 May 86 15:57:57 PDT
Date: Wed 21 May 86 15:49:39-PDT
From: Peter Karp <KARP@SUMEX-AIM.ARPA>
Subject: Review of the comp reading list
To: phdcom@SU-AI.ARPA, jlh@SU-SONOMA.ARPA
Message-ID: <12208557561.22.KARP@SUMEX-AIM.ARPA>
Since some of you reviewers were not at yesterday's meeting, I'd like to
repeat my observation that the length of the reading list has grown by
approximately 50% over its length before the Phd Committee got hold of
it. I ask the reviewers to consider each addition to the list very
carefully because I feel (and I think most other students do as well)
that the comp covers a lot of material already, and we do not want to
over-burden the students with this drastic increase in the amount of
material. When we set out to revise the comp I do not recall that
anyone felt that the reading list did not cover enough material (except
perhaps the systems people?). In any event, it's a simple matter to add
a line of text to a file because "it's not a bad idea for people to read
this", but that line of text will translate into a lot of time spent by
a lot of people.
Peter
-------
∂21-May-86 1901 CLT shopping list
spray n wash
strained fruits and veg 2 each of
squash
applesauce+apricots
rice cereal
beer guiness and XX
paper towels
cottage cheese
veg.
∂21-May-86 2000 JMC
read
∂22-May-86 0030 VAL Reminder: Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
AUTOEPISTEMIC LOGIC, STRATIFIED PROGRAMS AND CIRCUMSCRIPTION
Michael Gelfond and Halina Przymusinska
University of Texas at El Paso
Thursday, May 22, 4pm
MJH 252
In Moore's autoepistemic logic, a set of beliefs of a rational agent is described
by a "stable expansion" of his set of premises T. If this expansion is unique then
it can be viewed as the set of theorems which follow from T in autoepistemic logic.
Marek gave a simple syntactic condition on T which guarantees the existence of
a unique stable expansion. We will propose another sufficient condition, which
is suggested by the definition of "stratified" programs in logic programming.
The declarative semantics of such programs can be defined using fixed points
of non-monotonic operators (Apt, Blair and Walker; Van Gelder) or by means of
circumscription (Lifschitz). We show how this semantics can be interpreted in
terms of autoepistemic logic.
∂22-May-86 0609 F68349%BARILAN.BITNET@WISCVM.WISC.EDU Logician
Received: from WISCVM.WISC.EDU by SU-AI.ARPA with TCP; 22 May 86 06:09:09 PDT
Received: from (MAILER)BARILAN.BITNET by WISCVM.WISC.EDU on 05/22/86 at
07:13:31 CDT
Date: Thu, 22 May 86 15:13 P
From: Larry Manevitz
<F68349%BARILAN.BITNET@WISCVM.WISC.EDU>
To: jmc@su-ai.arpa
Subject: Logician
Dear Prof. McCarthy
I am writing at the suggestion of Yuri Gurevich (of the U. of
Michigan). I am a logician who has started to work
in computer science (in particular areas related to aspects of
artificial intelligence). Prof. Gurevich indicated that you might
be interested in someone with my background.
If you have an interest I will send you (via email or hardcopy)
my CV. However, briefly my background is as follows:
My Ph.D. was completed as a student of Abraham Robinson in model
theory. Since then most of my work has been in applications of model
theory to other (rather diverse) areas of mathematics.
(I have given solutions (or partial solutions) to problems of
Gaifman, Kreisel, Sikorski and shown how to use methods of model theory
in group theory, ring theory and stochastic processes. Much of this
work was with different collaborators.)
*********
I have now decided to work in subjects related to
artificial intelligence. Over the past couple of years (at my
request) I have been teaching many C.S. courses at both first and
second degree levels (artificial intelligence, computability theory
complexity theory, formal languages and automata theory, as well
logic courses). I think you can guess the general scope of my
background from this; I know pretty well the "theoretical" side
of C.S. but would be somewhat weak on say architecture. I should
also point out that I worked for IBM as a consulting prog/analyst
for more than 3 years and from this experience know what it is to
bring fairly large projects to completion.
You should also note that over the years I've had an interest
in AI related subjects. For example, at Yale I organized a special
crossdisciplinary seminar on brain modeling.
At the moment I am (besides finishing up the semester here )
studying annealing heuristics with a graduate student in physics
here and discussing foundations of probablistic inference with
another logician. I hope eventually to try and implement these
ideas in e.g. some expert system, or certain NP problems.
********
I quite realize that I am starting a new field and it seems
to me very important to work at a major center for some period.
Probably the wisest thing will be to collaborate with some
more experienced workers.
Thus I would be very pleased to come to Stanford in some capacity
if something could be arranged. I could come almost immediately.
********
For references I would suggest the following:
1.Prof. David Harel Weizmann Institute Rehovot Israel
2.Prof. Jonathan Stavi Dept Comp.Sci. Bar Ilan Univ. Ramat Gan Israel
3. Prof. Dov Gabbay Dept Comp. Sci. Imperial College London
4. Prof. Angus Macintyre Dept. Math. Oxford University
*********
I thank you for your consideration. I realize that a letter
of this length steals a fair amount of time.
********
Sincerely yours
Larry Michael Manevitz
e-mail F68349@BARILAN (on BITNET)
tel: ISRAEL (03) 718407 work
ISRAEL (08) 463767 home
addr: Dept Math and Comp Sci
Bar Ilan University
Ramat Gan 52100
Israel
∂22-May-86 0800 JMC
call acm
∂22-May-86 0900 JMC
atenolol
∂22-May-86 1006 CHEADLE@SU-SCORE.ARPA Friday's AI Qual
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 22 May 86 10:06:44 PDT
Date: Thu 22 May 86 10:05:31-PDT
From: Victoria Cheadle <CHEADLE@SU-SCORE.ARPA>
Subject: Friday's AI Qual
To: jmc@SU-AI.ARPA, tob@SU-AI.ARPA, genesereth@SUMEX-AIM.ARPA,
buchanan@SUMEX-AIM.ARPA, rosenbloom@SUMEX-AIM.ARPA, tenenbaum@SRI-KL.ARPA,
waldinger@SRI-AI.ARPA, green@KESTREL.ARPA
Office: Margaret Jacks 258, 723-1519
Message-ID: <12208757058.20.CHEADLE@SU-SCORE.ARPA>
This is to remind all of you that the second half of the AI quals
is scheduled for this Friday morning, 8:30-11:30 am, with a lunch
meeting immediately following in Jacks 352. The specific schedule
giving room and student assignments will be posted on the door of
Jacks 252. There is also a packet of information that has been
prepared for each of you available in my office. Please pick it up
before proceeding to your particular exam room.
Feel free to contact me if you have any questions.
Victoria
-------
∂22-May-86 1122 RA Jack Harper
Jack Harper from Daedalux Machines called (303) 988 3383. He will call again
tomorrow.
∂22-May-86 1200 JMC
mad
∂22-May-86 1254 GRP Files on IBM RT
To: JMC@SU-AI.ARPA, LES@SU-AI.ARPA, JJW@SU-AI.ARPA, DAN@SU-AI.ARPA,
RWW@SU-AI.ARPA
CC: GRP@SU-AI.ARPA, CLT@SU-AI.ARPA
I restored your files from LaBrea to the IBM. However, I have
reason to believe that the new version of the IBM software may
have errors in the network code, and the files may not all have
been transferred correctly. Carolyn has also mentioned
seeing evidence of network problems since the new version was
installed.
∂22-May-86 1331 LES Qlisp task status
∂22-May-86 1246 PUCCI@USC-ISI.ARPA Qlisp task status
Received: from USC-ISI.ARPA by SU-AI.ARPA with TCP; 22 May 86 12:46:38 PDT
Date: 22 May 1986 15:23:50 EDT
From: PUCCI@USC-ISI.ARPA
Subject: Qlisp task status
To: les@SU-AI.ARPA
cc: pucci@USC-ISI.ARPA
Les,
I haven't forgotten you. We are in the midst of reorganizing and are
quite busy. I will send the draft of the task tomorrow. Please look it
over and suggest any changes. The only hangup at this time is getting ADP
approval for the workstations. This will obviously slow things up and a 1
June start would be a miracle. Hopefully all will be completed by 15
June. I will do my best to keep things moving.
John
-------
∂22-May-86 1550 VAL Perlis on Circ. with sets
I read the paper and my impression was the same as yours. Two more comments:
1. The basic version of his system seems almost isomorphic to the second-order
circ'n. He merely writes xεP instead of Px and treats the postulate of the
second-order logic which guarantees the existence of predicates as a non-logical
axiom, and this is the question of notation and terminology.
2. He claims that he can write axioms like Min(p) to express that p is to be
minimized. This cannot be done the way he wants (he admitted that as a result
of our exchange of messages).
∂22-May-86 1701 SJM book
I have compiled the essay drafts as we discussed. Recall that you are
going to look at them carefully (making many marks with writing
implements) and then we are going to discuss them lengthily and
profoundly Saturday. I have placed said compilation on top of your
little office refrigerator, since there was no other paper there for it
to blend in with. Do not neglect this! Otherwise I shall have to write
the book by myself, making factual errors and expressing opinions you
wouldn't express at a dogfight.
Susie
∂22-May-86 2041 AN02@A.CS.CMU.EDU Re: reference to GPS
Received: from A.CS.CMU.EDU by SU-AI.ARPA with TCP; 22 May 86 20:40:48 PDT
Date: Thu, 22 May 86 23:40 EDT
From: Allen.Newell@A.CS.CMU.EDU
To: John McCarthy <JMC@SU-AI.ARPA>
Subject: Re: reference to GPS
In-Reply-To: "John McCarthy's message of 22 May 86 17:13-EST"
Message-Id: <22May86.234005.AN02@A.CS.CMU.EDU>
John: The first unpublished paper on GPS was:
Newell, A., Shaw, J. C. & Simon, H. A., "Preliminary Description of General
Problem Solving Program-I (GPS-I)", CIP Working Paper #7, Carnegie Institute
of Technology, Dec 1957.
The first published paper on GPS was:
Newell, A., Shaw, J. C., & Simon, H. A., "Report on a general problem-solving
program for a computer", Information Processing: Proceedings of the
International Conference on Information Processing, UNESCO, Paris, 1960,
pp 256-264. (RAND P-1584, and reprinted in Computers and Automation, July
1959).
The remarks on learning refer no doubt to:
Newell, A. Shaw, J. C., & Simon, H. A. "A variety of intelligent learning
in a General Probelm Solver", in Self-Organizing Systems, Yovits, M. C., &
Cameron, S. eds., Pergammon, 1960, pp 153-189 (RAND P-1742).
My recollection (without going and digging up the papers to reread them)
is that the notion of generality that we expressed explicity was that of
separating the problem solving structure from the task definition. All prior
efforts (what few there had been, such as LT) were built around the data
structures of the domain. That is, the search tree was a tree of theorems,
or chess positions or whatever. With GPS it was a tree of goals and subgoals.
(It was also expressed, I am almost sure, as "GPS has pretensions of
generality".) There was no notion that all problems were expressible as the
transformation of one situation into another. Indeed, there were subgoals
(ie, subproblems) of form "Reduce the difference D on X" and "Apply operator
Q to X", neither of which were in the form "Transform X to Y". Again, I
would have to check, but I do not believe that we ever address the scope of
the three goals types of GPS to express all problems. In fact, since we
considered adding additional goal types (and did at one time add "Select
X from set {Yi}" -- maybe even in the "intelligent learning" paper) I don't
think there was any discussion at all of the scope problem.
I'm not sure that any of the latter discussion should inhibit you from passing
your judgment of why GPS failed, but you might want to distinguish what it
failed at -- i.e., whose aspiration got scotched.
On the other hand -- and I guess alas for us, I would have to conclude --
the current work on Soar (my current candidate for a general intelligent
architecture, along with John Laird and Paul Rosenbloom, but one that Herb
doesn't have to bear responsibiltiy for) does take exactly the presumption
that all problems can be expressed in problem spaces, ie, can be taken
as starting from some initial state in a space and applying operators to
that state until a desired (goal) state is discovered. So we seem to have
come around to where your injunction does apply. When your injunction is
correct or not is beside the point.
The work in production systems is usually traced to A. Newell and H. A. Simon,
Human Problem Solving, Prentice-Hall, 1972. But it goes pack to some earlier
papers in 1965 and 1967, but you don't seem like you need those references
(holler if you want them). Productions systems were not widely used in
the CMU envioronment until after running PS systems were created in the early
1970s.
Ok? AN
∂23-May-86 0130 GRP Unix system calls
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
At Carolyn's suggestion, I organized the names of the Unix system
calls by topic, to make it easier to get some idea of what services
the Unix kernel provides. This is in UNXCAL.TXT[1,GRP], which also
includes a short introduction. The names and one-line descriptions
came straight out of the manual pages.
∂23-May-86 0131 GRP files as active objects
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
I revised part of some notes I wrote a year or two ago about file
operations, which I think is relevant. It's in FILSYS.TXT[1,GRP].
∂23-May-86 0609 jeh@gvax.cs.cornell.edu Memo about the NSF report on scientific computing
Received: from CU-ARPA.CS.CORNELL.EDU by SU-AI.ARPA with TCP; 23 May 86 06:09:18 PDT
Received: by cu-arpa.cs.cornell.edu (5.31/4.30)
id AA10731; Fri, 23 May 86 09:07:09 EDT
Date: Fri, 23 May 86 09:07:02 edt
From: jeh@gvax.cs.cornell.edu (John E. Hopcroft)
Message-Id: <8605231307.AA05779@gvax.cs.cornell.edu>
Received: by gvax.cs.cornell.edu (4.12/4.30)
id AA05779; Fri, 23 May 86 09:07:02 edt
To: dana.scott@c.cs.cmu.edu, lampson@decwrl.dec.com, mccarthy@su-ai.arpa,
rabin%lbl-ux7.arpa@csnet-relay.arpa, schwartz@nyu-cmcl1.arpa
Subject: Memo about the NSF report on scientific computing
Please get your comments back to me so that I can see if we are
in agreement. I need to get on the telephone to potential authors and
see if I can get them lined up.
John Hopcroft
∂23-May-86 0950 lantz@su-gregorio.arpa comp
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 23 May 86 09:50:50 PDT
Received: by su-gregorio.arpa with Sendmail; Fri, 23 May 86 09:49:52 pdt
Date: 23 May 1986 0949-PDT (Friday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: phdcom@sail
Cc: jlh@sonoma
Subject: comp
In response to my query as to how John felt about the weighting of
systems -- relative to overall computer science preparation, etc. --
he sent the attached. This inspired me to think (or, more correctly,
remember) that defining the basic preparation in terms of courses,
first, and then constructing a suitable syllabus (based presumably on
the content of those courses), is what we really should be doing.
AND the weight on the exam should indeed reflect the distribution of
"required" courses!
Somehow the "list of courses" aspect has been lost. So I propose to
reopen the discussion from scratch and come up with a list of courses
that everyone agrees to. Not having a catalog in front of me, I'll
just side with John's list for now.
Keith
------- Forwarded Message
Replied: 23 May 86 09:44
Return-Path: <@SU-SCORE.ARPA:jlh@su-sonoma.arpa>
Received: from SU-SCORE.ARPA by su-gregorio.arpa with Sendmail; Thu, 22 May 86 21:04:57 pdt
Received: from su-sonoma.arpa by SU-SCORE.ARPA with TCP; Thu 22 May 86 21:04:42-PDT
Received: by su-sonoma.arpa with TCP; Thu, 22 May 86 14:52:10 pdt
Date: 22 May 1986 1452-PDT (Thursday)
From: John Hennessy <jlh@su-sonoma.arpa>
To: lantz@score
Subject: courses and % of comp
Here's my list of what folks should have:
Systems = 8 courses noty including 108 material
112, 140, 143A, 212, 242, 240 A,B, 243
Applications = 4-5 courses
237A,(B,C?), 225, 248, 245
Theory = 4 courses including basic AI course
223?, 254, 261, 257A
So we have to cover roughly twice as many courses assuming this is the
right level of systems material....
------- End of Forwarded Message
∂23-May-86 1013 G.GINNIE@LOTS-B.STANFORD.EDU Party Corrections
Received: from LOTS-B by SU-AI.ARPA with TCP; 23 May 86 10:13:23 PDT
Date: Fri 23 May 86 10:04:11-PDT
From: Virginia Steuber <G.GINNIE@LOTS-B.STANFORD.EDU>
Subject: Party Corrections
To: "party animals": ;
Message-ID: <12209018959.14.G.GINNIE@LOTS-B.STANFORD.EDU>
Good news for all you good folks!
The LOTS picnic will actually begin at *4:00*
with drinks, music and a volleyball game.
Dinner will not be served until 5:00
and the food will be left out until 7:00, for all you late arrivers.
Master's Grove is located directly behind the sunken diamond. It is the
intramural field closest to El Camino. Just listen for Pierre's music...
Laurie will be leading a caravan to Master's Grove at 4:00 from 124 CERAS.
You are all invited to join her.
Cheers!
Ginnie
-------
∂23-May-86 1016 G.GINNIE@LOTS-B.STANFORD.EDU *PARTY*
Received: from LOTS-B by SU-AI.ARPA with TCP; 23 May 86 10:16:35 PDT
Date: Fri 23 May 86 10:16:34-PDT
From: Virginia Steuber <G.GINNIE@LOTS-B.STANFORD.EDU>
Subject: *PARTY*
To: "Party Animals": ;
Message-ID: <12209021212.14.G.GINNIE@LOTS-B.STANFORD.EDU>
O O O
O O O O O
O O O O O O
WHO: All you wonderful folks who help to BBBBBBBBOBBBBBBBBOBBBBBBB
make LOTS such a wonderful place! BB BB
BB---------------------BB
WHAT: Great food! Great music! Great BB---------------------BB
drinks! Great volley ball! Great BB BB
frisbee! Great people! Great BB BBBBBBB
foo... BB BB
BB BB
WHEN: Friday, May 30th (that's *next* BB BB
Friday!) at 5:00 P.M. BB BB
BB BB
WHERE: Master's Grove BB BBBBBB
BB BB
RSVP: Tuesday, May 27th (that's *next* BB BB
Tuesday!) at 3:00 SHARP! BB BB
To G.Ginnie@LOTSA. BBBBBBBBBBBBBBBBBBBBBBBBBB
-------
∂23-May-86 1515 SJM Saturday schedule
When shall we meet Saturday? Here, I assume.
Susie
∂23-May-86 1627 KARP@SUMEX-AIM.ARPA Comp readings library
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 23 May 86 16:26:51 PDT
Date: Fri 23 May 86 16:27:06-PDT
From: Peter Karp <KARP@SUMEX-AIM.ARPA>
Subject: Comp readings library
To: phdcom@SU-AI.ARPA, jlh@SU-SONOMA.ARPA
cc: grosof@SUMEX-AIM.ARPA
Message-ID: <12209088667.32.KARP@SUMEX-AIM.ARPA>
Ben Grosof has suggested that we assemble a library of the proposed
additions to the comp reading list so that students may evaluate books
with which they are unfamiliar. I have already gotten some valuable
comments from students who were already familiar with some of the
texts, so this seems like a reasonable idea. I propose that we:
1) Try and round up a copy of each book from members of the committee
or anyone else we may know who has them.
2) Put them in Victoria's office with a sign-out sheet nearby so people
can check them out to check them out.
Victoria, I hope this doesn't impose on you too much.
Please send me a note if you can lend one of the books.
Peter
-------
∂23-May-86 1624 VAL pointwise circ'n paper
I'd like to know your opinion about the new section on applications at the end.
Does it sound convincing? If yes, how important do you think it is to have it
in the short version that will appear in Proc. AAAI-86? (I have only 5 pages
there, so I'll have to decide between mathematics and applications).
∂23-May-86 1636 SJM saturday schedule
2pm it is. I might be loitering around here earlier than that, but I
will think about other things.
Susie
∂24-May-86 1437 JIML@SU-CSLI.ARPA re: Turning off call-waiting (from SAIL's BBOARD)
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 24 May 86 14:37:24 PDT
Date: Sat 24 May 86 14:36:35-PDT
From: Jim Lewinson <Jiml@SU-CSLI.ARPA>
Subject: re: Turning off call-waiting (from SAIL's BBOARD)
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Sun 18 May 86 15:01:00-PDT
*70 was recently added. I think I saw it in the stuff they send with
one's phone bill.
Jim
-------
∂24-May-86 1444 LES Mason trip request
To: JMC
CC: CLT
Does Ian Mason's request for support to attend the logic conference at
M.I.T. in June and/or the workshop at Harvard in July sound reasonable to
you?
∂24-May-86 1455 LES
To: IAM, JMC
CC: CLT
∂18-May-86 1703 IAM
To: LES, JMC, CLT
LICS
As you know I am giving a paper
on my recent work at the first symposium
of logic in computer science, to be held at M.I.T
in June. I am have also been invited to participate
in a program transformation workshop taking place at Harvard in July.
I was wondering if there is any travel money available
for me to help with this galavanting about?
Ian
LES - The answer is "Yes" for both conferences, but Carolyn would like to
get the workshop promoters to cover some or all of the cost of attending
that event, if possible.
∂24-May-86 1458 IAM
To: LES, JMC, CLT
transformation workshop
okey dokey, i will ask scherlis
Ian
∂25-May-86 0827 GZT.TDF%OZ.AI.MIT.EDU@AI.AI.MIT.EDU
Received: from AI.AI.MIT.EDU by SU-AI.ARPA with TCP; 25 May 86 08:27:15 PDT
Date: Sun, 25 May 1986 11:27 EDT
Message-ID: <GZT.TDF.12209525663.BABYL@MIT-OZ>
Sender: GZT.TDF%OZ.AI.MIT.EDU@XX.LCS.MIT.EDU
From: "David D. Story" <FTD%MIT-OZ @ MC.LCS.MIT.EDU>
To: mccarthy@SU-AI.ARPA
Phase-Of-The-Moon: FM+2D.8H.18M.59S.
Hi John,
I am in need of a name and would like to take you
back to AAAI '83 for an instant, where we met not just
in the upstairs lounge but at the Army Conference just
previous (April ?). There was a tall blondish gentlemen
that you were talking with at the conference outside of
the lower auditoriums. I was hoping you could give me his
name. He was approx. 6'2'', Blondish but balding hair, 220
pounds, aggressive in assuredance. We met just previous to
your meeting over coffee and I have lost his name. Can you
help ? Thanks,
Dave Story
∂25-May-86 0900 JMC
pills
∂25-May-86 0907 GZT.TDF%OZ.AI.MIT.EDU@AI.AI.MIT.EDU reply to message
Received: from AI.AI.MIT.EDU by SU-AI.ARPA with TCP; 25 May 86 09:07:40 PDT
Date: Sun, 25 May 1986 12:07 EDT
Message-ID: <GZT.TDF.12209532932.BABYL@MIT-OZ>
Sender: GZT.TDF%OZ.AI.MIT.EDU@XX.LCS.MIT.EDU
From: "David D. Story" <FTD%MIT-OZ @ MC.LCS.MIT.EDU>
To: John McCarthy <JMC@SU-AI.ARPA>
Cc: FTD%MIT-OZ@MC.LCS.MIT.EDU
Subject: reply to message
Phase-Of-The-Moon: FM+2D.9H.1M.14S.
In-reply-to: Msg of 25 May 1986 11:57-EDT from John McCarthy <JMC at SU-AI.ARPA>
No that was me. I was wondering about him. He was speaking
with you at the table at the top of the stairs of the lower
auditoriums.
X
-------------
|||| ||||
---| |--| |--
Except somewhat semi-circular if I remeber that fleabag
in Washington Correctly. See you in Philadelphia anyway.
Should be a nice time. Need restaurants ? Hotels ? Recreation
? Just send me mail ?
Dave Story
∂25-May-86 1832 vijay@ernie.Berkeley.EDU Re: any sign of life?
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 25 May 86 18:32:28 PDT
Received: by ernie.Berkeley.EDU (5.51/1.12)
id AA02268; Sun, 25 May 86 18:32:30 PDT
Date: Sun, 25 May 86 18:32:30 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605260132.AA02268@ernie.Berkeley.EDU>
To: JMC@su-ai.arpa
Subject: Re: any sign of life?
Hi,
I was just waiting for material from Pappert and Bobrow, but since
you'll be leaving, I'm going to send you the responses to you from
Winograd and Dreyfus (Searle has been in Venice... just got back
yesterday and is preparting a response to you and Charniak).
Thanks for letting me know,
Vijay
∂25-May-86 1838 vijay@ernie.Berkeley.EDU Correction...
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 25 May 86 18:37:58 PDT
Received: by ernie.Berkeley.EDU (5.51/1.12)
id AA02396; Sun, 25 May 86 18:37:59 PDT
Date: Sun, 25 May 86 18:37:59 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605260137.AA02396@ernie.Berkeley.EDU>
To: jmc@SU-AI.ARPA
Subject: Correction...
I meant Weizenbaum, not Winograd. -VR
∂25-May-86 1849 vijay@ernie.Berkeley.EDU Weizenbaum's stage 3 comments
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 25 May 86 18:49:28 PDT
Received: by ernie.Berkeley.EDU (5.51/1.12)
id AA02562; Sun, 25 May 86 18:49:20 PDT
Date: Sun, 25 May 86 18:49:20 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605260149.AA02562@ernie.Berkeley.EDU>
To: jmc@SU-AI.ARPA
Subject: Weizenbaum's stage 3 comments
Preface:
I am well aware that readers of these remarks - other's, as well as mine -
expect a discussion of 'scientific' and 'technical' issues, such as
speculations about the role of mathematics in AI or whether connectionism
will lead to machines exhibiting self awareness, and so on. As I
understood the invitation to participate in this forum, each of us was to
discuss whatever he thought was important and interesting about AI. I
think there are two matters related to AI that overshadow all others in
importance.
One has to do with what awaits us if we succumb to the temptation so
seductively dangled before us by, I would say, irresponsible or vastly
ill-educated AI practitioners, the temptation to believe that artificial
intelligence can be, or indeed is, equivalent to human intelligence in the
sense that it can understand and produce wisdom with respect to
interpersonal, social and cultural human affairs, as occasionaly human
intelligence can.
The other matter of overriding importance is that the work which is or is
paraded as being part of AI, whether correctly or not, has become an
indispensable component, even one of the foundations of the military
armamentarium which threatens the very survival of the human race. The
crucial role the military has assigned to AI in such programs as the
Strategic Defense Initiative (SDI) and the Strategic Computing Initiative
(SCI) casts a mantle of responsibility over the entire AI community. It is
now inescapably clear that the insane pace of the technological arms race
cannot be maintained without the willing cooperation of virtually the
entire AI community, academic and industrial. (Most of the rest of the
computer community is not exempt from this judgment. We happen here to be
concerned with AI.) In other words, the AI community has the power to
brake the mad cycle. It can rightly say "Not without us !" but
does not say it. We must all live with our consciences.
You may say we can't just discuss these two points forever. Correct. But
we can't evade them forever either.
End of preface.
McCarthy's two pointers, one to his long ago review of my book CP&HR and
one to something Berliner wrote in who knows what journal, don't help me
very much. True, there are things about which I haven't changed my mind.
Well then, JMC still refers to papers he wrote in 1958 and book reviews
over a decade old. There is some merit in consistency.
I reiterate :
1) I see no way to put a limit on the degree of intelligence artifacts may
eventually gain.
2) I am convinced that machine intelligence must always be alien to human
intelligence and that
3) the necessarily alien nature of machine intelligence has some
implications with respect to what we ought and ought not to do with machine
intelligence.
My disagreement is with those who insist that machines can be made to
understand human events in human terms. I have never understood, and
perhaps one of you can help me here, what it is that the true believers in
AI cannot accept in the above propositions.
To Hofstadter I want to say that I don't know what the phrase "the human
soul is infinite" means, hence whether it corresponds to anything I
believe. I do think that there are human experiences, lots of them, that
cannot be fully expressed by bit strings of any length. There are, in
other words, things humans know that they cannot say. Is it really the
position of mature AI researchers that that isn't true ? That it isn't
true of them as individuals ? That they can say everything they know? Can
Hofstadter express the whole of his experience of hearing a rhyme in any
way at all, in writing, for example ? If one of you has a dream so
beautiful that you want to hold on to it, perhaps to sleep again and
continue it, can you write it down or tell it to someone and know that you
have captured all of it ? Doesn't the very attempt to hold on to a dream
destroy it ? (Am I unique in this respect !?) Can even the events of a
single day be recorded completely - let alone the whole of a person's
history. That is why we have art - art is the attempt to transcend such
boundaries and, in the absolute sense, it is doomed to failure in that
attempt. Not so ?
All of this says that there are limits to what a non-human organism (if you
will) can be made to understand, and it says something about the nature of
these limits. If I were to argue anlogously with respect to Dolphins,
would anyone disagree ?
Natural language understanding is the arena in which issues of this kind
arise in very sharp form. Perhaps if and when computer vision has
progressed very far, scene understanding will become another such arena.
(To what extent will machines ever understand the look on that Frenchman's
face as he watched German troops march into Paris, as it was recorded in
that famous photograph ? Or the scene of two naked children, their backs
covered with burning napalm, running toward the camera in another famous
photograph ?) But, back to language. There seems to be general agreement
that understanding of an utterance depends on the understander sharing the
speaker's (or writer's) context.
JMC and others often make a distinction between "full understanding" and
"useful lesser levels of understanding." I would make the same
distinction. The reason we can make ourselves understood to, say, a Tokyo
(or, for that matter, New York) cabbie is that our very situation, e.g.,
having just hailed the cab and climbed in, together with the purpose of our
attempt to communicate, so drastically restricts the contextual framework
that communication amounts to little more than engaging in a very short
multiple choice test. Yes, such lesser levels of understanding are useful
and, I have no doubt, we can get computers to understand the kind of things
one says to waiters, cabbies, and so on, in the ordianry run of doing
business with them. But it doesn't follow from that there exists a
'default context', i.e., a context in which the sentence to be understood
has exactly one correct interpretation, a "literal meaning". History
enters into context formation. There are no 'standard conditions' that
apply to human affairs. What is given up when one accepts that ?
Dan Dennet coined the very telling term 'Cognitive Wheel'. He points out
that wheels do not occur in nature, that human beings invented wheels and
that this 'unnatural' invention has permitted human beings to do better
than nature in several respects. He suggests that artificial intelligence
may well come up with ways to "think" (why not ?) very different from the
way brains think and, in many respects, in ways superior to human thought.
I can't prove otherwise. Whether or not the attempt should be made is a
question of research priorities. But, whatever the decisions, it seems to
me important that questions of what can and what cannot be done be
confronted in earnest and soberly. Perhaps that is in part what we're
doing here.
A few quotes:
DH: "Scientific discoveries have fantastic potential in either
direction, bad or good, and so I would not curtail science in any
way."
JMC: "The scientist isn't the boss of society and can neither
force society to use the results of science nor prevent it from
doing so. This divison of labor between science and politics is
socially apporpriate."
DH: "...AI as a scientific activity is quite benign, and
therefore [I] feel that this forum ought to concentrate on AI as
science rather than on AI as social activity."
TW: "It is often claimed that concerns of 'social impact' should
be left to the political process, or perhaps to engineers who are
direct;y developing products, but should be ignored in pursuing
'pure science.' These (often self-serving) claims are based on a
rationalistic (and narrow) understanding of science as a human
enterprise. They might be true for some idealized scientist
living selfsufficiently and incommunicado on an isolated island,
but are irrelevant to the real world. The contiuing enterprise
of any science depends on a public consensus that support the
allocation of resources to it. This concensus is maintained by a
process of publication and 'eductaion' in which the ideology of
the science is promulgated and justified."
I agree with the last quotation, of course. But I would add that science
is a social enterprise in another sense as well. What is and what is not
to be counted as scientific, as fact is decided by a consensus of members
of the relevant section of the scientific community. Who is and who is not
a member of a particular section is similarly a social decision. It makes
a difference when someone who was once out is, as the communists say,
rehabilitated. (And vice-versa. A third of the members of this panel have
reasons to know about that.) Whether or not it is a legitimate AI task to
attempt to create a computer system that will take over the office of the
President of the United States under certain conditions, or to work on an
AI based expert system to do, say, astrological fortune telling, these are
questions answered by socially arrived at consensus.
The first three quotations above present the doing of science and science
itself as value free. They assert that the products of science can after
all be used for good or evil. Which way they are used is somebody else's
department. That's the socially appropriate division of labor. Perhaps
so. Perhaps on that isolated island Winograd mentions, scientists just
couldn't predict to what end use, good or evil, their work would be
put.vil. But we live in a concrete society which has a long and vivid
record of co-opting virtually every scientific and technical achievement of
the human genius to the purpose of creating and deploying ever more
efficient killing machines. (Forgive me for not using the usual
euphemisms.)
Some of us even celebrate that fact as, for example, when they tell us
that, if society generously supports AI and an American 5th Generation
effort to match Japan's, the so called 'smart weapons' of today will seems
like simple 'wind-up toys' compared to what we will then be able to
produce.
We know very well to what end uses, say, advances in computer vision, or
speech understanding, or robotics, or ... and so on and on, will be put.
In the concrete social and political reality in which we live, scientists
and engineers who work to improve the computers' ability to see, for
example, tacitly give their assent to the use of their work to help guide
nuclear tipped cruise type missiles on their way to mass murder. Except
for acts of self deception of heroic porportions, we know. We cannot then
deny our responsibility by arguing that our work is neutral, our science
value free, that somebody else decided, and so on.
I disagree with those who believe it necessary to maintain present levels
of weaponry, support bases overseas, and so on. But I don't think holding
that belief necessarily involves an abdication of responsibility. I do
think, on the other hand, that scientists and technologists who take the
position with respect to their work that
1) they don't want to contribute to weapon technology, and
2) that, while they know that their work can be used in ways they don't
like, it could equally well be used benignly,
are gravely mistaken. Science and politics simply can't be disentangled.
Responsibilities can't be shed by simply closing one's eyes. It just isn't
possible to discuss AI and what's to become of it without looking at and
deciding about the context in which it is embedded - not, that is, without
a constant effort of self censorship.
Finally, I didn't suggest that current chess playing programs don't embody
psychological principles. What I asked was what new psychological
principles have been discovered and deployed THAT ACCOUNT FOR THE STRENGTH
of today's machine chess play. I mean, of course, that account for
improvement beyond that brought about by the increase in raw computing
power deployed these days. Perhaps Berliner's paper answers this question.
I'd like to be able to look.
Joseph Weizenbaum
-------
∂25-May-86 1851 vijay@ernie.Berkeley.EDU Dreyfus's stage 3 comments
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 25 May 86 18:50:56 PDT
Received: by ernie.Berkeley.EDU (5.51/1.12)
id AA02577; Sun, 25 May 86 18:50:56 PDT
Date: Sun, 25 May 86 18:50:56 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605260150.AA02577@ernie.Berkeley.EDU>
To: jmc@SU-AI.ARPA
Subject: Dreyfus's stage 3 comments
Dreyfuses to John McCarthy
You raised interesting questions concerning our views
on how the brain detects similarity and how typical situa-
tions are constructed.
Until recently, we had no clear idea about how the
brain recognizes the similarity of a current situation to
some stored memory, but the phenomenology led us to believe
that it was somehow done in terms of stored records of brain
activity patterns rather than in terms of abstract symbolic
descriptions. Recent connectionist work has provided us
with a reasonable explanation at the brain level quite dif-
ferent from what we had in mind, although this explanation
too is no doubt wrong in its details. Rather than store
brain-records of memories (of which one would presumably
need a long list), each experience (sensory input and an
associated output that experience shows to be appropriate)
leaves behind a "trace" by affecting the strengths of con-
nections between neurons. These strength modifications occur
in such a way that the input will tend to produce the
appropriate output when the neuron net "does its thing."
Then, later, the sensory input to the brain in some then-
current situation produces an output based on these connec-
- 2 -
tions. When the input is similar to one or more previously
seen inputs, the output will be similar to the appropriate
output due to the memory trace (i.e., appropriate connection
strengths).
Hence, memories themselves are not stored, nor are typ-
ical ones created. Many similar experiences merely produce
certain strong connections that then reproduce the appropri-
ate learned output when presented with a new, but similar,
input.
This account was motivated by Chapter 17 of Rumelhart's
and McClelland's book Parallel Distributed Processes, MIT
Press, May 1986.
While the input and output may, in some sense, be
interpretable as meaningful facts and beliefs about the
problem domain, if the neural net has hidden nodes and feed-
back among them, the process of getting from input to output
by "settling" of a neuron net will result in changing pat-
terns of activation of hidden nodes that have no interpreta-
tion in terms of facts and hypotheses about the problem
domain, and the strengths of connections among hidden nodes
have no interpretation as correlations or inferences con-
cerning any domain-meaningful facts. This is the sense in
which we believe that a process at the brain level need have
no correlate at the level of symbol manipulation where the
symbols have meanings concerning the domain and these mean-
ings are taken account of in the manipulation process.
- 3 -
While there may be meaningful symbol-manipulation
interpretations of certain inferential activities, mainly
those of non-experts, we believe that most expertise is of
the associative, pattern completion, type that can be
explained only at the brain (neuron net) level.
Dreyfuses to Hofstadter
We do indeed believe that expertise is largely the pro-
duct of the people, not leaders, to use your governmental
metaphor. See our note to John McCarthy. The fact that
this belief is clearly untenable concerning governmental
structure implies nothing about its correctness for cogni-
tion.
∂25-May-86 1859 vijay@ernie.Berkeley.EDU Papert's name...
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 25 May 86 18:59:22 PDT
Received: by ernie.Berkeley.EDU (5.51/1.12)
id AA02667; Sun, 25 May 86 18:59:24 PDT
Date: Sun, 25 May 86 18:59:24 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605260159.AA02667@ernie.Berkeley.EDU>
To: jmc@SU-AI.ARPA
Subject: Papert's name...
Thanks for the spelling corr. on his name... this has been a constant
source of confusion for me ever since I called his "assistant" to tell
me how to spell his name once and for all... she simply said there
were " two p's " - we both took it to be a satisfactory answer, and hung
up. This was, of course, ambiguous...
-Vijay
∂26-May-86 0900 JMC
Atenolol
∂26-May-86 1124 NILSSON@SU-SCORE.ARPA Tom Mitchell
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 26 May 86 11:24:16 PDT
Date: Mon 26 May 86 11:22:14-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Tom Mitchell
To: feigenbaum@SUMEX-AIM.ARPA, buchanan@SUMEX-AIM.ARPA,
genesereth@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA, shah@SU-SIERRA.ARPA,
grosof@SU-SUSHI.ARPA, hirsh@SU-SUSHI.ARPA
cc: nilsson@SU-SCORE.ARPA, gibbons@SU-SIERRA.ARPA
Message-ID: <12209819599.13.NILSSON@SU-SCORE.ARPA>
Both Jim Meindl and I have talked with Tom Mitchell by phone over this
weekend. Jim is encouraged that Tom might be appropriate for a CIS
billet and suggests that Tom visit Stanford again to talk to John Hennessy
and others at CSL/CIS. Tom is planning to be at Stanford on Wednesday,
May 28. Hennessy is making time available to see him and will arrange
for him to see others who he has not yet seen at Stanford. Whether or
not Tom ends up at Stanford now depends heavily on how Tom and John Hennessy
see their areas of interest overlapping on matters of AI applications to
VLSI. I think if John decides that Tom would be an appropriate appointment,
then Jim Meindl will agree also.
