perm filename F86.IN[LET,JMC] blob
sn#832037 filedate 1987-01-08 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00668 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00071 00002
C00073 00003 ∂01-Oct-86 1641 RA new phone
C00074 00004 ∂01-Oct-86 1713 ME Prancing Pony Bill
C00076 00005 ∂01-Oct-86 1701 VAL section for proposal
C00077 00006 ∂01-Oct-86 1720 VAL Commonsense Reasoning Seminar
C00078 00007 ∂02-Oct-86 0900 JMC
C00079 00008 ∂02-Oct-86 1042 RA ofind command
C00080 00009 ∂02-Oct-86 1123 VAL reply to message
C00081 00010 ∂02-Oct-86 1234 EVE@CSLI.STANFORD.EDU Symbolic Systems Faculty Roster
C00082 00011 ∂02-Oct-86 1241 VAL
C00083 00012 ∂02-Oct-86 1446 SJG cryptograms with multiple decodings
C00084 00013 ∂02-Oct-86 1454 SJG re: cryptograms with multiple decodings
C00085 00014 ∂02-Oct-86 1554 RA leaving early
C00086 00015 ∂02-Oct-86 1605 RA ofind command
C00087 00016 ∂03-Oct-86 0318 rwsh%computer-lab.cambridge.ac.uk@Cs.Ucl.AC.UK Forwarding: The ELEPHANT Language
C00090 00017 ∂03-Oct-86 0939 RA audit a course
C00091 00018 ∂03-Oct-86 0945 VAL milestones
C00093 00019 ∂03-Oct-86 0946 CLT jjw,air,AMEHTA@Sushi.Stanford.EDU
C00094 00020 ∂03-Oct-86 0947 RA [Reply to message recvd: 23 Sep 86 19:14 Pacific Time]
C00095 00021 ∂03-Oct-86 0950 SJG cryptogram
C00096 00022 ∂03-Oct-86 1111 SJG re: cryptogram
C00097 00023 ∂03-Oct-86 1152 RWF Eric Drexler
C00098 00024 ∂03-Oct-86 1159 RA new schedule
C00099 00025 ∂03-Oct-86 1314 RA ijcai.tex
C00101 00026 ∂03-Oct-86 1434 RA would you mind?
C00102 00027 ∂03-Oct-86 1451 RA elephant paper
C00103 00028 ∂03-Oct-86 1458 RA elephant paper
C00104 00029 ∂03-Oct-86 2043 shoham@YALE.ARPA two things
C00110 00030 ∂04-Oct-86 1800 JMC
C00111 00031 ∂04-Oct-86 1820 RESTIVO@Score.Stanford.EDU picture request
C00112 00032 ∂04-Oct-86 1926 BRINK@Sushi.Stanford.EDU Tenure track
C00114 00033 ∂05-Oct-86 0456 @DIAMOND.S4CC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM Knotted Doughnuts
C00122 00034 ∂06-Oct-86 1255 BRINK@Sushi.Stanford.EDU re: Tenure track
C00124 00035 ∂06-Oct-86 1346 ANDY@Sushi.Stanford.EDU Re: Einstein, etc.
C00125 00036 ∂06-Oct-86 1418 RA going to the bookstore
C00126 00037 ∂06-Oct-86 1500 JMC
C00127 00038 ∂06-Oct-86 1529 REGES@Score.Stanford.EDU Machines for CS306
C00129 00039 ∂06-Oct-86 1631 REGES@Score.Stanford.EDU re: Machines for CS306
C00131 00040 ∂06-Oct-86 1653 PANDREW@Score.Stanford.EDU CS306
C00133 00041 ∂06-Oct-86 1804 CLT shopping list
C00134 00042 ∂06-Oct-86 1851 R.REULING@LEAR.STANFORD.EDU re: Classes using LOTS
C00136 00043 ∂06-Oct-86 1939 LES Arkady
C00137 00044 ∂06-Oct-86 2013 PHYSICSLIB@Sierra.Stanford.EDU Meeting next Monday
C00140 00045 ∂06-Oct-86 2225 cramer@Sun.COM Paly International Festival
C00144 00046 ∂06-Oct-86 2250 cheriton@pescadero.stanford.edu Postdocs
C00146 00047 ∂07-Oct-86 0900 JMC
C00147 00048 ∂07-Oct-86 1041 AMEHTA@Sushi.Stanford.EDU Appearance in class
C00148 00049 ∂07-Oct-86 1226 JOSHI@cis.upenn.edu Workshop
C00154 00050 ∂07-Oct-86 1325 UDI@Score.Stanford.EDU Re: error
C00155 00051 ∂07-Oct-86 1442 CLT Qlisp
C00156 00052 ∂07-Oct-86 1533 AMEHTA@Sushi.Stanford.EDU TA Offices
C00157 00053 ∂07-Oct-86 1556 PHYSICSLIB@Sierra.Stanford.EDU re: Meeting next Monday
C00158 00054 ∂07-Oct-86 1626 AMEHTA@Sushi.Stanford.EDU Console time on LOTS
C00159 00055 ∂07-Oct-86 1634 HSU@Sushi.Stanford.EDU cs306
C00160 00056 ∂07-Oct-86 1738 @CSNET-RELAY.ARPA,@utterly.ai.toronto.edu:utterly.ai!hector McDermott again
C00164 00057 ∂07-Oct-86 1851 SINGH@Sierra.Stanford.EDU Many thanks for your response.
C00166 00058 ∂07-Oct-86 2342 FERNANDO@Sushi.Stanford.EDU RAship
C00171 00059 ∂08-Oct-86 0723 OTA-CIT@A.ISI.EDU Followup--Comments on Conference
C00174 00060 ∂08-Oct-86 0809 shoham@YALE.ARPA
C00176 00061 ∂08-Oct-86 0847 RA [Reply to message recvd: 07 Oct 86 23:02 Pacific Time]
C00179 00062 ∂08-Oct-86 1015 JMC
C00180 00063 ∂08-Oct-86 1046 porter%wldwst.DEC@decwrl.DEC.COM Student Advising Hours
C00182 00064 ∂08-Oct-86 1058 AMEHTA@Sushi.Stanford.EDU Balance
C00184 00065 ∂08-Oct-86 1132 RA TEXing
C00185 00066 ∂08-Oct-86 1156 SANKAR@Score.Stanford.EDU Compass, Gyros etc.
C00188 00067 ∂08-Oct-86 1224 SJG so long sucker?
C00189 00068 ∂08-Oct-86 1550 RA Nancy Staggs
C00190 00069 ∂08-Oct-86 1745 1F1BROWN%UKANVAX.BITNET@Forsythe.Stanford.Edu postponement
C00192 00070 ∂08-Oct-86 1945 BRINK@Sushi.Stanford.EDU CS306
C00194 00071 ∂08-Oct-86 2331 GLB
C00195 00072 ∂09-Oct-86 0924 BRINK@Sushi.Stanford.EDU re: CS306
C00197 00073 ∂09-Oct-86 0932 VAL McDermott again
C00198 00074 ∂09-Oct-86 0941 VAL re: RAship
C00199 00075 ∂09-Oct-86 1448 RA Nancy Skaggs
C00200 00076 ∂09-Oct-86 1602 THEODOROU@Sushi.Stanford.EDU Grade for CS206/306
C00202 00077 ∂10-Oct-86 0059 CLT arpa proposal
C00203 00078 ∂10-Oct-86 1213 RA CS306 tutors
C00204 00079 ∂10-Oct-86 1610 LES Robotics Exhibit Loan
C00208 00080 ∂10-Oct-86 1627 THEODOROU@Sushi.Stanford.EDU re: Grade for CS206/306
C00209 00081 ∂10-Oct-86 1647 RA Monday
C00210 00082 ∂12-Oct-86 2128 binford@su-whitney.arpa Robotics Exhibit Loan
C00212 00083 ∂13-Oct-86 1010 RA CS 306
C00213 00084 ∂13-Oct-86 1100 JMC
C00214 00085 ∂13-Oct-86 1139 RA gps[f86,jmc]
C00215 00086 ∂13-Oct-86 1212 RA going out
C00216 00087 ∂13-Oct-86 1357 RA Academician Ershov
C00217 00088 ∂13-Oct-86 1429 GLB Re: cs 306
C00219 00089 ∂13-Oct-86 1624 LIBRARY@Score.Stanford.EDU Tech Reports
C00221 00090 ∂13-Oct-86 1833 HSU@Sushi.Stanford.EDU cs306
C00223 00091 ∂14-Oct-86 0715 AI.PETRIE@MCC.COM Emycin Criticism
C00225 00092 ∂14-Oct-86 0800 EC.MA1@Forsythe.Stanford.Edu Math/CS": Socrates display
C00228 00093 ∂14-Oct-86 0803 AI.PETRIE@MCC.COM re: Emycin Criticism
C00229 00094 ∂14-Oct-86 0849 RA be in late
C00230 00095 ∂14-Oct-86 0907 VAL re: McDermott
C00231 00096 ∂14-Oct-86 0917 SJG what is the qualification problem?
C00233 00097 ∂14-Oct-86 1127 RINDFLEISCH@SUMEX-AIM.ARPA Radford AI Mkt Survey Input
C00241 00098 ∂14-Oct-86 1130 RA Nancy Staggs
C00242 00099 ∂14-Oct-86 1216 gref@nrl-aic AAAI support
C00245 00100 ∂14-Oct-86 1220 VAL semi-annual context talk
C00246 00101 ∂14-Oct-86 1226 RA Chris Rogers
C00247 00102 ∂14-Oct-86 1443 AMEHTA@Sushi.Stanford.EDU New TA hours
C00248 00103 ∂14-Oct-86 1451 DEK Ershov visit
C00249 00104 ∂14-Oct-86 1517 RA re AAAI workshop
C00250 00105 ∂14-Oct-86 1710 RA coming in late
C00251 00106 ∂14-Oct-86 2015 LES CS300 lecture
C00252 00107 ∂15-Oct-86 0900 JMC
C00253 00108 ∂15-Oct-86 0900 JMC
C00254 00109 ∂15-Oct-86 0956 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
C00257 00110 ∂15-Oct-86 1000 B.BELGIUM@OTHELLO.STANFORD.EDU Re: LOTS allocation for EKL
C00259 00111 ∂15-Oct-86 1022 SJG qualification problem problem
C00260 00112 ∂15-Oct-86 1124 CLT Qlisp reminder
C00261 00113 ∂15-Oct-86 1257 SJG re: what is the qualification problem?
C00263 00114 ∂15-Oct-86 1448 SJG qualification problem
C00264 00115 ∂15-Oct-86 1600 JMC
C00265 00116 ∂16-Oct-86 0934 ACT
C00266 00117 ∂16-Oct-86 1725 RA Joshi, Univ. of Pennsylvania
C00267 00118 ∂17-Oct-86 0022 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM difference of squair routes
C00269 00119 ∂17-Oct-86 0207 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM bleary-eyed guess:
C00271 00120 ∂17-Oct-86 0821 PHY Janos Komlos of UCSD
C00274 00121 ∂17-Oct-86 0900 JMC
C00275 00122 ∂17-Oct-86 1322 VAL draft available
C00276 00123 ∂17-Oct-86 1454 VAL non-monotonic seminar
C00277 00124 ∂17-Oct-86 1627 AIR
C00278 00125 ∂17-Oct-86 1810 UDI@Score.Stanford.EDU program
C00284 00126 ∂18-Oct-86 0417 @DIAMOND.S4CC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM sqrt-sqrt, I'm wet.
C00287 00127 ∂18-Oct-86 1217 CLT rivin
C00288 00128 ∂18-Oct-86 1345 shoham-yoav@YALE.ARPA Re: inaccurate history
C00290 00129 ∂18-Oct-86 1525 CERF@A.ISI.EDU [CERF@A.ISI.EDU: Next Digital Library Meeting]
C00298 00130 ∂19-Oct-86 0334 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM would you believe just damp?
C00302 00131 ∂19-Oct-86 0413 CERF@A.ISI.EDU re: [CERF@A.ISI.EDU: Next Digital Library Meeting]
C00304 00132 ∂20-Oct-86 1050 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu Sorry, can't make it on Friday
C00306 00133 ∂20-Oct-86 1141 VAL Commonsense and Non-Monotonic Reasoning Seminar
C00309 00134 ∂20-Oct-86 1456 guibas@decwrl.DEC.COM Komlos
C00312 00135 ∂20-Oct-86 1608 Mailer failed mail returned
C00314 00136 ∂20-Oct-86 1730 NORDMARK@Sushi.Stanford.EDU Re: Gene Golub said you might help
C00316 00137 ∂20-Oct-86 1748 shoham-yoav@YALE.ARPA Re: inclusion in proposal
C00319 00138 ∂20-Oct-86 1803 GLB Re: CS 306
C00321 00139 ∂20-Oct-86 1845 GLB
C00322 00140 ∂21-Oct-86 0003 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM Halmos
C00324 00141 ∂21-Oct-86 0813 BRONSTEIN@Score.Stanford.EDU re: Airline REGULATORY organization?
C00325 00142 ∂21-Oct-86 0900 JMC
C00326 00143 ∂21-Oct-86 1111 RA 11:00 appointment
C00327 00144 ∂21-Oct-86 1122 NORDMARK@Sushi.Stanford.EDU re: Gene Golub said you might help
C00328 00145 ∂21-Oct-86 1717 roy@navajo.stanford.edu Your copy of "The New Republic".