Perhaps it would also be possible for Tom to meet Jim Gibbons while he
is here on Wednesday. Anne Richardson will check with Mary Cloutier
(won't you, Anne) to see if that can be done.
Stay tuned. -Nils
-------
∂26-May-86 1732 vijay@ernie.Berkeley.EDU Replies....
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 26 May 86 17:32:07 PDT
Received: by ernie.Berkeley.EDU (5.51/1.12)
id AA05301; Mon, 26 May 86 17:32:01 PDT
Date: Mon, 26 May 86 17:32:01 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605270032.AA05301@ernie.Berkeley.EDU>
To: jmc@SU-AI.ARPA
Subject: Replies....
to Dreyfus and Weizenbaum received. -VR
∂26-May-86 1800 JMC
Pagels for references
∂27-May-86 0700 JMC
Time photographer
∂27-May-86 0922 JMC
tux,cap and gown, phone Sarah, phone Erik
∂27-May-86 1006 CONTRERAS@SU-SCORE.ARPA message
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 27 May 86 10:06:01 PDT
Date: Tue 27 May 86 10:04:39-PDT
From: Tina Contreras <CONTRERAS@SU-SCORE.ARPA>
Subject: message
To: JMC@SU-AI.ARPA
Message-ID: <12210067621.31.CONTRERAS@SU-SCORE.ARPA>
Chris Goad phoned, please call. 967-5878.
Tina
-------
∂27-May-86 1030 JMC
Gozani
∂27-May-86 1123 RA References
Fikes R. and Nils Nilsson, STRIPS: A new approach to the application of
theorem proving to problem solving. Artificial Intelligence 2 3,4 (1971),
pp. 189-208.
Friedberg, R.M., B. Dunham and J.H. North. A learning machine, Part II, IBM
Journal of Research, July 1959 pp. 282-287.
Friedberg, R.M.. A learning machine, IBM Journal of Research, January 1958,
pp. 2-13.
∂27-May-86 1205 GRP Reminder: next meeting June 10 at 2pm
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
∂27-May-86 1209 JJW Alliant now on Ethernet
To: JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, RPG@SU-AI.ARPA, LES@SU-AI.ARPA,
kolk@SU-NAVAJO.ARPA, GRP@SU-AI.ARPA
CC: Bosack@SU-SCORE.ARPA, Tom@SU-SCORE.ARPA,
THarvey@SU-SCORE.ARPA, kent@SU-NAVAJO.ARPA, ME@SU-AI.ARPA,
mogul@SU-NAVAJO.ARPA
Gang-of-Four is now working on the Ethernet. Their software people
fixed a bug this morning that was causing it to put incorrect
addresses in output packets.
At present, however, you should NOT try to Telnet to or from a TOPS-20
host. There is a bug in their Telnet user and server programs that
causes it to dump core when TOPS-20 sends it a flush output command
(which it does when you log out). It is especially bad if you're
logged in on the console since it seems to leave that terminal in a
weird state. They are sending us a floppy with a new version that
will fix this.
Here's my understanding of what was wrong: in some of their previous
installations, they were on networks that didn't use ARP, and they
indicated this by using host address greater than 0x1000 (after
removing the network part of the address). They've never run on a
class A network before, especially one like ours where the subnet
field causes all host addresses to be greater than 0x1000, so they
thought we were using the non-ARP address mapping. By patching out
this code, the problem was fixed.
∂27-May-86 1349 RA Jack Harper
Harper from Daedalux Machines Inc. (303) 988 3383 would like you to call him
re a letter he had sent you. If he doesn't hear from you, he'll call back this
afternoon.
∂27-May-86 1444 RA [Reply to message recvd: 27 May 86 14:41 Pacific Time]
Do you know whether it is in the Math Library or Green Library?
∂27-May-86 1609 RA New seats
Franklin called that you have the seats you wanted on the SAS flight, going
there. The new seats are 2H and 2J.
∂27-May-86 1619 VAL Common Sense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
The last meeting this quarter:
APPLICATIONS OF NON-MONOTONIC REASONING TO INDUCTIVE LEARNING
Benjamin Grosof
Stanford University, Computer Science Department
Thursday, May 29, 4pm
MJH 252
"Bias" is a term used in the machine learning literature to denote the
basis for selection among competing, consistent, explanatory,
inductive hypotheses. I will begin by discussing at a very abstract
level how non-monotonic reasoning can be applied to represent bias in
inductive learning.
I will then show how to formulate non-monotonically some types of
bias in the Version Space [cf. Tom Mitchell] approach to inductive
concept learning. One kind is maximal specificity/ maximal
generality. A second kind is the "linguistic" bias embodied by the
initial version space. A third is the completeness of the description
of classified instances. We show how non-monotonic reasoning can
provide a relatively declarative and semantic account of how to "shift
bias" in inductive concept learning, as opposed to the relatively
procedural and syntactic accounts heretofore explored [e.g. Utgoff's
thesis]. This enables a concise and principled specification of how
to relax overly strong biases, a open problem in machine learning.
∂27-May-86 1648 LES
Stanford University proposal to IBM
for
Editor-based Operating System
9 months beginning 1 July 1986.
Personnel
Prof. John McCarthy, Principal Investigator n/c
Tepper, Stephan, Research Programmer (100%) 28,500
-----, Student Research Assistant 8,348
(50% acad. yr., 100% summer)
-----, Secretary (5% time) 783
-------
Salary subtotals 37,631
Allowance for salary increases 2,341
(8% beginning 9/1/86)
-------
Salary totals 39,972
Staff benefits (25.4% till 9/1/86, 10,215
25.6% thereafter)
Travel (2 East Coast trips. @ $1000) 2,000
Computer time costs 3,000
Other direct costs 3,000
-------
Subtotal 58,187
Indirect costs (69% of direct costs 41,959
initially, 73% beginning 9/1/86)
=======
Total $ 100,146
∂27-May-86 1713 GRP editor ramblings
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
See RAMBL1.TXT[1,GRP]
∂28-May-86 0102 LES Facilities Committee Meeting Today
To: facil@SU-AI.ARPA, Nilsson@SU-SCORE.ARPA,
BScott@SU-SCORE.ARPA, Reges@SU-SCORE.ARPA
At noon in MJH 220. Topics are review of a draft policy on "ownership"
of computer equipment (see below) and development of aids to financial
control and planning within the department.
Formally, equipment purchased in the name of of the University belongs to
the Board of Trustees. Equipment purchased under research contracts is
assigned to the Board and is often donated to the University later.
In both cases, administrative responsibility for the equipment is assigned
by the University to the Department.
The Department has relied on common-sense assumptions of responsibility
for administering equipment assigned to it. The purpose of this policy is
to formalize these assumptions so as to avoid potential misunderstandings.
Les Earnest
--------------------------------------------------------------------------
Department of Computer Science
Draft Policy on
Administration of Computer Equipment
PURPOSE. This policy assigns responsibilities for determining how
computer equipment will be used and accounted for within the Department of
Computer Science. Its purpose is to formalize these responsiblities in a
way that is both consistent with legal responsibilities of the University
and common sense practices within the Department.
SCOPE. This policy specifies normal assignment of responsibilities within
the Department. It is subordinate to general University policies as well
as the legal and contractual obligations of the University.
CAPITAL EQUIPMENT REGISTER. A central register of all capital equipment
assigned to the department will be maintained by the Computer Facilities
Group. This register will be audited annually by the Department to confirm
accuracy and completeness.
CONTROL. Administrative control of equipment purchased with general
departmental funds rests with the Department Chairman. Control of
research equipment acquired under grants and contracts shall rest with
the Principal Investigator who obtained the funds under which it was
acquired. If the Principal Investigator leaves, the Department Chairman
assumes control of the equipment.
OFF-CAMPUS USE. Computer terminals, modems and other such equipment can
be assigned to someone for use off-campus only by the controlling
authority. The recipient must sign an agreement accepting responsibility
for the equipment, promising to report any damage or theft promptly, and
promising to return it on demand of the controlling authority or upon the
recipient's separation from the project for which it was assigned. The
controlling authority is responsible for seeing that complete and timely
records are kept of such equipment loans, returns and transfers, including
verification at least annually that the equipment is still in the hands of
the person who accepted it.
∂28-May-86 0836 ROSENBLOOM@SUMEX-AIM.ARPA Re: typo
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 28 May 86 08:36:16 PDT
Date: Wed 28 May 86 08:36:28-PDT
From: Paul Rosenbloom <ROSENBLOOM@SUMEX-AIM.ARPA>
Subject: Re: typo
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Tue 27 May 86 21:46:00-PDT
Message-ID: <12210313711.30.ROSENBLOOM@SUMEX-AIM.ARPA>
John: Thanks for letting me know. I'm not sure how that dropped out.
-- Paul
-------
∂28-May-86 0900 JMC
Gozani
∂28-May-86 1011 RA Academic Costume
The academic costume we ordered did not arrive. I was told that "it is somewhere
in the UPS truck. The bookstore can get you a Stanford gown and cap
and they can deliver it here. Would you like a Stanford one instead? They
are waiting for my answer at the bookstore.
∂28-May-86 1033 RA Re: Academic Costume
[Reply to message recvd: 28 May 86 10:13 Pacific Time]
The bookstore will deliver the costume to your house shortly. Please tell
me when it gets there.
Thanks,
∂28-May-86 1103 tgd%oregon-state.csnet@CSNET-RELAY.ARPA Proposal for "Knowledge Compilation" Workshop
Received: from CSNET-RELAY.ARPA by SU-AI.ARPA with TCP; 28 May 86 11:03:42 PDT
Received: from oregon-state by csnet-relay.csnet id ab18425; 28 May 86 13:56 EDT
Received: by orstcs.UUCP (4.12/6.3.ORST)
id AA15726; Mon, 26 May 86 12:17:13 pdt
Date: Mon, 26 May 86 12:17:13 pdt
From: Tom Dietterich <tgd%oregon-state.csnet@CSNET-RELAY.ARPA>
Posted-Date: Mon, 26 May 86 12:17:13 pdt
To: jmc%su-ai@CSNET-RELAY.ARPA
Subject: Proposal for "Knowledge Compilation" Workshop
Cc: aaai-office%sumex-aim.stanford.edu@CSNET-RELAY.ARPA,
tgd%oregon-state.csnet@CSNET-RELAY.ARPA
Prof. McCarthy,
Thank-you for agreeing to provide $5,000 in support of our workshop
on knowledge compilation. Here is a copy of the proposal that I sent
you along with the cover letter that I sent. We will take care to
treat all of the expert systems companies equally.
Regards,
Tom Dietterich
=========
Cover letter:
\documentstyle[11pt]{osuletter}
%\setlength{\topmargin}{1in}
%\setlength{\oddsidemargin}{1.25in}
%\setlength{\evensidemargin}{1.25in}
%\setlength{\headheight}{0in}
%\setlength{\headsep}{0in}
%\setlength{\textheight}{9in}
%\setlength{\textwidth}{6in}
\name{Dr. Thomas G. Dietterich}
\telephone{(503) 754-4466}
\signature{Tom Dietterich}
\begin{document}
\begin{letter}{Prof. John McCarthy \\
Department of Computer Science \\
Margaret Jacks Hall \\
Stanford University \\
Stanford, California 94305}
\opening{Dear Prof. McCarthy:}
Please find enclosed a proposal for AAAI to sponsor a scientific workshop on
the topic of ``Knowledge Compilation.'' Jim Bennett, Jack Mostow, and
myself have felt for some time that it would be valuable to hold a workshop
on this topic---a workshop in which we focused on the computational issues
involved in compiling explicit knowledge into implicit forms. Although
there are many people working on various aspects of this problem, they are
spread across a number of sub-areas of artificial intelligence and computer
science. The purpose of the workshop is to bring these people together and
focus attention on this topic.
To prevent the workshop from growing into a large conference, we
have decided that attendence will be by invitation only. We have developed
a procedure for selecting the invitees that we hope will guarantee broad
participation. To make the results of the workshop accessible to the larger
AI community, we will be publishing the proceedings through Morgan-Kaufmann
using the procedure that they developed for Joe Halpern's conference.
I've tried to present a complete budget for the workshop. Please contact me
if you need any additional information. I can be reached on CSNET as
\verb"tgd%oregon-state@csnet-relay".
\closing{Sincerely,}
\end{letter}
\end{document}
========
Proposal:
\documentstyle[11pt]{article}
% titlepage causes separate title page
%\input{eqnhack} % make numbered equations work
\setlength{\topmargin}{1in}
\setlength{\oddsidemargin}{1in}
\setlength{\evensidemargin}{1in}
\setlength{\headheight}{0in}
\setlength{\headsep}{0in}
\setlength{\textheight}{9in}
\setlength{\textwidth}{6.5in}
%\renewcommand{\baselinestretch}{1.1} % interline spacing
%\setlength{\parindent}{0in}
%\parskip=10pt plus2pt minus2pt
\setlength{\unitlength}{0.1in}
%\pagestyle{empty} % no page numbering
\newcommand{\aq}{$A↑q$} %A superscript q
\newtheorem{axiom}{Axiom} % create axiom environment.
\newtheorem{definition}{Definition} % create definition env.
\newenvironment{ttexample}{\envbegin{quote}\renewcommand{\baselinestretch}{0.9}\small\tt\flushleft\obeycr}{\restorecr\renewcommand{\baselinestretch}{1.0}\envend{quote}}
\newenvironment{bibparagraph}{\envbegin{list}{}{ %
\setlength{\labelsep}{-\leftmargin} %
\setlength{\labelwidth}{0pt} %
\setlength{\itemindent}{-\leftmargin} %
\setlength{\listparindent}{-\leftmargin}}}{\envend{list}}
\begin{document}
\title{Proposal for a Workshop on Knowledge Compilation}
\author{Thomas G. Dietterich\\
Oregon State University \\
\\
James S. Bennett \\
Teknowledge, Inc. \\
\\
Jack Mostow \\
Rutgers University \\[.3in]
Inn at Otter Crest \\
September 24-26, 1986}
\date{}
\maketitle
\section{Purpose}
Most work in artificial intelligence is based on the explicit representation
of knowledge coupled with the run-time interpretation of this explicit
knowledge to yield appropriate behavior. This methodology possesses many
advantages (e.g., logical precision and modifiability), but it also
confronts a major difficulty: there is a direct tradeoff between the
explicitness of the representation and the efficiency of the computation.
To counter this, a number of researchers have been investigating methods for
automatically ``compiling'' explicit, but inefficient, knowledge
representations into implicit, but more efficient, forms.
The purpose of this workshop is to bring together researchers who have been
investigating issues and methods of knowledge compilation. It is our hope
that the workshop will lead to (a) increased cohesion and communication
among researchers in this area, (b) increased research attention to this
area, (c) clarification of methodological approaches to developing knowledge
compilation systems, and (d) exploration of the implications of knowledge
compilation for computer science, artificial intelligence, and cognitive
science.
\section{Who will attend?}
At the present time, few researchers would describe their work as ``knowledge
compilation.'' Indeed, work in this area has proceeded under a variety of
headings, for example:
\begin{enumerate}
\item Automatic programming and transformational development. Work in this
area has shown how to convert high-level specifications into efficient
implementations. This is a kind of knowledge compilation in which two kinds
of knowledge are being applied to produce an efficient problem solver.
First, there is the knowledge about the desired behavior of the system.
This knowledge is explicit in the high level program specification. Second,
there is the knowledge about the available program transformations and the
situations in which they should be applied. This knowledge is usually
employed implicitly by programmers when they guide the program
transformation process. Fickas has shown how to make some of this knowledge
explicit.
\item Algorithm design and improvement. Work by people such as Tappel and
Mostow has provided one paradigm for designing algorithms by converting
inefficient, but executable algorithms into more efficient ones. The key
operation of knowledge compilation in these systems involves incorporating
``solution tests'' into ``generators of candidate solutions''. The knowledge in
the ``test'' is made implicit in the ``generator''. Other approaches to
algorithm design include Doug Smith's schema-based approach and Elaine
Kant's psychological investigations into the methods employed by algorithm
designers.
\item Problem reformulation. In this area, Amarel's work has focused on
developing specialized representations that incorporate domain knowledge in
order to make the solution process more efficient. Hence, problem
reformulation is another form of knowledge compilation.
\item Knowledge engineering of expert systems. Experience in expert system
development has shown that knowledge engineers frequently ``compile''
knowledge into the architecture and knowledge bases of expert systems. In
order to understand this process, some researchers have attempted to analyze
existing expert systems and ``de-compile'' them. This process involves
identifying the knowledge that was known to the system designers but which
is implicit in the finished expert system. Clancey and Swartout provide
good examples of this approach. Barstow and other workers at Schlumberger
have also been making progress in this area.
\item Machine learning. Recent work in machine learning has pursued the idea
of analytical or explanation-based learning. In this paradigm, learning
involves converting knowledge from an explicit, but inefficient, form into a
more efficient form. This form of knowledge compilation has been applied to
several problems including the development of a learning apprentice for VLSI
design. Related work in learning has investigated chunking and rule
composition methods of improving the performance of problem solving
programs.
\item Meta-level control of problem solvers. A few researchers (e.g., D. E.
Smith) have developed methods for exploiting knowledge about particular
problems to construct specialized control strategies for problem solvers.
With these techniques, knowledge is being ``compiled'' into the control
strategies of the problem solver.
\end{enumerate}
Leading researchers in all of these areas will be invited to the workshop.
An important goal of this workshop is to give a name and coherence to this
area of research.
\section{Format}
The first half of the workshop will be devoted to presentation of
papers covering substantive work in the various areas described above. The
purpose of this will be to describe the work that is already being done and
relate it to the overall problem of knowledge compilation. The second half
of the workshop will focus on position papers and panel discussions
concerning topics such as (a) methodologies for doing research on knowledge
compilation, (b) important directions for future funding and research, (c)
general overviews of knowledge compilation, (d) implications of knowledge
compilation for expert systems, (e) implications of knowledge compilation
for machine learning, etc.
Attendance at the workshop will be by invitation only. The list of invitees
will be developed in two stages. First, the three organizers will develop a
list of 20 people. Then, each of these people will be asked to nominate one
additional person to attend. Nominations for graduate students will be
particularly encouraged. Workshops like this tend to grow in size. We
recognize this fact, but nevertheless, we will enforce a strict limit of 50 on
the total size of the workshop.
Papers will be solicited from each of the invitees and published in a
proceedings through Morgan-Kaufmann (under the same kind of arrangement that
Joe Halpern used for the conference on reasoning about knowledge).
Panel discussion topics (and coordinators) will also be solicited. To make
the panel discussions more worthwhile, each panel coordinator will write a
short paper describing a list of questions and issues to be discussed, along
with background material explaining why these issues are important for
research in or application of knowledge compilation.
It is expected that all non-student participants in the workshop will
either present a paper or coordinate a panel discussion.
\section{Student Travel Grants}
A limited number of fellowships will be made available to cover travel
expenses for graduate students attending the workshop.
\section{Budget}
The following budget shows per person costs for the workshop.
\begin{verbatim}
Rooms (double-occupancy) $29.00
Meals 32.00
Meeting rooms, A/V equip. 0.00
Transportation (buses to and
from Portland Airport) 30.00
Proceedings 19.00
Secretarial (1 month) 30.00
-------------------------------------
Total per person $140.00
Total for 50 people $7,000.00
Student travel awards
(5 students x $500.00) 2,500.00
-------------------------------------
Total budget request: $9,500.00
\end{verbatim}
\end{document}
∂28-May-86 1204 SINGER@su-sushi.arpa maternity leave
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 28 May 86 12:03:58 PDT
Date: Wed 28 May 86 12:02:30-PDT
From: Stephanie F. Singer <SINGER@su-sushi.arpa>
Subject: maternity leave
To: jmc@su-ai.arpa
Message-ID: <12210351218.44.SINGER@su-sushi.arpa>
Thanks for the insurance suggestion, which I never considered. Is there any
such insurance available that you know of? Are there businesses which take
advantage of it?
Stephanie Singer
-------
∂28-May-86 1413 LANSKY@SRI-WARBUCKS.ARPA PLEASE RESPOND IMMEDIATELY
Received: from SRI-WARBUCKS.ARPA by SU-AI.ARPA with TCP; 28 May 86 14:12:44 PDT
Date: Wed 28 May 86 14:10:55-PDT
From: Amy Lansky <LANSKY@SRI-WARBUCKS.ARPA>
Subject: PLEASE RESPOND IMMEDIATELY
To: genesereth@SUMEX-AIM.ARPA, val@SU-AI.ARPA, jmc@SU-AI.ARPA,
rosenschein@SUMEX-AIM.ARPA, shoham@YALE.ARPA,
princeton!mind!ghh@SUN.COM, agre@AI.AI.MIT.EDU,
bonnie%upenn@CSNET-RELAY.ARPA, schabes%upenn@CSNET-RELAY.ARPA
Message-ID: <VAX-MM(194)+TOPSLIB(120)+PONY(0) 28-May-86 14:10:55.SRI-WARBUCKS.ARPA>
TIMBERLINERS:
All of you are the remaining few who have not sent in their yellow
form with your explicit plans for the workshop. YOU MUST SEND THESE
IN IMMEDIATELY! or else your rooms cannot be guaranteed nor can
we guarantee any stipend. Please indicate whatever your stipend
needs are and we will do our best to accommodate them.
Please send in the form, and IN ADDITION, respond to this message.
(Yves and Phil, you were on the waiting list and can now come if you like.)
Attached is the information form I sent out a while back, with the
info sheet for you to send in at the end.
PLEASE DO SO ASAP!!!!!
-Amy Lansky
------------------------------------------------------------------
Dear Workshop Attendees:
Now that the summer is rapidly approaching, we would like to
finalize our plans for the workshop. This letter is to get more
definite information from all of you about your stay at Timberline.
Please do not respond with indefinite answers -- we need to make
concrete room and food arrangements at this time.
Enclosed you will find a tentative workshop schedule as well as
more information about housing, payment, and stipends.
You will also find a yellow sheet with questions. PLEASE FILL THIS
OUT AS ACCURATELY AS POSSIBLE and return to the address given.
By the beginning of June we will be sending you a complete schedule
of talks as well as transportation details. Please feel free
to call us if you have further questions.
Amy Lansky Mike Georgeff
(415) 859-4376 (415) 859-4769
LANSKY@SRI-AI GEORGEFF@SRI-AI
Timberline Toll Free Number: (800) 547-1406
!
Arrival
-------
Attendees should fly into the Portland, Oregon airport. Timberline is
approximately a one hour drive from the airport. You may elect to
rent a car or arrange limousine service to and from Timberline. You
must make these arrangements on your own. The limousine service is
AES LIMOUSINE (telephone: (503)255-0325) and is $22.50 per person,
each way (there must be at least two passengers in the limo).
Directions to Timberline will be sent to you closer to the time of the
workshop.
Rooms
-----
We have reserved in advance 30 rooms for the nights of June 29,30 and
July 1. For those of you who want to stay on after the workshop, 10
rooms have been reserved for the nights of July 2,3,4,5. The cost of
the rooms is $61.60/per person (double occupancy), which includes 3
meals per day (starting on the night of the 29th). For those of you
bringing children (i.e. more than 2 people per room), each child is
$8.50 per day, which does NOT include meals. You will therefore have
to pay extra for meals for your children. However, children under 6
eat free.
The rooms are of a variety of types. Some have only one double or one
queen bed, some have two single beds, and the rest have a queen and a
single OR a double and a single. On the yellow form please note the
size and type of your party and exactly which nights you plan to
spend. If you are coming as a couple, a double or queen bed may be
assigned. Families with children will have priority in the rooms with
two beds. For those bringing 2 children, an extra cot can be brought
into your room. For those coming singly, double occupancy will be the
rule (obviously, only in those rooms with 2 beds!). If you have a
roommate preference, please indicate on the yellow form.
The 10 rooms for July 2-5 will be assigned first-come-first-serve.
If we need more rooms, we will try to make arrangements for those nights.
For those of you wanting to stay extra nights (besides June 29--July 5),
please make separate arrangements DIRECTLY WITH TIMBERLINE. Their toll
free number is (800) 547-1406. (One room has already been reserved
for Tate and Drummond for June 27-28). When you call, you should
indicate that you are part of the SRI workshop. (They may have you
talk to the sales office rather than the regular reservations desk).
Food
----
On the evening of your arrival (June 29), dinner will be served in
the public dining room (the Cascade Room). It will be open seating
and a limited menu. After that, breakfast, lunch, and dinner (up through
lunch on July 2) will be served in a private dining room. Only a fixed
menu will be available. If you have special food restrictions,
please advise on the yellow form.
Payment
-------
Each of you will be responsible for paying your full bill upon checking
out of Timberline. Those of you receiving stipends will receive a check
separately, which you can apply to your costs as you desire.
Stipends
--------
We have only a limited source of support for this workshop.
Therefore, we will be awarding stipends only for those who need them.
The approximate award will be $900 for those coming from England or
Australia, $400 for East Coasters or other non-West Coast origins, and
$200 for West Coasters. Our aim is to at least cover airfare costs.
Obviously only a limited number of these awards can be made, so please
indicate on the form the severity of your need!
!
TENTATIVE WORKSHOP SCHEDULE
---------------------------
Sunday June 29:
People arrive in the afternoon and evening
Dinner in the Cascade Dining Room -- public dining
Monday June 30:
Breakfast: 8:00-9:00
Session I: 9:00-10:30 (2 papers)
Coffee Break: 10:30-10:45
Session II: 10:45-12:15 (2 papers)
Lunch: 12:15-1:30
Session III: 1:30-3:45 (3 papers)
Discussion 4:00-5:00
Bar: 6:30-7:30
Dinner: 7:30-9
Tuesday July 1:
Breakfast: 8:00-9:00
Session IV: 9:00-10:30 (2 papers)
Coffee Break: 10:30-10:45
Session V: 10:45-12:15 (2 papers)
Lunch: 12:15-1:30
Session VI: 1:30-3:00 (2 papers)
Bar: 6:30-7:30
Dinner: 7:30-8:30
Discussion: 8:30-10
Wednesday July 2:
Breakfast: 8:30-9:30
Session VII: 9:30-11:00 (2 papers)
Coffee Break: 11:00-11:15
Session VIII: 11:15-12:45 (2 papers)
Lunch: 12:45-2
!
PLEASE RETURN THIS FORM ASAP TO:
PLANNING + ACTION WORKSHOP
MARGARET OLENDER
SRI INTERNATIONAL, AI CENTER
333 RAVENSWOOD AVE.
MENLO PARK, CA 94025
NAME/ADDRESS/PHONE:
Please circle which nights you want your room:
June July
29 30 1 2 3 4 5
How many people in your party?
Indicate form of your party:
(ex. couple, couple with 2 children, 2 singles, etc.)
For singles, roommate preference (optional):
Dietary Restrictions :
Stipend needs:
(Please break down into airfare, housing, transportation.
An attempt will be made to at least cover airfare costs.)
-------
∂28-May-86 1627 vijay@ernie.Berkeley.EDU
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 28 May 86 16:27:46 PDT
Received: by ernie.Berkeley.EDU (5.51/1.12)
id AA25133; Wed, 28 May 86 16:27:28 PDT
Date: Wed, 28 May 86 16:27:28 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605282327.AA25133@ernie.Berkeley.EDU>
To: jmc@SU-AI.ARPA
Hello all,
Slowly but steadily we are progressing...
So far we have:
Stage 1 papers:
D. Rumelhart
D. Hofstadter
H. & S. Dreyfus
T. Winograd
J. Weizenbaum
J. Searle
Stage 2 responses:
E. Charniak
J. McCarthy
D. Hofstadter
Stage 3: (Exchanges)
S. & H. Dreyfus
J. Weizenbaum
J. McCarthy
P.S. If you don't have any of the above materials, please let
us know immediately.
∂29-May-86 1210 VAL Reminder: Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SU-AI.ARPA
The last meeting this quarter:
APPLICATIONS OF NON-MONOTONIC REASONING TO INDUCTIVE LEARNING
Benjamin Grosof
Stanford University, Computer Science Department
Thursday, May 29, 4pm
MJH 252
"Bias" is a term used in the machine learning literature to denote the
basis for selection among competing, consistent, explanatory,
inductive hypotheses. I will begin by discussing at a very abstract
level how non-monotonic reasoning can be applied to represent bias in
inductive learning.
I will then show how to formulate non-monotonically some types of
bias in the Version Space [cf. Tom Mitchell] approach to inductive
concept learning. One kind is maximal specificity/ maximal
generality. A second kind is the "linguistic" bias embodied by the
initial version space. A third is the completeness of the description
of classified instances. We show how non-monotonic reasoning can
provide a relatively declarative and semantic account of how to "shift
bias" in inductive concept learning, as opposed to the relatively
procedural and syntactic accounts heretofore explored [e.g. Utgoff's
thesis]. This enables a concise and principled specification of how
to relax overly strong biases, a open problem in machine learning.
∂29-May-86 1321 WALDINGER@SRI-AI.ARPA circumscription
Received: from SRI-AI.ARPA by SU-AI.ARPA with TCP; 29 May 86 13:21:30 PDT
Date: Thu 29 May 86 13:20:27-PDT
From: Richard Waldinger <WALDINGER@SRI-AI.ARPA>
Subject: circumscription
To: jmc@SU-AI.ARPA
cc: val@SU-AI.ARPA
in my talk on clearing a block, i found it necessary
to strengthen the specification. in addition to saying
that the given block was to become clear, i had to say
that all the blocks underneath the given block were to
remain in the same relative position. you suggested that
this strengthening might be achieved by the circumscription
mechanism. on thinking it over, this seems unlikely to me.
if the mechanism could suggest this, why wouldn't it also
impose the additional condition that all the blocks on
top of the given block remain in the same relative position.
of course it is impossible to satisfy this last condition
if the given block is to become clear.
-------
∂29-May-86 1339 lantz@su-gregorio.arpa FYI: Changes to the Proposed systems (hardware) syllabus
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 29 May 86 13:39:05 PDT
Received: by su-gregorio.arpa with Sendmail; Thu, 29 May 86 13:37:31 pdt
Date: 29 May 1986 1337-PDT (Thursday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: phd-program@score, phdcom@sail
Cc:
Subject: FYI: Changes to the Proposed systems (hardware) syllabus
In fact, these changes reflect what John always had in mind but they
got lost in transit somewhere...
------- Forwarded Message
Return-Path: <@SU-SCORE.ARPA:jlh@su-sonoma.arpa>
Received: from SU-SCORE.ARPA by su-gregorio.arpa with Sendmail; Thu, 22 May 86 11:36:28 pdt
Received: from su-sonoma.arpa by SU-SCORE.ARPA with TCP; Thu 22 May 86 11:35:59-PDT
Received: by su-sonoma.arpa with TCP; Thu, 22 May 86 11:35:02 pdt
Date: 22 May 1986 1135-PDT (Thursday)
From: John Hennessy <jlh@su-sonoma.arpa>
To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Cc: lantz@score
Subject: Re: PhD Program revision
In-Reply-To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA> /
Wed 21 May 86 14:48:33-PDT.
Terry,
A few changes to the reading list that got dropped somewhere:
Bell, Mudge and MacNamara, Should be PART III ONLY.
Kogge, Should be Chapter 1 ONLY.
Tannenbaum really is a subset of Mano and is unneeded. If someone wants
to leave it on, I would suggest it be only chapters 1-5. But, in the
interest of keeping the reading list short, let's drop it!
John
------- End of Forwarded Message
∂29-May-86 1345 lantz@su-gregorio.arpa prod...
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 29 May 86 13:45:11 PDT
Received: by su-gregorio.arpa with Sendmail; Thu, 29 May 86 13:43:38 pdt
Date: 29 May 1986 1343-PDT (Thursday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: phdcom@sail
Cc:
Subject: prod...
This message didn't get a rise out of anyone, but the current
discussion on the phd-program bboard seems to reflect similar feeling
among the students at large. So, why NOT take this approach?
Keith
P.S. I've decided that the networks course (ref. A.S. Tanenbaum,
Computer Networks, Prentice-Hall, 1981) should be added to the
specialized areas. A LOT more people take it than take either graphics
or databases, for example.
------- Forwarded Message
Return-Path: <@SU-AI.ARPA:lantz@su-gregorio.arpa>
Received: from SU-AI.ARPA by su-gregorio.arpa with Sendmail; Fri, 23 May 86 09:50:35 pdt
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 23 May 86 09:50:50 PDT
Received: by su-gregorio.arpa with Sendmail; Fri, 23 May 86 09:49:52 pdt
Date: 23 May 1986 0949-PDT (Friday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: phdcom@sail
Cc: jlh@sonoma
Subject: comp
In response to my query as to how John felt about the weighting of
systems -- relative to overall computer science preparation, etc. --
he sent the attached. This inspired me to think (or, more correctly,
remember) that defining the basic preparation in terms of courses,
first, and then constructing a suitable syllabus (based presumably on
the content of those courses), is what we really should be doing.
AND the weight on the exam should indeed reflect the distribution of
"required" courses!
Somehow the "list of courses" aspect has been lost. So I propose to
reopen the discussion from scratch and come up with a list of courses
that everyone agrees to. Not having a catalog in front of me, I'll
just side with John's list for now.
Keith
------- Forwarded Message
Replied: 23 May 86 09:44
Return-Path: <@SU-SCORE.ARPA:jlh@su-sonoma.arpa>
Received: from SU-SCORE.ARPA by su-gregorio.arpa with Sendmail; Thu, 22 May 86 21:04:57 pdt
Received: from su-sonoma.arpa by SU-SCORE.ARPA with TCP; Thu 22 May 86 21:04:42-PDT
Received: by su-sonoma.arpa with TCP; Thu, 22 May 86 14:52:10 pdt
Date: 22 May 1986 1452-PDT (Thursday)
From: John Hennessy <jlh@su-sonoma.arpa>
To: lantz@score
Subject: courses and % of comp
Here's my list of what folks should have:
Systems = 8 courses noty including 108 material
112, 140, 143A, 212, 242, 240 A,B, 243
Applications = 4-5 courses
237A,(B,C?), 225, 248, 245
Theory = 4 courses including basic AI course
223?, 254, 261, 257A
So we have to cover roughly twice as many courses assuming this is the
right level of systems material....
------- End of Forwarded Message
------- End of Forwarded Message
∂29-May-86 1445 VAL re: circumscription
To: WALDINGER@SRI-AI.ARPA, JMC@SU-AI.ARPA
[In reply to message from WALDINGER@SRI-AI.ARPA sent Thu 29 May 86 13:20:27-PDT.]
Circumscription will suggest that the relative positions of blocks remain the
same as much as possible under the condition that the given block becomes clear.
In other words, we'll get (I think) this condition:
on1(x,y) iff (on0(x,y) and not x=A),
where on0 and on1 describe the states before and after the execution of the plan,
and A is the given block. This specification is possible to satisfy, but the
resulting plan will be inefficient: it will lead from
C
B to C
A A B
----- -----------
(with the redundant step of moving C onto B at the end).
Please let me know if you're planning to think more about it and if I can be
of any help. JMC is in Europe now.
Vladimir
∂30-May-86 1237 KIRSH%OZ.AI.MIT.EDU@MC.LCS.MIT.EDU [kirsh: AI Workshop]
Received: from MC.LCS.MIT.EDU by SU-AI.ARPA with TCP; 30 May 86 12:36:51 PDT
Date: 30 May 1986 15:34 EDT (Fri)
Message-ID: <KIRSH.12210881375.BABYL@MIT-OZ>
From: David Kirsh <KIRSH%OZ.AI.MIT.EDU@XX.LCS.MIT.EDU>
To: jmc@SU-AI.ARPA
Subject: [kirsh: AI Workshop]
Date: Wednesday, 26 March 1986 17:34-EST
From: kirsh at OZ.AI.MIT.EDU
Sender: David Kirsh <KIRSH at OZ.AI.MIT.EDU>
To: jmc at SU-AI.ARPA
Re: AI Workshop
John,
You asked me to write down our plans for a conference on AI
foundations. Our intent is to have a small workshop of around 55
people discuss what seem to us the major approaches in AI today. The
conference would be called AI PARADIGMS and would be held in June/July
87 at Endicott House, near Boston.
What do you think of these topics and speakers?
I Logic
chair McCarthy
main Hayes
comm Winograd
II Connectionism
chair Hopfield
main Hinton
comm Rumelhart/ Feldman/ ?
III Methodology (Marr approach)
chair Poggio
main Ullman
comm Kirsh/ ?
IV Organization (parallel Hewitt style)
chair Nilson
main Hewitt
comm Simon/ McCarthy
V Analogical Reasoning
chair Schank
main Winston
comm Gentner
VI Society Of Mind (Minsky's theory)
chair Papert
main Minsky
comm Newell
VII Production Systems
chair Simon
main Newell/ Feigenbaum
comm Norman/ ?
FORMAT 45 min Main presentation
15 min questions of clarification
←←←←←←←←←←←←←←←
15-30 min break
←←←←←←←←←←←←←←←
20 min commentary
10 min questions of clarification
30 min general discussion
←←←←←←←←←←←←←←←←←←←←←
2 1/2 hour total
MAJOR SPEAKER 45 min
1 State the basic thesis. Eg Logic plus extensions can do it all.
Explain the principles, the natural type of problems and tasks where
these principles win.
2 Show an exemplary application of theory. Use simple example
carefully chosen to highlight different aspects of method.
3 Explain where power resides.
4 Discuss scope and limitations of approach. Eg current hard
problems that can be overcome and why. Problems that are not likely to
succumb to this approach and why.
COMMENTATOR 20 min
1 Evaluate the source of power and explain why it works.
2 Comment on 3 and 4 above.
Drafts of major speakers' papers must be submitted to us by March so
that we can return them by mid March with our comments about adherence
to format and clarity. We would like finished drafts of major
speakers' essays by May 15 and finished drafts of commentators' by
early June.
We also request that each participant suggest 2 other people to invite
and 1 graduate student. We then will invite a subset of suggestions
and all grad students.
The workshop will last 3 full days. Fri will have 3 sessions. Sat 2
sessions plus a banquet. Sunday will have 2 sessions. Everything
will take place in Endicott House and will be timed to keep everyone
around, rather than eating off on their own.