C00329 00146 ∂22-Oct-86 0810 shoham@YALE.ARPA Re: addendum
C00333 00147 ∂22-Oct-86 0900 JMC
C00334 00148 ∂22-Oct-86 0939 CLT Qlisp reminder
C00335 00149 ∂22-Oct-86 1028 CLT Qlisp reminder
C00336 00150 ∂22-Oct-86 1251 shoham@YALE.ARPA proposal
C00357 00151 ∂22-Oct-86 1400 M.MMMIKEY@HAMLET.STANFORD.EDU re: Phony Status Symbols
C00359 00152 ∂22-Oct-86 1434 ILAN@Score.Stanford.EDU Automatic Status
C00361 00153 ∂22-Oct-86 1542 CRISPIN@SUMEX-AIM.ARPA re: Phony Status Symbols
C00363 00154 ∂22-Oct-86 1600 JMC
C00364 00155 ∂22-Oct-86 1800 JMC
C00365 00156 ∂23-Oct-86 0100 JMC
C00366 00157 ∂23-Oct-86 0949 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
C00369 00158 ∂23-Oct-86 1208 PERLAKI@Score.Stanford.EDU message
C00370 00159 ∂23-Oct-86 1214 RA Invitation to speak
C00371 00160 ∂23-Oct-86 1232 UDI@Score.Stanford.EDU RE: partial solution
C00373 00161 ∂23-Oct-86 1649 MATHY@Score.Stanford.EDU People count
C00374 00162 ∂23-Oct-86 2000 JMC
C00375 00163 ∂23-Oct-86 2102 YOM Thanks
C00376 00164 ∂24-Oct-86 0153 NILSSON@Score.Stanford.EDU Binford
C00378 00165 ∂24-Oct-86 1037 RA Amir Naqwi's orals
C00379 00166 ∂24-Oct-86 1158 CRISPIN@SUMEX-AIM.ARPA cretinism
C00388 00167 ∂24-Oct-86 1330 VAL Judea Pearl on causal reasoning
C00393 00168 ∂24-Oct-86 1355 BSCOTT@Score.Stanford.EDU Binford's file
C00395 00169 ∂24-Oct-86 1426 BSCOTT@Score.Stanford.EDU re: Binford's file
C00396 00170 ∂24-Oct-86 1428 RICHARDSON@Score.Stanford.EDU Binford
C00398 00171 ∂24-Oct-86 1429 RICHARDSON@Score.Stanford.EDU Binford
C00399 00172 ∂24-Oct-86 1501 RICHARDSON@Score.Stanford.EDU re: Binford
C00400 00173 ∂24-Oct-86 1704 VAL Commonsense and Non-Monotonic Reasoning Seminar
C00402 00174 ∂24-Oct-86 1740 CERF@A.ISI.EDU Nov 7-9 meeting - update
C00405 00175 ∂25-Oct-86 0452 HSU@Sushi.Stanford.EDU Re: question
C00407 00176 ∂25-Oct-86 1025 ASHOK@Sushi.Stanford.EDU [John McCarthy <JMC@SAIL.STANFORD.EDU>: Conway problem]
C00410 00177 ∂25-Oct-86 1203 CRISPIN@SUMEX-AIM.ARPA Re: textbooks
C00412 00178 ∂25-Oct-86 1637 CRISPIN@SUMEX-AIM.ARPA re: textbooks
C00416 00179 ∂26-Oct-86 0900 JMC
C00417 00180 ∂26-Oct-86 1221 vijay@ernie.Berkeley.EDU AI DISC: Searle replies to Hofstadter
C00435 00181 ∂26-Oct-86 1606 CLT
C00436 00182 ∂26-Oct-86 1609 CLT
C00437 00183 ∂26-Oct-86 2137 CLT
C00438 00184 ∂26-Oct-86 2336 vijay@ernie.Berkeley.EDU AI DISC: A restatement of the Dreyfus Position
C00515 00185 ∂27-Oct-86 0100 JMC
C00516 00186 ∂27-Oct-86 0900 JMC
C00517 00187 ∂27-Oct-86 1042 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu lunch this week?
C00519 00188 ∂27-Oct-86 1353 AIR lectures
C00520 00189 ∂27-Oct-86 1428 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu Re: re: lunch this week?
C00522 00190 ∂27-Oct-86 1424 MRC cellular
C00524 00191 ∂27-Oct-86 1604 TEICH@Sushi.Stanford.EDU [David Teich <TEICH@Sushi.Stanford.EDU>: degree plan]
C00528 00192 ∂28-Oct-86 0651 JOSHI@cis.upenn.edu supporting your proposal
C00530 00193 ∂28-Oct-86 0800 JMC
C00531 00194 ∂28-Oct-86 1027 VAL reply to message
C00532 00195 ∂28-Oct-86 1045 VAL re: reply to message
C00533 00196 ∂28-Oct-86 1100 JMC
C00534 00197 ∂28-Oct-86 1137 TAJNAI@Score.Stanford.EDU Inference Corporation
C00536 00198 ∂28-Oct-86 1151 AI.ELLIE@MCC.COM group picture
C00537 00199 ∂28-Oct-86 1422 AI.ELLIE@MCC.COM ERROR
C00538 00200 ∂28-Oct-86 1500 JMC
C00539 00201 ∂28-Oct-86 1857 RWF
C00544 00202 ∂28-Oct-86 1856 POSER@CSLI.STANFORD.EDU Re: parent suit
C00547 00203 ∂29-Oct-86 0549 JUSTESON@Sushi.Stanford.EDU Core courses possibly to waive
C00549 00204 ∂29-Oct-86 0929 CLT Qlisp meeting reminder
C00550 00205 ∂29-Oct-86 1155 RA Dr. Bloom
C00551 00206 ∂29-Oct-86 1623 CLT arpa proposal
C00552 00207 ∂29-Oct-86 1708 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
C00554 00208 ∂29-Oct-86 1730 cramer@Sun.COM Paly International Festival
C00559 00209 ∂29-Oct-86 1852 VAL Kheifets
C00560 00210 ∂29-Oct-86 2100 ASHOK@Sushi.Stanford.EDU Re: Parent suit
C00562 00211 ∂29-Oct-86 2131 RWF
C00568 00212 ∂30-Oct-86 0134 CERF@A.ISI.EDU Digital Library Meeting
C00570 00213 ∂30-Oct-86 0140 CERF@A.ISI.EDU January meeting
C00571 00214 ∂30-Oct-86 0224 JUSTESON@Sushi.Stanford.EDU Courses to waive
C00572 00215 ∂30-Oct-86 1356 UDI@Score.Stanford.EDU CFL
C00573 00216 ∂30-Oct-86 1615 RA leaving
C00574 00217 ∂30-Oct-86 2154 NILSSON@Score.Stanford.EDU Binford Letter
C00579 00218 ∂30-Oct-86 2203 cheriton@pescadero.stanford.edu Room switching
C00580 00219 ∂30-Oct-86 2234 FULLERTON@Sierra.Stanford.EDU Interview
C00582 00220 ∂30-Oct-86 2236 UDI@Score.Stanford.EDU Re: another bug
C00584 00221 ∂31-Oct-86 0127 CERF@A.ISI.EDU re: Digital Library Meeting
C00586 00222 ∂31-Oct-86 0836 FULLERTON@Sierra.Stanford.EDU re: Interview
C00588 00223 ∂31-Oct-86 0900 JMC
C00589 00224 ∂31-Oct-86 0921 alliant!alliant.Alliant.COM!jat%mit-eddie.UUCP@harvard.harvard.edu.HARVARD.EDU visit
C00591 00225 ∂31-Oct-86 1018 NILSSON@Score.Stanford.EDU re: Binford Letter
C00592 00226 ∂31-Oct-86 1111 VAL Commonsense and Non-Monotonic Reasoning Seminar
C00593 00227 ∂31-Oct-86 1335 CERF@A.ISI.EDU Reservations
C00595 00228 ∂31-Oct-86 1435 POSER@CSLI.STANFORD.EDU Re: Is this stupid propaganda?
C00596 00229 ∂31-Oct-86 1442 CLT levy
C00597 00230 ∂31-Oct-86 1454 CLT tonight
C00598 00231 ∂31-Oct-86 1530 RA Inference meeting
C00599 00232 ∂31-Oct-86 1732 POSER@CSLI.STANFORD.EDU re: Is this stupid propaganda?
C00601 00233 ∂31-Oct-86 2026 GOLDBERG@CSLI.STANFORD.EDU re: Is this stupid propaganda?
C00606 00234 ∂31-Oct-86 2139 ASHOK@Sushi.Stanford.EDU Irrational Punch
C00613 00235 ∂31-Oct-86 2204 ASHOK@Sushi.Stanford.EDU Disjoint triangles in the plane
C00615 00236 ∂01-Nov-86 0716 DEK
C00616 00237 ∂01-Nov-86 1044 NILSSON@Score.Stanford.EDU AI Quals
C00619 00238 ∂01-Nov-86 1134 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu Lunch and more
C00621 00239 ∂01-Nov-86 1817 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu lunch
C00623 00240 ∂01-Nov-86 2330 CLT rivin
C00625 00241 ∂01-Nov-86 2340 CLT
C00626 00242 ∂02-Nov-86 1043 CLT
C00627 00243 ∂02-Nov-86 1927 binford@su-whitney.arpa AI Quals
C00629 00244 ∂03-Nov-86 1129 cramer@Sun.COM re: Paly International Festival
C00639 00245 ∂03-Nov-86 1219 coraki!pratt@Sun.COM Programming Languages
C00642 00246 ∂03-Nov-86 1319 YM
C00643 00247 ∂03-Nov-86 1401 WINOGRAD@CSLI.STANFORD.EDU Quals
C00646 00248 ∂03-Nov-86 1353 CLT calendar item
C00647 00249 ∂03-Nov-86 1512 G.BEESON@CSLI.STANFORD.EDU regrets
C00649 00250 ∂03-Nov-86 1515 GLB homework
C00650 00251 ∂03-Nov-86 1616 RA leaving at 4:30
C00654 00252 ∂04-Nov-86 0640 AI.BOYER@MCC.COM visit
C00655 00253 ∂04-Nov-86 0801 ZM Re: Programming Languages
C00657 00254 ∂04-Nov-86 1107 AIR EBOS
C00658 00255 ∂04-Nov-86 1144 RA vacation
C00659 00256 ∂04-Nov-86 1155 JUTTA@Score.Stanford.EDU advisor signature
C00661 00257 ∂04-Nov-86 1403 GLB homework
C00662 00258 ∂04-Nov-86 1417 RA correction on vacation dates
C00663 00259 ∂04-Nov-86 1423 RA Trip to Virginia
C00664 00260 ∂04-Nov-86 1515 ME Prancing Pony Bill
C00666 00261 ∂04-Nov-86 1526 CERF@A.ISI.EDU Confirmation Numbers for DC trip
C00668 00262 ∂04-Nov-86 1629 CLT Qlisp meeting reminder
C00669 00263 ∂04-Nov-86 1741 cramer@Sun.COM re: Paly International Festival
C00671 00264 ∂04-Nov-86 1849 vijay@ernie.Berkeley.EDU Statements...
C00672 00265 ∂04-Nov-86 2133 CLT friday's meeting
C00673 00266 ∂04-Nov-86 2359 coraki!pratt@Sun.COM lunch
C00675 00267 ∂05-Nov-86 0602 DEK Ershov
C00676 00268 ∂05-Nov-86 0900 UDI@Score.Stanford.EDU re: game of life
C00678 00269 ∂05-Nov-86 1112 RA Maddalana
C00679 00270 ∂05-Nov-86 1339 DEK right
C00680 00271 ∂05-Nov-86 1742 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
C00681 00272 ∂05-Nov-86 1750 VAL Ershov
C00682 00273 ∂05-Nov-86 1940 LES DARPA budget
C00687 00274 ∂06-Nov-86 1155 vijay@ernie.Berkeley.EDU AI DISC: HOFSTADTER REPLY TO SEARLE
C00692 00275 ∂06-Nov-86 1230 CLT igor
C00697 00276 ∂06-Nov-86 1430 RA proposal
C00698 00277 ∂06-Nov-86 1608 RA trip back
C00699 00278 ∂06-Nov-86 1617 RA leaving
C00700 00279 ∂06-Nov-86 1626 HSU@Sushi.Stanford.EDU discussion
C00702 00280 ∂06-Nov-86 2021 HSU@Sushi.Stanford.EDU re: discussion
C00703 00281 ∂06-Nov-86 2204 HSU@Sushi.Stanford.EDU re: discussion
C00704 00282 ∂07-Nov-86 0026 CERF@A.ISI.EDU Meeting This weekend
C00707 00283 ∂07-Nov-86 0600 JMC
C00708 00284 ∂07-Nov-86 0936 JJW EKL on Sushi
C00709 00285 ∂08-Nov-86 1355 CLT calendar item
C00710 00286 ∂08-Nov-86 1750 SJG Frame problem workshop
C00712 00287 ∂08-Nov-86 1836 JJW Qlisp questions (long message)
C00726 00288 ∂08-Nov-86 2127 kuo@eniac.seas.upenn.edu Happy Birthday to Timothy!
C00728 00289 ∂08-Nov-86 2315 BAUDINET@Sushi.Stanford.EDU re: death valley: question
C00729 00290 ∂09-Nov-86 0919 VAL
C00730 00291 ∂09-Nov-86 1234 vijay@ernie.Berkeley.EDU AI DISC: DEADLINE
C00732 00292 ∂09-Nov-86 1633 GLB
C00734 00293 ∂09-Nov-86 1705 coraki!pratt@sun.com fndsch@navajo = Foundations search
C00736 00294 ∂10-Nov-86 1027 CLT message for Ershov
C00737 00295 ∂10-Nov-86 1108 VAL Commonsense and Non-Monotonic Reasoning Seminar
C00738 00296 ∂10-Nov-86 1141 coraki!pratt@sun.com John Mitchell
C00740 00297 ∂10-Nov-86 1420 RA Heftler letter
C00741 00298 ∂10-Nov-86 1617 UDI@Score.Stanford.EDU israeli companies
C00742 00299 ∂10-Nov-86 1822 G.GORIN@HAMLET.STANFORD.EDU invitation to lunch
C00744 00300 ∂10-Nov-86 1833 HSU@Sushi.Stanford.EDU EKL demo
C00745 00301 ∂10-Nov-86 2226 JJW Qlisp questions (continued)
C00749 00302 ∂10-Nov-86 2306 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM stanford indians
C00751 00303 ∂11-Nov-86 0908 VAL re: Who is Frolov?