All presentations and discussions will be taped, which then will be
made available for distribution at a cheap rate. NSF will help us
cover the costs of videoing and publishing the proceedings.
We're very excited about our format. If the topics are covered
competently we think this workshop will be a genuine service to the
field. Look forward to your comments.
-- David
617 253-6569
∂30-May-86 1609 RA Msg. from Feigenbaum's office
Feigenbaum wants you to look at the letters concerning 30th anniversary
antology and make your selection.
∂31-May-86 0548 nttlab!masahiko@su-shasta.arpa my visit to Stanford
Received: from SU-SHASTA.ARPA by SU-AI.ARPA with TCP; 31 May 86 05:48:40 PDT
Received: by su-shasta.arpa; Sat, 31 May 86 05:46:59 PDT
Received: by ntt.junet (4.12/4.7JC-7) CHAOS with CHAOS-MAIL
id AA20266; Sat, 31 May 86 17:22:45 jst
Date: 31 May 1986 17:09-JST
From: Masahiko Sato <nttlab!masahiko@su-shasta.arpa>
Subject: my visit to Stanford
To: John McCarthy <Shasta!jmc@su-ai.arpa>
Cc: Carolyn Talcott <Shasta!clt@su-ai.arpa>
Message-Id: <517910978/masahiko@nttlab>
Dear Professor McCarthy,
How are you? It was nice that Richard visited us recently.
This year I am planning to visit Stanford from July 15 to August 23.
In order to get a visa for my family, I need a (formal) letter from you.
Will you kidly send me a letter at your earliest convinience stating
that:
* you welcome my visit to Stanford from July 15 to August 23, 1986
* my visit is part of our joint collaboration sponsored by JSPS
(Japan Society for the Promotion of Science) and NSF
My postal address is:
Masahiko Sato
Research Institute of Electrical Communication
Tohoku University
2-1-1 Katahira, Sendai 980
Japan
Thank you very much in advance. I am looking forward to seeing you at
Stanford.
** masahiko **
∂31-May-86 1234 vijay@ernie.Berkeley.EDU AI DISC: Daniel Bobrow
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 31 May 86 12:33:50 PDT
Received: by ernie.Berkeley.EDU (5.51/1.12)
id AA23586; Sat, 31 May 86 12:32:58 PDT
Date: Sat, 31 May 86 12:32:58 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8605311932.AA23586@ernie.Berkeley.EDU>
To: jmc@SU-AI.ARPA
Subject: AI DISC: Daniel Bobrow
*** Response to Les Dreyfus:
The Dreyfus start their critique by asking for some already occurring
"expectation adjustment". In the excitement engendered by the
enthusiasm of a new adventure, and the relative ease of achieving
superficially interesting results, a number of early researchers made
unwarranted extrapolations of the rate of future progress. As
illustrated by the Dreyfus' documentation of Minsky's statements, major
researchers in the field have a new, and hopefully more realistic
evaluation. It is a lesson for all of us to adjust our immediate
expectations, and not succumb to popular press extrapolation. What is
clear is that there will be no step-function success, with the discovery
of some single philospher's stone; there will be the usual slow
accumulation of knowledge, techniques and technology that will extend
the range of applicability of the science of Artificial Intelligence.
The Dreyfus' also make arguments of impossibility for Artificial
Intelligence based on their characterization of computation. I find
their terms very unclear. The notion of "rule-like" could mean simple
if-then rules (falsely) advertised to enable the building of any expert
system. Or it could refer to the program for any digital computer, from
signal processing, visual feature extraction, and holgraphic
construction to list processing and learning symbolic descriptions from
examples. If they mean to describe the limits of simple if-then rules, I
will join them. I think there we again suffer from press-agentitis with
respect to early success in achieving superficial results. If they
imply limits on computation in general, then a more thorough-going
technical discussion is called for.
I find it strange that use of prototypical cases -- a lot of them -- is
thought to be out of the realm of what could be done with computers.
Minsky made a suggestion for organizing programs that way in his famous
Frames paper over ten years ago. However, it has been easier (less work
for the researchers) to try to extract general rules than to gather
cases, and so we have seen that style flourish. As learning programs
get stronger, and we have mechanisms for bringing large numbers of cases
to computers (as we do with human apprentices), we may see this change.
There is no problem in principle.
Incompatible characterizations of computation are used by the Dreyfuses
for comparing human and machine capabilities. Time for computation is
used for humans: "recognizing two patterns as similar ... seems to be a
direct process for human beings"; but the computer algorithm is
(partially) described: "a logic machine [performs] a complicated process
of first defining each patern in terms of features, ...". Similarly, a
developmental argument ("unlikely in a child's mind") is used to counter
a particular suggested form of algorithm.
If we are to understand the limits of computation, then superficial
comparisons of behavior will not do. Dogmatic statements that
"computers functioning as analytic engines cannot ... recognize the
similarity of whole images" also need explication. In what sense is a
massively parallel machine not an analytic engine, and what similarities
will remain unseen. Can not such machines do local and global
comparisons of the image and find such similarities. Very similar
techniqes are used at the moment in matching points in stereo pairs
(though this is clearly a much simpler task.
After decrying the over-extapolation of results by AI people, the
Dreyfus brothers make a similar error, seeing technical problems as
necessary harbingers of ultimate failure. Hard problems do not "a
degenerating research paradigm" make. However, the Dreyfuses are
probably safe in predicting that the common sense problem will "never"
be solved or that AI will eventually fail, since the work required (we
now believe) is significantly longer than the academic lifetimes of
these predictors.
*** Response to Searle
Searle hides a rabid neo-vitalism under the pretense of making a formal
argument. He gives formal definitions, propositions, and logical
deductions, and challenges one to deny any of the propositions. My
responses are of three kinds -- pointing out the major hole in the
argument, mocking the the form of the polemic, and making a suggestion
as to what is the issue really raised by Searle.
The slippery part of Searle's argument comes between proposition 1 and
proposition 2 in the use of the term syntactic. In 1, he defines
programs (as opposed to computers) as formal and hence syntactic. In
proposition 2, syntax has taken on another meaning. It is put in
contrast (and not equivalent) to semantics. But semantics, as a
"foundational principle behind modern logic ..." is formal. Of course
the syntax of logic is not the semantics of logic, but this is a level
shift rather than a form change. Computers can embody many formal
systems, and transformations between them. Thus computers can embody
such formal semantics. The "inexorability of the syntax/semantics
distinction" is a bad play on words.
But more to the point, what is the meaning of semantics? Is it this
formal study? When we discuss the semantics embodied in minds (as in
proposition 3), we are surely not referring to this meaning of that
word. As Fodor pointed out, minds viewed in this way have no semantics
either. It must be from minds embedded in bodies, and those bodies
embedded in a social system that we get the sense of semantics in which
I can successfully refer to the terminal in front of me (and touch it)--
and have you understand the reference. And isn't this the crux of the
robot argument?
Vitalism again.
Searle seems to believe in Dennett's "wonder tissue" -- that miraculous
aspect of the brain that leads to thought. No separation of function
and structure for him. Although in the beginning of his statement
Searle states:
"I have repeatedly and explicitly denounced [the] view [that] only
carbon-based or perhaps only neuronal-based substances could have the
sorts of thoughts and feelings that humans ... have".
this constrasts with his later statements:
"specific biochemical features of the brain ... cause all of our
mental phenomena" "Does anyone really doubt that all of our mental
states are caused by low level (e.g.neuronal) processes in the brain?"
"Is there someone in AI so totally innocent of biological
knowledge that he thinks that the specfic biochemical powers of human
nervous systems can be duplicated in silicon chips."
This insistence on specific biochemical powers of the brain is
reminiscent of arguments made at the end of the 18th century about how
"organic chemicals" could only be produced in living organisms. It was
only after the test tube synthesis of urea that appropriate properties
of organic chemicals as a natural kind began to be explored. The
backbone of carbon atoms that characterizes what we now call organic
chemicals is found in substances never produced in a living thing, and
in fact poisonous to them.
But the real issue brought in to focus by Searle's arguemnt is how
should we think about the phrase: "have a mind in exactly the same
sense that human beings have minds". To me, the way my wife has a
mind (or how she thinks, as I would put it) is very different from how I
have a mind -- in many dimensions; because she is French by upbringing,
a woman, and a non-techologist to mention a few. Heidigger (among
others) points out how much our thinking (use of our minds) is affected
by the background of our social context and upbringing; he describes the
impossibility of anyone brought up in our culture understanding some
Japanese word. I think that no two creatures will ever have a mind in
the same sense as any other. We must consider many ways of dividing
minds into natural kinds.
The questions we must ask then are how are the similarities and
differences to be characterized. Factors must include task, background,
embodiment,... etc. Dangers include extrapolation from one context to
another. A significant danger is jumping from surface similarities to
projection of other independent capabilities. As Winograd mentions, a
worse danger is that careless use of language to describe current very
limited capabilities of machines can lead to redefinition of the word
intelligence, and perhaps lead us to "systematically undervalue
[certain] aspects of intelligent human action".
Perhaps it is this danger, fear of being judged by the presumed
rationality of computers, that leads Searle to make such strong claims
for the one true way to build minds.
*** Response to Winograd
I looked in vain for the controversial assertions promised by Winograd.
He recommends skepticism about uses of technology, care in our
descriptions of phenomena, and social responsibility of scientists -- I
agree whole-heartedly.
Winograd also worries about "computerization"; will not the use of AI
rather than conventional technology perhaps lead to more flexible (and
hence more "human") systems than most we have to deal with now? This I
believe is the hope of the Fifth Generation computing adventure, despite
its excessive hoopla.
*** Response to Hofstadter
Hofstadter argues for particular aspects of the current AI research
programme, and make predictions of what will be most useful. Only time
will tell what will win in the marketplace of ideas. Hofstadter also
wants us adjust our expectations for the long haul. So be it.
I found Hofstadters rhyme detector most amusing, and the alliteration
detector. Why not postulate going on for a while. Computers don't
count 1,2,3, infinity. Two or three dozen of these might cover the
spectrum (so to speak). Minsky has been proposing (and building
justifications for) an organization of mind based on many such
specialized detectors. They seem not entirely implausible to me either.
*** Response to Rumelhart
Connectionism is an interesting new way of looking at things; it can
probably be usefully guided by attempts to implement symbol processing
(See Touretsky, Symbols among the Neurons, for example), while
preserving good properties of nets. I think both levels of description
will be of significant use in the descripiton of human cognitive
process.
∂01-Jun-86 1213 REM@IMSSS Prize for Go program
Received: from IMSSS by SU-AI with PUP; 01-Jun-86 12:13 PDT
Date: 1 Jun 1986 1209-PDT
From: Rem@IMSSS
Subject: Prize for Go program
To: JMC@SU-AI
I seem to recall you had a bet/prize for first program to play Chess better
than grandmaster, or you were involved somehow? Now similar for Go:
29-May-86 21:38:16-PDT,3855;000000000001
Mail-from: SU-NET host Score rcvd at 29-May-86 2138-PDT
Received: from MC.LCS.MIT.EDU by SU-SCORE.ARPA with TCP; Thu 29 May 86 21:36:56-PDT
Received: from MX.LCS.MIT.EDU by MC.LCS.MIT.EDU via Chaosnet; 30 MAY 86 00:38:03 EDT
Date: Fri, 30 May 86 00:39:03 EDT
From: "Keith F. Lynch" <KFL%MX.LCS.MIT.EDU@MC.LCS.MIT.EDU>
Subject: I am sure you have seen this already, but...
To: rem%imsss@SU-SCORE.ARPA
Message-ID: <[MX.LCS.MIT.EDU].922759.860530.KFL>
Date: 22 May 86 19:05:00 GMT
From: pur-ee!uiucdcs!kadie@ucbvax.berkeley.edu
Subject: $1,000,000 Prize
This might be of general interest:
/* May 17, 1986 by chen@uiucdcsb.CS.UIUC.EDU in uiucdcs:uiuc.ai */
/* ---------- "$1,000,000 for a program" ---------- */
The following was posted in net.game.go. In case you don't know about Go,
it is an ancient oriental board game played between two players
on a 19 by 19 grid. The best Go program so far is no better than an
intellegent novice that has received only one week intensive training.
/* May 14, 1986 by alex@sdcrdcf.UUCP in uiucdcsb:net.games.go */
/* ---------- "Million $ prize" ---------- */
I think this is a big news for the go community. The Chinese
Wei Chi(go in Chinese) Association(TWCA) in Taipei, Taiwan and conjunction
with one of Taiwan's largest computer company have put 2 million US
dollar in trust as prize money of computer go games. The top standing
prize is 1 million dollar for any computer go game defeating reigning
junior champion in Taiwan. The prize offer is good for 15 years.
(BTW, if you are wondering how they raise the prize money, take a look
at all the cheap IBM PC clones around.) The prize money is much more
interesting the Fredkin's prize. They are other prizes for the computer
go champion, etc.
The TWCA is the first organization offering prize money for
computer-computer and computer-human competition, according to my
and the computer go game pioneer Bruce, who appeared in TWCA first
computer tournament last January. Bruce lost twice and did not place
in top five. That tournament offered 2 to 3 thousand price money to the
winner. His first loss was to a go game written in BASIC running on
an Apple. Bruce was winning convincingly until the Apple games made
a suicide move which is legal under Chinese rule but not under Japanese
rule. Bruce's game went into loop. The judge allowed Bruce to fix his
code on the spot as long as he can make the move before his time clock
runs out. (They did not want Bruce to lose because he was the main
attraction, and I believe they paid him some appearance fee.) But
Bruce did not fix it right within the 30 minutes he had. I
did not stick around for his second loss. Bruce's game was running on
a 8MHz PC clone.
If you are interested in entering the next competition which
is in November, you better get the rule book on the Chinese rules, which
differ slightly from Japanese in area like suicide moves and scoring.
Last competition was restricted to personal computer, although I
find big disparity in computer power between a MacIntosh and an Apple
II. However, I don't think computing power is the main bottleneck right
now.
If there are enough people interested, I can get additional
detail about the tournament.
Also, a junior champion in Taiwan is about 1 dan in Chinese
amateur rating, which is about 5-6 dan in US and Japanese amateur
rating. Bruce's game was last rated to be 19Q in Japan human
tournament. He said he may push it to 11-12Q by November. I think Bruce
has got a good technique but his potential is limited by his knowledge of
go. But at any rate, you have your work cut out for you.
Alex Hwang
/* End of text from uiucdcsb:net.games.go */
/* End of text from uiucdcs:uiuc.ai */
-------
∂01-Jun-86 1457 LES DARPA Update
To: JMC
CC: CLT
Here are recent communications from & to DARPA. I would like to review both
the proposed task and equipment needs with you and whoever at the earliest
opportunity.
RPG says he is planning to do the Alliant development work using Sun
workstations, so tool-sharing is more likely to work if we have some too.
Alliant says that they will shortly support NFS, which would make it
feasible to use diskless Suns ($5k each). Of course, we can make do with
SAIL terminals with some reduction in efficiency.
∂30-May-86 1415 squires@ipto.ARPA [Stephen Squires <SQUIRES@IPTO.ARPA>: Re: Qlisp messages]
Received: from IPTO.ARPA by SU-AI.ARPA with TCP; 30 May 86 14:15:08 PDT
Received: by ipto.ARPA (4.12/4.7)
id AA09251; Fri, 30 May 86 17:14:57 edt
Date: Fri 30 May 86 17:14:51-EDT
From: Stephen Squires <SQUIRES@IPTO.ARPA>
Subject: [Stephen Squires <SQUIRES@IPTO.ARPA>: Re: Qlisp messages]
To: les@SU-AI.ARPA
Message-Id: <VAX-MM(187)+TOPSLIB(118) 30-May-86 17:14:51.IPTO.ARPA>
In-Reply-To: Message from "PUCCI@USC-ISI.ARPA" of 27 May 1986 09:12:39 EDT
Received: by ipto.ARPA (4.12/4.7)
id AA09244; Fri, 30 May 86 17:13:26 edt
Date: Fri 30 May 86 17:13:19-EDT
From: Stephen Squires <SQUIRES@IPTO.ARPA>
Subject: Re: Qlisp messages
To: PUCCI@USC-ISI.ARPA
Cc: SQUIRES@IPTO.ARPA
Message-Id: <VAX-MM(187)+TOPSLIB(118) 30-May-86 17:13:19.IPTO.ARPA>
In-Reply-To: Message from "PUCCI@USC-ISI.ARPA" of 27 May 1986 09:12:39 EDT
I need the specific list of equipment, the reasons why this choice was made,
the lease verses purchace analysis, and a statement that the university cannot
afford to pay for the equipment. For less than $50K this can be handled
very quickly. Send the material over the net, I will approve it hand have
an ARPA Order ammendment prepared in parallel with the official letter
comming in.
The task can be put in place without the ADP approval providing they hold
back the ADP funds. This is what they should do.
-------
[Above message forwarded to RPG - Les]
∂30-May-86 1009 RPG task description
Sounds like Feigenbaum's crew got an idea thrown into our
task description - the default value after timeout. Also, the
debugging tools sound new, but not offensive. The
`informal methods of specification and reasoning about ...' sounds
new to me, also. I have no particular objection to it.
-rpg-
∂30-May-86 1434 LES re: Qlisp messages
To: SQUIRES@IPTO.ARPA
CC: Pucci@USC-ISI.ARPA
[In reply to message sent Fri 30 May 86 17:14:51-EDT.]
As luck would have it, the proposed task description arrived here at the
same time that John McCarthy was leaving for Sweden, to be inducted into
a national academy there. He will be back the middle of next week, but
will be jet-lagged of course.
I see no problems in the task description, but some of the words are new
and I would like John to have a look before we confirm. Regarding the
capital equipment budget, we now know more concretely what needs to be
acquired and where we can cut corners. I am confident that we can get
below $50k. We will feed back as soon as John returns.
∂02-Jun-86 1512 FEIGENBAUM@SUMEX-AIM.ARPA please respond!
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 2 Jun 86 15:12:25 PDT
Date: Mon 2 Jun 86 15:12:20-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: please respond!
To: jmc@SU-AI.ARPA
Message-ID: <12211696496.58.FEIGENBAUM@SUMEX-AIM.ARPA>
Dear John,
I'm getting close to the wire (with the publisher) on naming
the papers that will go into the "anniversary" anthology of
AI papers, i.e. things are getting desperate.
I would appreciate it if you would make the paper selection (or if you
can't decide, send me two choices, and I'll choose), and return
to me the signed permission form.
Please,please,please,
Ed
-------
∂03-Jun-86 1416 hitson@su-pescadero.arpa CSD-CFC possible action items...
Received: from SU-PESCADERO.ARPA by SU-AI.ARPA with TCP; 3 Jun 86 14:16:45 PDT
Received: by su-pescadero.arpa with Sendmail; Tue, 3 Jun 86 14:15:35 pdt
Date: 3 Jun 1986 1415-PDT (Tuesday)
From: Bruce Hitson <hitson@su-pescadero.arpa>
To: facil@sail, "@xtra.dis[1,les]"@sail
Cc: hitson@su-pescadero.arpa
Subject: CSD-CFC possible action items...
I've recently received a couple of questions/suggestions from students
that should be brought to the attention of CSD-CFC for possible action:
1. Maintenance of public workstations: it was pointed out that some
of the public Sun workstations were broken and that they had been
in disrepair for some time. Apparently, those who might fix them
did not know that they were broken (e.g., some are in the back of
room 408 and other places where they are unlikely to be noticed
unless someone from CSD-CF went specifically looking for them),
and students who used (or would like to have used) them did not
know who to report the hardware trouble to.
Specific suggestion: all CSD-CF "owned and operated" workstations
(and probably other hardware) be labeled with a sticker or similar
indication of who to contact in case of hardware (and in the
case of some systems possibly software) trouble.
2. Departmental postscript printers: increasing numbers of systems are
capable of producing PostScript output for high-quality
text/graphics, but there are no departmental printers on which
this output can be produced. Various individuals/groups own and
operate LaserWriters or similar, but there appears to be additional
demand within the department for PostScript printers. My
understanding of the "final" shopping list from the DARPA funds is
that the LaserWriters that had originally been budgeted were dropped
(see message from LES to FACIL dated May 6) - unless the gateway
was finally dropped in order to keep the LaserWriters. Even if this
happened, however, the two LaserWriters were allocated to groups,
not to general departmental use (as are Boise, Rover, Dover, etc).
Specific suggestion: investigate the possibility of acquiring a
CSD general-purpose Postscript printer (possibly a LaserWriter).
Perhaps this can be run off of LaBrea (since there is presumably
some spare capacity there), and charged for in the same way that
Boise, Rover, Dover, et al are charged for (anyone know details
of accounting software for running LaserWriters?).
That's all for now. --- Bruce
∂03-Jun-86 1445 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 3 June 1986
Previous Balance 12.39
Payment(s) 12.39 (check 5/13/86)
-------
Current Charges 6.00 (bicycle lockers)
0.60 (vending machine)
-------
TOTAL AMOUNT DUE 6.60
Please deliver payments to Debbie Woodward, room 040, Jacks Hall.
Make checks payable to: STANFORD UNIVERSITY.
To ensure proper crediting, please include your Pony account name on your check.
Note: Payment recordation takes up to three weeks after delivery of a payment
(but not beyond the next billing date).
Bills are payable upon presentation. Interest of 1.5% per month will be
charged on balances remaining unpaid 25 days after bill date above.
∂03-Jun-86 1740 JJW Gang-of-Four
I've created an account for you on Gang-of-Four, the Alliant, with the
name "jmc" and password "archcrown". Please change the password to
whatever you want it to be.
You should be able to Telnet into the machine using the name
"Gang-of-Four" or the nickname "go4". However, you should NOT try to
Telnet to or from a TOPS-20 system. It will cause your job to crash.
The Alliant folks have promised to fix this by sending us a new Telnet
program.
We have had several problems with network connections, and things
still don't always work quite right. What's most likely is that your
connection will hang, and for several minutes the machine will seem
unreachable but will then fix itself.
∂05-Jun-86 1408 NILSSON@SU-SCORE.ARPA [Tom Strat <STRAT@SRI-AI.ARPA>:]
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 5 Jun 86 14:07:54 PDT
Date: Thu 5 Jun 86 13:50:42-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: [Tom Strat <STRAT@SRI-AI.ARPA>:]
To: rindfleisch@SUMEX-AIM.ARPA, feigenbaum@SUMEX-AIM.ARPA,
buchanan@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA, genesereth@SUMEX-AIM.ARPA,
tob@SU-AI.ARPA
Message-ID: <12212468067.20.NILSSON@SU-SCORE.ARPA>
Have there been any problems of this sort on the Stanford AAAI papers? -Nils
---------------
Return-Path: <@SRI-AI.ARPA,@SRI-WARBUCKS.ARPA,@SRI-AI.ARPA,@SRI-BISHOP.ARPA:STRAT@SRI-AI.ARPA>
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Thu 5 Jun 86 13:18:44-PDT
Received: from SRI-WARBUCKS.ARPA by SRI-AI.ARPA with TCP; Thu 5 Jun 86 13:17:58-PDT
Received: from SRI-AI.ARPA by SRI-WARBUCKS.ARPA via SMTP with TCP; Thu,
5 Jun 86 13:17:41-PDT
Received: from SRI-BISHOP.ARPA by SRI-AI.ARPA with TCP; Thu 5 Jun 86
13:16:08-PDT
Received: from SRI-MENDEL.ARPA by SRI-BISHOP.ARPA via CHAOS with
CHAOS-MAIL id 24719; Thu 5-Jun-86 13:18:18-PDT
Date: Thu, 5 Jun 86 13:17 PDT
From: Tom Strat <STRAT@SRI-AI.ARPA>
To: aic-staff@SRI-AI.ARPA
cc: Strat@SRI-AI.ARPA
Message-ID: <860605131740.5.STRAT@SRI-MENDEL.ARPA>
Several of us have submitted requests for approval of publication of
AAAI papers, only to have DARPA turn them down. In particular, these
papers have been deemed to be unclassified, but distribution was limited
to DOD and DOD contractors. This is apparently a beaurocratic foul-up
and I am going to be contacting DARPA to get it straightened out. If
those of you who find yourselves in a similar predicament send me the
names of your papers, I will include it in my communications with them
and, hopefully, we can get the whole mess straightened out once and for
all.
--Tom
-------
∂05-Jun-86 1424 LES Minutes of Facilities Committee Meeting on 5/28
To: facil@SU-AI.ARPA, BScott@SU-SCORE.ARPA, Reges@SU-SCORE.ARPA,
Eppley@SU-SCORE.ARPA, Reuling@SU-SCORE.ARPA,
Nilsson@SU-SCORE.ARPA
The committee reviewed a draft policy on administration of computer
equipment and the question of what computer assistance should be provided
in support of departmental administration. Two such development projects
were identified as having high priority:
(1) computer-aided preparation of Stanford personnel forms and
(2) computerized financial planning and monitoring.
The first project would involve some software development and purchase of
a daisy printer suitable for imprinting multi-part forms. Stuart Reges
volunteered to specify the requirements more precisely. The second
project would involve development of spreadsheet-like budgeting software
and interfacing to the Stanford accounting system to semi-automatically
track expenditures against existing plans. Tom Rindfleisch agreed to
draft some specifications for this project, which is likely to require
a substantial effort.
The proposed "Policy on Administration of Computer Equipment" given below
incorporates suggestions made by committee members and certain additional
requirements that were subsequently identified through interviews with
Stanford administrators. A related draft form also is attached; it is
intended for use by non-exempt staff members who need to take terminals
home.
I request that any questions or recommended changes to the policy
statement given below be sent to me by Sunday, June 8. If there are no
substantial unresolved problems with this statement, it will be released
then to C.S. faculty for review and adoption at the next faculty meeting
(Tuesday).
Les Earnest
--------------------------------------------------------------------------
Department of Computer Science
Proposed Policy on Administration
of
Computer Equipment
SCOPE. This policy assigns normal responsibilities for administration of
computer equipment within the Department. It is subordinate to general
University policies as well as the legal and contractual obligations of
the University.
CAPITAL EQUIPMENT REGISTER. A central register of all capital equipment
assigned to the department will be maintained by the Computer Facilities
Group. All packing slips that include capital equipment purchased under
departmental accounts will be forwarded promptly by the recipient to the
Computer Facilities Group. The capital equipment register will be audited
annually by the Department to confirm accuracy and completeness.
CONTROL. Administrative control of equipment purchased with general
departmental funds rests with the Department Chairman. Control of
research equipment acquired under grants and contracts shall rest with the
Principal Investigator who obtained the funds under which it was acquired.
If there is more than one principal investigator on a given contract or
grant, one of them shall be designated as the controlling authority for each
piece of equipment at the time it is received; this must be done in a
written agreement signed by all PIs and sent to the Computer Facilities
Group. Failing that, the first-named PI in the grant or contract application
shall be the controlling authority.
TRANSFERS. When a Principal Investigator leaves Stanford, the disposition
of any of his or her equipment that still belongs to an outside agency
will, of course, be determined by that agency. Under certain
circumstances, Stanford-owned equipment may be transferred to the PI's new
employer; any requests of this sort will be decided by the Stanford
Provost. Otherwise, the disposition of the equipment of a departing PI
will be decided by the Department Chairman. If he determines that the
same research program is to be carried on by a different investigator,
then control of the equipment will be transferred to that investigator.
Otherwise, the equipment will be considered part of the departmental pool
that can be temporarily assigned by the Chairman to any projects.
OFF-CAMPUS USE. Computer terminals, modems and other such equipment can
be assigned to individuals for use off-campus only by the controlling
authority. The recipient must sign an agreement accepting responsibility
for the equipment, promising to report any damage or theft promptly, and
promising to return it on demand of the controlling authority or upon the
recipient's separation from the project for which it was assigned.
Non-exempt staff must sign a special agreement affirming compliance with
labor laws. The controlling authority is responsible for seeing that
complete and timely records are kept of all such equipment loans, returns
and transfers, including verification at least annually on June 1 that the
equipment is still in the hands of the person who accepted it.
--------------------------------------------------------------------------
Equipment Loan Agreement for Non-exempt Staff
I wish to temporary use the equipment listed below off-campus and
understand that I may use it only under the conditions checked below and
with the approval of my supervisor.
Equipment to be used: __________________________________________________.
[Check one.]
__ This equipment is to be used in support of overtime work that is part of
my responsibilities at Stanford. I will keep an accurate account of all
such time spent on my work and will forward time sheets showing this work
to my supervisor at the end of each month.
__ This equipment is to be used in support of my responsibilities at
Stanford. I understand that my working hours on-campus will be reduced by
any time spent working off-campus on a day-by-day basis. I will keep an
accurate account of all time spent on my work and will forward time sheets
showing this work to my supervisor at the end of each month.
__ I wish to use this equipment as part of a self-education program, to learn
about _____________________________. This activity is of my own choosing
and is not a part of my Stanford responsibilities; consequently, I expect
no compensation for time spent on this activity. I agree that I will NOT
use this equipment in connection with any of my Stanford responsibilities
without first getting written approval from my supervisor under one of the
alternative arrangements listed above.
Expected completion date: ________________________.
Employee signature: _______________________ Date: ___________________
Employee name (printed):___________________
APPROVED.
Supervisor Signature: _____________________ Date: ___________________
Supervisor name (printed):_________________
∂05-Jun-86 1741 LES Pucci visit
John Pucci called today to say that he needs to learn more about what we
are doing and have done. It seems that they have a new admiral taking over
and everyone in their organization is supposed to give him a briefing
on past accomplishments and future prospects.
He proposes to come visit on June 24. I said that I would send him a copy
of Gabriel's book on Lisp benchmarking and have arranged to do so. Are you
game to spend a little time with him? Or can you think of a graceful way
to shuck his request?
∂05-Jun-86 1957 SJG busy Friday evening? (6/6)
I'm having some people over either to play charades or to read
a play. Would you and Carolyn like to join us?
Matt
∂06-Jun-86 0859 CONTRERAS@SU-SCORE.ARPA message
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 6 Jun 86 08:58:54 PDT
Date: Fri 6 Jun 86 08:57:25-PDT
From: Tina Contreras <CONTRERAS@SU-SCORE.ARPA>
Subject: message
To: JMC@SU-AI.ARPA
Message-ID: <12212676821.29.CONTRERAS@SU-SCORE.ARPA>
Bill Griffith phoned. He would like to set up a time he could talk with
you, please call. 494-7190. If you are unable to reach him at this
number he asked that you please leave the time you will beable to speak
with him with Rutie.
Tina
-------
∂06-Jun-86 1043 RA Chiang
Minsky from MIT told Chiang to get in touch with you re Chiang's new computer
architecture in hardware and software. Chiang's tel. 856 9562. He will call back
Monday.
∂06-Jun-86 1400 RPG NEXT Evaluation
To: hk.wfm@SU-FORSYTHE.ARPA
CC: nilsson@SU-SCORE.ARPA, JMC@SU-AI.ARPA
Prof. Massy:
I was asked to evaluate NEXT as a possible investment for Stanford. I am
a Consulting Associate Professor affiliated with CSD, and am president of
a local startup with 50 employees. I got my PhD here in 1980. Please treat
the following note as confidential within Stanford:
NEXT is a company dedicated to providing a high-speed, low-cost
educational machine with extensive educational software. This report will
outline the risk factors, with a slant towards the technical risks.
The design goals for the computer are:
1. 2 - 4 (or possibly 5) MIPS (millions of instructions per second).
Current Sun 3 workstations are approximately 2 MIPS, the LOTS computers
(DEC 2060's) are approximately 2 MIPS as well.
2. 4 - 16 megabytes of RAM. Sun 3 workstations have this capacity. LOTS
computers have approximately 12 megabytes.
3. Large capacity disk drives. No figures are available.
4. High resolution, 16 inch diagonal screen. Mouse interface.
5. Digital signal processing capability and software.
6. Unix 5.3 or better operating system.
7. Educational software.
8. $3000 selling price.
9. Available third quarter, calendar 1987.
10. Single board CPU and memory.
11. Ethernet hardware and TCP/IP.
Goals 1 - 7 and 9 - 11 are easily achievable, technically. Goal 8,
low-cost, is problematic. The following paragraphs discuss each in turn
with emphasis on the additional constraint of low-cost. A selling price
of $3000 implies a manufacturing cost of $1700 or so. This machine could
not, today, be manufactured at this cost.
Goal 1: Speed.
Using a 20 - 24 megahertz 68020 with a medium-sized cache memory (32
kbytes, 2- or 4-way), one could get 4 mips with little difficulty. Five
MIPS is achievable with a little faster clock, possibly with a little
bigger cache, or with another CPU. A possible CPU is the Mips, Inc
processor. The Mips, Inc. processor is not yet on the market. Good
compilers for it are a little more difficult than for the 68020. A large
body of code for this processor is not available. However, a C-based Unix
machine acquires software rapidly. My confidence in 4 mips is 95%, and in
5 mips it is 80%.
Goals 2 and 10: 4 - 16 megabytes on a single board.
Four megabytes on a single board with the CPU is trivially achievable.
Sixteen megabytes on a single board with the CPU and cache memory requires
using megabit ram chips. IBM and TI have them available now; the price
curve would naturally bring them down to the current cost-effectiveness of
the 256K ram chips around the first or second quarter of calendar 1988, so
there is a price risk associated with the high end (16 megabytes) of the
memory size goal at the low cost.
Here the success might depend on Steve Jobs's personal negotiation
strengths. At Apple he managed to achieve good pricing from outside
vendors through a combination of tenacity, messianic drive, and persuasive
powers. At Apple he negotiated very low disk drive prices. He might have
to do that here to achieve 16 megabytes. Four or 8 megabytes is easily
achievable. My confidence for 4 megabytes is 100%, for 8 megabytes 100%,
and for 16 megabytes 95%.
Goal 3: Large disk drive.
Again, the combination of a low price with large capacity is hard to
achieve. Steve Jobs had success with this at Apple. My confidence is 80%
Goal 4: Large, high-resolution screen.
NEXT has a prototype working, and it was demonstrated to me. The quality
was excellent, much better than the Sun 3 workstation. The electronics
were simple, the power requirements modest. Achieving the prototype
required a custom chip, which NEXT designed. It was working in the
circuit. The problem is with the cost of manufacturing. NEXT will need an
excellent price from the tube vendor. My confidence is 70% of achieving
the low-cost attribute here.
Goal 5: Digital signal processing (DSP).
NEXT has no expertise in this area. Several of my friends are interviewing
with NEXT. One, in particular, can achieve NEXT's goal of a
Bosendorfer-in-a-box. There is a rub here, though. This friend told me
about his compensation negotiation with NEXT. My friend, a Stanford PhD in
EE, with several patents in DSP registered through OTL, and with an
excellent record of achievement in DSP, was offered $50,000 in salary and
1000 shares of stock. Apparently this is the way Jobs hires everyone, even
at Apple. His promise is that the stock will split many times (it split
64 times at Apple), and that the 1000 shares will become 20,000 - 100,000
shares, at $20 per share when NEXT goes public. Not everyone belives in
Jobs enough to accept this, so perhaps NEXT will not attract and keep all
the talent it needs.
My friend is now negotiating a consulting arrangement with NEXT, rather
than submit to the low offer. My confidence in this aspect of the product
is 50%.
Goal 6: Unix.
Unix is being ported to this machine by one of the outside Unix vendors. I
believe it is Interactive, which is, perhaps, the leading such vendor. My
confidence is 100%.
Goal 7: Education Software (called Courseware).
NEXT relies on outside sources for this, with little or no royalties, from
what I can gather. The alliances with Stanford and the other educational
partners is important for this to be achieved. Part of this is a better
user interface than Unix provides. Software is often the hardest component
in a company that trying to meet an aggressive schedule. My confidence is
70%
Goal 8: Low Cost.
Achieving the goal requires the talents of Jobs to press good terms for
the parts, to put together an efficient manufacturing facility (on a par
with the Fremont Apply facility), and to spur his undercompensated staff
to extreme limits. My confidence is 30 - 40%.
Goal 9: Available October 1987.
A machine, perhaps without fancy software, can be delivered at that time.
My confidence in all of their goals being met in this time frame is 60%.
Adding a 6-month slip moves my confidence up to 80%.
Goal 10: Single board machine.
To do this, NEXT needs 5 custom chips; one of them works. The design team
is excellent, and I believe they can do the designs well. NEXT depends on
outside chip houses, who can slip schedules, etc. My confidence is 80%.
Goal 11: Ethernet.
No information was available on their plans. This is presumably easy.
I would say my overall confidence in the venture achieving its goals is
50%. They can achieve 75% of their goals with 90% confidence. The toughest
goal in the manufacturing cost.
Their marketing strategy seems to depend on placing this machine in the
hands of most every student. Thus, there is a high price-sensitivity. If
the $3000 price becomes $5000, they lose much more than the proportion of
price increase indicates. At $5000 NEXT might effectively lose the
educational market, because a $5000 price tag can be achieved easily by
Sun or the other workstation companies at that point. The edge NEXT has
over Sun and the others is a lower price, an emphasis on the educational
market through software, and Jobs's high visibility. Universities are very
price sensitive.
If they lose a large share of the educational market, they could sell to
the engineering market, but they seem to not be in a position to establish
a retail distribution network.
The price of the investment seems high to me, at least insofar as a
straight venture deal is concerned. I would have placed a $10m to possibly
$15m valuation on the company, even factoring in Jobs's name. They don't
have enough technical achievements to justify any higher. If Jobs were
not involved, I would place a $6m to $10m valuation on the company. I
understand their asking price is nearly three times higher than that.
Stanford will probably not lose money on the deal, but it can be argued
that there is no tangible benefit to Stanford from the deal aside from the
monetary aspect. NEXT must seed Stanford with early machines, it must get
software from Stanford, and it must make sure Stanford gets good pricing,
regardless of whether or not Stanford is an investor. On the other hand,
Jobs probably wants to line up the top universities to bolster the
prestige of NEXT.
Finally, it ought to be noted that very few successful entrepeneurs have
put together a successful second company.
Dick Gabriel
∂06-Jun-86 1533 RA
my vacation
It turns out that I need to go to Israel this summer. I will leave June 26,
and be back at work July 22. Sorry about the short notice, but June 26 was
the last date for the inexpensive flights.
Would you like Susan to replace me while I am gone, or would you like me to
find someone else?
∂06-Jun-86 1557 RINDFLEISCH@SUMEX-AIM.ARPA Re: [Tom Strat <STRAT@SRI-AI.ARPA>:]
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 6 Jun 86 15:57:00 PDT
Date: Fri 6 Jun 86 15:56:29-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Re: [Tom Strat <STRAT@SRI-AI.ARPA>:]
To: NILSSON@SU-SCORE.ARPA, feigenbaum@SUMEX-AIM.ARPA, buchanan@SUMEX-AIM.ARPA,
jmc@SU-AI.ARPA, genesereth@SUMEX-AIM.ARPA, tob@SU-AI.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA
In-Reply-To: <12212468067.20.NILSSON@SU-SCORE.ARPA>
Message-ID: <12212753109.59.RINDFLEISCH@SUMEX-AIM.ARPA>
There have been no DARPA review problems that I know of for KSL papers, Nils.
We have been very careful to negotiate out any "prior review" clauses from our
contracts and cannot engage in any classified research.
Tom R.
-------
∂07-Jun-86 1759 SJG re: busy Friday evening? (6/6)
[In reply to message rcvd 07-Jun-86 17:58-PT.]
No sweat. I hope you had a good time. Care for a raincheck?
Matt
∂07-Jun-86 1759 SJG re: busy Friday evening? (6/6)
[In reply to message rcvd 07-Jun-86 17:58-PT.]
P.S. Did you get the Nobel prize?
∂08-Jun-86 1153 FEIGENBAUM@SUMEX-AIM.ARPA re: please respond!
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 8 Jun 86 11:53:21 PDT
Date: Sun 8 Jun 86 11:53:13-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: re: please respond!
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Sat 7 Jun 86 17:55:00-PDT
Message-ID: <12213233111.32.FEIGENBAUM@SUMEX-AIM.ARPA>
John, will you fill out the letter-form I sent you, with the paper name
and yuour permission signature? Please?
Thanks,
Ed
-------
∂09-Jun-86 0007 100 (from: jmc on TTY44)
Utopias are a bad idea: for essays.
∂09-Jun-86 0814 JMC
return tux
∂09-Jun-86 0842 SJM Good News/Bad News
Yes, I still have The Good News Is The Bad News Is Wrong in my custody.
For book report on same, see gdnews[1,sjm]. I have put said book on
your desk.
Susie
∂09-Jun-86 0853 RA check
I put your check from Sprachen der Wiessenschaft in the drawer in your desk.
∂09-Jun-86 0901 VAL Waldinger on McDermott's "Critique"
To: JMC@SU-AI.ARPA, phayes@SRI-KL.ARPA
∂05-Jun-86 1239 WALDINGER@SRI-AI.ARPA message to mcdermott
Received: from SRI-AI.ARPA by SU-AI.ARPA with TCP; 5 Jun 86 12:39:15 PDT
Date: Thu 5 Jun 86 12:38:18-PDT
From: Richard Waldinger <WALDINGER@SRI-AI.ARPA>
Subject: message to mcdermott
To: val@SU-AI.ARPA
here is the text of my message to mcdermott.
Vladimir Lifschitz called my attention to your remarks
on Bob Moore's Bomb-in-the-Toilet Problem in your paper
"Critique of Pure Reason." You show that the Nilsson-
Luckham method for extracting answers is not adequate.
In fact the issues raised by the problem have been
dealt with, though not by the Nilsson-Luckham method.
First of all, it is possible to extract single answers
from resolution proofs: often these will be conditional.
However, having a conditional answer does not guarantee that
it will be operational; we would not be happy with a plan
if In(bomb, a) then carry(a, toilet) else carry(b toilet),
because presumably we cannot evaluate the "In" predicate.
To avoid such nonoperational solutions, we must impose
a sort of constructiveness on our proofs. We distinguish
between "primitive" and "nonprimitive" symbols. Primitive
symbols are the ones we know how to compute. In the Moore
problem, "In" is nonprimitive, the others are primitive.
We reject proofs yielding plans containing nonprimitive symbols.
Thus the proof yielding the above conditional plan would be rejected.
There is an easy four-step proof in situational logic
yielding the primitive plan
carry(a, toilet);carry(b, toilet).
The answer-extraction framework is in my 1980 TOPLAS paper
with Manna or in my 1969 IJCAI paper with Lee. I'm now working
on a paper (with Manna) on planning in situational logic, which
I'll send you.
richard
-------
∂09-Jun-86 0907 RA telex from Barbara Kuhn
This is a telex you got from Barbara Kuhn:
Having almost finished the documentation of the Hanover presentation, I'd
like to ask you for the proper spelling of 2 names. First, the Frenchman's
name who together with Robert Kowalski developed logic programing
(Alain Come(?rera) and the cartoonist with the abnormal penguin example
(Cyd Harris?). A speedy reply by telex would be very much appreciated.
With thanks in advance and regards,
Barbara Kuhn.
∂09-Jun-86 1000 JMC
Call Jussi about his return and Shankar.
∂09-Jun-86 1041 RA Please call
Please call Feigenbaum's office 3-4879 re: anthology
∂09-Jun-86 1041 OR.LUSTIG@su-sierra.arpa Re: I always thought musicians were fools.
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 9 Jun 86 10:41:15 PDT
Date: Mon 9 Jun 86 10:40:00-PDT
From: Irvin Lustig <OR.LUSTIG@su-sierra.arpa>
Subject: Re: I always thought musicians were fools.
To: JMC@su-ai.arpa
cc: su-bboards@su-ai.arpa
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Sun 8 Jun 86 23:21:00-PDT
Message-ID: <12213481926.12.OR.LUSTIG@su-sierra.arpa>
I think your header to this message is inappropriate. Because 2 musicians
made a decision which is foolish, you condemn ALL musicians as fools.
As a musician, I resent that remark. Your header generalizes all
musicians as fools, based on the actions of two.
If your header was "I always thought SOME musicians were fools", then
I would have no argument.
-Irv
-------
∂09-Jun-86 1058 GRP Tomorrow's meeting
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
We have a meeting scheduled for tomorrow at 2pm. Ralph says this may
conflict with a CSD faculty meeting. Shall we change the time? (Any
time is ok with me.)
∂09-Jun-86 1130 BARWISE@SU-CSLI.ARPA Invitation to Lunch
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 9 Jun 86 11:30:36 PDT
Date: Mon 9 Jun 86 11:29:57-PDT
From: Jon Barwise <BARWISE@SU-CSLI.ARPA>
Subject: Invitation to Lunch
To: jmc@SU-AI.ARPA, nilsson@SU-SCORE.ARPA, john@SU-CSLI.ARPA, sag@SU-CSLI.ARPA,
sf@SU-CSLI.ARPA
cc: Ingrid@SU-CSLI.ARPA, Nissenbaum@SU-CSLI.ARPA
Could you come to lunch this Friday, at noon, in the Faculty club, to
help kick of the SSP (Symbolic Systems Program)? We have a special
dining room. RSVP. Thanks, Jon
p.s., If anyone knows the email address of herb clark, would you
forward this to him. Otherwise I might have to break down and use the
phone.
-------
∂09-Jun-86 1144 RA Charles Wallace, NSF
Wallace would like you to call him re reuqest for extension of US-Japan
grant. His number (202) 357 9558. He will be there for another hour and a half;
tomorrow (our time) 11:00 -- 1:00 and Wednesday morning.
∂09-Jun-86 1306 RA John Hopcroft
Hopcroft called; his tel. (607) 255 7416.
∂09-Jun-86 1351 SJM
not swimming
∂09-Jun-86 1517 RA Trip to LA
You have reservation SJ to LA 8:35, Western 686, arr. 9:38.
On the way back, Air Cal 527, 4:27, arr. 5:27. Is this ok?
∂09-Jun-86 1520 RA ebos meeting
Will Thursday at 2:00 be ok for the meeting? Please let me know so that I can
check with the others.
Thanks,
∂09-Jun-86 1637 RA Spektrum bill
They paid you what you asked for ($4,486.66). However, we stated on the
invoice that you
would send them another invoice for the hotel once you get your american
express bill. Did you get your bill yet?
We were going to send the Italians an invoice for:
Airfare $1,052.22
Lecture fee $3,000.00
Taxi SF to SU 40.00
----------
$4,092.22
∂09-Jun-86 1755 LES EBOS Meeting
To: REG@SU-AI.ARPA, JMC@SU-AI.ARPA, CLT@SU-AI.ARPA, GRP@SU-AI.ARPA
It appears that we could meet after the Faculty meeting tomorrow (Wednesday)
if CLT can make it -- everyone else says it is OK. That would be some time
in the 3:30-5:30 time frame depending on how much discussion happens at the
faculty meeting.
∂09-Jun-86 1839 CLT EBOS meeting
To: LES@SU-AI.ARPA, REG@SU-AI.ARPA, JMC@SU-AI.ARPA, GRP@SU-AI.ARPA
Tomorrow (Tuesday not Wednesday) after the faculty meeting is
ok with me. The time constraint is that I will have to leave by 4:45.
Otherwise I can try to make it for a 1:30 meeting.
∂09-Jun-86 1846 LES EBOS meeting
To: CLT@SU-AI.ARPA, REG@SU-AI.ARPA, JMC@SU-AI.ARPA, GRP@SU-AI.ARPA
OK, lets do it after the faculty meeting. Nils estimates that it will
be over by 3:30pm.
∂09-Jun-86 1855 vijay@ernie.Berkeley.EDU AI DISC: Daniel Bobrow Response to Joseph Weizenbaum
Received: from ERNIE.Berkeley.EDU by SU-AI.ARPA with TCP; 9 Jun 86 18:55:17 PDT
Received: by ernie.Berkeley.EDU (5.51/1.14)
id AA20033; Mon, 9 Jun 86 18:54:14 PDT
Date: Mon, 9 Jun 86 18:54:14 PDT
From: vijay@ernie.berkeley.edu (Vijay Ramamoorthy)
Message-Id: <8606100154.AA20033@ernie.Berkeley.EDU>
To: jmc@SU-AI.ARPA
Subject: AI DISC: Daniel Bobrow Response to Joseph Weizenbaum
Response to Weizenbaum
Weizenbaum makes an impassioned plea for the inherent difference that
must always be there between an intelligence embedded in a human and
embedded in a computer. One cannot doubt that now, nor in the near
future. In the near term, the important issue is when this matters.
Weizenbaum, perhaps scared by the untoward "success" of the Eliza
program, believes that machines should not be applied to problems of
clinical psychiatry. But perhpas it is the very differences that would
lead much better versions of Eliza to be a useful first step, and/or
adjunct to a more extended psychiatric rehabilitation program. The even
"temper" and regualr response of a computer may lead to extremely
insecure people feeling they can build a model of some part of the
world. It is well known that better self evaluation can aid people in
self improvement -- the hallmark of the "talking cure". Similar
arguments can be raised about other tasks that Weizenbaum sets as
off-limits for computers.
In the longer run, one can where the differences between computer and
human cognition will lie, and what the will arise from. Will robots
brought up with people (have a long enduring relationship with them) and
interacting with the world be more or less different than a human
brought up in a very different culture (say as different as Japan and
the US today). How will the exact nature of the "inner life" of these
machines affect the relation they can have with people? Isaac Asimov is
a better predictor of possibilities than I; but long term predictions
ar I think irrelevant to how we try to do our AI today.
----- dgb:
∂09-Jun-86 1958 CLT calendar item
fri 23-jun 10:30 TTM to James L. Shively, MD
SF PEDI ORTHO CLinic 2425 Geary Blvd 929-4620
∂09-Jun-86 2000 JMC
Jerry about Mosher
∂09-Jun-86 2243 somewhere!nobuo@nttyrl.ntt.junet Common Business Communication Language
Received: from SU-SHASTA.ARPA by SU-AI.ARPA with TCP; 9 Jun 86 22:43:08 PDT
Received: by su-shasta.arpa; Mon, 9 Jun 86 22:41:36 PDT
Received: by nttlab.ntt.junet (4.12/4.7JC-7) with TCP
id AA07804; Tue, 10 Jun 86 14:17:22 jst
Received: by nttlab.ntt.junet (4.12/4.7JC-7) with TCP
id AA07712; Tue, 10 Jun 86 14:16:32 jst
Received: by nttyrl.ntt.junet (4.12/4.8Jb)
id AA22487; Tue, 10 Jun 86 10:01:41 jst
Message-Id: <8606100101.AA22487@nttyrl.ntt.junet>
Date: Tue 10 Jun 1986 09:54:02 JST
From: Nobuo Kawashima <nobuo@nttyrl.ntt.junet>
Subject: Common Business Communication Language
To: nttlab!Shasta!JMC@su-ai.arpa
Cc: NTT-20!goto@su-shasta.arpa, nttyrl!kawashima@su-shasta.arpa
Dear Prof. McCarthy:
Today I received your paper, "Applications of Circumscription to
Formalizing Common Sense Knowledge."
Thank you very much for sendig your paper.
Regards,
Nobuo Kawashima
∂10-Jun-86 0729 Mailer failed mail returned
To: JMC
In processing the following command:
MAIL
The command was aborted because these Host Name(s) are Unknown:
nttyrl.ntt.junet
------- Begin undelivered message: -------
∂10-Jun-86 0729 JMC re: Common Business Communication Language
[In reply to message sent Tue 10 Jun 1986 09:54:02 JST.]
"Common Business Communication Language" and "Applications of Circumscription
to Formalizing Common Sense Knowledge" are different papers. Did you
want both, and which do you have?
------- End undelivered message -------
∂10-Jun-86 0930 JMC
Nafeh 943-1711
∂10-Jun-86 0953 CONTRERAS@SU-SCORE.ARPA message
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 10 Jun 86 09:53:17 PDT
Date: Tue 10 Jun 86 09:51:36-PDT
From: Tina Contreras <CONTRERAS@SU-SCORE.ARPA>
Subject: message
To: JMC@SU-AI.ARPA
Message-ID: <12213735259.19.CONTRERAS@SU-SCORE.ARPA>
Rutie will be back at 11:00.
-------
∂10-Jun-86 1001 MDD Ernie Davis
To: JMC@SU-AI.ARPA
CC: davism@NYU-CSD1.ARPA
Your comments about him (cf. recent letter) would really be very helpful.
If you reply by computer mails please use this address:
davism@csd1
By the way, I''ll be in Berkeley, July 16-Aug. 31.
Thanks, Martin
∂10-Jun-86 1116 RA Ebos meeting
Gorin would rather not have the meeting on Thursday since
Thursday is his big push to grade finals. He suggested today after
the faculty meeting. Will this be ok with you?
Please let me know.
Thanks,
∂10-Jun-86 1121 TAJNAI@SU-SCORE.ARPA
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 10 Jun 86 11:21:41 PDT
Date: Tue 10 Jun 86 11:12:28-PDT
From: Carolyn Tajnai <TAJNAI@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Tue 10 Jun 86 11:05:00-PDT
Message-ID: <12213749980.13.TAJNAI@SU-SCORE.ARPA>
Between 12:30 and 1 p.m. Depends on how long it takes everyone to
get from the stadium to the Old Union Courtyard and for me to give
instructions and call everyone to order.
C.
-------
∂10-Jun-86 1426 CSL.GERLACH@su-sierra.arpa CSL fac cand--MacQueen
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 10 Jun 86 14:25:53 PDT
Date: Tue 10 Jun 86 14:20:33-PDT
From: Sharon Gerlach <CSL.GERLACH@su-sierra.arpa>
Subject: CSL fac cand--MacQueen
To: linton@su-amadeus.arpa, ullman@su-aimvax.arpa, guibas@DECWRL.DEC.COM,
pratt@su-navajo.arpa, jmc@su-ai.arpa, nilsson@su-score.arpa,
cheriton@su-pescadero.arpa, lantz@su-gregorio.arpa, ungar@su-sonoma.arpa
cc: richardson@su-score.arpa, csl.gerlach@su-sierra.arpa
Message-ID: <12213784221.28.CSL.GERLACH@su-sierra.arpa>
We have a CSL faculty candidate who will be interviewing on Mon, June 23
and the search committee would especially like you to talk with him.
David MacQueen is at Bell Labs & received his PhD in 1972 from MIT.
His research interests include:
1. Theory of data types:
semantic models for types
type checking algorithms and type inference systems
type abstraction
2. Design & implementation of functional programming languages
3. Programming environments & meta linguistic tools
4. Theory of domains & denotational semantics
5. Computational models & languages for concurrent programming
I do not have his abstract yet, but will send you a copy once I receive it.
Please let me know YES or NO if you are able to see him, and a time
preference.
Thanks
Sharon
-------
∂10-Jun-86 1559 RA Kathleen Martin
Kathleen Martin (415) 832 7430 called re United Cerebral Palsy Computer Access
project.
∂10-Jun-86 1616 SJM where?
Where are Doctor's Dilemma and America's Birthday Party stashed?
Susie
∂10-Jun-86 1657 G.GORIN@LOTS-A confirm meeting
Received: from LOTS-A by SU-AI.ARPA with TCP; 10 Jun 86 16:57:07 PDT
Date: Tue 10 Jun 86 16:55:18-PDT
From: Ralph Gorin <G.GORIN@LOTS-A>
Subject: confirm meeting
To: jmc@SU-AI.ARPA, greep@CAMELOT, clt@SU-AI.ARPA, les@SU-AI.ARPA
Message-ID: <12213812393.112.G.GORIN@LOTS-A>
This is to confirm that I can attend the EBOS meeting on 6/17 at 2PM
Ralph
-------
∂10-Jun-86 1724 CLT calendar item correction
fri 27-jun 10:30 TTM to James L. Shively, MD
SF PEDI ORTHO CLinic 2425 Geary Blvd 929-4620
∂10-Jun-86 PALLAS@su-sushi.arpa 10-Jun-86 JMC re: Evaluation of CS Profs @ Stanford (from SAIL's BBOARD)
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 10 Jun 86 18:45:28 PDT
Date: Tue 10 Jun 86 18:43:25-PDT
From: Joseph I. Pallas <PALLAS@su-sushi.arpa>
Subject: re: Evaluation of CS Profs @ Stanford (from SAIL's BBOARD)
To: SU-BBOARDS@su-ai.arpa
cc: JMC@su-ai.arpa
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Tue 10 Jun 86 18:16:00-PDT
Message-ID: <12213832073.9.PALLAS@su-sushi.arpa>
JMC says:
The Tau Beta Pi survey is correctly entitled an opinion survey.
The implication is that this somehow makes the survey less useful for
the task of evaluating professors. If JMC has some objective criteria
by which one can judge the teaching abilities of professors, I would
be delighted to hear them (and so, I think, would the rest of the
University). Otherwise, I can't think of anything more useful than an
opinion survey.
joe
-------
jmc - It calls itself an opinion survey, and it is indeed useful in
evaluating teaching. However, to call it an evaluation is to imply
that student opinion is authoritative. I have no special ways to
propose for evaluating teaching. The way that is used prospectively
is to invite each job candidate to give a lecture. There
appears to be a high correlation between the quality of such lectures
and subsequent teaching. Of course, teaching ability is the second
criterion at a research university.
[JMC - Forwarded to OTHER-SU-BBOARDS from line 25]
∂11-Jun-86 0946 davism@nyu-csd1.arpa
Received: from NYU-CSD1.ARPA by SU-AI.ARPA with TCP; 11 Jun 86 09:44:22 PDT
Date: Wed, 11 Jun 86 11:47:30 edt
From: davism@nyu-csd1.arpa (Martin Davis)
Message-Id: <8606111547.AA00171@nyu-csd1.arpa>
Received: by nyu-csd1.arpa; Wed, 11 Jun 86 11:47:30 edt
To: jmc@su-ai
From: John McCarthy <JMC@SU-AI.ARPA>
Subject: Ernest Davis
To: MDD@SU-AI.ARPA
I don't know him. What does he do that I might have opinions about?
In general I'm reluctant to spend much time on evaluations other than
for appointments to tenure, since we evaluate or own non-tenure
re-appointments locally.
Ernie got his Ph.D. with McDermott at Yale 3 years ago.
His interest has been in formalizing common sense reasoning,
especially common sense physics. He is our only faculty member
who is in some sense a product of the AI community. He wrote the
article on common sense reasoning for the forthcoming
Encyclopedia of AI.
His approach to common sense reasoning has been to formalize
the geometric properties of physical bodies at a very abstract
level. So we have simply connected sets with sectionally smooth
boundaries, all expressed in first order logic. We fear that this
approach is going nowhere. I wanted your opinion because I know
of your long standing interest in computer implementations of
formal reasoning. But perhaps your not being aware of his work is
already a negative comment.
∂11-Jun-86 1000 JMC
Sandewall letter
∂11-Jun-86 1021 RA Invitation to 2nd World Basque Congress
There is going to be a congress, 8/30 - 9/4/87, in the San Sebastian, Spain,
which will include a conference on AI. Professor Goirizelaya who, if I understood
it correctly, was a student of yours at Stanfords, is one of the organizers.
They were wondering whether you would be willing to come. They will
contact you tomorrow around 10:30 in the morning. In case you are not here,
what shall I tell them?
∂11-Jun-86 1030 VAL Boyer's problem
Since you seemed to be interested in the problem Boyer had asked me to work on,
let me tell you how the things stand now. We are given a predicate defined by
a predicate formula, and some elementary predicates are known to have finite
extensions; the problem is to determine whether the given predicate has a finite
extension which can be effectively computed. During my visit to MCC last week, I
simplified and substantially generalized my condition and wrote a rather
straightforward algorithm for actually producing a list of objects satisfying
a given predicate (possibly with repetitions). Boyer apparently believes that
even simple-minded algorithms like that can be useful and thinks of coding it.
∂11-Jun-86 1105 SJM bday essay
All I can find under bday.lit is aone paragraph. There is more somewhere
else, isn't there?
Susie
∂11-Jun-86 1157 GIVAN@su-sushi.arpa Starting summer work
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 11 Jun 86 11:57:08 PDT
Date: Wed 11 Jun 86 11:52:12-PDT
From: Bob Givan <GIVAN@su-sushi.arpa>
Subject: Starting summer work
To: jmc@su-ai.arpa
cc: les@su-ai.arpa
Message-ID: <12214019358.34.GIVAN@su-sushi.arpa>
Tomorrow (Thurs) is the day i had specified a month ago as the day i would
start to work for the summer full time.
I am available to work tomorrow.
However, from the local perspective of today as i have just returned from
an unexpected trip to the east coast, it looks infinitely better for
me to start Monday (~5 days from now). But only if it's all the same to
you--if you were in any way counting on, or could especially use me
tomorrow, i'll be there.
i'll be reading my mail,
Thanx,
Bob Givan
p.s. i also need an idea of when and where to report to first get started!
-------
∂11-Jun-86 2003 LES Last (?) shot at Qlisp
To: JMC, RPG, CLT
The problem with the nominally 100kb modem was found by Tom D. to be in
the phone line -- it grossly misses specifications. It will allegedly
be fixed in the next day or so, so that the new modem can be tested again.
If it continues not to work for some reason, then it will certainly still
be possible to use traditional 56kb modems with a specially conditioned line.
Based on this expectation, Dick agrees that we can drop the microwave link
from the Qlisp budget, which puts capital equipment under the magic level
of $50k, making approval (supposedly) instantaneous. I have altered the
budget below accordingly and have changed the statement in the original
Lucid proposal about having a Sequent machine there to having access to an
Alliant through a data link of at least 56 kb/s.
Otherwise, the only change from what we received from Pucci is a twiddle of
the second paragraph -- it used to put Dick in charge of the "design and
implementation of Qlisp," which is true in a sense, but to avoid ambiguity
in the improbable event of a dispute I think it best to state that he is
solely responsible for just the implementation.
So here is the whole caboodle -- please take your last shot. I aim to try
to walk it through the bureaucrats and get it shipped officially to Pucci
tomorrow afternoon.
Les
BACKGROUND
This research is for the implememntation and demonstration
of applications of an extension of Lisp for parallel processing
called Qlisp. Qlisp allows the programmer to specify tasks that
may be performed in parallel by creating queues of tasks. This
project involves developing Qlisp for a parallel processor,
developing test applications for it, and testing the performance
of the applications on substantial symbolic computation tasks.
The project will be carried out with both formal and
informal collaboration among several groups. The principal investigator
for the overall project will be Professor John McCarthy. The Qlisp
implementation shall be done under subcontract to Lucid, Inc. The
implementation of Qlisp shall be supervised by Richard P. Gabriel of
Lucid. Development of parallel programming tools and programs for testing
applications will be supervised by Carolyn Talcott.
It is generally agreed that the main hope for large
increases in computer speed, whether for numerical work or
symbolic computation as in artificial intelligence, lies in
massive parallelism. Projects are being undertaken that will
involve hundreds or even thousands of processors. Therefore, it
is important to explore a variety of approaches to getting good
performance from parallel processors on a variety of problems.
This project's approach is queue-based multi-processing. It
has the following features:
Each processor can address the whole of memory, and a
processor may execute programs anywhere in memory on data located
anywhere in memory.
The programmer indicates when parallelism is possible by
using parallel constructions in the source language, which is an
extension of Lisp.
When a program comes to a statement allowing parallelism and
decides (according to the computed value of an allow-parallelism
parameter in the statement) that parallelism is to be invoked, it
adds a collection of tasks to a queue and starts on the first.
When a processor completes a task it goes to the queue for
its next task. It may execute some queue management code to
decide what to do next.
Basing parallelism on run-time queues means that a program
isn't written or compiled for any specific number of processors.
The number available can even change during the course of a
computation. Tasks need not be of similar length, because a
processor finishing a short task merely takes another queue.
It is necessary to follow through on the development and
implementation of Qlisp by trying it out on substantial
applications. These trials will most likely lead to improvements
in Qlisp and in the parallel processing hardware. They may also
determine whether it is possible to relax the requirement that
all memory be equally accessible to all processors, since the
hardware people find this expensive to implement.
There are two kinds of problems to which parallelism can be
applied. In one case the goal is to make what you already have
run faster. This improves interaction with the computer and
allows larger problems to be addressed. In the second case,
speed is of the esssence. Part of the algorithm is devoted to
insuring speed and some sort of real time behavior. For example,
a process may want to time out or use some default value if a
subprocess does not reply within a given time limit. Qlisp
primitives directly address problems of the first sort.
Additional primitives may be needed for the real time
applications.
TOOLS FOR PARALLEL SYMBOLIC PROGRAMMING
This work involves an entirely new style of programming and
will require the development of new methods and tools to support
the programming activity. This development will be carried out
in parallel with and in support of the chosen application.
Algorithm Design
Part of the development of parallel Lisp involves
discovering what primitives are needed to carry out various
tasks. This involves classifying problems as to the sort of
parallelism required and the design of control structures
appropriate for the various classifications. Two well known
classifications are AND versus OR parallelism and asynchronous
tasks with occasional interactions versus synchronous tasks
operating on a shared data structure. Somewhat newer ideas
include pipelines and other geometric structures that make it
easy to visualize the interactions among a set of processes. In
some cases successful use of parallelism will require design of
data structures with information making them suited to parallel
processing.
Methods For Debugging
Traditional Lisp debugging methods such as stepping and
tracing are not going to be adequate for debugging Qlisp
programs. Programmers will need to learn where to look for bugs
and to devise methods for identifying meaningful checkpoints.
This effort shall include the building of tools for monitoring
the progress of a set of processes working on a problem. It
shall also include work on informal methods of specification and
reasoning about Qlisp programs as aids to the process of program
development and testing.
Benchmarks and Measures of Success
Benchmarks for testing parallel Lisp and for comparing the
relative efficiency of parallel and sequential algorithms shall
be designed. Also tools for measuring factors such as degree of
parallelism achieved, queue management overhead, and memory
contention shall be developed. Some measuring tools have already
been developed by Gabriel in a sophisticated system for
simulating execution of Qlisp programs. This system was used in
initial experiments testing Qlisp primitives.
MILESTONES
7 Months After Initiation of Task (MAIT) - Have a monoprocessor
version of a subset of Common Lisp suitable for parallel
computation working on a chosen parallel computer.
Preliminary specification of Qlisp extensions to Common Lisp
completed.
12 MAIT - Preliminary version of Qlisp working, with preliminary
documentation. Specification of test problems to be run shall be
completed.
18 MAIT - Improved version of Qlisp completed together with
documentation. Test results of sample problems run on the
parallel computer documented.
DELIVERABLES
12 MAIT - Copy of preliminary documentation.
Report detailing the specification of the test problems to
be run.
18 MAIT - Copy of documentation on the improved Qlisp.
Report detailing test results of sample problems.
19 MAIT - Final report on task, including progress, lessons
learned, and future research to be considered.
In addition to the above stated milestones and deliverables,
quarterly progress reports shall be submitted in accordance with
DARPA/SPAWAR 613 reporting instructions.
BUDGET
Personnel
Prof. John McCarthy, Principal Investigator 35,467
(15% acad. yr., 50% summer)
Lester Earnest, Senior Res. Assoc. (50%) 50,625
Carolyn Talcott, Research Associate (100%) 64,500
-----, Research Associate (100%) 61,500
-----, Computer Systems Analyst (100%) 60,000
-----, Computer Technician (25%) 12,000
Ian Mason, Student Research Assistant 20,160
(50% acad. yr., 100% summer)
-----, Student Research Assistant 20,160
(50% acad. yr., 100% summer)
-----, Student Research Assistant 7,560
(beg. 6/1/87; 50% acad. yr.,
100% summer)
-----, Student Research Assistant 7,560
(beg. 6/1/87; 50% acad. yr.,
100% summer)
-----, Secretary (50% time) 15,660
-------
Salary subtotals 355,192
Allowance for salary increases 28,415
(8% beginning 9/1/86,
16% beginning 9/1/87)
-------
Salary totals 383,607
Staff benefits (25.4% till 9/1/86, 98,459
25.6% till 9/1/87, then 26.2%)
Consultants (10 days/yr. @ $500) 7,500
Travel (8 East Coast trips/yr. @ $1000, 22,500
14 Western trips/yr. @ $500)
Computer maintenance 64,548
Computer time costs 40,000
Other direct costs 36,100
-------
Subtotal 652,714
Capital equipment
4 Sun 3/50M-4 Workstations 22,120
Ethernet page printer 11,937
Interfacing equipment & peripherals 12,000
-------
Subtotal 46,057
Lucid subcontract 722,600
Indirect costs (69% of direct costs 490,213
initially, 73% beginning 9/1/86,
excluding all capital equipment and
subcontract amounts over $25,000) -------
Total $1,911,584
Lucid
Project Cost Estimate
($000's)
Half-
Year 1 Year 2
Direct Labor (Sch. A) 109.2 56.7
Salary Related Expense @19.8% 21.6 11.2
-----------------
Salary Expense 130.8 67.9
Overhead @148% 193.6 100.5
Other Direct Project Costs
(Sch. B) 126.0 9.5
-----------------
450.4 177.9
Fee @15% 67.6 26.7
-----------------
Total Project Cost 518.0 204.6
Notes:
1. Project cost assumes access to Alliant machine provided to Lucid
at no additional cost through a communication link to Lucid facilities
with a bidirectional transfer rate of at least 56 K bits/s. for duration
of project.
2. Salary Related Expense and overhead are provisional rates
based on FY86 (year end June 30) business plan projections.
3. Direct Labor rates escalated at 10% per year.
Schedule A
Direct Labor
($000's)
Half-
Year 1 Year 2
Principal Investigator
man-months 3 1
rate per month 6.0 6.6
total dollars 18.0 6.6
Senior Scientist
man-months 12 6
rate per month 5.0 5.5
total dollars 60.0 33.0
Scientist
man-months 6 3
rate per month 4.2 4.6
total dollars 25.2 13.8
Technician
man-months 2 1
rate per month 3.0 3.3
total dollars 6.0 3.3
Total Man-Months 23 11
Total Dollars 109.2 56.7
Schedule B
Other Direct Project Costs
($000's)
Half-
Year 1 Year 2
Computer Equipment & Supplies
Symbolics 3670 108
Supplies @ 3K/yr 3 1.5
Stanford Communication 2 1
Travel & Subsistence
Air
2 trips to Washington, D.C. @1.5 ea.
1 trip to National Conference @1.5
4 trips to East Coast, technical workshop @1.5 ea.
Local travel @.5/yr 11 6
escalate @10%/yr
Printing and Reproduction @2K/yr 2 1
---------------
Total 126 9.5
∂11-Jun-86 2016 LES Oops
To: JMC, RPG, CLT
I neglected to finish fiddling the capital equipment budget. Here is what
I propose we use. Of course, we can later decide to spend the money any
way we want, for all practical purposes, though it helps to make this
correspond to expected reality as much as possible.
Capital equipment
4 Sun 3/50M-4 Workstations 22,120
2 Fujitsu "Eagle" disk drives 15,600
1 Alliant K002 cabinet 7,500
Interfacing equipment & peripherals 4,000
-------
Subtotal 49,220
∂12-Jun-86 1112 RA Dr. Perris
Dr. Perris wanted to remind you about your appointment at 2:00pm, Monday,
June 16.
∂12-Jun-86 1212 LANSKY@SRI-WARBUCKS.ARPA timberline workshop
Received: from SRI-WARBUCKS.ARPA by SU-AI.ARPA with TCP; 12 Jun 86 12:11:55 PDT
Date: Thu 12 Jun 86 12:11:31-PDT
From: Amy Lansky <LANSKY@SRI-WARBUCKS.ARPA>
Subject: timberline workshop
To: jmc@SU-AI.ARPA
Message-ID: <VAX-MM(194)+TOPSLIB(120)+PONY(0) 12-Jun-86 12:11:31.SRI-WARBUCKS.ARPA>
John,
Could you please send me your finalized plans for the workshop?
I need to finalize the rooming arrangements. Please send me a completed
version of form at the end of this message. The rest of the message
is some more detail about the workshop. You should be receiving
the final schedule next week sometime.
Thanks. I really need your response ASAP.
-Amy
←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←
Dear Workshop Attendees:
Now that the summer is rapidly approaching, we would like to
finalize our plans for the workshop. This letter is to get more
definite information from all of you about your stay at Timberline.
Please do not respond with indefinite answers -- we need to make
concrete room and food arrangements at this time.
Enclosed you will find a tentative workshop schedule as well as
more information about housing, payment, and stipends.
You will also find a yellow sheet with questions. PLEASE FILL THIS
OUT AS ACCURATELY AS POSSIBLE and return to the address given.
By the beginning of June we will be sending you a complete schedule
of talks as well as transportation details. Please feel free
to call us if you have further questions.
Amy Lansky Mike Georgeff
(415) 859-4376 (415) 859-4769
LANSKY@SRI-AI GEORGEFF@SRI-AI
Timberline Toll Free Number: (800) 547-1406
!
Arrival
-------
Attendees should fly into the Portland, Oregon airport. Timberline is
approximately a one hour drive from the airport. You may elect to
rent a car or arrange limousine service to and from Timberline. You
must make these arrangements on your own. The limousine service is
AES LIMOUSINE (telephone: (503)255-0325) and is $22.50 per person,
each way (there must be at least two passengers in the limo).
Directions to Timberline will be sent to you closer to the time of the
workshop.
Rooms
-----
We have reserved in advance 30 rooms for the nights of June 29,30 and
July 1. For those of you who want to stay on after the workshop, 10
rooms have been reserved for the nights of July 2,3,4,5. The cost of
the rooms is $61.60/per person (double occupancy), which includes 3
meals per day (starting on the night of the 29th). For those of you
bringing children (i.e. more than 2 people per room), each child is
$8.50 per day, which does NOT include meals. You will therefore have
to pay extra for meals for your children. However, children under 6
eat free.
The rooms are of a variety of types. Some have only one double or one
queen bed, some have two single beds, and the rest have a queen and a
single OR a double and a single. On the yellow form please note the
size and type of your party and exactly which nights you plan to
spend. If you are coming as a couple, a double or queen bed may be
assigned. Families with children will have priority in the rooms with
two beds. For those bringing 2 children, an extra cot can be brought
into your room. For those coming singly, double occupancy will be the
rule (obviously, only in those rooms with 2 beds!). If you have a
roommate preference, please indicate on the yellow form.
The 10 rooms for July 2-5 will be assigned first-come-first-serve.
If we need more rooms, we will try to make arrangements for those nights.
For those of you wanting to stay extra nights (besides June 29--July 5),
please make separate arrangements DIRECTLY WITH TIMBERLINE. Their toll
free number is (800) 547-1406. (One room has already been reserved
for Tate and Drummond for June 27-28). When you call, you should
indicate that you are part of the SRI workshop. (They may have you
talk to the sales office rather than the regular reservations desk).
Food
----
On the evening of your arrival (June 29), dinner will be served in
the public dining room (the Cascade Room). It will be open seating
and a limited menu. After that, breakfast, lunch, and dinner (up through
lunch on July 2) will be served in a private dining room. Only a fixed
menu will be available. If you have special food restrictions,
please advise on the yellow form.
Payment
-------
Each of you will be responsible for paying your full bill upon checking
out of Timberline. Those of you receiving stipends will receive a check
separately, which you can apply to your costs as you desire.
Stipends
--------
We have only a limited source of support for this workshop.
Therefore, we will be awarding stipends only for those who need them.
The approximate award will be $900 for those coming from England or
Australia, $400 for East Coasters or other non-West Coast origins, and
$200 for West Coasters. Our aim is to at least cover airfare costs.
Obviously only a limited number of these awards can be made, so please
indicate on the form the severity of your need!
!
TENTATIVE WORKSHOP SCHEDULE
---------------------------
Sunday June 29:
People arrive in the afternoon and evening
Dinner in the Cascade Dining Room -- public dining
Monday June 30:
Breakfast: 8:00-9:00
Session I: 9:00-10:30 (2 papers)
Coffee Break: 10:30-10:45
Session II: 10:45-12:15 (2 papers)
Lunch: 12:15-1:30
Session III: 1:30-3:45 (3 papers)
Discussion 4:00-5:00
Bar: 6:30-7:30
Dinner: 7:30-9
Tuesday July 1:
Breakfast: 8:00-9:00
Session IV: 9:00-10:30 (2 papers)
Coffee Break: 10:30-10:45
Session V: 10:45-12:15 (2 papers)
Lunch: 12:15-1:30
Session VI: 1:30-3:00 (2 papers)
Bar: 6:30-7:30
Dinner: 7:30-8:30
Discussion: 8:30-10
Wednesday July 2:
Breakfast: 8:30-9:30
Session VII: 9:30-11:00 (2 papers)
Coffee Break: 11:00-11:15
Session VIII: 11:15-12:45 (2 papers)
Lunch: 12:45-2
!
PLEASE RETURN THIS FORM ASAP TO:
PLANNING + ACTION WORKSHOP
MARGARET OLENDER
SRI INTERNATIONAL, AI CENTER
333 RAVENSWOOD AVE.
MENLO PARK, CA 94025
Please circle which nights you want your room:
June July
29 30 1 2 3 4 5
How many people in your party?
Indicate form of your party:
(ex. couple, couple with 2 children, 2 singles, etc.)
For singles, roommate preference (optional):
Dietary Restrictions :
Stipend needs:
(Please break down into airfare, housing, transportation.
An attempt will be made to at least cover airfare costs.)
-------
∂12-Jun-86 1215 SJM sjm's movements
I am going off to get some lunch. Back fairly soon. Mary Fischer
called and told me that the Domesday Book shows that Peeping Tom had
a brother called Tittering Philip.