C00752 00304 ∂11-Nov-86 0910 CLT igor
C00753 00305 ∂11-Nov-86 0927 BOAZ@CSLI.STANFORD.EDU
C00754 00306 ∂11-Nov-86 1153 CLT
C00765 00307 ∂11-Nov-86 1239 RLG job
C00766 00308 ∂11-Nov-86 1445 RA leaving?
C00767 00309 ∂12-Nov-86 0957 RA Marylin Salmansohn, ACM
C00768 00310 ∂12-Nov-86 1004 JIMENEZ@Score.Stanford.EDU message
C00769 00311 ∂12-Nov-86 1054 PERLAKI@Score.Stanford.EDU message
C00770 00312 ∂12-Nov-86 1100 JMC
C00771 00313 ∂12-Nov-86 1117 RA leaving
C00772 00314 ∂12-Nov-86 1307 VAL re: Here's a draft.
C00773 00315 ∂12-Nov-86 1627 CLT Qlisp meeting reminder
C00774 00316 ∂12-Nov-86 1650 6058598%PUCC.PRINCETON.EDU@WISCVM.WISC.EDU SDI debate
C00785 00317 ∂12-Nov-86 1711 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
C00786 00318 ∂12-Nov-86 1734 HADDAD@sushi.stanford.edu procedural questions about the faculty search
C00789 00319 ∂12-Nov-86 1830 ullman@navajo.stanford.edu Re: procedural questions about the faculty search
C00791 00320 ∂13-Nov-86 0907 TAJNAI@Score.Stanford.EDU Nobel Prize
C00792 00321 ∂13-Nov-86 1502 WIEDERHOLD@SUMEX-AIM.ARPA John Sowa - Industrial lecturer Fall quarter 87/88
C00794 00322 ∂13-Nov-86 1517 CERF@A.ISI.EDU Thanks
C00796 00323 ∂13-Nov-86 1557 RA leaving
C00797 00324 ∂13-Nov-86 2230 BRINK@Sushi.Stanford.EDU Thanks
C00799 00325 ∂14-Nov-86 0538 CERF@A.ISI.EDU re: Thanks
C00802 00326 ∂14-Nov-86 1253 RA Walter Rosenblith
C00803 00327 ∂14-Nov-86 1331 shoham@YALE.ARPA update
C00805 00328 ∂14-Nov-86 1421 CLT calendar item
C00806 00329 ∂14-Nov-86 1446 AI.NOVAK@R20.UTEXAS.EDU Lisp and Ada
C00812 00330 ∂14-Nov-86 1624 RA Staff TGIF
C00813 00331 ∂14-Nov-86 1636 shoham@YALE.ARPA re: update
C00815 00332 ∂15-Nov-86 0900 JMC
C00816 00333 ∂15-Nov-86 0900 JMC
C00817 00334 ∂15-Nov-86 1334 ILAN@Score.Stanford.EDU re: saying for today
C00818 00335 ∂15-Nov-86 1557 GOLDBERG@CSLI.STANFORD.EDU Re: Saying for today
C00820 00336 ∂15-Nov-86 1702 CLT
C00821 00337 ∂15-Nov-86 1824 WIEDERHOLD@SUMEX-AIM.ARPA Request for sabbatical style Leave-of-absence
C00824 00338 ∂15-Nov-86 2134 niz@lindy.STANFORD.EDU forsythe connection... crash recovery
C00826 00339 ∂15-Nov-86 2145 niz@lindy.STANFORD.EDU Ok... I took ttyj5 out of the tty list
C00827 00340 ∂16-Nov-86 1250 CLT tonight
C00828 00341 ∂17-Nov-86 0900 CLT
C00829 00342 ∂17-Nov-86 0917 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu lunch this week
C00831 00343 ∂17-Nov-86 0926 wbm@su-whitney.arpa Thanks for Reply: integrals
C00832 00344 ∂17-Nov-86 1024 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu Re: re: lunch this week
C00834 00345 ∂17-Nov-86 1024 RA Paula Simons, Communication Dept.
C00835 00346 ∂17-Nov-86 1037 VAL Commonsense and Non-Monotonic Reasoning Seminar
C00838 00347 ∂17-Nov-86 1111 RA Re: inclusion
C00839 00348 ∂17-Nov-86 1241 RA your trip to Virginia, Nov. 7
C00840 00349 ∂17-Nov-86 1401 RA expenses for Kahn
C00841 00350 ∂17-Nov-86 1403 RA new room for the seminar
C00842 00351 ∂17-Nov-86 1410 RA Ellie Gray
C00843 00352 ∂17-Nov-86 1449 ILAN@Score.Stanford.EDU Psych paper
C00854 00353 ∂18-Nov-86 0111 rivin@STONY-BROOK.SCRC.Symbolics.COM teaching
C00857 00354 ∂18-Nov-86 0925 VAL re: new room for the seminar
C00858 00355 ∂18-Nov-86 0943 RA Ellie Gray
C00859 00356 ∂18-Nov-86 1030 CLT msg from SF (missing a J)
C00862 00357 ∂18-Nov-86 1259 AAAI-OFFICE@SUMEX-AIM.ARPA workshop request
C00864 00358 ∂18-Nov-86 1510 VAL Przymusinski
C00865 00359 ∂18-Nov-86 1713 VAL re: Przymusinski
C00866 00360 ∂18-Nov-86 1748 @REAGAN.AI.MIT.EDU:JCMA@AI.AI.MIT.EDU Seminar - Formalizing the Notion of Context (SU)
C00869 00361 ∂18-Nov-86 2115 FULLERTON@Sierra.Stanford.EDU TV
C00870 00362 ∂19-Nov-86 1300 RA Binford
C00871 00363 ∂19-Nov-86 1306 RA Ellie Gray
C00872 00364 ∂19-Nov-86 1634 VAL New Place for the Commonsense Reasoning Seminar
C00875 00365 ∂19-Nov-86 1642 OKUNO@SUMEX-AIM.ARPA Catch and Qcatch
C00878 00366 ∂19-Nov-86 2020 FULLERTON@Sierra.Stanford.EDU re: TV
C00880 00367 ∂19-Nov-86 2058 CHANDRA%OSU-20@ohio-state.ARPA Workshop on High Levl Tools
C00883 00368 ∂20-Nov-86 0747 JJW Re: Catch and Qcatch
C00884 00369 ∂20-Nov-86 0902 MAYR@Score.Stanford.EDU Re: industry lecturers and visiting faculty for 1987-88
C00886 00370 ∂20-Nov-86 1006 CHANDRA%OSU-20@ohio-state.ARPA re: Workshop on High Levl Tools
C00888 00371 Nils:
C00890 00372 ∂20-Nov-86 1140 @REAGAN.AI.MIT.EDU:JCMA@AI.AI.MIT.EDU re: Seminar - Formalizing the Notion of Context (SU)
C00892 00373 ∂20-Nov-86 1157 VAL re: Przymusinski
C00893 00374 ∂20-Nov-86 1346 RLG seminar
C00894 00375 ∂20-Nov-86 1408 RICHARDSON@Score.Stanford.EDU CSD Booklet
C00896 00376 ∂20-Nov-86 1424 RA to do before I leave?