Susie
∂12-Jun-86 1322 CL.SHANKAR@R20.UTEXAS.EDU Re: possible research associate position
Received: from R20.UTEXAS.EDU by SU-AI.ARPA with TCP; 12 Jun 86 13:22:23 PDT
Date: Thu 12 Jun 86 15:22:34-CDT
From: CL.SHANKAR@R20.UTEXAS.EDU
Subject: Re: possible research associate position
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Wed 11 Jun 86 18:52:00-CDT
Message-ID: <12214297952.13.CL.SHANKAR@R20.UTEXAS.EDU>
I would definitely be interested. It would be a wonderful opportunity.
I can visit there some time in July after my dissertation has been
completed. I can get in touch with Ketonen directly and work things
out.
Regards,
Shankar
-------
∂12-Jun-86 1427 RWW
John try this!! If it doesn't work let me know
"nttlab!nttyrl!nobuo"@Shasta
Richard
∂12-Jun-86 1418 LES EBOS progress
To: JMC
CC: CLT, GRP
We received confirmation from Bob Eustis that the RAMUS review committee
will provide one IBM RT, with a prospect of also hanging onto the one we
already have. A copy of the response is in JMC's mailbox. We need to
tell Bob Semans exactly what configuration we want.
Dr. Calo of IBM called to say that they were reviewing our informal
proposal. I mentioned the fact that the money would go farther (i.e. the
overhead could be avoided) if the money were provided in the form of a
"no-strings" grant. He said that might be interesting and wanted to know
about the probability of its actually being spent in the intended way.
I said that if the grant were made directly to JMC that the probability
was extremely high.
Calo said that 1987 money was easier to program than 1986 money. I suggested
that they give us some now and some more in January (they work on a calendar
year budget). He asked if a September 1 start would work rather than
July 1 as we had requested. I said that we were currently funding the work
out of departmental funds and that sooner was better, but that we could
probably survive until then if necessary.
∂12-Jun-86 1446 LES Alliant
To: JMC
CC: CLT, JJW, RPG
I talked to Bob Nikora of Alliant on Tuesday about our interest in the
concurrency instructions. He said that they now have a new C compiler
that includes concurrency constructs that they will make available.
He also now says that he will deliver descriptions of the hardware
concurrency features tomorrow (Friday) provided that we will sign their
nondisclosure agreement.
It just dawned on me that I had not budgeted for ethernet-datalink
gateways between here and Lucid. It turns out that we can tap into an
existing gateway at this end, but maybe not at the Lucid end. I sent
a message to RPG about that. If we need one at that end, it will cost
about $6k according to Len.
∂12-Jun-86 1548 john@lbl-csam.ARPA mail probably intended for you
Received: from LBL-CSAM.ARPA by SU-AI.ARPA with TCP; 12 Jun 86 15:48:17 PDT
Received: by lbl-csam.ARPA (4.24/Experimental)
id AA07229; Thu, 12 Jun 86 15:47:49 pdt
Date: Thu, 12 Jun 86 15:47:49 pdt
From: john@lbl-csam.ARPA (John Mccarthy [csam])
Message-Id: <8606122247.AA07229@lbl-csam.ARPA>
To: jmc@su-ai.arpa
Subject: mail probably intended for you
From OTA@USC-ISI.ARPA Wed Jun 11 09:20:05 1986
Return-Path: <OTA@USC-ISI.ARPA>
Received: from USC-ISI.ARPA by lbl-csam.ARPA (4.24/Experimental)
id AA13709; Wed, 11 Jun 86 09:19:55 pdt
Message-Id: <8606111619.AA13709@lbl-csam.ARPA>
Date: 11 Jun 1986 12:16:44 EDT
Subject: Congressional Case Study of AI
From: Earl Dowdy <OTA@USC-ISI.ARPA>
To: john@lbl-csam.ARPA
Cc: ota-cit@USC-ISI.ARPA
Status: RO
Dear Prof. McCarthy:
The Office of Technology Assessment, U.S. Congress, has been
asked by the House Science and Technology Committee to examine
the field of artificial intelligence as part of our ongoing
series of case studies in different areas of technology. In
particular, the committee is interested in assessing the adequacy
of the Federal role in AI, and the general health of AI R&D in
the United States. You may be familiar with our previous case
study of artificial intelligence that was a part of OTA's report,
Information Technology Research and Development, released
in February, 1985. That case study and other OTA work in AI
will serve as background for this effort.
We're contacting you because as one of the research tools for
this case study, we are going to hold an informal discussion via
ARPANET with AI researchers who have good insights about the
field. Our plan is to send participants in this discussion about
one question per week for about eight weeks, beginning in mid-
June. Each participant would respond as often (and in as much
volume) as he or she felt was appropriate; we would also
occasionally share the responses with others in order to get
feedback and contrasting views. We will of course honor any
requests from participants that a particular comment be kept
private or not quoted by name. Ultimately, you will also have an
opportunity to review the final case study document.
This is a new medium for us, and we believe that it could be a
productive and interesting experience for all the participants.
Please tell us if you would not like to participate; otherwise we
will begin sending our questions in the next week or so. Also
please send along any suggestions you have about others who might
be good participants or topics you think should be explored.
We will confirm this invitation by mail and send along some
additional information.
Thank you.
Earl Dowdy (ota@usc-isi), and Jim Dray (ota-cit@usc-isi)
Office of Technology Assessment
United States Congress
Washington, D.C. 20510
Phone: (202) 226-3953
-------
∂12-Jun-86 1649 CLT
don't forget to pick up the pictures you left for copies
∂12-Jun-86 1747 @MIT-MULTICS.ARPA,@UMich-MTS.Mailnet:Jeff←Pelletier@UQV-MTS
Received: from MIT-MULTICS.ARPA by SU-AI.ARPA with TCP; 12 Jun 86 17:47:06 PDT
Received: from UMich-MTS.Mailnet by MIT-MULTICS.ARPA with Mailnet id <2696459813326006@MIT-MULTICS.ARPA>; 12 Jun 1986 20:36:53 edt
Received: from UQV-MTS by UMich-MTS.Mailnet via MTS-Net; Thu, 12 Jun 86 20:00:58 EDT
Date: Thu, 12 Jun 86 17:11:48 MDT
From: Jeff←Pelletier%UQV-MTS%UMich-MTS.Mailnet@MIT-MULTICS.ARPA
To: jmc@sail.arpa
Message-ID: <625849@UQV-MTS>
Dear Dr. McCarthy,
Ivan Sag suggested that I contact you about the possibility
of AAAI giving a grant to support a research group at next
year's Summer Linguistic Institute. The topic of the group
would be a study of the syntax/semantics/computational problems
involved with generic statements, such as Birds fly, Dogs are
loyal, Dogs are barking, etc. As you will easily recognize,
this area has received rather a lot of attention both from the
computational and the linguistic sides ("default reasoning",
"truth maintenence", etc.). The idea of the group is to gather
a group of leading researchers in the area and try to hammer out
a consensus. I have been elected (by the group) to try to
figure out a way to support those who need it. Sag thinks that
the Summer Institute might waive all fees, etc., and give us space.
This leaves only expenses plus (for some) salary replacement.
Before going on with the proposal and so forth, I thought I'd
just ask you whether there is any principle that would automatically
reject such a request from AAAI. The people who I have contacted
about joining in this group are:
Jeff Pelletier (myself): CompSci & Philosophy, U of Alberta
Greg Carlson: Linguistics, U of Iowa
Godehart Link: CSLI & Munich
Alice ter Meulen: Linguistics, U Washington
Erhard Hinrichs: BBN
Angelika Kratzer: Linguistics, U Mass
Gennaro Chierchia: Linguistics, Cornell
I hope this message gets to you--I've had difficulties from Alberta
before. I look forward to hearing from you.
Sincerely,
Francis Jeffry Pelletier
∂12-Jun-86 1828 "Lwd at vanderbilt%vanderbilt.csnet"@CSNET-RELAY.ARPA Why no AI?
Received: from CSNET-RELAY.ARPA by SU-AI.ARPA with TCP; 12 Jun 86 18:28:08 PDT
Received: from vanderbilt by csnet-relay.csnet id a015250; 12 Jun 86 21:16 EDT
Date: 12 Jun 86 10:30:16-CDT (Thu)
From: "Lwd at vanderbilt"@CSNET-RELAY.ARPA
To: "JMC%su-ai.arpa at Csnet-Relay"@CSNET-RELAY.ARPA
cc: "lwd at Vanderbilt"@CSNET-RELAY.ARPA
Subject: Why no AI?
John,
There is AI. The classification I gave was straight
from the new computing reviews classification tree.
AI is classified under section I (i.e., computing
methodologies). More specifically, subsection I.2 is
entitled "Artificial Intelligence". I apologize that
I didn't make things more clear, but I couldn't include
the entire CR tree for space reasons so I just gave
the major section titles.
Larry Dowdy
∂12-Jun-86 2000 SJM
finish with objectivity
∂13-Jun-86 0700 JMC
Emilio J. Nunez, Shell Houston (office: 713 493-8866) (home: 556-8840)
Norman Heitkamp, admin. 713 493-8364
∂13-Jun-86 0904 CSL.GERLACH@su-sierra.arpa re: CSL fac cand--MacQueen
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 13 Jun 86 09:04:12 PDT
Date: Fri 13 Jun 86 09:03:45-PDT
From: Sharon Gerlach <CSL.GERLACH@su-sierra.arpa>
Subject: re: CSL fac cand--MacQueen
To: JMC@su-ai.arpa
cc: csl.gerlach@su-sierra.arpa
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Tue 10 Jun 86 14:57:00-PDT
Message-ID: <12214512981.21.CSL.GERLACH@su-sierra.arpa>
I've put you down for 4:00 on 6/23 and will send MacQueen to your office
(is that MJH 356?). Would you also be interested in lunch or dinner with him, we have openings for both?
Sharon
-------
∂13-Jun-86 0958 RPG Concurrency Features
To: LES, JMC, CLT, JJW
Hm. Nikura is coming over to Lucid today to xerox our copy of the concurrency
instruction manual. I guess that's where it's coming from!
-rpg-
∂13-Jun-86 1124 VAL Przymusinski's paper
I agree that the results are very interesting. It seems that his semantics is
applicable to the program for even numbers that we discussed in Austin; I'll
ask him about it and suggest that he include that example in the paper. One
strange thing about his definition is that his definition of having a higher
priority isn't transitive. But maybe my reluctance to admit this kind of
non-transitivity was one of the reasons why I didn't know what to do with the
even numbers example.
∂13-Jun-86 1258 CLT japan collaboration -- time extension
To: JMC@SU-AI.ARPA, sf@SU-CSLI.ARPA, RWW@SU-AI.ARPA, JK@SU-AI.ARPA,
CG@SU-AI.ARPA, CLT@SU-AI.ARPA
We sent a letter to the NSF U.S.-Japanese Cooperative Science Program
requesting a no cost extension to Sept 88 (current expiration date is Sept
87). We were told that we can get a six month extension by applying to
the "Grants Research Office" here at Stanford, but betting an extension
beyond 1988 March 31 is probably impossible because of the agreement
between the U.S. and Japan.
∂13-Jun-86 1606 AAAI-OFFICE@SUMEX-AIM.ARPA Strategic Planning Meeting
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 13 Jun 86 16:06:49 PDT
Date: Fri 13 Jun 86 16:04:23-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: Strategic Planning Meeting
To: JMC@SU-AI.ARPA, Nilsson@SU-SCORE.ARPA, Minsky@MC.LCS.MIT.EDU,
Feigenbaum@SUMEX-AIM.ARPA, ai.woody@MCC.ARPA, Newell@A.CS.CMU.EDU,
MCK%mit-oz@MC.LCS.MIT.EDU, Reddy@A.CS.CMU.EDU
cc: aaai-OFFICE@SUMEX-AIM.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Message-ID: <12214589554.41.AAAI-OFFICE@SUMEX-AIM.ARPA>
Pat and Raj have suggested that ex-presidents of the AAAI meet to
discuss long-term directives for the association during the
conference. They have suggested Monday, August 11, from 1:00-4:00 pm
in the Franklin Plaza Hotel in Salon 10 (we'll serve lunch).
If you can attend the meeting, please rsvp.
Thank you for your attention to this matter.Claudia
-------
∂13-Jun-86 1641 VAL reply to message
[In reply to message rcvd 13-Jun-86 16:35-PT.]
OK. I'll referee the paper, if you'd like me to (and if you're sure Brady
doesn't mind: I understand that the "primary" referee is supposed to be a member
of the board). Also, I assume that this is not a "research note", for which they
want reports very quickly.
∂13-Jun-86 1906 LES Alliant etc.
To: JMC
CC: CLT, JJW, RPG
John Pucci will not be coming June 24 after all. Perhaps two or three
weeks later.
Bob Nikora delivered the concurrency instruction documentation today
(having copied it at Lucid) and did not ask me to sign a nondisclosure
agreement. Nevertheless, I had agreed to treat this material as
proprietary so we should do so. It is in JJW's hands.
I asked what their purpose was in trying to keep this material
confidential, pointing out the potential conflict with our desire to
publish results. I noted that if they are protecting their patent rights,
they need not guard the information after the patent application is filed.
Nikora said he didn't know and referred me to Dave Rolm in Acton.
I called Rolm and asked the same question. He said he didn't know but
would find out early next week. He said that Jack Test would be back then
and is planning to visit us soon. He also said that Jack would be filled
in on this issue and authorized to find a solution.
∂13-Jun-86 1940 OTA-CIT@USC-ISI.ARPA Re: [John McCarthy <JMC@SU-AI.ARPA>: misdirected mail from Earl Dowdy and Jim Dray ]
Received: from USC-ISI.ARPA by SU-AI.ARPA with TCP; 13 Jun 86 19:40:28 PDT
Date: 13 Jun 1986 22:35:25 EDT
Subject: Re: [John McCarthy <JMC@SU-AI.ARPA>: misdirected mail from Earl Dowdy and Jim Dray ]
From: Jim Dray <OTA-CIT@USC-ISI.ARPA>
To: Jim Dray <OTA-CIT@USC-ISI.ARPA>, jmc@SU-AI.ARPA
cc: ota@USC-ISI.ARPA
In-Reply-To: (Message from "Jim Dray <OTA-CIT@USC-ISI.ARPA>" of 13 Jun 1986 18:55:03 EDT)
Sorry about the wrong address, and thank you for being willing to
participate. We'll begin the conference within the next week.
Incidentally, we are particularly interested in additional women participants
or researchers who are outside the U.S. (but reachable through the network).
If you have any suggestions, we'd be very appreciative.
Thanks,
Jim Dray
-------
∂14-Jun-86 1613 HK.WFM@Lindy NEXT
Received: from SU-LINDY.ARPA by SU-AI.ARPA with TCP; 14 Jun 86 16:12:57 PDT
Date: Sat, 14 Jun 86 16:12:32 PDT
From: BILL MASSY <HK.WFM@SU-Forsythe.ARPA>
To: JMC@SU-AI
Subject: NEXT
Dear Nils, John, and Dick:
I'm sorry to tell you that we have not been able to consumate a deal
with Steve Jobs to invest in NEXT, Inc.. There is good news, however:
we parted the best of friends and there seems to be no loss of interest
on Steve's part in owrking with Stanford on the technical and
applications side.
The problem came down to money -- specifically, the price at which
Steve was willing to offer and we were willing to buy a share of the
company. As you know, Steve's valuation was $40 million. We enlisted
the aid of Pitch Johnson, a preeminant venture capitalist and good
friend of Stanford, and he and Rod Adams came up with a value of $20 to
$24 million. We learned later that Harvard, working through completely
independent channels, came p with $20 to $25 million.) We suggested $24
million to Steve who, after consideration, decided not to proceed
further. All other questions (e.g. the buyout provision) had been
worked out successfully priior to that point, so it was purely a matter
of being too far apart on the price.
Harvard and MIT are following our lead, and I have a call in to Dick
Cyert of CMU who I think will do likewise. Steve knows we have been in
touch with one another, and indeed wanted to negotiate with us on
behalf of the group. I don't think any other universities are in the
picture.Steve had nothing but good things to say about Pitch, and I'm
told he feels the same about Rod. He understood where we were coming
from, and respected our position even though not agreeing with it.
We're looking forward to staying in touch.
Thank you again for doing the technical evaluation, and also for your
comments about valuation (which were right on the mark). I hope the
time will prove to have been worthwhile in terms of your future
relations with Steve and NEXT. If there's anything Rod or I can do to
help, please let us know right away.
All the best... Bill
cc. Don Kennedy, Jim Rosse, Bob Street, Rod Adams
To: RPG@SU-AI.ARPA, JMC@SU-AI.ARPA, NILLSON@SU-SCORE
cc: HK.DXK, RLS(S.STREET@LOTS-A), HK.RHA
∂14-Jun-86 1625 HK.WFM@Lindy re: NEXT
Received: from SU-LINDY.ARPA by SU-AI.ARPA with TCP; 14 Jun 86 16:25:04 PDT
Date: Sat, 14 Jun 86 16:24:43 PDT
From: BILL MASSY <HK.WFM@SU-Forsythe.ARPA>
To: JMC@SU-AI
Subject: re: NEXT
REPLY TO 06/14/86 16:22 FROM JMC@SU-AI.ARPA: re: NEXT
Dear John... Glad the visit was interesting.. No, the
mailer didn't trap my mistyping -- I'll send it again just to be
sure. All the best... Bill
To: JMC@SU-AI.ARPA
∂15-Jun-86 2000 SJM
objectivity should be complete
∂16-Jun-86 0004 LES re: Interpress
[In reply to message rcvd 14-Jun-86 13:56-PT.]
You are not mistaken. It was designed by a committee that was integrating
a number of existing standards. Fortunately, thatt doesn't matter much --
you can organize the symbols any way you want for presentation as long as you
keep the codes straight.
∂16-Jun-86 0600 JMC
Phone Graubard about other essays. 491-2600
∂16-Jun-86 0911 BSCOTT@SU-SCORE.ARPA Summer Supplemental Research Salaries
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 16 Jun 86 09:11:15 PDT
Date: Mon 16 Jun 86 09:09:52-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Summer Supplemental Research Salaries
To: Cheriton@SU-PESCADERO.ARPA, Golub@SU-SCORE.ARPA, Lantz@SU-SCORE.ARPA,
ZM@SU-AI.ARPA, Mayr@SU-SCORE.ARPA, JMC@SU-AI.ARPA, Oliger@SU-SCORE.ARPA,
Ullman@SU-SCORE.ARPA, Winograd@SU-CSLI.ARPA, Yao@SU-SCORE.ARPA
cc: BScott@SU-SCORE.ARPA
Message-ID: <12215300526.18.BSCOTT@SU-SCORE.ARPA>
Over the next couple of weeks we'll be processing summer research salaries.
Will appreciate any comments you wish to make about how you want your summer
salaries charged; otherwise, I'll assume that you want them charged as
originally planned in your proposal budgets.
Betty
-------
∂16-Jun-86 0914 CLT Cate
I called cate and had him transfer money to the Keogh
to make the tax form legal. You need to send him
a note formally requesting the transfer for there records.
The amount was $6301.
∂16-Jun-86 0929 CSL.GERLACH@su-sierra.arpa [dbm.sola%btl.csnet@CSNET-RELAY.ARPA: Re: Abstract]
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 16 Jun 86 09:29:34 PDT
Date: Mon 16 Jun 86 09:26:13-PDT
From: Sharon Gerlach <CSL.GERLACH@su-sierra.arpa>
Subject: [dbm.sola%btl.csnet@CSNET-RELAY.ARPA: Re: Abstract]
To: jmc@su-ai.arpa
cc: csl.gerlach@su-sierra.arpa
Message-ID: <12215303504.19.CSL.GERLACH@SU-SIERRA.ARPA>
Here's Macqueen's abstract fyi...
---------------
Return-Path: <dbm.sola%btl.csnet@CSNET-RELAY.ARPA>
Received: from CSNET-RELAY.ARPA by SU-SIERRA.ARPA with TCP; Fri 13 Jun 86 15:28:06-PDT
Received: from btl by csnet-relay.csnet id ae00965; 13 Jun 86 18:20 EDT
To: CSL.GERLACH%su-sierra.arpa%CSNET-RELAY@CSNET-RELAY.ARPA
From: dbm.sola%btl.csnet@CSNET-RELAY.ARPA
Date: Fri 13 Jun EDT 1986 14:38
Original-From: sola!dbm
Original-To: Sharon Gerlach <CSL.GERLACH%su-sierra.arpa@CSNET-RELAY>
Subject: Re: Abstract
Here is a title and abstract for my talk on June 23:
TITLE: "Using dependent types to express modular structure"
ABSTRACT
Several related typed languages for modular programming and data
abstraction have been proposed recently, including Pebble, SOL, and
ML modules. We review and compare the basic type-theoretic ideas
behind these languages and evaluate how they meet the pragmatic
requirements of "programming in the large".
-------
∂16-Jun-86 0936 RPG Common Lisp
Have you talked to Stoyan about Common Lisp versus EuLisp?
What are his thouhts?
-rpg-
∂16-Jun-86 1012 SJM objectivity
Two suggestions only on this essay---why not delete remark on people
getting professorships in their twenties, since it resembles a remark
made a paragraph or two earlier on the recognition of the work of
quite young people? And why not put in the flashy description you
have given me in the past of allowing programs to `jump to conclusions'?
Susie
∂16-Jun-86 1035 RA George Marotta
George was wondering whether you can have lunch with him and David Marotta
Wed. or Friday this week. His tel. is 3-1007
∂16-Jun-86 1136 RA Bob Givan
Bob Givan told me you want to appoint him for this summer. He said that you want
to pay him more than an RA; could you please tell me what you want to pay him
and what will his title be?
Thanks,
∂16-Jun-86 1152 davism@nyu-csd1.arpa
Received: from NYU-CSD1.ARPA by SU-AI.ARPA with TCP; 16 Jun 86 11:52:35 PDT
Date: Mon, 16 Jun 86 14:13:24 edt
From: davism@nyu-csd1.arpa (Martin Davis)
Message-Id: <8606161813.AA19614@nyu-csd1.arpa>
Received: by nyu-csd1.arpa; Mon, 16 Jun 86 14:13:24 edt
To: jmc@su-ai
Thanks for agreeing to read Ernie Davis' papers. I've sent them to you.
-Martin
∂16-Jun-86 1427 RA Danzig
Please call George Danzig at 3-1304.
∂16-Jun-86 1621 GIVAN@su-sushi.arpa me
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 16 Jun 86 16:20:00 PDT
Date: Mon 16 Jun 86 16:18:45-PDT
From: Bob Givan <GIVAN@su-sushi.arpa>
Subject: me
To: jmc@su-ai.arpa
Message-ID: <12215378602.14.GIVAN@su-sushi.arpa>
i will be best reached by E-mail until i have a place to work. (I may be
working in the Math/CS library, or wherever i find most productive in the
meantime)
E-mail to GIVAN @ SUSHI.
thanx,
Bob.
p.s. Les will let me know where to work by E-mail.
-------
∂16-Jun-86 1942 FAHLMAN@C.CS.CMU.EDU further development of Lisp
Received: from C.CS.CMU.EDU by SU-AI.ARPA with TCP; 16 Jun 86 19:42:04 PDT
Received: ID <FAHLMAN@C.CS.CMU.EDU>; Mon 16 Jun 86 22:41:59-EDT
Date: Mon, 16 Jun 1986 22:41 EDT
Message-ID: <FAHLMAN.12215415589.BABYL@C.CS.CMU.EDU>
Sender: FAHLMAN@C.CS.CMU.EDU
From: "Scott E. Fahlman" <Fahlman@C.CS.CMU.EDU>
To: John McCarthy <JMC@SU-AI.ARPA>
Cc: cl-steering@SU-AI.ARPA
Subject: further development of Lisp
In-reply-to: Msg of 16 Jun 1986 14:40-EDT from John McCarthy <JMC at SU-AI.ARPA>
I agree that standardizing Common Lisp and developing new ideas for Lisp
are distinct activities, both worthy. I wish the Europeans agreed. As
Weinreb has pointed out, they seem to muddle together the ideas of
"standard" and "perfect".
I think that their effort would go much better if they were to forget
all about the existence of Common Lisp. They seem to want to build on
Scheme, but are making some concessions in the direction of Common Lisp
(but not the right ones from our point of view).
I think that the Europeans are afraid that they will come up with a
shiny new Lisp, and that this will be ignored by everyone else. They're
probably right. So the plan is to make it THE Lisp standard at ISO, and
thereby to force everyone to use their dialect. Standards organizations
don't work this way -- unpopular standards have always been ignored.
I think that the creation of an official standard for Common Lisp is
important (though not essential). We need some mechanism for resolving
the ambiguities in CLtL and for fixing a few glaring problems, and a
standards effort is one good way of organizing this. If everyone
beleives that we're headed toward a standard and then we get blocked at
the end, I think that people will tend to follow what we've done.
-- Scott
∂16-Jun-86 2250 hector%utai.toronto.edu@CSNET-RELAY.ARPA A proposal
Received: from CSNET-RELAY.ARPA by SU-AI.ARPA with TCP; 16 Jun 86 22:50:36 PDT
Received: from toronto by csnet-relay.csnet id au08611; 17 Jun 86 1:38 EDT
Received: from utai.uucp by utcsri.uucp id AA12026; Fri, 13 Jun 86 10:51:58 edt
Received: by utai.uucp id AA03528; Fri, 13 Jun 86 10:41:01 edt
Message-Id: <8606131441.AA03528@utai.uucp>
Received: by utai.uucp id AA03525; Fri, 13 Jun 86 10:40:53 edt
Date: 13 Jun 86 10:40:50 EDT (Fri)
From: Hector Levesque <hector%utai.toronto.edu@CSNET-RELAY.ARPA>
To: jmc@SU-AI.ARPA
Subject: A proposal
Dear fellow researcher,
As you may already know, Drew McDermott has recently written a paper called "A
Critique of Pure Reason" where he explains why he thinks the "logicist"
programme in AI (as in the work of Hayes, McCarthy, Moore etc.) is bound to
fail. The conclusion is certainly not original, and you probably consider it
obviously correct, or maybe obviously incorrect. Either way, it's obvious, and
you've no doubt lost interest in the whole issue by now, and want to get on
with your work. What makes Drew's paper more than just the trouncing of some
plainly defunct equine specimen is that he has very clearly changed his mind
after a decade or so of work within the logicist programme (including research
papers, a textbook, and methodology papers insisting that this was the *only*
way to go). Right or wrong, he has unquestionably amassed considerable
experience in actually doing AI, and his arguments are street-smart in a way
that so much armchair philosophizing about AI simply is not.
So what does this have to do with you? I will be guest editing a special
issue of the journal COMPUTATIONAL INTELLIGENCE that will be a public forum
devoted to the issues raised by this paper. The model we will follow is that
of The Behavioral and Brain Sciences: the paper in question, followed by
commentary from two dozen or so luminaries in the field, and a final reply and
summary from Drew himself. He has agreed to all of this, so the only thing
left is to find suitable luminaries willing to illuminate. Like you, say.
I can guess what you're starting to think at this point (i.e. something like
"Groan!"), but let me try to convince you. First, I know you already have
opinions in this area, as you have expressed some of them publicly before.
And it's not that big a job, really. All I would need from you is a thousand
words (or less), and you'll have lots of time: a couple of months after
receiving the final draft of the paper (which should be ready in a month or
so). Finally, I would encourage you to share the modest workload involved
with a like-minded colleague or student. At best, you'll just send me an
electronic mail message with your commentary and I'll do the rest. With luck,
by the end of the year, we should have something of interest to a wide variety
of AI researchers. So if you feel (as I do) that scientific discussions are
impossible at our AI mega-conferences, and cringe at the thought of the AIList
as our only public forum, then now's the time to change all this. If nothing
else, it should be fun (I am grateful to Pat Hayes for this last observation).
Obviously this project won't get off the ground if I can't solicit suitable
commentary. I'm not asking you today if you have such commentary; you may not
have even seen the paper. What I'm asking is whether or not you're willing to
make an honest effort in this direction once I send you the final draft.
Ultimately, it will be my responsibility to select for publication and edit
those commentaries that I find the most interesting, all the while minimizing
redundancy, irrelevancy, ad hominem remarks, and the like. So I really can't
even guarantee that what you send will appear in print since, for instance, I
assured Drew that only a small number of "Nyah! Nyah! I told you so!!" would
actually be published.
Since many of you are probably away for at least part of the summer, I need to
know who exactly received this message so please PLEASE *PLEASE* answer this
immediately. (I'll have to track down by phone those of you who do not
answer.) To make it even easier for you, feel free to use the following codes:
A -- Got your message. Don't know what to think. Will send another
message within a couple of days.
B -- I'm tempted. But I absolutely must see a draft of the paper before
I decide. Can't get one around here. Please mail me a xerox.
C -- Sounds good to me. I'm not sure I'll have anything very relevant
to say, but I'm willing to give it the old college try.
D -- I don't think I want to get involved. If you're really stuck, I
suppose I could be convinced. But for now, the answer is no.
E -- Sorry, no. I have many excuses, but you probably don't want
to hear even the best of them.
If your response is other than E please also send me a phone number, a
preferred electronic address, and a regular address for future use. My
addresses are (don't trust "reply" on your mail system):
ARPA: hector%toronto.csnet@CSNET-RELAY
CSNET: hector@toronto
UUCP: ...{utai,utcsri}!hector
REGULAR: Dept. of Computer Science
University of Toronto
Toronto, Ont. M5S 1A4
CANADA.
PHONE: (416) 978-3618
Thanks for your attention and I await your reply.
Hector Levesque
∂17-Jun-86 0859 DLW@ELEPHANT-BUTTE.SCRC.Symbolics.COM further development of Lisp
Received: from [192.10.41.41] by SU-AI.ARPA with TCP; 17 Jun 86 08:58:53 PDT
Received: from CHICOPEE.SCRC.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 23317; Tue 17-Jun-86 10:59:14 EDT
Date: Tue, 17 Jun 86 11:01 EDT
From: Daniel L. Weinreb <DLW@QUABBIN.SCRC.Symbolics.COM>
Subject: further development of Lisp
To: Fahlman@CMU-CS-C.ARPA, JMC@SU-AI.ARPA
cc: cl-steering@SU-AI.ARPA
In-Reply-To: <FAHLMAN.12215415589.BABYL@C.CS.CMU.EDU>
Message-ID: <860617110152.9.DLW@CHICOPEE.SCRC.Symbolics.COM>
Date: Mon, 16 Jun 1986 22:41 EDT
From: "Scott E. Fahlman" <Fahlman@C.CS.CMU.EDU>
I think that the creation of an official standard for Common Lisp is
important (though not essential). We need some mechanism for resolving
the ambiguities in CLtL and for fixing a few glaring problems, and a
standards effort is one good way of organizing this. If everyone
beleives that we're headed toward a standard and then we get blocked at
the end, I think that people will tend to follow what we've done.
I agree. Scott, I am extremely encouraged to see how much initiative
you're taking and how much work you're putting in. I'd really hate to
see you discouraged or exhausted by political battles within ISO. In
fact, anticipation of such problems is exactly the reason we originally
felt we should not get involved with ISO in the first place. I would
much rather see your energy, skill, and experience directed towards
pulling together the mechanisms for resolving ambiguities in CLtL, than
see you get so put off by all the politics that you would rather have
nothing to do with CL any more! So, I respectfully recommend that you
give the ISO/Europe problem a reasonable amount of attention but be
careful not to let it get overwhelming.
In fact, if a good, working process emerges for fixing the ambiguities,
that will further increase people's confidence in the CL standard. In
the long run, that kind of confidence will lead to the kind of
popularity that will make the standard a success.
∂17-Jun-86 0921 VAL re: A proposal
[In reply to message rcvd 16-Jun-86 23:07-PT.]
I wonder if Waldinger has been invited to take part in the project. He has
a very definite refutation of McDermott's criticism of the logic approach to
planning (in connection with the "Bomb in the Toilet" example). I've sent you
a copy of his message to Waldinger on that subject.
∂17-Jun-86 1000 JMC
Cate about III.
∂17-Jun-86 1059 GRP Reminder: meeting today at 2pm
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
∂17-Jun-86 1439 RA George Morotta
George will meet you at the Faculty Club tomorrow at noon.
∂17-Jun-86 1518 RA reservation
I made reservation at the Faculty Club, under Morotta, for 3 people, noon, tomorrow.
∂17-Jun-86 1538 CLT alliant
To: JMC, LES, JJW, RPG
I talked to Jack Test this afternoon.
He plans to be here for a day sometime in the next two weeks -
tentatively Thursday June 26. Test will let me know definitely
later this week.
As to proprietariness of the concurrency instructions --
they have filed for a patent. When (if) that is granted
the instructions will be publishable. Meanwhile they
feel these instructions are unique in the market place and wish
to protect their product. For our purposes, Test says he
and RPG discussed what would be needed for Qlisp and
only the cstart and synchronization instructions are likely
to be needed. They are not protective of these instructions
so there is probably no problem.
About the OS, the new version is now out to beta test sites and Test is
trying to get us included as one.
The new version is claimed to have various fixes and small
improvements to the c-compiler. He will check to see about
the network problems we have had and whether the fixes are
incorporated. Berkeley has released UNIX4.3 with new (hopefully
greatly improved network code). It will be the basis of the
next Alliant OS release (end of the year).
They are working on a ``from scratch'' c-compiler that will
be a big improvement over the existing one. The claim is that
it will be available in 6 to 9 months.
∂17-Jun-86 1703 LES Alliant acceptance
To: JMC
CC: CLT, JJW
I have the Alliant acceptance report. It appears to me that they have
delivered a working system and should be paid. Any dissent?
∂17-Jun-86 1856 CWR
[cwr]
Professor McCarthy,
Would it be alright with you if I stop in to see you
tomorrow and ask for some advice about how to approach
a situation I've been chewing over?
∂17-Jun-86 2108 CLT Alliant acceptance
To: LES
CC: JMC, JJW
no dissent here
∂18-Jun-86 0700 JMC
car
∂18-Jun-86 1142 RA leaving for lunch
I am leaving for lunch; will be back around 1:30.
∂18-Jun-86 1327 SJM reservations
It's about time for you to make reservations for going to the U.K.
(You asked me to remind you.)
Susie
∂18-Jun-86 1659 RA Urgent from John Kwapisz
Please call John Kwapisz, Science and Engineering for Secure World,
(202) 547 5607 today. It is urgent.
∂18-Jun-86 1728 CLANCEY@SUMEX-AIM.ARPA Re: two proposed workshops
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 18 Jun 86 17:28:27 PDT
Date: Wed 18 Jun 86 17:26:34-PDT
From: William J. Clancey <CLANCEY@SUMEX-AIM.ARPA>
Subject: Re: two proposed workshops
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Tue 17 Jun 86 17:58:00-PDT
Message-ID: <12215915235.53.CLANCEY@SUMEX-AIM.ARPA>
John,
I don't have any information about Self's conference. In
education circles, "complex learning" refers to scientific and
technical reasoning, in contrast with nonsense patterns (and
perhaps simple mathematics). Therefore, Self's conference appears
to be aimed at a non-AI audience or perhaps a cognitive science
audience, and not specifically people who are writing programs.
Self's April conference was very well run. The work reported by
the British was disappointing, but the group was small enough for a good
dialogue. American work was well-represented. I think our good showing
earned us some points and might have a favorable influence on British
work. They have always been active in educational applications of
computing. Self chose his invited speakers well. There were very good
people from the Netherlands, Germany, and Italy.
I would need to see an abstract of Self's new conference to pass
judgement. In general, I think it was a good use of AAAI money to send
people to England for his last conference, but it is not clear that
"complex learning" fits under the AAAI charter.
I have been invited to speak at Bonar's conference and have
some information. It pertains specifically to tutoring programs, so fits
within AI interests. The emphasis is on requirements beyond knowledge
representation, including: learning, pedagogy, modeling, interface, and
evaluation. Thus, it is quite broad. I think Bonar and Ohlsson are trying
to discourage problem area descriptions and shift the focus to the problem
of teaching. The list of invited speakers is impressive. Note that this
conference and ones like it have only been held in England in the past.
This is shaping up to be the most important AI & Education conference
in the US to date. The Pittsburgh location is perfect; it has the support
of Glaser (LRDC-U Pitt), Larkin, and JS Brown. Bonar is not senior, but
is competent. Ohlsson is sharp. It's an easy decision to fund this
proposal, rather than Self's if you can only choose one.
I'd recommend funding Bonar and asking Self for more information.
Bill
-------
∂18-Jun-86 2000 SJM
restful space
∂18-Jun-86 2042 GENESERETH@SUMEX-AIM.ARPA
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 18 Jun 86 20:42:11 PDT
Date: Wed 18 Jun 86 20:40:27-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Wed 18 Jun 86 00:34:00-PDT
Message-ID: <12215950532.18.GENESERETH@SUMEX-AIM.ARPA>
John,
Yes after a little prodding I got the check for the full amount,
as we discussed. You should probably get you rsecretary to call Garo.
mrg
-------
∂18-Jun-86 2147 SF@SU-CSLI.ARPA Re: possible Freiling visit
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 18 Jun 86 21:46:59 PDT
Date: Wed 18 Jun 86 21:45:50-PDT
From: Sol Feferman <SF@SU-CSLI.ARPA>
Subject: Re: possible Freiling visit
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Wed 18 Jun 86 00:09:00-PDT
My regular seminar is not running anymore, but we have Logic Lunch, once
a week on Mondays, which is a good place for more casual presentations.
(Actually, Ian Mason gave a talk about Freilings JSL article at one of
these lunches.) I'd be happy to have him talk to this group. Currently
the next two slots are spoken for. Of course we could also schedule a
special meeting some time. Do you have funds to bring him?
-------
∂19-Jun-86 0605 CLT japan collaboration
To: JMC@SU-AI.ARPA, sf@SU-CSLI.ARPA, RWW@SU-AI.ARPA, JK@SU-AI.ARPA,
CG@SU-AI.ARPA, CLT@SU-AI.ARPA
Mr. Nakahara and Mr. Hagiya will visit Stanford from July 1 to August 31.
Sato will be here from July 15 to August 21.
We should probably have them give seminars fairly soon, so
we know what they are interested in and can proceed from there.
I will talk to them when they arrive to arrange something.
If there are any preferences as to good or bad seminar days let me know.
∂19-Jun-86 0845 VAL blocks
I told Teodor about the example we played with yesterday. Would you mind if
he uses it in his paper?
∂19-Jun-86 0854 VAL
Do you know by any chance whether there is a summer school in AI this year
like the one at Santa Cruz where we met three years ago? A friend of mine
is interested.
∂19-Jun-86 1003 RPG Weening
To: JMC@SU-AI.ARPA, ullman@SU-SHASTA.ARPA
Are either or both of you going to Black Friday? If one of you
is, and if I can express my opinions about Weening's progress
through that one, I would prefer to make a trip on that day.
If neither is going, or if neither would like to carry my opinion,
I will attend without inconvenience. Thanks.
-rpg-
∂19-Jun-86 1039 RA Kate Kelly, Sapiens Software
Kelly called re documents she gave you; she wants to give you an update
and set a meeting with you. Her tel. (408) 458 1990.
∂19-Jun-86 1050 RA my replacement
Taleen Marashian, who generally helps Kathy Berg, will replace me while
I am gone. She is TAL@SAIL. You would need to give her access to your files.
There will be 3 days she will not be here at work: July 10, 11, and 14.
I already started showing things to her. If there is anything you want me to
do before I leave (Wed. is my last day), please let me know.
Thanks,
∂19-Jun-86 1143 SJG busy Saturday night?
Either charades, or possibly a play reading. If you (and Carolyn?)
are free and have a preference, please let me know.
Matt
∂19-Jun-86 1422 VAL Recursive negation
To: ullman@SU-AIMVAX.ARPA, avg@SU-AIMVAX.ARPA
CC: JMC@SU-AI.ARPA, "@UTEP.[NET,VAL]"@SU-AI.ARPA
I'd like to convince you that NAIL!'s restriction on the use of negation is too
strong and eliminates some reasonable programs that are handled correctly by
Prolog. Until yesterday, my only example was this rather artificial definition
of even numbers:
even(0).
even(s(N)) :- \+ even(N).
But now John McCarthy came up with this program describing the effect of
moving blocks, in the spirit of his work on applications of circumscription:
on(X,Y,r(m(X,Y),S)) :- \+ ab(X,Y,S).
ab(X,Y,S) :- on(Z,X,S).
Intuitively,
on(X,Y,S) means that block X is on block Y in situation S;
r(E,S) is the result of event E happening in situation S;
m(X,Y) is the action of moving X onto Y;
ab(X,Y,S) means that action m(X,Y) in situation S is abnormal
(i.e., doesn't lead to the normally expected result).
The first rule describes the effect of m(X,Y), the second rule says that
moving X is abnormal if X isn't clear.
Ray Reiter has observed that programs of this kind are handled by Prolog
correctly. This fact may become crucial for future work on implementing
circumscription. But this program isn't free from recursive negation.
Teodor Przymusinski generalizes the class of programs free from recursive
negation and defines the class of "locally stratified" programs, which
covers these two examples. I wonder whether the ideas of NAIL! can be
possibly applied to his generalization.
Vladimir
∂19-Jun-86 1449 CHEADLE@SU-SCORE.ARPA progress?
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 19 Jun 86 14:45:21 PDT
Date: Thu 19 Jun 86 11:26:18-PDT
From: Victoria Cheadle <CHEADLE@SU-SCORE.ARPA>
Subject: progress?
To: jjw@SU-AI.ARPA
cc: jmc@SU-AI.ARPA, ullman@SU-AIMVAX.ARPA, rpg@SU-AI.ARPA
Office: Margaret Jacks 258, 723-1519
Message-ID: <12216111795.11.CHEADLE@SU-SCORE.ARPA>
Joe,
As you know, your Gray Tuesday letter requested that you submit some
tangible work to your committee by Black Friday. Please let me
know what the status of your progress is in this matter, as the
meeting is next week.
Victoria
-------
∂19-Jun-86 1452 JJW Re: progress?
To: CHEADLE@SU-SCORE.ARPA
CC: JMC@SU-AI.ARPA, ullman@SU-AIMVAX.ARPA, RPG@SU-AI.ARPA
I am currently writing a thesis outline/draft, that should show
the state of my progress. It should be ready by early next week
and I'll give a copy to each member of my committee.
Joe
∂19-Jun-86 1703 SJM space salvage
I suggest you examine PRISPA.ESS, SPACE.ESS and SPACE1.ESS (all are
ess.jmc) for paragraphs that can be installed verbatim in SPACE2. Even
if these essays go in inappropriate directions, I bet there are parts
that could be salvaged and save you some work.
Susie
∂20-Jun-86 0126 SJM
space
∂20-Jun-86 0135 GIVAN@su-sushi.arpa locating Bob
Received: from SU-SUSHI.ARPA by SU-AI.ARPA with TCP; 19 Jun 86 14:54:35 PDT
Date: Thu 19 Jun 86 14:52:56-PDT
From: Bob Givan <GIVAN@su-sushi.arpa>
Subject: locating Bob
To: jmc@su-ai.arpa
cc: val@su-ai.arpa
Message-ID: <12216149411.14.GIVAN@su-sushi.arpa>
here is what anyone with a normal complement of common sense axioms and
inference/conjecture rules needs to know in order to reach me:
office: Room 402 (back right corner)(normally John Mohammed's)
office phone: 723-3605
home phone: 853-9766
Bob Givan
-------
∂20-Jun-86 0136 @SU-SCORE.ARPA:shoham@YALE.ARPA Arrangements
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 19 Jun 86 14:53:49 PDT
Received: from yale-bulldog by SU-SCORE.ARPA with TCP; Thu 19 Jun 86 12:31:50-PDT
Received: by Yale-Bulldog.YALE.ARPA; 19 Jun 86 09:50:59 EDT (Thu)
Date: 19 Jun 86 09:50:59 EDT (Thu)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8606191350.AA12180@Yale-Bulldog.YALE.ARPA>
Subject: Arrangements
To: nilsson@su-score.ARPA
Cc: jmc@su-score.ARPA
Nils, everything you propose in your message sounds great. I appreciate your
help and JMC's generousity.
Yoav.
-------
∂20-Jun-86 0137 avg@su-aimvax.arpa Re: Recursive negation
Received: from SU-AIMVAX.ARPA by SU-AI.ARPA with TCP; 19 Jun 86 15:22:25 PDT
Received: by su-aimvax.arpa with Sendmail; Thu, 19 Jun 86 15:20:54 pdt
Date: Thu, 19 Jun 86 15:20:54 pdt
From: Allen VanGelder <avg@diablo>
Subject: Re: Recursive negation
To: VAL@Sail, jmc@sail, ullman@diablo
I don't know what your standards for correct execution are, but I'm
sure the following test would not meet them:
Script started on Thu Jun 19 15:11:53 1986
cpl
% C Prolog version 1.4e.edai
yes
| ?- +block.
block.pl consulted 360 bytes 0.066667 sec.
yes
| ?- listing([ab,on]).
ab(A, B, C) :-
on(D, A, C).
on(A, B, r(m(A, B), C)) :-
\+ ab(A, B, C).
on(1, 2, s0).
on(2, 3, s0).
yes
| ?- trace.
yes
| ?- on(1,2,S).
( 1) 1 call: on(1,2,←0) ?
( 2) 2 call: \+ab(1,2,←18) ?
( 3) 3 call: ab(1,2,←18) ?
( 4) 4 call: on(L128,1,←18) ?
( 5) 5 call: \+ab(←23,1,←25) ?
( 6) 6 call: ab(←23,1,←25) ?
( 7) 7 call: on(L164,←23,←25) ?
( 8) 8 call: \+ab(←30,←23,←32) ?
( 9) 9 call: ab(←30,←23,←32) ? l
! Out of local stack during execution.
% execution aborted
| ?- halt.
% Prolog execution halted
%
script done on Thu Jun 19 15:13:20 1986
-----------------------------------------------------------------------------
Sure, Prolog will work on a few cases of lots of toy examples, but
you should really try to break them before jumping to conclusions.
-- Allen
∂20-Jun-86 0148 WINOGRAD@SU-CSLI.ARPA Chinese visitor
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 19 Jun 86 17:48:35 PDT
Date: Thu 19 Jun 86 17:46:37-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Chinese visitor
To: stucky@SU-CSLI.ARPA, grosz@SRI-AI.ARPA, nilsson@SU-SCORE.ARPA,
weiderhold@SU-SCORE.ARPA, feigenbaum@SUMEX-AIM.ARPA, dek@SU-AI.ARPA,
jmc@SU-AI.ARPA, bobrow@XEROX.COM, briansmith@XEROX.COM
cc: Winograd@SU-CSLI.ARPA
I have agreed to be host (no facilities provided) to two Chinese
visitors in September. Prof. Gao Qingshi, Director of computer
architecture of the Institute of Computing Technology will be coming,
along with a colleague. He requests:
"We wish you would make arrangements for us to meet Prof.
Feigenbaum, Prof. Knuth and others who are investigating A.I.,
especially in N.L.P. and to visit the CSLI of your university"
I don't know much about him, but he came recommended by someone I know,
and generally those senior Chinese scholars who manage to get over here
are interesting and surprisingly knowledgeable.
I ask of each of you the following:
1) Will you be around and willing to meet with them between Sept.
11 and 19?
2) Who else might be interested?
Thanks --t
-------
∂20-Jun-86 0154 SLEEMAN@SUMEX-AIM.ARPA AAAI-sponsored scientific meeting?
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 19 Jun 86 18:14:36 PDT
Date: Thu 19 Jun 86 18:12:47-PDT
From: Derek Sleeman <SLEEMAN@SUMEX-AIM.ARPA>
Subject: AAAI-sponsored scientific meeting?
To: jmc@SU-AI.ARPA
cc: SLEEMAN@SUMEX-AIM.ARPA, buchanan@SUMEX-AIM.ARPA
Message-ID: <12216185795.63.SLEEMAN@SUMEX-AIM.ARPA>
Below I give an outline for a meeting which Bruce Buchanan and I
have talked about organizing next year. Unfortunately, Bruce left for
his vacation before he had commented on my draft and I will be gone
by the time he returns.
Would it be possible to get AAAI support for such a meeting?
Derek Sleeman
------------------------------------------------------------------------
Proposal for a meeting on
AI & SCIENTIFIC DISCOVERY
Learning - and particularly Discovery - have been seen as major challenges
to the AI community. In this workshop we will review the early work
in the area of scientific reasoning in AI (DENDRAL & Meta-DENDRAL),
discuss current projects, and attempt to highlight critical research
issues. In the last two decades, AI has made substantial progress
with representational schemas, learning algorithms etc; what
relevance do these have for building systems which can make scientific
discoveries?
Additionally, to help the AI workers focus on the REAL issues of
scientific discovery we plan to have psychologists, history and
philosophy of science specialists, emminent scientists from the
fields of chemistry and molecular biology, and maybe statasticians
participating in the meeting.
Proposed location: Aberdeen, Scotland
Size: upto 50 persons - invitational meeting.
Timing: Just prior or just after IJCAI-87.
Organizers: Derek Sleeman(Aberdeen/Stanford), Bruce Buchanan(Stanford).
-------
∂20-Jun-86 0335 a32275%ccut.u-tokyo.junet%utokyo-relay.csnet@CSNET-RELAY.ARPA US-Japan collaboration
Received: from CSNET-RELAY.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 03:33:54 PDT
Received: from utokyo-relay by csnet-relay.csnet id af12970; 20 Jun 86 6:30 EDT
Received: by u-tokyo.junet (4.12/4.9J-1[JUNET-CSNET])
id AA01382; Fri, 20 Jun 86 18:11:20+0900
Received: by ccut.u-tokyo.junet (4.12/6.1Junet)
id AA09070; Fri, 20 Jun 86 18:05:38+0900
Date: Fri, 20 Jun 86 18:05:38+0900
From: NAKAHARA Hayao <a32275%ccut.u-tokyo.junet%utokyo-relay.csnet@CSNET-RELAY.ARPA>
Message-Id: <8606200905.AA09070@ccut.u-tokyo.junet>
To: jmc%su-ai.arpa%u-tokyo.junet@CSNET-RELAY.ARPA
Cc: clt%su-ai.arpa%u-tokyo.junet@CSNET-RELAY.ARPA
Subject: US-Japan collaboration
Dear Professor McCarthy:
I send a mail to you on June 7 via nttlab!Shasta. Today, I heard that
the Shasta machine has been very unhealty, and that some mails via
Shasta machine were lost. So I include the copy of the mail dated
June 7 (some mistakes are corrected). If you did not send the letter
or something mentioned the mail dated June 7, please send it to me
and Mr. Hagiya, as soon as possible.
I am looking forward to seeing you.
Sinceary Yours
Nakahara, Hayao
P.S.
My email address via CS-NET is the following.
a32275%ccut.u-tokyo.junet%utokyo-relay@csnet-relay.arpa
**** Here, I put the copy of the mail dated June 7. ****
Dear Professor McCarthy:
I am very pleased that I shall join in the join US-Japanese
collaboration on logic and computer science and visit to Stanford
this summer. I suppose that Professor Sato would inform our schedule
to you, but I inform our schdule here again. I and Hagiya (at RIMS,
Kyoto) are plannig to stay at Stanford from July 1 to August 31, and
Sato from July 17 to August 19. I would appreciate it if you would
arrange for our office and using computer during our stay at Stanford.
Sato said me that we will need your invitation letter or some
certificate which proves our position as a visiting researcher at
Stanford University and shows we will join US-Japanes collaboration
supported by NSF and JSPS. Without it, we would have some trouble
with the immigration officer. For this reason, please send your
invitation letter or certificate to me and Hagiya (and to Sato, if
you did not send it to him), as soon as posible.
I am looking forward to seeing you soon at Stanford.
Sinceary Yours
Nakahara, Hayao
Department of Information Science
Faculty of Science
University Tokyo
Tokyo, Japan, 113
P.S.
Hagiya's address:
Masami Hagiya
Research Institute of Mathematical Sciences
Kyoto University
Kyoto, Japan, 606
∂20-Jun-86 0956 OR.DANTZIG@su-sierra.arpa Re: Nunez
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 09:56:23 PDT
Date: Fri 20 Jun 86 09:55:32-PDT
From: GEORGE DANTZIG / GAIL STEIN <OR.DANTZIG@su-sierra.arpa>
Subject: Re: Nunez
To: JMC@su-ai.arpa
cc: OR.DANTZIG@su-sierra.arpa
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Wed 18 Jun 86 13:59:00-PDT
Message-ID: <12216357415.17.OR.DANTZIG@su-sierra.arpa>
Thanks for the message -- I have informed GBD. Gail
-------
∂20-Jun-86 1206 RA Thelma, Inference
Thelma called to let you know that there is going to be a special BOD meeting
July 11, 9:00-3:00; the purpose of the meeting is to interview 5 investment
banking firms to choose who will take them public. Thelma's tel. (213) 417 7997.
∂20-Jun-86 1351 hpm@rover.ri.cmu.edu robot vehicle update
Received: from ROVER.RI.CMU.EDU by SU-AI.ARPA with TCP; 20 Jun 86 13:51:31 PDT
Date: 20 Jun 1986 16:49-EDT
From: Hans.Moravec@rover.ri.cmu.edu
To: jmc@su-ai
Subject: robot vehicle update
Message-Id: <519684544/hpm@rover.ri.cmu.edu>
This is a report on a demo done yesterday in Denver:
-----------------------------------
Date: 20 Jun 1986 12:27-EDT
From: Jon.Webb@ius2.cs.cmu.edu
To: alv-group
Subject: ALV demo
The Martin Marietta demo was a complete success. The ALV travelled
about 2 km down an average road, around a turn, then turned around and
travelled back, at speeds up to 10 km/hr, under visual control
(exclusively -- they used almost the same algorithm as last May, on the
same hardware). It was pretty impressive: the vehicle tracked the
center of the road well, slowed down when it approached areas where it
couldn't see the road (such as when it passed through a gate, or went
around the turn, or went from dark road to light road), and speeded up
(even shifted, with a jerk) when it came to clear straight, stretches,
of road.
Jon
∂20-Jun-86 1355 WINOGRAD@SU-CSLI.ARPA Comp systems Syllabus
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 13:54:59 PDT
Date: Fri 20 Jun 86 13:53:09-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Comp systems Syllabus
To: jlh@SU-SONOMA.ARPA
cc: phdcom@SU-AI.ARPA
I am finally getting around to cleaning up the syllabus. There weren't
many specific responses, but some general one from students about
trimming out the "fat". In particular we need to designate specific
chapters and sections rather than just list books. Please indicate any
further specification and I'll put together a final proposal to get back
to you for one more look before the meeting next Friday (when all of
this should be finally done). I'm especially asking John for this,
since he put together most of the list, but comments from everyone are
solicited.--t
-------- {\bf LANGUAGES AND SOFTWARE}
Harold Abelson and Gerald Sussman, {\sl Structure and
Interpretation of Computer Programs}, MIT, 1985.
There was a lot of grumbling about this being too long and general. Are
there specific chapters that can be recommended independently, or should
we just drop it?
Alfred V. Aho, Sethi, R., and Jeffrey D. Ullman, {\sl Compilers --
Principles, Techniques, and Tools}, Addison-Wesley, 1986. All
except sections 9.11-9.12, 10.9-10.13.
M. Ben-Ari, Principles of Concurrent Programming, Prentice-Hall
International, 1982. Material on concurrent programming.
What are the specific chapter/page numbers for the relevant material?
Terrence W. Pratt, {\sl Programming Languages: Design and
Implementation}, Second edition, Prentice-Hall, 1984.
Whole thing?
{\bf HARDWARE AND OPERATING SYSTEMS}
Bell, C.G., Mudge, J.C., and McNamara, J.E., Computer Engineering
-- A DEC View of Hardware Systems Design, Digital Press, 1978. Part
III only.
Chapter numbers?
Kogge, P., The Architecture of Pipelined Computers, McGraw-Hill,
1981. Chapter 1 only.
M. Morris Mano, {\sl Computer System Architecture}, Second Edition,
Prentice-Hall, 1982. Basic logic design, data representation, and
computer organization. Material taught in CS 108 and CS 112.
Chapter numbers?
James L. Peterson and Abraham Silberschatz, {\sl Operating System
Concepts}, Addison-Wesley, Second Edition, 1985.
Whole thing?
--- Thanks. --t
-------
∂20-Jun-86 1429 WINOGRAD@SU-CSLI.ARPA Comp Theory Syllabus
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 14:28:48 PDT
Date: Fri 20 Jun 86 14:26:59-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Comp Theory Syllabus
To: phdcom@SU-AI.ARPA
I am finally getting around to cleaning up the syllabus. There weren't
many specific responses, but some general ones from students about
trimming out the "fat". In particular we need to designate specific
chapters and sections rather than just list books. Please indicate any
further specification and I'll put together a final proposal to get back
to you for one more look before the meeting next Friday (when all of
this should be finally done). I'm especially asking LEO for this, since
he put together the list, but comments from everyone are solicited.--t
Alfred V. Aho, John E. Hopcroft, and Jeffrey D. Ullman, {\sl Data
Structures and Algorithms}, Addison-Wesley, 1983.
Whole thing?
Herbert B. Enderton, {\sl A Mathematical Introduction to Logic},
Academic Press, 1972, Chapters 1--2.
Should this be on the list or as some kind of recommended prerequisite?
Michael R. Garey and David S. Johnson, {\sl Computers and
Intractability}, Freeman, 1979, Chapters 1--3, 7.
John E. Hopcroft and Jeffrey D. Ullman, {\sl Introduction to
Automata Theory, Languages, and Computation}, Addison-Wesley, 1979,
Chapters 1--3, 4.1--4.6, 5--7, 8.1--8.5.
Donald E. Knuth, {\sl The Art of Computer Programming}, Volume 1,
Second Edition, Addison-Wesley, 1973, Section 1.2 (except for ...
Leo suggested we leave out only 1.2.11.1. Oren Patashnik wrote an
impassioned note (see the PHDPROG BBOARD) arguing that we should keep
the old exclusions of subsections 1.2.9, 1.2.10, 1.2.11.2, and 1.2.11.3.
Which is the choice?
Zohar Manna, {\sl Introduction to Mathematical Theory of
Computation}, McGraw-Hill, 1973, Chapters 1--3.
John McCarthy and Carolyn Talcott, {\sl LISP: Programming and
Proving}, (available from McCarthy's secretary) 1980, Chapters
1--3.
Leo suggested we replace this with a real book if it doesn't come out
soon. Can we?
Nils Nilsson, {\sl Principles of Artificial Intelligence}, Kaufman,
1980, Chapters 4--6.
It was suggested that we replace this with drafts of the forthcoming
Nilsson/Genesereth book. I propose we wait until it is done.
Sedgewick, Robert, {\sl Algorithms}, Addison-Wesley, 1983.
Chapters/sections?
Overall, this list is long. Are there other places, other than the
specifics mentioned above, where it should be cut?
-------
∂20-Jun-86 1439 WINOGRAD@SU-CSLI.ARPA Applicatons and techniques Syllabus
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 14:39:13 PDT
Date: Fri 20 Jun 86 14:37:26-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Applicatons and techniques Syllabus
To: gio@SU-SCORE.ARPA, ullman@SU-SCORE.ARPA, oliger@SU-SCORE.ARPA,
golub@SU-SCORE.ARPA, phdcom@SU-AI.ARPA
I am finally getting around to cleaning up the syllabus. There weren't
many specific responses, but some general ones from students about
trimming out the "fat". In particular we need to designate specific
chapters and sections rather than just list books. Please indicate any
further specification and I'll put together a final proposal to get back
to you for one more look before the meeting next Friday (when all of
this should be finally done). I'm especially asking Jeff and Gio for
help on the Databases, Gene and Joe on the NA, and Leo on the graphics,
but comments from everyone are solicited.--t
{\bf Artificial intelligence}
Elaine Rich, {\sl Artificial Intelligence}, McGraw-Hill, 1983.
{\bf DATABASES}
Jeffrey Ullman, {/sl Principles of Data Base Systems,} Computer
Science Press, 1982
Chapters?
{\bf GRAPHICS}
Newman and Sproull, Principles of Interactive Computer Graphics,
Chapters 1--5, and 15--18.
Grosof thought this was a lot. Is there an appropriate subset?
{\bf NUMERICAL ANALYSIS}
Kendall E. Atkinson, {\sl An Introduction to Numerical Analysis},
Wiley, 1978, Chapters 1--3, 5, 7, 8 (except Sections 2.8, 2.10,
5.4).
Two people complained that this was not a good book, but didn't have an
alternative to suggest. Is it the best we can do?
{\bf NETWORKS}
A.S. Tanenbaum, Computer Networks, Prentice-Hall, 1981.
Keith suggested this be added here. Is that a good idea? The whole
book?
-------
∂20-Jun-86 1442 WINOGRAD@SU-CSLI.ARPA Courses for Comp Syllabus
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 14:42:08 PDT
Date: Fri 20 Jun 86 14:40:07-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Courses for Comp Syllabus
To: phdcom@SU-AI.ARPA
The old syllabus includes the following paragraph:
The comprehensive exam is meant generally to cover the material
from the following courses: 260, 261, 262 (algorithms and data
structures); 223 [520 is recommended in addition] (artificial
intelligence); 108, 112, 212 (hardware systems); 237A (numerical
analysis); 242, 243, 246 [346 is recommended in addition] (software
systems); and 254, 257A [306 is recommended in addition] (theory of
computation). Since the precise content of these courses varies,
the actual scope of the exam will be determined by the references
above.
Should we include an equivalent? John Henessey proposed the following:
Systems = 8 courses noty including 108 material
112, 140, 143A, 212, 242, 240 A,B, 243
Applications = 4-5 courses
237A,(B,C?), 225, 248, 245
Theory = 4 courses including basic AI course
223?, 254, 261, 257A
Comments? --t
-------
∂20-Jun-86 1437 WIEDERHOLD@SUMEX-AIM.ARPA John Sowa industrial lectureship
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 14:37:48 PDT
Date: Fri 20 Jun 86 14:35:01-PDT
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: John Sowa industrial lectureship
To: jmc@SU-AI.ARPA
Message-ID: <12216408295.74.WIEDERHOLD@SUMEX-AIM.ARPA>
John is interested and thinks he can get complementary IBM support by a reassignment
in this area for three months.
A potential quarter is Fall 1987.
Gio
-------
∂20-Jun-86 1509 lantz@su-gregorio.arpa Re: Applicatons and techniques Syllabus
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 15:09:15 PDT
Received: by su-gregorio.arpa with Sendmail; Fri, 20 Jun 86 15:07:55 pdt
Date: 20 Jun 1986 1507-PDT (Friday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Cc: phdcom@SU-AI.ARPA, jlh@sonoma
Subject: Re: Applicatons and techniques Syllabus
In-Reply-To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA> /
Fri 20 Jun 86 14:37:26-PDT.
Re networks, all I can say is that more people take the networks
course(s) than take graphics, databases, or n.a; and that there are
more courses related to networks than to graphics or databases.
For the third time (without any response whatsoever from any other
member of the committee!), I repeat my plea that we decide first what
courses should provide the basis for the comp and then decide on the
reading list! I previously sanctioned John's list of systems courses,
but would obviously add networks as well.
Keith
∂20-Jun-86 1514 lantz@su-gregorio.arpa Re: Courses for Comp Syllabus
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 15:13:15 PDT
Received: by su-gregorio.arpa with Sendmail; Fri, 20 Jun 86 15:12:11 pdt
Date: 20 Jun 1986 1512-PDT (Friday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Cc: phdcom@SU-AI.ARPA
Subject: Re: Courses for Comp Syllabus
In-Reply-To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA> /
Fri 20 Jun 86 14:40:07-PDT.
At last, a response... and simultaneous with my third "plea" at that!
I'd add whatever number networks now has to "applications".
And, by the way: Yes, the entire o.s. book should be on the comp. I'd
like to find a good reference on reliability as well -- which is the
major whole in that book...
Keith
∂20-Jun-86 1514 SJG tomorrow evening
To: myers@SU-SUSHI.ARPA, nsingh@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA
8.00 at my place. Probably charades unless I hear from someone ...
See (some subset of?) you there ...
Matt
∂20-Jun-86 1447 WINOGRAD@SU-CSLI.ARPA
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 14:47:34 PDT
Date: Fri 20 Jun 86 14:45:37-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
To: pratt@SU-NAVAJO.ARPA
cc: phdcom@SU-AI.ARPA
Comp Syllabus Vaughan,
Having just sent out a bunch of mail dealing with the comp syllabus, I
realized that the PhD committee had slipped into doing a job we didn't
set out to fo. The PhD committee was to set the general guidelines, and
leave it up to the comp and qual committees to generate their respective
syllabi. Unfortunately in order to get approval of the general plan we
needed to come up with a plausible syllabus. Since the comp committee
is in some sense done for the year, I'm not sure whether they should be
involved in this or what. Since you are on both, the distinction isn't
clear. Anyway, I would be happy for whatever help and suggestions you
(and others) can give on:
1) Getting the overall program approved next week with a minimum of
hassling over details
2) Getting an actual syllabus published this summer that the students
can count on for next year's comps.
Also, it was mentioned in one of the responses that the comp committee
should not have the power to alter the syllabus without faculty vote,
and that I had suggested otherwise in my message. I think that is
right, and barring complaints will include in the final proposal that
the Comp committee prepares a modification to the syllabus (whenever
they want) for full faculty approval.
-------
∂20-Jun-86 1654 WINOGRAD@SU-CSLI.ARPA Teaching specialization
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 16:54:47 PDT
Date: Fri 20 Jun 86 16:52:51-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Teaching specialization
To: bscott@SU-SCORE.ARPA, reges@SU-SCORE.ARPA, phdcom@SU-AI.ARPA
cc: peyton@SU-CSLI.ARPA
I propose the following requirement for the "distinction in teaching"
specialization for the PhD program (which will make the corresponding
TAing tax deductible). I hope to get a final version in time to be
voted on at next Friday's faculty meeting.
-----
For the designation of "distinction in teaching" a student must do the
equivalent of 5 units (one quarter, 50% time) of TA/TF, with the
following restrictions:
1) At least one unit must be as a TF with primary responsibility for
organizing and teaching a course (not just tutored video).
2) At least one unit must be in an advanced course (numbered 200 or
higher, under the current numbering system).
-------
This is a first cut, based on informal discussions. The number 5 seemed
reasonable, but could potentially be made larger. The advanced course
requirement could be increased, although the number of people following
this track will probably be small enough that it won't have a
significant impact on the more general problem of getting TAs for those
courses.
Comments? --t
-------
∂20-Jun-86 1716 SJM other things you could mess with
If you finsish with the home terminal essay over the weekend and still
don't feel like working on Space or Adhominem Remarks (see ADHOM[1,sjm])
you could work on Productivity or Construction. Or you could begin
proposed but unwritten essays on medical Ideology and What to Do With
Increased Leisure Time.
Susie
∂20-Jun-86 1721 VAL
I'd like to take a few days off in the form of driving to and from the Portland
workshop at a leisurely pace. Then I'll leave next Wednesday and return on Monday
or Tuesday after the holiday weekend.
∂20-Jun-86 1727 berglund@su-pescadero.arpa Re: Teaching specialization
Received: from SU-PESCADERO.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 17:24:47 PDT
Received: by su-pescadero.arpa with Sendmail; Fri, 20 Jun 86 17:23:48 pdt
Date: Fri, 20 Jun 86 17:23:48 pdt
From: Eric Berglund <berglund@su-pescadero.arpa>
Subject: Re: Teaching specialization
To: WINOGRAD@SU-CSLI.ARPA, bscott@SU-Score, phdcom@Sail, reges@SU-Score
Cc: peyton@SU-CSLI.ARPA
In our current teaching requirement, isn't a unit equal to one quarter,
25% time? If so, the teaching certification requirement should probably
use the same definition.
--Eric
∂20-Jun-86 2344 SF@SU-CSLI.ARPA re: possible Freiling visit
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 20 Jun 86 23:44:19 PDT
Date: Fri 20 Jun 86 23:42:26-PDT
From: Sol Feferman <SF@SU-CSLI.ARPA>
Subject: re: possible Freiling visit
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Fri 20 Jun 86 14:30:00-PDT
I'll discuss the July 7 date with the logic group next monday. Dont forget
that is July 4 weekend and therefore attendance is likely to be sparse.
in any case, I'll get back to you on Monday.
-------
∂21-Jun-86 0750 lantz@su-gregorio.arpa Re: Teaching specialization
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 21 Jun 86 07:49:59 PDT
Received: by su-gregorio.arpa with Sendmail; Sat, 21 Jun 86 07:48:57 pdt
Date: 21 Jun 1986 0748-PDT (Saturday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Cc: peyton@SU-CSLI.ARPA, bscott@SU-SCORE.ARPA, reges@SU-SCORE.ARPA,
phdcom@SU-AI.ARPA
Subject: Re: Teaching specialization
In-Reply-To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA> /
Fri 20 Jun 86 16:52:51-PDT.
2xx courses are not generally all that advanced. For example, both
mainstream operating systems courses and the networks course are 2xx
courses. I would prefer required student to have to TA/TF at least
N courses, with at least one in the 200's and one in the 300's.
The reason for not letting them do two 300's instead of one 200, is
that the nature of 200- and 300-level courses is typically quite
different (e.g. larger and not quite as advanced) and anyone with a
teaching specialization should be exposed to the different levels.
Keith
∂21-Jun-86 0937 SLEEMAN@SUMEX-AIM.ARPA re: AAAI-sponsored scientific meeting?
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 21 Jun 86 09:37:04 PDT
Date: Sat 21 Jun 86 09:34:16-PDT
From: Derek Sleeman <SLEEMAN@SUMEX-AIM.ARPA>
Subject: re: AAAI-sponsored scientific meeting?
To: JMC@SU-AI.ARPA
cc: buchanan@SUMEX-AIM.ARPA, SLEEMAN@SUMEX-AIM.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Fri 20 Jun 86 14:06:00-PDT
Message-ID: <12216615689.10.SLEEMAN@SUMEX-AIM.ARPA>
What additional information should we include? Possible participants?
Titles of sessions/themes? DS
-------
∂21-Jun-86 1213 SJM
huh?
∂21-Jun-86 1341 SJM kitty michaelson
I called Kitty Michaelson about the possibility of our visit and she
assents enthusiastically, claiming she and Sid will be in town then,
although the younger Michaelsons will probably be in Africa, Dundee,
the lake District and London. I have promised to call her with more
details when I have them.
Susie
∂21-Jun-86 1741 RESTIVO@SU-SCORE.ARPA Alliant
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 21 Jun 86 17:40:54 PDT
Date: Sat 21 Jun 86 17:39:20-PDT
From: Chuck Restivo <Restivo@SU-SCORE.ARPA>
Subject: Alliant
To: jmc@SU-AI.ARPA
Message-ID: <12216703992.16.RESTIVO@SU-SCORE.ARPA>
[cwr]
What is Alliant being used for? Are you planning on crushing
HiTech with it?
-------
∂21-Jun-86 1828 OTA-CIT@USC-ISI.ARPA First discussion topic
Received: from USC-ISI.ARPA by SU-AI.ARPA with TCP; 21 Jun 86 18:28:24 PDT
Date: 21 Jun 1986 21:18:00 EDT
Subject: First discussion topic
From: Jim Dray <OTA-CIT@USC-ISI.ARPA>
To: davis@HT.AI.MIT.EDU, tk@MC.LCS.MIT.EDU, allen.newell@A.CS.CMU.EDU,
feigenbaum@SUMEX-AIM.ARPA, schank@YALE.ARPA, rhayes-roth@SRI-KL.ARPA,
ohlander@USC-ISIB.ARPA, winograd@SU-CSLI.ARPA, waltz@GODOT.THINK.COM,
brown.pa@XEROX.COM, kay@MIT-MULTICS.ARPA, phw%mit-oz@MC.LCS.MIT.EDU,
raj.reddy@A.CS.CMU.EDU, jmc@SU-AI.ARPA, nilsson@SU-SCORE.ARPA,
hector%utai.toronto.edu@CSNET-RELAY.ARPA, lenat@MCC.ARPA,
hollister%ti-csl@CSNET-RELAY.ARPA, herb.simon@A.CS.CMU.EDU,
bundy%aiva.edinburgh.ac.uk@CS.UCL.AC.UK, lehnert%umass@CSNET-RELAY.ARPA,
forbus@P.CS.UIUC.EDU, wilensky@DALI.BERKELEY.EDU, kolodner@Gatech,
feldman@ROCHESTER.ARPA, norman@NPRDC.ARPA, collins@BBNA.ARPA,
jsl@ROCKEFELLER.ARPA
cc: ota@USC-ISI.ARPA, ota-cit@USC-ISI.ARPA
Here's the first discussion topic for OTA's informal computer conference.
We've been gratified by the enthusiastic reaction to our invitation to
participate. Incidentally, a copy of OTA's previous case study on AI
and other information should be on its way to you shortly. Please contact
us ASAP if it should be sent to a different address than the one that is
registered at the Network Information Center, or if an address is not
available through the NIC.
An initial caveat: All of these questions will be difficult and, to varying
extents, controversial. I am sure there will be cases where we have not
worded the question correctly, or have simply asked the wrong question.
Please tell us so. We are interested in as frank and complete responses
as you are willing to provide. As we noted earlier, we will be studying
the responses and asking other participants to provide supporting or
contrasting views when it seems appropriate. Also, please feel free
to suggest additional references or experts if they would be
especially helpful in analyzing the topic at hand.
FIRST TOPIC:
Some AI researchers have said that there has been little progress in basic
research over the past few years because of the intense interest in
commercialization of AI. Do you agree? Is the situation getting better or
worse? Do you have any ideas that could stimulate interest in basic research?
There is no time limit for responses to this topic, although we are
planning to introduce one new topic per week. Thus a response within
the next week (by June 28) would be best if possible.
Thank you,
Earl Dowdy and Jim Dray at OTA
-------
∂22-Jun-86 1646 @SU-SCORE.ARPA:shoham@YALE.ARPA historical question
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 22 Jun 86 16:46:34 PDT
Received: from yale-bulldog by SU-SCORE.ARPA with TCP; Sun 22 Jun 86 16:44:55-PDT
Received: by Yale-Bulldog.YALE.ARPA; 22 Jun 86 19:36:25 EDT (Sun)
Date: 22 Jun 86 19:36:25 EDT (Sun)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8606222336.AA03733@Yale-Bulldog.YALE.ARPA>
Subject: historical question
To: jmc@su-score.ARPA
Dear John,
In writing my current thesis chapter I'd like to make sure I get the historical facts
right. When and where did you first introduce the banana/tailpipe problem? (It was a banana
and not a potato, right?) And, if you happen to remember my terminology,
would you say that my initiation problem is precisely your qualification
problem, or perhaps a specialization of it?
Thanks,
Yoav.
-------
∂23-Jun-86 0948 AI.ELLIE@MCC.COM visit to MCC
Received: from MCC.ARPA by SU-AI.ARPA with TCP; 23 Jun 86 09:46:17 PDT
Date: Mon 23 Jun 86 11:45:31-CDT
From: Ellie Huck <AI.ELLIE@MCC.COM>
Subject: visit to MCC
To: jmc@SU-AI.ARPA
Message-ID: <12217142025.73.AI.ELLIE@MCC.COM>
Bob has told me you are coming this week -- do you need hotel
reservations?
-------
∂23-Jun-86 1000 JMC
Terry at Planlunch
∂23-Jun-86 1148 RA John Nafeh
John Nafeh called, (408) 943 1711. Please call him.
∂23-Jun-86 1339 ME hosts
∂23-Jun-86 1334 JMC
∂23-Jun-86 1326 Mailer failed mail returned
To: JMC
In processing the following command:
MAIL
The command was aborted because these Host Name(s) are Unknown:
MCC.COM
------- Begin undelivered message: -------
∂23-Jun-86 1326 JMC re: visit to MCC
[In reply to message sent Mon 23 Jun 86 11:45:31-CDT.]
Yes, please. For Tuesday and Wednesday nights.
------- End undelivered message -------
ME - There are numerous hosts in MCC.COM, so you must specify the
particular host you want. The command HOST MCC will show you the
possibilities.
∂23-Jun-86 1343 AI.ELLIE@MCC.COM
Received: from MCC.ARPA by SU-AI.ARPA with TCP; 23 Jun 86 13:43:35 PDT
Date: Mon 23 Jun 86 15:38:37-CDT
From: Ellie Huck <AI.ELLIE@MCC.COM>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Mon 23 Jun 86 13:33:00-CDT
Message-ID: <12217184460.73.AI.ELLIE@MCC.COM>
I have confirmed your reservations at the Brookhollow for a guaranteed
late arrival, direct bill to MCC, Tuesday, June 24. Bob has scheduled
your talk for Thursday at 2:00. Have a good flight. Ellie Huck
-------
∂23-Jun-86 1405 davism@nyu-csd1.arpa
Received: from NYU-CSD1.ARPA by SU-AI.ARPA with TCP; 23 Jun 86 14:05:11 PDT
Date: Mon, 23 Jun 86 17:00:22 edt
From: davism@nyu-csd1.arpa (Martin Davis)
Message-Id: <8606232100.AA02212@nyu-csd1.arpa>
Received: by nyu-csd1.arpa; Mon, 23 Jun 86 17:00:22 edt
To: jmc@su-ai
Thanks for your comments on Ernest Davis which were really very
helpful. Actually, it is not a matter of reappointment. The
purpose of the third year review is to make a preliminary
assessment of tenure prospects. It is intended that the results
be useful to the faculty member as well as to the department.