C00897 00377 ∂20-Nov-86 1456 CHIOU@STAR.STANFORD.EDU request funding for Space Station Automation Workshop
C00900 00378 ∂20-Nov-86 1510 WINOGRAD@CSLI.STANFORD.EDU re: industry lecturers and visiting faculty for 1987-88
C00901 00379 ∂20-Nov-86 1529 CLT Qlisp meeting reminder
C00902 00380 ∂20-Nov-86 1549 RLG notes on context seminar
C00903 00381 ∂21-Nov-86 0049 ILAN@Score.Stanford.EDU isolde
C00908 00382 ∂21-Nov-86 0741 RICHARDSON@Score.Stanford.EDU re: CSD Booklet
C00909 00383 ∂21-Nov-86 0923 MAYR@Score.Stanford.EDU industrial lectureship
C00912 00384 ∂21-Nov-86 1044 VAL Counterexample
C00913 00385 ∂21-Nov-86 1233 AIR RT and ANDREW
C00914 00386 ∂22-Nov-86 1415 RLG contexts notes
C00915 00387 ∂23-Nov-86 1659 NILSSON@Score.Stanford.EDU AI Quals
C00917 00388 ∂23-Nov-86 2203 hpm@rover.ri.cmu.edu "Mind Children" book
C00919 00389 ∂24-Nov-86 0742 CLT igor
C00922 00390 ∂24-Nov-86 1208 ceci@portia office hours
C00924 00391 ∂24-Nov-86 1944 LES DARPA Proposal
C00927 00392 ∂24-Nov-86 2131 CLT DARPA Proposal
C00928 00393 ∂24-Nov-86 2209 OKUNO@SUMEX-AIM.ARPA One more question and Re: Qcatch and Catch
C00931 00394 ∂24-Nov-86 2227 @RELAY.CS.NET,@utterly.ai.toronto.edu:hector@AI.AI.MIT.EDU Update on McDermott critique
C00934 00395 ∂24-Nov-86 2334 HSU@Sushi.Stanford.EDU cs306
C00935 00396 ∂25-Nov-86 0206 binford@su-whitney.arpa references
C00936 00397 ∂25-Nov-86 0538 JJW Process switching
C00939 00398 ∂25-Nov-86 0837 AI.JMC@MCC.COM Franklin's whale
C00941 00399 ∂25-Nov-86 1230 shoham@YALE.ARPA Re: reference to chronological minimization
C00943 00400 ∂25-Nov-86 1403 AAAI-OFFICE@SUMEX-AIM.ARPA Mathilda R. Wilson Foundation
C00946 00401 ∂26-Nov-86 0629 AI.WOODY@MCC.COM Re: Mathilda R. Wilson Foundation
C00948 00402 ∂26-Nov-86 0847 M.MRIZEK@LEAR.STANFORD.EDU LOTS 10th Birthday Celebration
C00950 00403 ∂26-Nov-86 1139 rms@prep.ai.mit.edu
C00952 00404 ∂26-Nov-86 1158 RICHARDSON@Score.Stanford.EDU AI Courses
C00953 00405 ∂26-Nov-86 1316 AAAI-OFFICE@SUMEX-AIM.ARPA Re: new Dahl proposal
C00955 00406 ∂26-Nov-86 1537 rms@prep.ai.mit.edu reply to message
C00956 00407 ∂26-Nov-86 1537 rms@prep.ai.mit.edu reply to message
C00960 00408 ∂26-Nov-86 2000 JMC
C00961 00409 ∂26-Nov-86 2100 JMC
C00962 00410 ∂26-Nov-86 2100 JMC
C00963 00411 ∂26-Nov-86 2314 SINGH@Sierra.Stanford.EDU Re: Lynch mob spirit
C00967 00412 ∂28-Nov-86 1317 WINSLETT@Score.Stanford.EDU Reagan Library
C00969 00413 ∂28-Nov-86 1555 SJG re: closest possible world
C00970 00414 ∂28-Nov-86 1618 SJG re: closest possible world
C00972 00415 ∂28-Nov-86 1632 SJG re: closest possible world
C00974 00416 ∂28-Nov-86 1637 SJG re: closest possible world
C00976 00417 ∂28-Nov-86 2239 JSW Scope and Extent rules for Qlisp
C00983 00418 ∂29-Nov-86 1157 SJG re: closest possible world
C00984 00419 ∂29-Nov-86 1442 WINSLETT@Score.Stanford.EDU re: Reagan Library
C00986 00420 ∂29-Nov-86 2031 YM Thesis
C00989 00421 ∂30-Nov-86 1542 M.MCD@LEAR.STANFORD.EDU re: Iran and the Contras
C00992 00422 ∂30-Nov-86 2342 LES re: dial costs
C00993 00423 ∂01-Dec-86 0854 VAL the context seminar
C00994 00424 ∂01-Dec-86 0900 JMC
C00995 00425 ∂01-Dec-86 0927 RICHARDSON@Score.Stanford.EDU AI Courses
C00996 00426 ∂01-Dec-86 1129 VAL
C00997 00427 ∂01-Dec-86 1140 RA lunch
C00998 00428 ∂01-Dec-86 1313 VAL your book
C01000 00429 ∂01-Dec-86 1501 VAL Nonmonotonic and Commonsense Reasoning Seminar
C01001 00430 ∂01-Dec-86 2159 GROSOF@Score.Stanford.EDU 2 thoughts about contexts
C01009 00431 ∂02-Dec-86 0800 JMC
C01010 00432 ∂02-Dec-86 0848 TAJNAI@Score.Stanford.EDU SOE Adv.Council
C01013 00433 ∂02-Dec-86 0953 CLT Qlisp meeting reminder
C01014 00434 ∂02-Dec-86 1047 @Score.Stanford.EDU:GSMITH@SUMEX-AIM.ARPA AI Planning Retreat
C01016 00435 ∂02-Dec-86 1056 RA
C01017 00436 ∂02-Dec-86 1058 CLT Qlisp meeting reminder
C01018 00437 ∂02-Dec-86 1124 AAAI-OFFICE@SUMEX-AIM.ARPA List of Sponsored Workshops
C01022 00438 ∂02-Dec-86 1130 JMC re: Qlisp meeting reminder
C01023 00439 ∂02-Dec-86 1153 JJW Seminar time
C01024 00440 ∂02-Dec-86 1153 OKUNO@SUMEX-AIM.ARPA re: Qlisp meeting reminder
C01025 00441 ∂02-Dec-86 1258 CLT Qlisp update
C01026 00442 ∂02-Dec-86 1329 LES Qlisp Meeting
C01027 00443 ∂02-Dec-86 1331 CLT Qlisp Meeting
C01028 00444 ∂02-Dec-86 1447 VAL the context seminar
C01029 00445 ∂02-Dec-86 1455 CHEADLE@Score.Stanford.EDU AI Qual prep course
C01031 00446 ∂02-Dec-86 1656 ME Prancing Pony Bill
C01033 00447 ∂03-Dec-86 0940 RA Bibel
C01034 00448 ∂03-Dec-86 1523 AIR
C01035 00449 ∂03-Dec-86 1615 LAMPING@Sushi.Stanford.EDU Re: Bug
C01037 00450 ∂03-Dec-86 1647 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
C01038 00451 ∂03-Dec-86 1703 GOLDBERG@CSLI.STANFORD.EDU Re: bug appreciation
C01040 00452 ∂03-Dec-86 1723 2F09ULL%UKANVAX.BITNET@WISCVM.WISC.EDU Frame workshop.