Would you be kind enough to send your comments in the form of a
signed letter, so we can include it in the file sent to the Dean
(who probably has only the vaguest notions about electronic
mail). Since I will be leaving New York for the summer in a few
days, please send it to the department chairman, Olof Widlund,
Courant Instiute, 251 Mercer Street, New York, NY 10012. The
letter could begin "Dear Martin" or just be headed something
like:
COMMENTS ON THE RESEARCH OF ERNEST DAVIS
In case you don't have a copy of your comments, I'm
retransmitting them:
Well, I suppose I'd favor reappointment, assuming that this
doesn't carry tenure. He tends to approach problems top down,
and this results in long lists of requirements. His opinions
are generally sensible, but in his discussion of common sense
he avoids formal detail including detailed discussion of the
formalisms introduced by others. I was particularly put off
by the vagueness of his dismissal of non-monotonic reasoning.
I wonder if he really understands the papers he refers to.
Sorry not to be more help.
∂23-Jun-86 1423 VAL
Please tell me if you have any further suggestions regarding my logic programming
paper, I'll be sending it out tomorrow.
∂23-Jun-86 1435 SJG who should handle this?
John -
It appears that Mike's group is not paying for my SAIL account.
Are you somehow still paying for it from when I was a visiting
scholar?
Thanks. (You're only too welcome to ignore this message ... !)
Matt
∂23-Jun-86 1437 RA Frank, Dina Bolla
Frank wanted to let you know that for your trip to Austin tomorrow, the South
West flight was sold out so he booked you on Western 650 which leaves at 10:30
from San Jose and arrives at Austin at 6:28. There are two stops, at Salt Lake
and San Antonio. This is the best he could do. Note that you are leaving from
San Jose and returning to San Francisco.
∂23-Jun-86 1618 GRP Reminder: meeting at 11 tomorrow
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
We have an ebos meeting scheduled for 11am tomorrow.
∂23-Jun-86 1651 WINOGRAD@SU-CSLI.ARPA Going, going, ...
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 23 Jun 86 16:51:21 PDT
Date: Mon 23 Jun 86 16:48:31-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Going, going, ...
To: phdcom@SU-AI.ARPA
cc: phdprog@SU-AI.ARPA, guibas@DECWRL.DEC.COM
I have now revised the proposal document to reflect all of the
discussions I know of to date. It is available from Victoria in
hardcopy, or in a deformatted version as
{score}<winograd>phd-program.txt. I plan to propose it for faculty
approval at the meeting THIS Friday. If there are any further changes,
let me know right away.
Most of it is as we discussed. You might want to take a look at the
following changes:
1. The new teaching specialization requirement is detailed (as
modified from my recent message, after comments by Eric and Keith).
It is in Section VB4.
2. The new PhD minor requirement (as per previous documents on it)
is detailed in Section VA5.
3. Section VI is new, stating the implementation of the policy. In
particular, it makes everything applicable to all students
(including those already here) except that those who have already
filed a G81 will not be required to do the thesis proposal. (Note,
for example that this means that students already here but pre-G81
will have to).
4. It makes it explicit that the Comp syllabus must be approved by
the faculty as a whole (VA2d), not changed by just vote of the Comp
committee. Also, new qual areas are PROPOSED by a group of three
faculty, but must be approved by faculty vote (VA3a). (This is in
response to various concerns voiced earlier).
5. There is an explicit statement that a student's qualifying exam
area must be relevant to the dissertation topic (as determined by
the dissertation committee) (VA4).
6. It explictly refers to the old policy statement (introductory
paragraph) and says that everything there holds unless explictly
revised. We should probably put out another document that
incorporates the requirements and guidelines from old and new
(without all the discussion and justification).
I plan to make the approval of the comprehensive syllabus a separate
vote. That is, the general phd program document is approved on the
basis that there seems to be the potential for a reasonable syllabus
given the structure, then people can vote on whether a particular one is
right (in fact, I'd like to get Vaughan to propose it with his Comp
committee hat on).
--t
-------
∂23-Jun-86 1758 LES Proposed Meeting
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
JMC proposes meeting at 5:00pm this Friday, June 27. Yes, it conflicts
with volleyball but I'll bear up. How 'bout the rest of you?
∂23-Jun-86 1831 AI.BOYER@MCC.COM Gnu & Editor based operating system
Received: from MCC.ARPA by SU-AI.ARPA with TCP; 23 Jun 86 18:31:02 PDT
Date: Mon, 23 Jun 1986 20:27 CDT
Message-ID: <AI.BOYER.12217237108.BABYL@MCC.COM>
From: AI.BOYER@MCC.COM
To: jmc@su-ai
Subject: Gnu & Editor based operating system
By the way, if you have not spent time hacking with
Stallman's "shell" hack for Gnu Emacs, please let me show it
to you when you are here. It is a really excellent
improvement over the interfaces to Lisp on the Lisp
machines, and yet it doesn't have anything to do with Lisp
per se. That is, it provides also an excellent interface to
any other Unix process which is willing to refrain from
cursor control and to take input line at a time. Not a
complete solution, but a major step forward, I think.
∂23-Jun-86 2000 JMC
mcc badge and keys
∂23-Jun-86 2221 SF@SU-CSLI.ARPA Freiling visit
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 23 Jun 86 22:21:27 PDT
Date: Mon 23 Jun 86 22:18:33-PDT
From: Sol Feferman <SF@SU-CSLI.ARPA>
Subject: Freiling visit
To: JMC@SU-AI.ARPA
We did not have a big turn-out today, but of those I spoke to, Monday
July 7 at noon is alright. I'll let you work out the arrangements with
Freiling. If that time is no good for him we could work for some
special time on a Monday or Tuesday afternoon.
Sol
-------
∂23-Jun-86 2237 ME domain names
∂23-Jun-86 1834 JMC re: hosts
[In reply to message rcvd 23-Jun-86 13:39-PT.]
Is it then an error on their part to furnish MCC.COM as a return
address?
ME - Well, not exactly. The Internet is currently in transition
from using centralized host tables to decentralized name servers.
Some places have changed over, but most such hosts have gotten
names like MCC.COM listed in the NIC host table so that hosts still
using the host table (like SAIL) can continue to reach them.
You can READ DOMAINS for a description of the basics of the transition.
∂23-Jun-86 2250 ME domains
∂23-Jun-86 2242 JMC re: domain names
[In reply to message rcvd 23-Jun-86 22:37-PT.]
For now then, the reply macro doesn't work. However, since MCC is
an abbreviation for MCC.AI, which is their main computer, it seems
to work just to strip off the .COM.
I assume we'll eventually have a domain server, but actually wouldn't
it work if we just put MCC.COM in our host table?
ME - Yes, eventually we be able to handle domains (we'll have some
sort of domain resolver, which will talk to name servers elsewhere).
It would work to add MCC.COM to our host table, but our host table is
generated by merging the Stanford host table (maintained by ITS these
days) with the NIC Internet host table. Additions like MCC.COM have to be
programmed into our merging program somehow.
You can feel free, however, to ask your friends at MCC to have MCC.COM
included in the NIC host table -- this would solve your reply problem.
∂24-Jun-86 0110 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK Paper on circumscription and induction
Received: from CS.UCL.AC.UK by SU-AI.ARPA with TCP; 24 Jun 86 01:10:41 PDT
Received: from aiva.edinburgh.ac.uk by 44d.Cs.Ucl.AC.UK via Janet with NIFTP
id a013055; 23 Jun 86 10:47 BST
From: Alan Bundy <bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK>
Date: Mon, 23 Jun 86 10:41:46 -0100
Message-Id: <17458.8606230941@aiva.ed.ac.uk>
To: jmc <@Cs.Ucl.AC.UK:jmc@su-ai.arpa>
Subject: Paper on circumscription and induction
Cc: bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK
John
YOu promised to send me a paper on circumscription
and induction when you were in Edinburgh. This is the 2nd
reminder.
Alan
∂24-Jun-86 0845 CSL.GERLACH@su-sierra.arpa Macqueen impressions
Received: from SU-SIERRA.ARPA by SU-AI.ARPA with TCP; 24 Jun 86 08:45:45 PDT
Date: Tue 24 Jun 86 08:44:31-PDT
From: Sharon Gerlach <CSL.GERLACH@su-sierra.arpa>
Subject: Macqueen impressions
To: linton@su-amadeus.arpa, pratt@su-navajo.arpa, lantz@su-gregorio.arpa,
jmc@su-ai.arpa, nilsson@su-score.arpa
cc: csl.gerlach@su-sierra.arpa
Message-ID: <12217393063.17.CSL.GERLACH@su-sierra.arpa>
Thank you for taking the time to interview David MacQueen yesterday.
The search committee would like your input, and appreciates any comments
you may have.
Sharon
-------
∂24-Jun-86 1202 G.GORIN@LOTS-B.STANFORD.EDU re: proposed meeting
Received: from LOTS-B by SU-AI.ARPA with TCP; 24 Jun 86 12:01:44 PDT
Date: Tue 24 Jun 86 11:56:55-PDT
From: Ralph Gorin <G.GORIN@LOTS-B.STANFORD.EDU>
Subject: re: proposed meeting
To: "@ebos.dis[p,doc]"@SU-AI.ARPA
Message-ID: <12217428088.12.G.GORIN@LOTS-B.STANFORD.EDU>
I can't meet Friday at 5PM.
I'd like to get onto a schedule with the meetings either Monday or
Thursday for the summer. May I propose Monday June 30 at 11 AM?
Ralph
-------
∂24-Jun-86 1225 CLT meeting
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
Generally Monday's are fine for me (except July 7 when I will be away).
Thursday's I would have to bring Timothy!
Monday June 30 JMC will be away, but maybe we should meet anyway.
∂24-Jun-86 1352 LES re: proposed meeting
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
[In reply to message sent Tue 24 Jun 86 11:56:55-PDT.]
Monday at 11am is fine with me.
∂24-Jun-86 1450 KARP@SUMEX-AIM.ARPA PhD program
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 24 Jun 86 14:49:58 PDT
Date: Tue 24 Jun 86 14:50:02-PDT
From: Peter Karp <KARP@SUMEX-AIM.ARPA>
Subject: PhD program
To: winograd@SU-CSLI.ARPA
cc: phdcom@SU-AI.ARPA, baudinet@SU-SUSHI.ARPA, pallas@SU-SUSHI.ARPA,
gangolli@SU-SUSHI.ARPA
Message-ID: <12217459605.60.KARP@SUMEX-AIM.ARPA>
I too was surprised to see there's no grandfather clause on the
thesis proposal. This sounds wrong to me. When was this decided?
Also, I received no copy of that message on the phd-program bboard,
nor have I received any announcement of a meeting I thought you
were scheduling for the phd committee before the faculty meeting.
Is the mailing list broken or something? When is this meeting
planned?
Peter
-------
∂24-Jun-86 1503 berglund@su-pescadero.arpa Uproar
Received: from SU-PESCADERO.ARPA by SU-AI.ARPA with TCP; 24 Jun 86 15:03:47 PDT
Received: by su-pescadero.arpa with Sendmail; Tue, 24 Jun 86 15:02:21 pdt
Date: Tue, 24 Jun 86 15:02:21 pdt
From: Eric Berglund <berglund@su-pescadero.arpa>
Subject: Uproar
To: phdcom@sail
Have y'all seen the uproar on the phdprogram bboard? I remember thinking
that the discussion of "no grandfathers" was fairly short but it seemed we
were on something else important at the time. At this point, I'd have to
agree with the students, but I'm not in an unbiased position here.
--Eric
∂24-Jun-86 1609 RA Trip to Portland
You have a reservation for June 29, Sunday, from SJ, AirCal 527 5:45pm., arr. 7:22.
On the way back, Wed., July 2, AirCal 510, 5:15pm arr. 6:53.
∂24-Jun-86 1734 WINOGRAD@SU-CSLI.ARPA [Andy Freeman <ANDY@su-sushi.arpa>: PhD changes]
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 24 Jun 86 17:33:57 PDT
Date: Tue 24 Jun 86 17:26:11-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: [Andy Freeman <ANDY@su-sushi.arpa>: PhD changes]
To: phdcom@SU-AI.ARPA
cc: WINOGRAD@SU-CSLI.ARPA
I referred to this in my message and don't know if you all have seen it.
--t
---------------
Return-Path: <@SU-SCORE.ARPA:ANDY@su-sushi.arpa>
Received: from SU-SCORE.ARPA by SU-CSLI.ARPA with TCP; Tue 24 Jun 86 14:14:34-PDT
Received: from su-sushi.arpa by SU-SCORE.ARPA with TCP; Tue 24 Jun 86 14:16:06-PDT
Date: Tue 24 Jun 86 14:15:40-PDT
From: Andy Freeman <ANDY@su-sushi.arpa>
Subject: PhD changes
To: winograd@su-score.arpa, phd-program@su-sushi.arpa
Message-ID: <12217453347.8.ANDY@su-sushi.arpa>
In late January, Eric Berglund (a student member of the Ph.D.
Committee) sent a message containing the following section to the
phd-program bboard.
>Anil--
>I don't blame you for being pissed off about the proposed addition of
>yet another requirement for the Ph.D. Of course you understand that
>the change won't affect you or me since we're both already in the
>program under the old degree requirements, but it seems important that
>the discussion be conducted as if it did for the sake of the future
>generations of students who will be affected.
Winograd's latest message says something different.
> 3. Section VI is new, stating the implementation of the policy. In
> particular, it makes everything applicable to all students
> (including those already here) except that those who have already
> filed a G81 will not be required to do the thesis proposal. (Note,
> for example that this means that students already here but pre-G81
> will have to).
The person I talked to at the graduate program office says that
candidacy is a contract. If a student satisfies the requirements that
exist when it is granted, the degree will be awarded. (Both parties
can, of course, alter the terms of the contract but neither can do so
unilaterally.) Dean Traugott is gone until Wednesday, so I don't have
the exact wording in hand yet. If it supports the contract
interpretation, students with candidacy can ignore the new
requirements.
I have to do a thesis proposal as my qual conditional so that doesn't
bother me. What does bother me is that I only have 4 days to look for
zingers in the new requirements.
The proposal isn't ready to be approved (it has many spelling and
grammar errors) and is incomplete (some sections basically say "we
want to change this because xxxx, details to be worked out later").
It shouldn't be submitted until (at least) these problems are
rectified and there is sufficient time for comment and discussion.
If the committee is attempting to beat some deadline, they've missed.
Rushing the current document through gives several wrong impressions.
-andy
ps - I suspect that the requirements for "with Distinction in
Teaching" will be disallowed as an abusive tax shelter by the IRS.
There should be something discussing the non-tax rationale.
-------
-------
∂24-Jun-86 1746 WINOGRAD@SU-CSLI.ARPA [Andy Freeman <ANDY@su-sushi.arpa>: Re: New Proposals for the Ph.D. Program]
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 24 Jun 86 17:46:21 PDT
Date: Tue 24 Jun 86 17:43:11-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: [Andy Freeman <ANDY@su-sushi.arpa>: Re: New Proposals for the Ph.D. Program]
To: phdcom@SU-AI.ARPA
cc: WINOGRAD@SU-CSLI.ARPA
1) 24-Jun Andy Freeman Re: New Proposals for the Ph.D. Program
2) 24-Jun To: ANDY@SU-SUSHI.AR Re: New Proposals for the Ph.D. Program
Message 1 -- ************************
Return-Path: <ANDY@su-sushi.arpa>
Received: from su-sushi.arpa by SU-CSLI.ARPA with TCP; Tue 24 Jun 86 17:35:12-PDT
Date: Tue 24 Jun 86 17:33:55-PDT
From: Andy Freeman <ANDY@su-sushi.arpa>
Subject: Re: New Proposals for the Ph.D. Program
To: WINOGRAD@su-csli.arpa
In-Reply-To: Message from "Terry Winograd <WINOGRAD@SU-CSLI.ARPA>" of Tue 24 Jun 86 17:13:06-PDT
Message-ID: <12217489438.8.ANDY@su-sushi.arpa>
Does this mean that there will be NO review of the final proposal
by the students before it is submitted to a faculty vote? I wonder
what surprises there are this time....
-andy
-------
Message 2 -- ************************
Mail-From: WINOGRAD created at 24-Jun-86 17:42:51
Date: Tue 24 Jun 86 17:42:51-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Re: New Proposals for the Ph.D. Program
To: ANDY@SU-SUSHI.ARPA
cc: WINOGRAD@SU-CSLI.ARPA
In-Reply-To: Message from "Andy Freeman <ANDY@su-sushi.arpa>" of Tue 24 Jun 86 17:35:59-PDT
Yes, it means NO review of the final proposal, formally. As far as I know,
there are student representatives both on the committee and on at the faculty
meeting, and I (to speak for myself as well as the committee) am not trying
to put anything over on them. In sending out the note with the previous
draft, I EXPLICITLY pointed to all of the places where changes were made,
precisely so that people wouldn't have surprises. I'm sorry if you are
feeling that process has passed you by, and indeed it did get a little
rushed at the end (something dissertations have a habit of doing as well).
If your upset is based on the feeling that the version you saw was
intended for approval without further comment, I apologize for the confusion.
If you really feel that the process of fixing it up and getting it approved
this week is unfair, exclusionary, etc. let's set up a time to talk. --t
-------
-------
∂24-Jun-86 1748 WINOGRAD@SU-CSLI.ARPA
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 24 Jun 86 17:48:06 PDT
Date: Tue 24 Jun 86 17:44:50-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
To: phdcom@SU-AI.ARPA
cc: guibas@DECWRL.DEC.COM
My apologies.
Due to my hurry to get things done after being away just after
end-of-quarter rush, I didn't make things clear. My intention was to do
a new draft, along with pointers to the changes, and then to get a round
of feedback from the committee (and BBOARD) before doing a final
version. Rather than calling a meeting of the committee I had hoped it
could be done by EMAIL since most of the changes were simply confirming
the things that had been discussed before, or were relatively minor.
In particular, the grandfather clause was a first attempt at clarifying.
It was not based on a strong principle, and I have no attachment to
making current students do a proposal (although I believe it is within
the "contractual" power, since it can be seen as a standard part of
doing a dissertation -- but there's no point in getting into a big
argument about it). Since it was the only thing in the entire proposal
that drew substantive comment on the BBOARD, it seems that changing that
one paragraph will more-or-less get the document into acceptable form.
I like Hitson's version:
"Students who were admitted to the PhD program prior to 1986 will not be
required to submit and present a thesis proposal. Nevertheless, these
students are strongly encouraged to satisfy the thesis proposal
requirements in addition to the other requirements. Individual advisors
may want to recommend this to students who are still at an early stage
in their research and will benefit from doing the paper and/or
presentation."
One other issue that came up in a student's question to me was the
definition of "3 faculty" for proposing a new qual area. I propose
that it be stated as 3 Academic Council members with appointments (full,
part, or courtesy, etc.) in CS or EE-CSL (the latter will become an
unnecessary distinction before long, I hope). Academic council is the
main University distinction used in deciding whether someone is
qualified to deal with academic (as opposed to research) matters. The
qual can be conducted with the help of others (research associates,
outsiders, etc.) but the 3 regular faculty members carry the
responsibility. This may be a bit stringent in areas where we are
faculty-short, but makes the qual stay a somewhat more adequate
reflection of department balance on the academic side (not just on the
basis of research-funding).
I was also being a bit lazy and hoping that having everyone on the
committee read through the document would help catch those last
remaining spelling errors, xxx's, etc. that Andy seems so upset about.
If it really seems so far away from being ready for approval as he
suggests, we should take stock, but certainly spelling, etc. can be
corrected in 3 days. Also the comments he suggests (e.g., on the
non-tax-realted aspects of the taching proposal) should be added.
Unfortunately due to my own unclarity in talking to Victoria, the draft
I wanted her to circulate to the committee got circulated to the world.
I've sent out a general message to that effect, and we'll just have to
duplicate with an improved one.
Questions for you (the committee):
Are there responses to any of the specific changes?
Is the proposed grandfathering clause OK?
Are there other matters that require further discussion?
What typos, etc. can you point out for correction?
Do we need a meeting (say on Wed. or Thurs) or can we handle
whatever is left by mail?
Thanks --t
-------
∂24-Jun-86 1836 LES Alliant visit
To: JMC
CC: CLT
I learned this afternoon that Ron Gruner, President of Alliant, will be
here Thursday afternoon, 6/26, at 3:30pm. He would like to meet you if
possible, but I gather you will not be back from your trip by then.
∂24-Jun-86 1951 LES Facilities Update
To: "@FACIL.DIS[P,DOC]"@SU-AI.ARPA
Nearly all the equipment orders based on the DARPA funds are now on their
way. As we were about to send off the Imagen ethernet printer order,
however, I learned of an alternative that should be considered. Imagen
now offers a new 20 page/minute printer called the 3320 that can be
purchased with an ethernet interface for $15,500 plus an Imprint-10
trade-in. We happen to have an Imprint-10 in the basement that could be
put into this deal.
The net effect of this change would be to obtain a printer that is 2.5
times as fast as the 8/300 and with a much larger ultimate life for a 50%
higher price plus the trade-in. If we got the 3320, it would make sense
to move the existing 8/300 to the teaching lab and install the 3320 in its
place on the 3rd floor in MJH. The 3320 has a resolution of 300 dots/inch,
is generally software compatible with the 8/300 and its 20 page/minute speed
is about 2/3 the effective speed of the Dover (60 ppm peak, 30 average).
So why am I telling you all this instead of doing it?
(1) I have a gross conflict of interest, being the largest shareholder in
Imagen.
(2) It requires a budget increase of $5k, which would have to come out of
the current reserve of $22k ($41k if we don't buy the Proteon gateway).
I invite comments on this proposition. I will not participate in making
the decision.
On another front, we are reviewing departmental policies on the use of CSD
computers by organizations outside Stanford. For various legal and
practical reasons it appears that a number of existing outside accounts
should be closed. I am engaged in drafting a new departmental policy
covering this issue. If anyone is interested in this problem and would
like to know more or make suggestions, please contact me.
Les Earnest
∂24-Jun-86 2019 lantz@su-gregorio.arpa
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 24 Jun 86 20:18:49 PDT
Received: by su-gregorio.arpa with Sendmail; Tue, 24 Jun 86 20:17:46 pdt
Date: 24 Jun 1986 2017-PDT (Tuesday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Cc: guibas@DECWRL.DEC.COM, phdcom@SU-AI.ARPA
In-Reply-To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA> /
Tue 24 Jun 86 17:44:50-PDT.
I am strongly opposed to letting current students get out of these
requirements altogether. Certainly anyone who has not passed the comp
cannot expect us to give them an old format comp, for example. As for
thesis proposal, I still do not subscribe to anyone getting a G81
signed without having done so. Nevertheless, there are situations even
in CS where a student hasn't filed a G81 but has already done
significant amounts of research (equivalent to a formal thesis
proposal). Therefore, the Systems PhD Program Committee used signing
of the G34 (Advancement to Candidacy) as the basis for grandfathering:
"<These requirements> will become effective on June 12, 1986 (that is,
after the end of the current quarter). Anyone who has not filed
University form G34 (Advancement to Candidacy) by this date must
satisfy the new requirements. Students who have already filed form G34
are strongly encouraged to satisfy the new requirements as well. For
example, students who have not yet selected a formal thesis topic are
encouraged to satisfy the thesis proposal requirement."
I really cannot imagine a good reason to be more "lenient" than this.
(Bending to the will of a vocal minority of loophole-seekers does not
constitute a good reason.) Fortunately, no matter what CSD approves,
any of my students (whether CS or EE) will be required to abide by the
new requirements since those same requirements have been approved by
CSL -- whose students, by the way, did not object in any way to the
grandfather clause as stated above.
Keith
∂24-Jun-86 2135 a32275%ccut.u-tokyo.junet%utokyo-relay.csnet@CSNET-RELAY.ARPA re: US-Japan collaboration
Received: from CSNET-RELAY.ARPA by SU-AI.ARPA with TCP; 24 Jun 86 21:35:51 PDT
Received: from utokyo-relay by csnet-relay.csnet id af04641; 25 Jun 86 0:32 EDT
Received: by u-tokyo.junet (4.12/4.9J-1[JUNET-CSNET])
id AA18007; Wed, 25 Jun 86 12:49:06+0900
Received: by ccut.u-tokyo.junet (4.12/6.1Junet)
id AA18125; Tue, 24 Jun 86 13:34:39+0900
Date: Tue, 24 Jun 86 13:34:39+0900
From: NAKAHARA Hayao <a32275%ccut.u-tokyo.junet%utokyo-relay.csnet@CSNET-RELAY.ARPA>
Message-Id: <8606240434.AA18125@ccut.u-tokyo.junet>
To: JMC%SU-AI.ARPA%u-tokyo.junet@CSNET-RELAY.ARPA
In-Reply-To: John McCarthy's message of 20 Jun 86 1130 PDT
Subject: re: US-Japan collaboration
Dear Professor McCarthy:
I got your message. Thanks so much for taking the time to write the
letter to me.
∂25-Jun-86 0000 JMC
ota
∂25-Jun-86 0024 HST 1st idea
To: lisp30@SU-AI.ARPA
in the fall of 1988 there were 30 years that john mccartthy made the first
design for lisp. therefore we propose a "30 years lisp conference".
the place should be cambridge, mass., the time september or octoiber 1988.
we think, the conference should be based on invited contributions only.
if, however, somebody has an important subject (he believes so) then he may
apply for a talk.
every talk should be mainly historical: covereing events, people, development
of ideas, implementations, solutions etc.
there are two possibilities for the basic idea of this conference: we could
concentrate on subjects or on countries.
the first means that we cover important lisp related subjects and make our
speakers to tell us something about their development over the years.
the second means that we cover countries and make our speakers to tell us how
the national development of lisp has been.
i believe that the former will be more interesting than the latter.
in the next mail i will send my first proposal for interesting subjects and
some ideas on people who could take them. the purpose is, to discuss the mat-
ter.
another point is the program committee. according to the plan, program commit-
tee members are not thought to read a lot of material but act as sources of
ideas. sometimes they should help to find speakers.
my current idea is: Danny Bobrow, Tony Hearn,
John McCarthy, Herbert Stoyan, Dan Weinreb.
Dan Weinreb is not very enthusiastic for being the local organisator at MIT.
this seems to be an important point - we should find one. Who has a proposal?
herbert
∂25-Jun-86 0039 HST first idea for an announcement and call for proposals:
To: lisp30@SU-AI.ARPA
The 30-years-LISP Conference
To be held at October 1988 at MIT Cambridge USA.
Proposed Program:
1. Basic Ideas and their Historic Relation to LISP
a) Lambda-Calculus
b) Recursive Functions
c) Functional Programming
2. The Work on LISP at MIT 1958-1963
a) LISP design and LISP Reality 1958-1959
b) The LISP1 System
c) The LISP1.5 System
3. The Spread-out of LISP in Dialects
a) The Missing Next Step - the LISP2 Project
b) From LISP1.6 to MacLISP
c) From BBN-LISP to InterLISP
4. The Question of Language Description and Standardisation
a) The LISP1.5 Description as First Formal Description of Programming Language
Semantics (...an Approach to ...)
b) Formal Specifications of Scheme and LISP
c) Virtual LISP Machines
d) Standard LISP
e) CommonLISP
5. 29 years LISP application
a) Formula Manipulation - Computer Algebra
b) LISP as Language Implementation Language
c) The Dendral System as first big Application of LISP
d) Expert-System Shells - The Big Business
6. LISP Machines and Parallel LISPs
a) The MIT (LMI/SYMBOLICS) LISP Machine
b) The XEROX Dolphin as LISP Machine (or: LISP Machines at XEROX)
c) The Scheme Chip
d) Parallel LISPs
7. The Development of the art of LISP Implementation
a) Binding Strategies in LISP
b) LISP Compilers
c) LISP Transport Systems
d) Storage Management in LISP
∂25-Jun-86 0052 HST possible speakers for the 30-years-lisp conference
To: lisp30@SU-AI.ARPA
Theme 1a: Dana Scott (A.Church, B.Rosser, S.Kleene)
Theme 1b: Hartley Rogers (D.Park, Z.Manna, J.Vuillemin)
Theme 1c: Dan Friedman (J.Henderson, D.Wise, M.Wand)
Theme 2a: Herbert Stoyan (J.McCarthy, S.B.Russell)
Theme 2b: Stephen Russell (D.J.Edwards, K.Maling, H.Stoyan)
Theme 2c: Timothy Hart (M.Levin, H.Stoyan)
Theme 3a: Paul Abrahams (C.Weissman, L.Hawkinson, M.Levin)
Theme 3b: JonL White (P.Nelson, P.Samson, R.Greenblatt, G.Steele)
Theme 3c: Danny Bobrow (W.Teitelman, A.Hartley)
Theme 4a: W.Clinger (D.Friedman, D.Wise, J.McCarthy)
Theme 4b: W.Clinger (G.Sussman, D.Friedman)
Theme 4c: John S.Moore (R.Fateman)
Theme 4d: Tony Hearn (J.Marti, M.Griss, J.Fitch)
Theme 4e: Guy Steele (R.Gabriel, S.Fahlman, D.Moon, D.Weinreb)
Theme 5a: Joel Moses (R.Fateman, A.C.Hearn etc.)
Theme 5b: M.Stefik (D.Bobrow, C.Hewitt, G.Sussman, D.McDermott etc.)
Theme 5c: Bruce Buchanan (E.Feigenbaum, W.White)
Theme 5d: R.Fikes (D.Bobrow, B.Clancey, expert system shell people...)
Theme 6a: R.Greenblatt (D.Moon, D.Weinreb, R.Baker)
Theme 6b: Peter Deutsch (L.Masinter, XEROX-people...)
Theme 6d: Richard Gabriel (R.Halstead)
Theme 7a: R.Baker (G.Steele, D.Wise, M.Wand)
Theme 7b: Fred Blair (A.C.Hearn, J.Gosper, H.Stoyan)
Theme 7c: Martin Griss (J.Chailloux, R.Gabriel)
Theme 7d: John Fitch (D.Bobrow, J.Chailloux, M.Griss)
∂25-Jun-86 0739 TOM@SU-SCORE.ARPA Re: Facilities Update
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 07:39:51 PDT
Date: Wed 25 Jun 86 07:30:55-PDT
From: Thomas Dienstbier <TOM@SU-SCORE.ARPA>
Subject: Re: Facilities Update
To: LES@SU-AI.ARPA
cc: "@FACIL.DIS[P,DOC]"@SU-AI.ARPA, TOM@SU-SCORE.ARPA
In-Reply-To: Message from "Les Earnest <LES@SU-AI.ARPA>" of Tue 24 Jun 86 19:51:00-PDT
Message-ID: <12217641811.9.TOM@SU-SCORE.ARPA>
I go along with the suggestion on purchasing the new Imagin 3320.
We have to think about replacing the dovers and this may be one way
to slowly move them out. The deal that Imagin is giving us on the old
Imprint-10 is to good to turn down. Imprint-10's are basically boat
anchors and it is hard to find anyone who might want it. For an
additional 5K I don't see how we can go wrong here.
tom
-------
∂25-Jun-86 1101 WINOGRAD@SU-CSLI.ARPA [Andy Freeman <ANDY@su-sushi.arpa>: Re: New Proposals for the Ph.D. Program]
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 11:01:21 PDT
Date: Wed 25 Jun 86 09:58:38-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: [Andy Freeman <ANDY@su-sushi.arpa>: Re: New Proposals for the Ph.D. Program]
To: phdcom@SU-AI.ARPA
cc: WINOGRAD@SU-CSLI.ARPA
1) 24-Jun Andy Freeman Re: New Proposals for the Ph.D. Program
2) 25-Jun To: ANDY@SU-SUSHI.AR Re: New Proposals for the Ph.D. Program
Message 1 -- ************************
Return-Path: <ANDY@su-sushi.arpa>
Received: from su-sushi.arpa by SU-CSLI.ARPA with TCP; Tue 24 Jun 86 18:11:14-PDT
Date: Tue 24 Jun 86 18:12:32-PDT
From: Andy Freeman <ANDY@su-sushi.arpa>
Subject: Re: New Proposals for the Ph.D. Program
To: WINOGRAD@su-csli.arpa
In-Reply-To: Message from "Terry Winograd <WINOGRAD@SU-CSLI.ARPA>" of Tue 24 Jun 86 17:44:29-PDT
Message-ID: <12217496469.8.ANDY@su-sushi.arpa>
I'm not asking for formal review; I'm asking for timely notice. (My
last message confused the two.) Given the trivial amount of feedback
that we've been getting from the PhD committee (on phd-program bboard
or via mail to phd), that notice is very important. The Systems
people changed fewer things but kept us much better informed.
From the beginning, our complaint has been that the committee is so
concerned with input that it neglects output. We hear nothing for a
long time until one day, a proposal is being submitted in three days
to the faculty for approval. Moreover, the changes which we were told
wouldn't apply to us, do. (I'll find out tomorrow whether that is
legal for those of us who have candidacy.)
In short, what's the rush now? We've been waiting for a couple of
months to see what a near-final draft was going to look like. Why not
let us have a couple of weeks to look at it?
-andy
ps - I think we'd agree that the proposal drastically changes the
programming requirement. The draft says nothing about the coursework
option for satisfying it. Is the problem seminar an example of a
class that will satisfy it or not? (I don't think it should be as the
projects are all done by small groups.)
-------
Message 2 -- ************************
Mail-From: WINOGRAD created at 25-Jun-86 09:57:55
Date: Wed 25 Jun 86 09:57:55-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Re: New Proposals for the Ph.D. Program
To: ANDY@SU-SUSHI.ARPA
cc: WINOGRAD@SU-CSLI.ARPA
In-Reply-To: Message from "Andy Freeman <ANDY@su-sushi.arpa>" of Tue 24 Jun 86 18:13:25-PDT
I accept your criticism that more could have been done to inform the
students at more steps along the way. From the beginning, I have assumed
that student interests were primarily represented by the students on the
committee, not by some kind of universal poll. I tried to post things
onto the relevant BBOARD whenever they were put into writing. But more things
that were verbal could have been put into writing sooner to facilitate this
communication.
The rush is that if the faculty does not vote on this at the meeting this
Friday, there will be no time to vote on it until the faculty meets
after Fall Quarter begins, which means it can't go into effect next year.
This is not disastrous, but certainly should be avoided unless there are
substantive reasons (not just procedural dissatisfactions).
As to the programming requirement, it was put down several months ago and
nothing changed in the last revision. Why has it suddenly become an issue?
We chose not to try to get to precise on the exact nature of the courses,
since there will always be funny cases and exceptions. The reason for having
an ongoing committee to "interpret and administer" the policy is that these
kinds of decisions can be made dynamically instead of being cast in concrete.
I am a bit put out by the tone (at least) of your comments about running to
the Dean for legal judgments. It creates an impression that the department
is an adversarial arena, in which the students need to go to the boss to
make sure their rights are protected against their opponents, the faculty. You
could have asked me or Victoria or any of a number of other people about the
policy without "I'll find out tomorrow whether that is legal..." In fact,
the proposal should have been (and a minor correction would have acheived that)
...people who have filed for candidacy..." rather than "...people who
have filed a G81..." I confused the two. But since students seem so up
in arms about the possibility of having to do a proposal, and since it is in
the spirit of the system to treat the pre-advertisement when you got here
(rather than just the candidacy) as determining what you have to do,
the current proposal is to make it apply only to new students anyway.
As I said in the previous note, I would be glad to get together and talk
about these or any other of the policies before doing the final draft.
As with all policies, everything that gets decided there is still subject
to further amendment and interpretation, so although we should do our best,
I don't feel that this an exercise in closing every conceivable loophole.
As I said above, that is needed in contracts where the two parties assume
a pure commercial relationship with no shared common goals and interests.
I sincerely hope that isn't what we are engaged in here.
--t
-------
-------
∂25-Jun-86 1128 hitson@su-pescadero.arpa Re: Facilities Update
Received: from SU-PESCADERO.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 11:28:44 PDT
Received: by su-pescadero.arpa with Sendmail; Wed, 25 Jun 86 11:27:24 pdt
Date: 25 Jun 1986 1127-PDT (Wednesday)
From: Bruce Hitson <hitson@su-pescadero.arpa>
To: facil@sail
Cc: hitson@su-pescadero.arpa
Subject: Re: Facilities Update
----- Forwarded message follows -----
Return-Path: <hitson>
Date: 25 Jun 1986 1123-PDT (Wednesday)
From: Bruce Hitson <hitson>
To: Les Earnest <LES@SU-AI.ARPA>
Cc: "@FACIL.DIS[P, DOC]"@SU-AI.ARPA, hitson
Subject: Re: Facilities Update
In-Reply-To: Les Earnest <LES@SU-AI.ARPA> / 24 Jun 86 1951 PDT.
Les,
I think that putting in an additional $5k for the larger Imagen is
probably a good idea, as long as this does not affect...
GATEWAY ACQUISITION: You mention the following in your message:
> (2) It requires a budget increase of $5k, which would have to come out of
> the current reserve of $22k ($41k if we don't buy the Proteon gateway).
I continue to be confused about whether acquisition of a new gateway is
IN or OUT of the budget. I feel fairly strongly that it should be IN.
The curret gateway is on it's last legs (and has been for a while). It
needs a fair amount of "tending" to keep it running. It is an
important link to the rest of the Internet. If we can spend a
reasonable amount of money to get a reliable box that does this mundane
but important task, I think everyone wins. If instead we spend money
and staff/student time trying to put together something from spare
parts in the basement (or whatever - as had occasionally been
suggested), I think that we fritter away time and energy on less
interesting tasks when there are plenty of more interesting and
important things to be done. Whatever approach is taken, we will need
a "new" gateway fairly soon.
So, is there some concensus as to whether the GATEWAY budget item is
IN or OUT, and further, what type of gateway (e.g., buy one or try to
put one together from spare parts)?
--- Bruce
------- End of Forwarded Message
∂25-Jun-86 1257 mcdermott-drew@yale.ARPA Reply to your message
Received: from YALE-VENUS.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 12:57:24 PDT
Date: 25 JUN 1986 15:59:48 EST
From: <MCDERMOTT-DREW@YALE.ARPA>
To: JMC@SU-AI.ARPA
Subject: Reply to your message
Reply-To: <MCDERMOTT-DREW@YALE.ARPA>
John:
Sorry it took so long to reply to your message. I had
decided to write a paper summarizing my pessimistic conclusions
about some ways of applying logic to AI, and it took me longer
than expected to finish it. A draft of the paper leaked out,
and I expect you have seen it. If not, Hector Levesque will be
sending you a copy. He is devoting an issue of Computational
Intelligence to the paper and replies, and obviously wants a
reply from you. The final version is rather different in some
of its points from the old version, and a lot better, I think.