C01042 00453 ∂03-Dec-86 1742 6058598%PUCC.BITNET@forsythe.stanford.edu SDI reply
C01044 00454 ∂03-Dec-86 1745 HEWETT@SUMEX-AIM.ARPA Re: bug appreciation
C01045 00455 ∂03-Dec-86 2303 JACOBS@Sushi.Stanford.EDU Re: "America Bashing"
C01047 00456 ∂04-Dec-86 0942 RA going to the bookstore
C01048 00457 ∂04-Dec-86 1100 PALLAS@Sushi.Stanford.EDU Re: "America Bashing"
C01051 00458 ∂04-Dec-86 1126 VAL Przymusinski
C01052 00459 ∂04-Dec-86 1400 JMC
C01053 00460 ∂04-Dec-86 1435 RA telephone
C01054 00461 ∂04-Dec-86 1544 RA leaving
C01055 00462 ∂04-Dec-86 2323 ILAN@Score.Stanford.EDU Island dishes
C01057 00463 ∂05-Dec-86 0457 Yuri_Gurevich@um.cc.umich.edu e-address correction
C01059 00464 ∂05-Dec-86 0900 JMC
C01060 00465 ∂05-Dec-86 0909 cramer@Sun.COM re: news story
C01062 00466 ∂05-Dec-86 0918 EPPLEY@Score.Stanford.EDU Rutie
C01063 00467 ∂05-Dec-86 0936 AAAI-OFFICE@SUMEX-AIM.ARPA [Aravind Joshi <Joshi@cis.upenn.edu>: Workshop proposal]
C01072 00468 ∂05-Dec-86 0953 stevens@amadeus.stanford.edu re: news story
C01074 00469 ∂05-Dec-86 0954 RA Shankar
C01075 00470 ∂05-Dec-86 1112 RA mcc invoice
C01076 00471 ∂05-Dec-86 1150 RA Staff x-mas lunch
C01077 00472 ∂05-Dec-86 1157 CHEADLE@Score.Stanford.EDU Gray Tuesday
C01095 00473 ∂07-Dec-86 1800 GLB students and EKL
C01096 00474 ∂07-Dec-86 2213 @SUMEX-AIM.ARPA:MRC@PANDA [Ian Merritt <ihnp4!nrcvax!minnie!ihm@ucbvax.Berkeley.EDU>: Russian PDP-11 ADA compiler for sale?!?]
C01109 00475 ∂08-Dec-86 0048 RLG NSF application
C01110 00476 ∂08-Dec-86 0115 GLB
C01111 00477 ∂08-Dec-86 0907 NILSSON@Score.Stanford.EDU DARPA
C01114 00478 ∂08-Dec-86 0937 AI.ELLIE@MCC.COM 1986 Expenses
C01115 00479 ∂08-Dec-86 1131 GLB
C01116 00480 ∂08-Dec-86 1240 RA going out
C01117 00481 ∂08-Dec-86 1322 CLT Qlisp meeting reminder
C01119 00482 ∂08-Dec-86 1455 RA seminar room
C01120 00483 ∂08-Dec-86 1526 RICHARDSON@Score.Stanford.EDU AI Mtg.
C01121 00484 ∂08-Dec-86 1601 RA tired
C01122 00485 ∂08-Dec-86 1606 LES Gould number cruncher
C01123 00486 ∂08-Dec-86 1751 CLT shopping list
C01124 00487 ∂08-Dec-86 2000 JMC
C01125 00488 ∂09-Dec-86 0800 JMC
C01126 00489 ∂09-Dec-86 1031 RA credit
C01127 00490 ∂09-Dec-86 1203 GARDNER@CSLI.STANFORD.EDU Book promotion
C01129 00491 ∂09-Dec-86 1205 RA lunch
C01130 00492 ∂09-Dec-86 1519 VAL
C01131 00493 ∂09-Dec-86 1857 DEK Ed Fredkin
C01132 00494 ∂09-Dec-86 2145 MODET@Sushi.Stanford.EDU Some questions/comments about the final
C01136 00495 ∂10-Dec-86 0129 HSU@Sushi.Stanford.EDU gmat
C01138 00496 ∂10-Dec-86 0441 KE425%RPICICGE.BITNET@WISCVM.WISC.EDU
C01140 00497 ∂10-Dec-86 0449 KE425%RPICICGE.BITNET@WISCVM.WISC.EDU
C01184 00498 ∂10-Dec-86 0800 JMC
C01185 00499 ∂10-Dec-86 0914 RA procedural question
C01186 00500 ∂10-Dec-86 0954 RA Joan Treichel
C01187 00501 ∂10-Dec-86 1234 PHY
C01189 00502 ∂10-Dec-86 1242 VAL Commonsense and Non-Monotonic Reasoning Seminar
C01191 00503 ∂10-Dec-86 1532 RA CS 306 final
C01192 00504 ∂10-Dec-86 1603 DEK Pierre Baldi
C01194 00505 ∂10-Dec-86 1640 PHYSICSLIB@Sierra.Stanford.EDU Re: Preserving information
C01197 00506 ∂11-Dec-86 0900 JMC
C01198 00507 ∂11-Dec-86 1100 JMC
C01199 00508 ∂11-Dec-86 1209 RA contex[f86,jmc]
C01200 00509 ∂11-Dec-86 1439 RA
C01201 00510 ∂11-Dec-86 1540 RINDFLEISCH@SUMEX-AIM.ARPA Computer Museum AI Exhibit
C01205 00511 ∂11-Dec-86 1540 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU Star Wars
C01243 00512 ∂11-Dec-86 1600 RA leaving
C01244 00513 ∂11-Dec-86 1721 HSU@Sushi.Stanford.EDU Re: problem 2
C01245 00514 ∂11-Dec-86 1904 HSU@Sushi.Stanford.EDU re: problem 2
C01246 00515 ∂11-Dec-86 1936 LES re: Computer Museum AI Exhibit
C01249 00516 ∂12-Dec-86 0236 DEGENHARDT@Sushi.Stanford.EDU CS 306: Questions about the final
C01253 00517 ∂12-Dec-86 0800 JMC
C01254 00518 ∂12-Dec-86 0901 PERLAKI@Score.Stanford.EDU Rutie
C01255 00519 ∂12-Dec-86 1209 RLG reminder
C01256 00520 ∂12-Dec-86 1253 RA Barbara Vrabec
C01257 00521 ∂12-Dec-86 1314 RA going out
C01258 00522 ∂12-Dec-86 1439 NILSSON@Score.Stanford.EDU
C01259 00523 ∂12-Dec-86 1531 RA leaving
C01260 00524 ∂12-Dec-86 1605 VAL Nonmonotonic reasoning seminar
C01261 00525 ∂13-Dec-86 1113 CLT igor
C01262 00526 ∂13-Dec-86 1114 CLT qlisp mailing list
C01263 00527 ∂13-Dec-86 1355 LES re: igor
C01264 00528 ∂13-Dec-86 1401 CLT igor
C01265 00529 ∂13-Dec-86 1421 bulwinkl@rocky.stanford.edu re: don't want old magazines??