My paper is fairly hard on nonmonotonic logic generally, including
circumscription. I won't repeat my reasons here, but I guess you
could summarize them by saying I've come to agree with David Israel
that nonmonotonic reasoning and nondeductive reasoning are the same
thing -- a huge topic that is not going to be encompassed by a minor,
uniform change to logic.
-- Drew
-------
∂25-Jun-86 1321 LES re: Facilities Update
To: hitson@SU-PESCADERO.ARPA
CC: "@FACIL.DIS[P,DOC]"@SU-AI.ARPA
[In reply to message sent 25 Jun 1986 1123-PDT.]
The Proteon gateway is still in the budget and approval to buy it has
been obtained from the government. I still have not ordered it pending
the outcome of negotiations regarding the Stanford/cicso gateway.
There is also another prospective source of an ARPAnet gateway: DECWRL
would like to put one at Stanford to connect a local branch of CSnet to
ARPAnet. Such a gateway could also serve Stanford.
∂25-Jun-86 1411 BOSACK@SU-SCORE.ARPA Re: Facilities Update
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 14:11:23 PDT
Date: Wed 25 Jun 86 13:33:57-PDT
From: Len Bosack <BOSACK@SU-SCORE.ARPA>
Subject: Re: Facilities Update
To: facil@SU-AI.ARPA
In-Reply-To: Message from "Les Earnest <LES@SU-AI.ARPA>" of Tue 24 Jun 86 19:51:00-PDT
Message-ID: <12217707898.15.BOSACK@SU-SCORE.ARPA>
I think we should take the deal on the new Imagen. In general, we have
regretted buying low-capacity anything. If the device is in demand at all,
the bigger ones are usually a better match to our needs in terms of
long-term durability and cost of ownership.
Len
-------
∂25-Jun-86 1422 AI.ELLIE@MCC.COM Abstract - McCarthy's Thursday talk
Received: from MCC.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 14:22:15 PDT
Date: Wed 25 Jun 86 16:18:37-CDT
From: Ellie Huck <AI.ELLIE@MCC.COM>
Subject: Abstract - McCarthy's Thursday talk
To: MCC-people@MCC.COM, Bboard@R20.UTEXAS.EDU, JMC@SU-AI.ARPA
Message-ID: <12217716028.35.AI.ELLIE@MCC.COM>
Abstract for McCarthy's talk on Thursday, June 26, 2:00pm in Echelon
Room 409:
Problems with Non-Monotonic Reasoning
McDermott's "Critique of Pure Reason" presents understanding the
following story as a reason why the "logicist" approach to AI will
fail. I think it's an interesting but technical problem in
non-monotonic logic and its model theory.
At t=0 Fred is alive and his gun is loaded; at t=1, he points the
gun at his head and pulls the trigger. We have the non-monotonic
rules that a living person stays alive till something happens and a
loaded gun stays loaded till something happens. On the basis of this
we expect the death of Fred. However, in all the published
non-monotonic formalisms the model in which the gun becomes unloaded
before t=1 is equally treated. The problem is how to formalize the
problem so as to regard the model in which Fred dies as standard.
Analogous problems arise in the semantics of logic programs with
negation. A possible brute force solution will be discussed.
-------
∂25-Jun-86 1514 WINOGRAD@SU-CSLI.ARPA Re: phdcom
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 15:14:31 PDT
Date: Wed 25 Jun 86 15:10:46-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Re: phdcom
To: KARP@SUMEX-AIM.ARPA
cc: WINOGRAD@SU-CSLI.ARPA, phdcom@SU-AI.ARPA
In-Reply-To: Message from "Peter Karp <KARP@SUMEX-AIM.ARPA>" of Wed 25 Jun 86 13:02:27-PDT
Did you get my last message about what is going on? Maybe it got there after
you sent yours. In any case, it asks the committee whether or not they
think there needs to be a meeting, and I take this as a positive answer
from you. How about tomorrow at 10, 11 or 4? If I don't hear from you
by the end of the day I'll pick one of those and announce it.
(I'll CC this to the committee so everyone can comment on the times).
--t
-------
∂25-Jun-86 1730 WINOGRAD@SU-CSLI.ARPA New draft and meeting tomorrow at 11
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 17:30:43 PDT
Date: Wed 25 Jun 86 17:27:04-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: New draft and meeting tomorrow at 11
To: phdcom@SU-AI.ARPA
cc: Winograd@SU-CSLI.ARPA
I have edited the proposal on the basis of the various comments and
discussions that followed its last (prematurely public) release. It is
<winograd>phd-program.scribe on SCORE (Victoria can get it run off
tomorrow). In addition to fixing up bugs (e.g., the comp was described
as having 3 sections in one place and 4 in another) it has the following
substantive revisions:
1) The grandparent clause is according to Keith's proposal (and in
agreement with various later discussions with Andy Freeman).
Students are required to do a dissertation proposal if they have
not yet reached candidacy, but are only urged to if they are
post-candidacy pre-G81.
2) The programming project clause that classwork can be used
includes a clarifying remark that it must satisfy the requirements
for individual work, as set down in the old policy.
3) Comments on taxability of TA salaries have been expurgated. The
justification for the teaching distinction simply says it is a good
way to learn how to teach.
4) A number of minor wording changes (e.g., mentioning that the
cooperation suggested with CSL is evidenced by the report of the
systems committee, changing "will propose" to "do propose",
replacing references to Victoria with "appropriate staff", etc.
Peter voted for having a meeting, and nobody else (on the committee) has
responded to much of anything (except Keith's comment). Let's meet at
11 tomorrow (Thursday) in 252 to discuss whatever of this needs further
work. If you have opinions but can't attend, send a message.
Note that all of this doesn't deal with a final version of the syllabus.
I will do a draft of that for our meeting, but as mentioned earlier I
want to treat it as a feasibility demonstration, not a final version.
That is, it will be used to indicate the kind of syllabus that would go
along with the new structure, but final details are to be determined by
the Comp committee and approved next fall. If you feel strongly that we
should finalize it for this meeting, come and discuss it.
--t
-------
∂25-Jun-86 1734 WINOGRAD@SU-CSLI.ARPA [Terry Winograd <WINOGRAD@SU-CSLI.ARPA>: New version]
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 25 Jun 86 17:34:01 PDT
Date: Wed 25 Jun 86 17:30:29-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: [Terry Winograd <WINOGRAD@SU-CSLI.ARPA>: New version]
To: phdcom@SU-AI.ARPA
cc: WINOGRAD@SU-CSLI.ARPA
Mail-From: WINOGRAD created at 25-Jun-86 17:30:12
Date: Wed 25 Jun 86 17:30:12-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: New version
To: phd-program@SU-SCORE.ARPA
cc: WINOGRAD@SU-CSLI.ARPA
A new version of the draft is now on <winograd>phd-program.scribe at
score. It fixes a variety of minor problems and adds clarifications.
The one policy issue in which it differs from the previous version is the
application of new requirements to old students. The dissertation proposal
will be required of everyone who has not been promoted to CANDIDACY,
and recommended (at the option of advisers) for those beyond that but still
pre G-81. If you have strong feelings on this, lobby your committee members
and/or faculty (in the end it is the full faculty vote, not the committee
that determines policy).
--t
-------
-------
∂25-Jun-86 1817 RA David Kirsh
Kirsh from MIT AI lab called (617) 253 6569, re: Foundation of AI conference
next year.
∂26-Jun-86 0002 lantz@su-gregorio.arpa Re: phdcom
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 26 Jun 86 00:02:41 PDT
Received: by su-gregorio.arpa with Sendmail; Thu, 26 Jun 86 00:01:46 pdt
Date: 26 Jun 1986 0001-PDT (Thursday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: phdcom@sail
Subject: Re: phdcom
In-Reply-To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA> /
Wed 25 Jun 86 15:10:46-PDT.
I can't make any meetings this week. I was out of town today at a
funeral and am busy tomorrow and Friday at a workshop. I am quite
happy with the the proposal as it stands (semantically). Since I can't
be there in person to vote on anything, the safest thing to do is
assume that I vote in favor of every clause in the version I received
(Tuesday @ 5) -- including the one about making the requirements hold
for anyone who hasn't filed a G81. As I have already implied, I would regard
it as wimpish in the extreme for us not to demand, at least, that
anyone who hasn't filed a G34 be subject to the new requirements. I
don't mind stating that I think it reflects rather badly on the
students who are complaining about that.
Keith
∂26-Jun-86 0108 bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK re: Paper on circumscription and induction
Received: from CS.UCL.AC.UK by SU-AI.ARPA with TCP; 26 Jun 86 01:08:12 PDT
Received: from aiva.edinburgh.ac.uk by 44d.Cs.Ucl.AC.UK via Janet with NIFTP
id a010548; 26 Jun 86 8:51 BST
From: Alan Bundy <bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK>
Date: Thu, 26 Jun 86 08:50:26 -0100
Message-Id: <6163.8606260750@aiva.ed.ac.uk>
To: JMC@su-ai.arpa
Subject: re: Paper on circumscription and induction
Cc: bundy%aiva.edinburgh.ac.uk@Cs.Ucl.AC.UK
John
The issue I was interested in was the formal representation
of all the various forms of induction arising from different datastructures
and different forms of recursion, and whether Circumscription was a neat
way of representing them all (inspired by your derivation of bog-standard
induction from C). When I put this to you you said you had a paper onthe
subject. I think it was unpublished, but don't remember anything more.
Does this ring any bells?
Alan
∂26-Jun-86 0600 JMC
Call III about stock.
∂26-Jun-86 1357 ME MCC.COM
MCC.COM has shown up in the host table, so you can use that address
directly for mail now.
∂26-Jun-86 1423 WINOGRAD@SU-CSLI.ARPA Qual areas
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 26 Jun 86 14:23:31 PDT
Date: Thu 26 Jun 86 14:19:33-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Qual areas
To: healey@SU-SUSHI.ARPA
cc: phdcom@SU-AI.ARPA
Under the current version of the PhD program changes, which will go to
the faculty for vote tomorrow, the requirement is that three academic
council members with appointments in CS or EE/CSL propose the syllabus
for a new qual area, and it must be approved by the faculty as a whole.
Once it is approved, the exam can be conducted with outside help of
whatever kind seems appropriate to the three responsible faculty. The
reasoning in setting it up this way is that any area that is so
tangential to the academic program that three reasonably-related faculty
can't be found is not likely to be a valid area of academic breadth for
a Stanford CS PhD (although it may be a fine area for individual
research). In the case of robotics, this will all become more feasible
when our robotics faculty slot is filled, hopefully next year. I'd be
glad to discuss it further with you. --t
-------
∂26-Jun-86 1610 WINOGRAD@SU-CSLI.ARPA Last changes
Received: from SU-CSLI.ARPA by SU-AI.ARPA with TCP; 26 Jun 86 16:10:23 PDT
Date: Thu 26 Jun 86 16:06:09-PDT
From: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Subject: Last changes
To: cheadle@SU-SCORE.ARPA
cc: phdcom@SU-AI.ARPA, WINOGRAD@SU-CSLI.ARPA
page 8, section VA3c - "responsibile"
page 8 VA4 last sentence: "For further guidelines...." replace with:
"The Report of the Systems PhD Committee, dated April 1986, gives some
guidelines for the proposal requirement for Systems students. These
express the general orientation of the departmental requirement. The
faculty at the presentation will be responsible for determining the
exact form of written material and presentation most appropriate to the
student's area and research topic, within the overall spirit of the
guidelines.
[[[Not great, but maybe an improvement?]]]
page 12
Students who entered the PhD program before these program modifications
were approved will not be required to submit and present a dissertation
proposal. However we encourage individual advisors to recommend this to
students who are at an early stage in their research and will benefit
from doing the paper and/or presentation.
-------
∂26-Jun-86 2049 lantz@su-gregorio.arpa Re: Last changes
Received: from SU-GREGORIO.ARPA by SU-AI.ARPA with TCP; 26 Jun 86 20:47:47 PDT
Received: by su-gregorio.arpa with Sendmail; Thu, 26 Jun 86 20:45:52 pdt
Date: 26 Jun 1986 2045-PDT (Thursday)
From: Keith Lantz <lantz@su-gregorio.arpa>
To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA>
Cc: phdcom@sail
Subject: Re: Last changes
In-Reply-To: Terry Winograd <WINOGRAD@SU-CSLI.ARPA> /
Thu 26 Jun 86 16:06:09-PDT.
I cannot believe the grandfather clause has been changed as you note.
I'll give up at this point, but believe it should be made clear that
the requirement DOES hold for existing CSL students as per the "Report
of the ..." ... or, more generally, that these requirements (plural)
specify the MINIMUM set of requirements for ALL PhD students and that
individual areas may choose to impose additional requirements, etc.,
etc. Obviously, I am trying to avoid an inevitable clash down the line
when someone says "But, that isn't a CS requirement...".
Is it really true, as several people have been saying, that a draft has
not been circulated to the students at large since January????? If so,
then I guess we really did drop the ball.
Keith
∂26-Jun-86 2326 nttlab!nttyrl!nobuo@su-shasta.arpa Common Business Communication Langauge
Received: from SU-SHASTA.ARPA by SU-AI.ARPA with TCP; 26 Jun 86 23:26:26 PDT
Received: by su-shasta.arpa; Thu, 26 Jun 86 23:26:31 PDT
Received: by nttlab.ntt.junet (4.12/5.0) with TCP; Fri, 27 Jun 86 14:12:42 jst
Received: by nttlab.ntt.junet (4.12/5.0) with TCP; Fri, 27 Jun 86 14:11:42 jst
Received: by nttyrl.ntt.junet (4.12/4.8Jd)
id AA01655; Fri, 27 Jun 86 09:14:14 jst
Message-Id: <8606270014.AA01655@nttyrl.ntt.junet>
Date: Fri 27 Jun 1986 09:05:42 JST
From: Nobuo Kawashima <nttlab!nttyrl!nobuo@su-shasta.arpa>
Subject: Common Business Communication Langauge
To: nttlab!Shasta!JMC@su-ai.arpa
Cc: NTT-20!goto@su-shasta.arpa, nttyrl!nobuo@su-shasta.arpa
Dear Peof. McCarthy:
You sent me your paper, "Application of Circumscription to
Formalizing Common Knowledge," at my request. Since I wanted
to know what Common Business Communicattion Language looks like,
I tried to find your paper appeared in Textverarbeitung und
Burosysteme. However, I could not find it in Japan.
I feel badly asking you again to send me the paper.
Thank you in advance.
Regards,
Nobuo Kawashima
Commun. Proc. Programs Sect.
NTT Commun.'s & Inf. Proc. Labs.
1-2356, Take
Yokosuka, Kanagawa
238-03 JAPAN
∂27-Jun-86 0913 ROSENBLOOM@SUMEX-AIM.ARPA Re: Last changes
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 27 Jun 86 09:13:20 PDT
Date: Fri 27 Jun 86 09:10:04-PDT
From: Paul Rosenbloom <ROSENBLOOM@SUMEX-AIM.ARPA>
Subject: Re: Last changes
To: lantz@SU-GREGORIO.ARPA
cc: phdcom@SU-AI.ARPA
Message-ID: <12218184148.34.ROSENBLOOM@SUMEX-AIM.ARPA>
Keith: I was one of the people who argued strongly in the PhD
committee meeting for the change in the grandfather clause. Below are
some of the points that are relevant to the issue of grandfathering.
-- Paul
(1) Grandfathering is a fair way of dealing with the students who are
already part way through the program and have already made
preparations relative to the changed portion of the program.
(2) In discussions about changes to the programs, grandfathering
allows the students to pull up a bit from worrying about how the
changes are going to affect them, and think about what is best for the
program and the students in the abstract.
(3) It helps to preserve the sense of community among faculty and PhD
students. Changing the program significantly in mid-stream leads to
the attitude on the part of students that the faculty are doing things
to them rather than with them.
(4) Grandfathering more students does not weaken the set of
requirements in any way. It just includes some existing students with
those who have already gone through the program without the new
requirement. This is a relatively small price to pay.
(5) In general I believe in grandfathering all enrolled students
unless: (a) the change is too trivial to worry about; or (b) having
two versions would be infeasible.
-------
∂27-Jun-86 0952 TAL call todorovich
To: JMC
CC: TAL
Please call Professor Miro Todorovich at (212) 840-6595.
Taleen
∂27-Jun-86 1101 langley@ICSE.UCI.EDU machine learning workshop
Received: from ICSE.UCI.EDU by SU-AI.ARPA with TCP; 27 Jun 86 10:57:20 PDT
Received: from localhost by ICSE.UCI.EDU id a008477; 27 Jun 86 10:55 PDT
To: jmc@su-ai.arpa
cc: aaai-office@sumex-aim.arpa, langley@ICSE.UCI.EDU
Subject: machine learning workshop
Date: Fri, 27 Jun 86 10:54:27 -0800
From: Pat Langley <langley@ICSE.UCI.EDU>
John - You probably know that every two years, the machine learning
community has organized a summer workshop. The next one is scheduled
for June 1987 and will be held at UCI. Together with Jaime Carbonell,
Tom Mitchell, and Ryszard Michalski, I've been working out the details
and trying to secure funding for the meeting. I've already written up
a brief proposal that summarizes our plans, and I can forward that to
you if think it's appropriate for AAAI to help out.
We're planning on 150 attendees, which is probably larger than most
workshops, but it seems necessary if we want to be representative of the
field. Given so many people, we need considerable support just to cover
people's room and board, much less their travel costs. I think we can get
NSF IST to cover local expenses for the 30 speakers, but I'd like to give
support to some graduate students as well. We'll also need secretarial
support and funds to pay for room and audio-visual rental.
I'm not sure what you consider an appropriate budget for a AAAI workshop
proposal, which is the main reason for this message. Given the size of
the meeting, I'm tempted to ask for $10K, but will certainly ask for less
if you advise it. I'm also not sure whether I should include secretarial
support and other administrative costs, or whether to limit the budget to
travel and lodging. Finally, I'm not sure whether AAAI will cover university
overhead charges, or whether it acts more like a foundation in this respect.
I'll appreciate whatever information you can give me, and I can have the
proposal off to you very quickly if you give an encouraging response.
Sincerely, Pat Langley
P.S. If by chance you're no longer in charge of AAAI workshops, maybe you
can forward this to your successor.
∂27-Jun-86 1317 AI.WOODY@MCC.COM Hi
Received: from MCC.ARPA by SU-AI.ARPA with TCP; 27 Jun 86 13:17:13 PDT
Date: Fri 27 Jun 86 15:14:11-CDT
From: Woody Bledsoe <AI.Woody@MCC.COM>
Subject: Hi
To: jmc@SU-AI.ARPA
cc: AI.Woody@MCC.COM
Message-ID: <12218228588.25.AI.WOODY@MCC.COM>
John,
I hope that you are feeling better. You had me concerned when you
were here.
As usual we much enjoyed your visit and hope you return soon.
Woody
-------
∂27-Jun-86 1340 OTA-CIT@USC-ISI.ARPA Split in content of AI
Received: from USC-ISI.ARPA by SU-AI.ARPA with TCP; 27 Jun 86 13:39:56 PDT
Date: 27 Jun 1986 16:28:03 EDT
Subject: Split in content of AI
From: Jim Dray <OTA-CIT@USC-ISI.ARPA>
To: davis@HT.AI.MIT.EDU, feigenbaum@SUMEX-AIM.ARPA, schank@YALE.ARPA,
raj.reddy@A.CS.CMU.EDU, jmc@SU-AI.ARPA, nilsson@SU-SCORE.ARPA,
hector%utai.toronto.edu@CSNET-RELAY.ARPA, herb.simon@A.CS.CMU.EDU,
lehnert%umass@CSNET-RELAY.ARPA, forbus@P.CS.UIUC.EDU,
wilensky@DALI.BERKELEY.EDU
cc: ota@USC-ISI.ARPA, OTA-CIT@USC-ISI.ARPA
Could we get your response to a concern expressed by Jerry Feldman in
answer to the first discussion topic? His concern is that the term
AI is losing significance, and that this could have substantive effects
on the field by fragmenting it:
"The big commercialization push has been in expert systems which has
dropped what little research content it had. If the entrepreneurs succeed
in capturing the name "artificial intelligence," scientific work will need
to find a new one. Cognitive science is doing extremely well and may
take over much of the ingellectual territory. A possible split into
expert systems, cognitive science and robotics would threaten the basic
computer science core of AI and reduce the effectiveness of the field
greatly....It could well happen that AAAI becomes almost exclusively
expert systems and commercial activities. There are separate cognitive
science and vision and robotics conferences already. If the term AI loses
significance and there is no common scientific forum, you bet the
substance will be affected."
Do you agree with Feldman about the implications of the term AI losing
significance? We'd appreciate any comments.
Thanks,
Jim
-------
∂27-Jun-86 1526 SJM UK trip
I have caused Frank to make airline reservations as follows
July 12 Pan Am 124 lv SF 5:40pm arr London 11:25am July 13
July 18 British Midland 52 lv L 9am arr Ed 10:10am
July 21 British Air 4713 lv Ed 7:10am arr L 8:20am
Pam Am 125 lv L 10:30am
(Excuse duplication of first line.) These are tourist class. He claims
you have to find out about upgrades with your own personal magic numbers.
I also made reservations through the people whose number you
gave me for July 13-17. The cost, however, L40/night, seems high to me.
If you have fault to find with these arrangements, it would
probably be best to find it promptly.
Susie
∂27-Jun-86 1656 langley@CIP.UCI.EDU Re: machine learning workshop
Received: from CIP.UCI.EDU by SU-AI.ARPA with TCP; 27 Jun 86 16:56:38 PDT
Received: from cip2.uci.edu by CIP.UCI.EDU id a001201; 27 Jun 86 16:55 PDT
To: John McCarthy <JMC@su-ai.arpa>
cc: langley@CIP.UCI.EDU
Subject: Re: machine learning workshop
Date: Fri, 27 Jun 86 16:55:29 -0800
From: Pat Langley <langley@CIP.UCI.EDU>
Thanks, John. I'll write up a $10K budget and get it to you as soon as
possible. The no-overhead restriction is good news; it will make the
money go farther. I'd like to budget some of the AAAI money for areas
for which other groups (like NSF) WILL pay overhead (like secretarial
support), and let them cover categories for which they DON'T pay overhead
(like travel). This strategy will let us keep as much of the funds for
the workshop as possible. I will write up the budget this way (say 50% for
administrative support and 50% for travel) unless I hear otherwise from
you. Thanks for the rapid response and the encouragement. Pat
∂28-Jun-86 1524 NILSSON@SU-SCORE.ARPA Re: First discussion topic
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 28 Jun 86 15:24:04 PDT
Date: Sat 28 Jun 86 14:08:45-PDT
From: Nils Nilsson <NILSSON@SU-SCORE.ARPA>
Subject: Re: First discussion topic
To: OTA-CIT@USC-ISI.ARPA
cc: davis@HT.AI.MIT.EDU, tk@MC.LCS.MIT.EDU, allen.newell@A.CS.CMU.EDU,
feigenbaum@SUMEX-AIM.ARPA, schank@YALE.ARPA, rhayes-roth@SRI-KL.ARPA,
ohlander@USC-ISIB.ARPA, winograd@SU-CSLI.ARPA, waltz@GODOT.THINK.COM,
brown.pa@XEROX.COM, kay@MIT-MULTICS.ARPA, phw%mit-oz@MC.LCS.MIT.EDU,
raj.reddy@A.CS.CMU.EDU, jmc@SU-AI.ARPA,
hector%utai.toronto.edu@CSNET-RELAY.ARPA, lenat@MCC.ARPA,
hollister%ti-csl@CSNET-RELAY.ARPA, herb.simon@A.CS.CMU.EDU,
bundy%aiva.edinburgh.ac.uk@CS.UCL.AC.UK, lehnert%umass@CSNET-RELAY.ARPA,
forbus@P.CS.UIUC.EDU, wilensky@DALI.BERKELEY.EDU,
kolodner.gatech@CSNET-RELAY.ARPA, feldman@ROCHESTER.ARPA,
norman@NPRDC.ARPA, collins@BBNA.ARPA, jsl@ROCKEFELLER.ARPA,
ota@USC-ISI.ARPA, NILSSON@SU-SCORE.ARPA
In-Reply-To: Message from "Jim Dray <OTA-CIT@USC-ISI.ARPA>" of Sat 21 Jun 86 21:18:00-PDT
Message-ID: <12218500665.13.NILSSON@SU-SCORE.ARPA>
I think there has been "satisfactory progress" in basic research in AI
during the past few years. But a lot more progress has to be made
before the people who are now buying expert systems, etc., will be able
to get what they really want.
Here are the areas where I think good progress has been made:
1) nonmonotonic reasoning: several important papers, an excellent
AAAI-supported workshop, the area is attracting increasing attention
and has produced some good results.
2) reasoning about other agents: several important papers about
reasoning about knowledge and belief (Moore, Konolige, Halperin, ...),
an excellent AAAI-supported workshop, people beginning to incorporate
these ideas in proposals for new systems.
3) commonsense reasoning: work on reasoning about time, events, causation.
4) natural language processing: fundamental work on new grammars (lfg,
unification grammars), processing that takes into account the
"intentions" and plans of other agents (Cohen, Levesque), discourse
structure (Grosz), planning utterances (Appelt). Some of this work is
being done in conjunction with CSLI at Stanford
5) Reasoning with uncertain information: several new papers (Pearl,
...), AAAI-supported workshops.
I could probably name other areas too, but it seems to me that the people
who claim no basic work is being done just aren't looking. Now I wouldn't
want to claim that breakthroughs are being made, but the work seems to have
the kind of momentum that suggests that people are not running out of ideas.
I am worried about the funding situation for basic research. If progress
is slowed down, it won't be for lack of good people having good ideas to
pursue. It will be because funding sources aren't adequate to support these
people.
As far as Jerry's worry about the expert systems and applied people
"stealing our name," I'm not worried. Other scientific fields have
survived commercial exploitation of their names. I know what Artificial
Intelligence means, there are courses in Artificial Intelligence, there
are textbooks, there is a fine journal with that name. I think the name
will remain. As I said in my AAAI President's address in 1983, I think
the name will come to denote a somewhat narrower field of activity than
it did then. Then and now, there are still people claiming to do AI who
are really doing "systems" or "operations research" or "cognitive
psychology" or ..... Since AI cannot effectively defend all the
territory it has spilled into against the rightful claims to those
subject areas by the people who got their first, I think AI will come to
denote that aspect of computer science in which knowledge is represented
declaratively (by logical languages, of course!) and in which this
knowledge is manipulated by inference processes of one sort or another
(not always sound ones).
The development of "intelligent programs" is a much bigger problem than
can be dealt with by this declarative approach. Special procedures
shaped more by the niche in which they are to be applied than by the
knowledge used in them, will continue to be developed by computer
scientists generally. AI will sooner or later retreat from the task of
filling all of these niches and concentrate on the task of developing
reasoning systems that represent and reason with knowledge using
techniques that are independent of the exact use to which that knowledge
might be put.
Ideas for promoting basic research in AI: As I said, I don't think there
is a shortage of ideas. Basic research would be promoted if NSF had
a larger budget or if DARPA returned to the practice of finding good people
and funding them without being overly concerned about their "milestones"
and attempting to "manage" them.
-Nils Nilsson
-------
∂29-Jun-86 0457 100 (from: jmc on TTY44)
The u.s. version will run on a Cray-3.
∂29-Jun-86 1300 CRISPIN@SUMEX-AIM.ARPA re: Contra aid (from SAIL's BBOARD)
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 29 Jun 86 13:00:18 PDT
Date: Sun 29 Jun 86 13:00:52-PDT
From: Mark Crispin <Crispin@SUMEX-AIM.ARPA>
Subject: re: Contra aid (from SAIL's BBOARD)
To: JMC@SU-AI.ARPA
cc: SU-BBOARDS@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Sat 28 Jun 86 19:01:00-PDT
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12218750452.20.CRISPIN@SUMEX-AIM.ARPA>
A question for JMC: how is the "North Vietnamese invasion of South
Vietnam" any different from the US invasion of the Confederacy? The
US government was quite upset at aid given by the British to the
Confederacy, and after the Civil War collected reparations from
London.
It is commonly ignored (especially by friends of mine who were
officials in the South Vietnamese government) that historically the
government of the North was the legitimate government of the entire
country. Ho Chi Minh's independence movement defeated the French in
1954 (in spite of US aid to the French -- what in the name of sanity
were we doing helping the French hold on to colonies 9 years after
WWII ended?). Vietnam's partition was established by the Geneva
treaty only until nationwide elections were held. The temporary
government in the South refused, with US support, to hold these
elections, since it was obvious that Ho Chi Minh would win.
That says nothing about whether or not the government of the North
was/is a good, pleasant, free, what-have-you government. It simply
says it was the legitimate government.
-------
∂29-Jun-86 2123 ULLMAN@SU-SCORE.ARPA Database reading
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 29 Jun 86 21:23:51 PDT
Date: Sun 29 Jun 86 21:22:18-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: Database reading
To: phdcom@SU-AI.ARPA, gio@SU-AI.ARPA, tw@SU-AI.ARPA
Message-ID: <12218841734.20.ULLMAN@SU-SCORE.ARPA>
I would be comfortable eliminating from the DB reading list:
Ch. 9
Sect. 10.5
I wouldn't complain if Sect. 7.6, 8.5, 8.6, and 11.4 were dropped.
One could argue that Ch. 2 is really data structures material,
and that could be cut. If not cut, I would still be happy to
eliminate 2.8 and 2.9.
Everything else I consider rather fundamental, and I would not
like to see it eliminated.
---jeff
-------
∂29-Jun-86 2202 ullman@su-aimvax.arpa Re: Recursive negation
Received: from SU-AIMVAX.ARPA by SU-AI.ARPA with TCP; 29 Jun 86 22:02:43 PDT
Received: by su-aimvax.arpa with Sendmail; Sun, 29 Jun 86 22:01:38 pdt
Date: Sun, 29 Jun 86 22:01:38 pdt
From: Jeff Ullman <ullman@diablo>
Subject: Re: Recursive negation
To: VAL@Sail, avg@diablo
Cc: "@UTEP.[NET, JMC@Sail, VAL]"@Sail, ullman@diablo
1. Interesting. Why don't we broadcast the discussion to the nail@diablo
list?
2. I'm not happy with the example, because it assumes that "situtations"
are expressible only by terms that tell how you got to the "situation,"
and situations with identical "states" are not necessarily recognized
as the same. The consequence is that the recursion terminates because
there is a structural induction going on, just as in the "even" example.
I would prefer to express things as if "situations' were states,
and have a DB relation:
result(X,Y,S,S') = the result of moving X on top of Y in
state S is state S'
Then, your rule becomes (with ab eliminated):
on(X,Y,S') :- result(X,Y,S,S') & \+ on(Z,X,S)
Now, it is not so clear that this recursion terminates.
i think this recusion (and yours) requires a basis case, but
assuming that taken care of, there still could be cycles
in the state transition graph. You, effectively, eliminate such
cycles by not allowing something to BE a state unless you can
explain how to construct it (via a term with your r and m functions).
Fine, but that is messy computationally, since in order to
pose your program a question, I have to figure out how the state
was obtained by a sequence of moves from some [unspecified] initial
conditions.
---jeff
∂30-Jun-86 0937 OTA-CIT@USC-ISI.ARPA Second Discussion Topic
Received: from A.ISI.EDU by SU-AI.ARPA with TCP; 30 Jun 86 09:35:56 PDT
Date: 30 Jun 1986 12:21:58 EDT
Subject: Second Discussion Topic
From: Jim Dray <OTA-CIT@A.ISI.EDU>
To: davis@HT.AI.MIT.EDU, tk@MC.LCS.MIT.EDU, allen.newell@A.CS.CMU.EDU,
feigenbaum@SUMEX-AIM.ARPA, schank@YALE.ARPA, rhayes-roth@SRI-KL.ARPA,
ohlander@B.ISI.EDU, winograd@SU-CSLI.ARPA, waltz@GODOT.THINK.COM,
brown.pa@XEROX.COM, kay@MIT-MULTICS.ARPA, raj.reddy@A.CS.CMU.EDU,
jmc@SU-AI.ARPA, nilsson@SU-SCORE.ARPA,
hector%utai.toronto.edu@CSNET-RELAY.ARPA, lenat@MCC.ARPA,
hollister%ti-csl@CSNET-RELAY.ARPA, herb.simon@A.CS.CMU.EDU,
bundy%aiva.edinburgh.ac.uk@CS.UCL.AC.UK, lehnert%umass@CSNET-RELAY.ARPA,
forbus@P.CS.UIUC.EDU, wilensky@DALI.BERKELEY.EDU, kolodner@Gatech,
feldman@ROCHESTER.ARPA, norman@NPRDC.ARPA, collins@BBNA.ARPA,
lederberg@SUMEX-AIM.ARPA, weldon@NOSC-COD.ARPA, ar@ALV.UMD.EDU,
sebrecht%weslyn.bitnet@WISCVM.WISC.EDU,
hthompson%uk.ac.ed.eusip@CS.UCL.AC.UK,
aarons%cvaxa.sussex.ac.uk@CS.UCL.AC.UK
cc: ota@A.ISI.EDU, OTA-CIT@A.ISI.EDU
We've received several dozen interesting and worthwhile responses to
the first discussion topic, and we appreciate the thoughtfulness and
time that went into the responses. While we still welcome comments
on the first topic, we're going to present a new topic roughly once
each week to keep the discussion moving. The same caveats that we
included with the first topic also apply to the second.
SECOND DISCUSSION TOPIC
Most of those who responded to the first discussion topic suggested
that while commercial AI activity may be drawing some researchers away
from basic work, that effect is likely to be transient, and may even
be healthy. However, several researchers expressed a much more deep
concern about possible changes in funding for basic AI research.
Specifically, ARPA has been the patron of basic research in AI for many years,
and the AI field seems almost irreversibly committed to continuing ARPA
support. However, there are current concerns that ARPA is becoming
substantially less benign toward development of the field, and is
steering the field toward problems that have military payoffs,
even (some would argue) with funding that is tagged "basic research."
Do you agree with these concerns, and can you share any examples from
your work or your relations with funding agencies that demonstrate
such changes in priorities? How has research funding (either from
ARPA or other sources) shifted the goals of the field, both recently
and historically?
As before, there is no time limit on responses to this topic, though
the next topic will be presented in approximately one week.
Thank you,
Earl Dowdy and Jim Dray at OTA
-------
∂30-Jun-86 0948 CONTRERAS@SU-SCORE.ARPA message
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 30 Jun 86 09:43:36 PDT
Date: Mon 30 Jun 86 09:41:58-PDT
From: Tina Contreras <CONTRERAS@SU-SCORE.ARPA>
Subject: message
To: JMC@SU-AI.ARPA
Message-ID: <12218976387.20.CONTRERAS@SU-SCORE.ARPA>
Sue Corneille phoned regarding the Logic Programming Conference. She
said it is very URGENT. 1-871-2016.
Tina
-------
∂30-Jun-86 1054 TAL Chris Freiling called
To: JMC
CC: TAL
Chris Freiling can make it on July 7, but he needs info such as how long the talk
is, where it is, etc... May try to reach you later, but if possible, call (714)
887-7375. Home number is: (714)881-2291. Taleen
∂30-Jun-86 1134 KARP@SUMEX-AIM.ARPA test
Received: from SUMEX-AIM.ARPA by SU-AI.ARPA with TCP; 30 Jun 86 11:29:39 PDT
Date: Mon 30 Jun 86 11:27:14-PDT
From: Peter Karp <KARP@SUMEX-AIM.ARPA>
Subject: test
To: phdcom@SU-AI.ARPA
Message-ID: <12218995549.55.KARP@SUMEX-AIM.ARPA>
This mailing list has been avoiding me lately...
-------
∂30-Jun-86 1530 CLT WICS
To: JMC, IAM
Joleen B. called to ask about the state of materials to be
reproduced. She would like to get them Monday July 14.
I think I will use my thesis possibly minus the chapters
on comparisons and abstract machines and the draft of the derivations
paper.
Joleen will send us a copy of the info letter that is sent
to students telling when and where everything is.
Students arrive Sunday afternoon and get all the materials
If desired they can be given a reading assignment for Monday.
There are 13 enrolled sofar, and people are still enrolling.
Class room - probably History 205
Schedule 9-5
break at 10:30 with refreshments provided by WICS
lunch break 1.5 hour at around 12:15
break at 3
∂30-Jun-86 1821 G.MDP@SU-SCORE.ARPA Re: Anti-American punctuation
Received: from SU-SCORE.ARPA by SU-AI.ARPA with TCP; 30 Jun 86 18:21:37 PDT
Date: Mon 30 Jun 86 18:19:35-PDT
From: Mike Peeler <G.MDP@SU-SCORE.ARPA>
Subject: Re: Anti-American punctuation
To: JMC@SU-AI.ARPA
cc: SU-BBOARDS@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI.ARPA>" of Fri 27 Jun 86 22:03:00-PDT
Message-ID: <12219070617.54.G.MDP@SU-SCORE.ARPA>
In case I contributed to JMC's distress, I didn't mean the
disambiguation can't be done. I meant it isn't done by
punctuation.
-------
∂30-Jun-86 1846 Bobrow.pa@Xerox.COM Re: Lifschitz for editorial board
Received: from XEROX.COM by SU-AI.ARPA with TCP; 30 Jun 86 18:46:33 PDT
Received: from Cabernet.ms by ArpaGateway.ms ; 30 JUN 86 18:31:24 PDT
Date: 30 Jun 86 18:31 PDT
From: Bobrow.pa@Xerox.COM
Subject: Re: Lifschitz for editorial board
In-reply-to: John McCarthy <JMC@SU-AI.ARPA>'s message of 30 Dec 85 12:34
PST
To: JMC@SU-AI.ARPA
Message-ID: <860630-183124-3393@Xerox>
What is his address? I would like to ask him to join, and to attend the
AAAI breakfast meeting.
-- danny
∂30-Jun-86 1846 GRP graphics on apa16
To: "@EBOS.DIS[P,DOC]"@SU-AI.ARPA
I called Alex Morrow to ask whether graphics support is planned for
the APA-16; he referred me to Carl von Loewenfeldt who "can't tell" me
but referred me to Doris Stoessel. She gave me the "standard answer",
which is that IBM does not discuss anything until it is officially
announced. She was somewhat upset that I called the people there, and
said that all such inquiries should go through the customer support
(or whatever they call it) number. She did say that if another
university is working on graphics for the APA-16 they could tell us
about that, but she didn't know of any such development. Her guess
was that if IBM does provide graphics support for the APA-16, it will
probably not be compatible with the routines for the AED.