C01266 00530 ∂13-Dec-86 1807 BRONSTEIN@Score.Stanford.EDU I'm always amazed...
C01268 00531 ∂14-Dec-86 1450 bulwinkl@rocky.stanford.edu re: don't want old magazines??
C01269 00532 ∂14-Dec-86 1652 PHYSICSLIB@Sierra.Stanford.EDU Re: same author
C01271 00533 ∂15-Dec-86 0804 RICHARDSON@Score.Stanford.EDU AI Courses
C01272 00534 ∂15-Dec-86 0900 JMC
C01273 00535 ∂15-Dec-86 0936 RA telephone answering machine
C01274 00536 ∂15-Dec-86 1040 @Score.Stanford.EDU:GSMITH@SUMEX-AIM.ARPA AI Planning Retreat
C01276 00537 ∂15-Dec-86 1103 REGES@Score.Stanford.EDU couple of things
C01278 00538 ∂15-Dec-86 1141 RA Taleen
C01279 00539 ∂15-Dec-86 1212 REGES@Score.Stanford.EDU CS326
C01281 00540 ∂15-Dec-86 1527 CRISPIN@SUMEX-AIM.ARPA Re: regulating packet switching
C01284 00541 ∂16-Dec-86 0837 AAAI-OFFICE@SUMEX-AIM.ARPA Re: Space Station Automation Workshop
C01286 00542 ∂16-Dec-86 0932 RA Maria Finitzo
C01287 00543 ∂16-Dec-86 0952 RA Dina Bolla
C01288 00544 ∂16-Dec-86 0957 RA tel. ansering machine
C01289 00545 ∂16-Dec-86 1053 STEINBERGER@SRI-KL.ARPA re: Decommissioning reactors
C01293 00546 ∂16-Dec-86 1126 RPG Gray Tuesday Reminder
C01295 00547 ∂16-Dec-86 1131 RPG Weening
C01296 00548 ∂16-Dec-86 1227 RA Dr. Flaherty
C01297 00549 ∂16-Dec-86 1314 KOHEN@Sushi.Stanford.EDU re: Decommissioning reactors
C01300 00550 ∂16-Dec-86 1630 RPG State of Affairs
C01301 00551 ∂16-Dec-86 1700 RA be late tomorrow
C01302 00552 ∂16-Dec-86 1822 LES re: dial costs
C01304 00553 ∂17-Dec-86 0800 JMC
C01305 00554 ∂17-Dec-86 0820 STEINBERGER@SRI-KL.ARPA re: Nuclear Power
C01307 00555 ∂17-Dec-86 0905 coraki!pratt@Sun.COM
C01309 00556 ∂17-Dec-86 0948 TREITEL@Sushi.Stanford.EDU re: NATO bucks
C01312 00557 ∂17-Dec-86 1156 YEFF@Sushi.Stanford.EDU NatGeo
C01313 00558 ∂17-Dec-86 1157 RICHARDSON@Score.Stanford.EDU Annual Faculty Report
C01314 00559 ∂17-Dec-86 1237 @Score.Stanford.EDU:ROSENBLOOM@SUMEX-AIM.STANFORD.EDU Graduate Cognitive Science Program
C01318 00560 ∂17-Dec-86 1358 RA Chicago conference operator
C01319 00561 ∂17-Dec-86 1409 RA trip to Portland
C01320 00562 ∂17-Dec-86 1418 berke@CS.UCLA.EDU approximate theories and naming
C01330 00563 ∂17-Dec-86 1458 LIBRARY@Score.Stanford.EDU book on hold
C01332 00564 ∂17-Dec-86 2210 ILAN@Score.Stanford.EDU re: I don't know about you guys
C01334 00565 ∂18-Dec-86 0734 CHANDRA%OSU-20@ohio-state.ARPA Support Requested for Workshop
C01339 00566 ∂18-Dec-86 0926 RA Re: answering machine
C01340 00567 ∂18-Dec-86 0948 VAL
C01341 00568 ∂18-Dec-86 0957 RA Hotel in Chicago
C01342 00569 ∂18-Dec-86 1023 VAL re: reply to message
C01343 00570 ∂18-Dec-86 1041 edsel!sunvalleymall!jlz@navajo.stanford.edu Lisp Journal announcement
C01345 00571 ∂18-Dec-86 1448 RA IBM RT loan
C01346 00572 ∂18-Dec-86 1439 NSH Munindar Paul Singh
C01347 00573 ∂18-Dec-86 1553 RA leaving
C01348 00574 ∂18-Dec-86 1621 RICHARDSON@Score.Stanford.EDU Annual Faculty Report
C01356 00575 ∂18-Dec-86 1700 JMC
C01357 00576 ∂18-Dec-86 1716 BUCHANAN@SUMEX-AIM.STANFORD.EDU Re: Computer Museum AI Exhibit
C01360 00577 ∂18-Dec-86 1751 LES re: Computer Museum AI Exhibit
C01363 00578 ∂18-Dec-86 1821 RINDFLEISCH@SUMEX-AIM.STANFORD.EDU re: Computer Museum AI Exhibit
C01365 00579 ∂18-Dec-86 2122 G.BEESON@CSLI.STANFORD.EDU commonsense reasoning
C01367 00580 ∂18-Dec-86 2129 G.BEESON@CSLI.STANFORD.EDU commonsense reasoning & knowledge
C01369 00581 ∂18-Dec-86 2134 G.BEESON@CSLI.STANFORD.EDU commonsense and natural language
C01371 00582 ∂18-Dec-86 2135 WINSLETT@Score.Stanford.EDU more reagan library
C01373 00583 ∂18-Dec-86 2206 CHANDRA%OSU-20@ohio-state.ARPA Workshop grant request again
C01375 00584 ∂19-Dec-86 0658 hendler@brillig.umd.edu ~m
C01377 00585 ∂19-Dec-86 0941 ANDY@Sushi.Stanford.EDU Re: animal rights
C01379 00586 [In reply to message rcvd 18-Dec-86 16:09-PT.]
C01380 00587 ∂19-Dec-86 1127 RA Dr. Gibbon
C01381 00588 ∂19-Dec-86 1358 RA meeting with Gibbon
C01382 00589 ∂19-Dec-86 1419 RA leave early
C01383 00590 ∂19-Dec-86 1422 HSU@Sushi.Stanford.EDU finals and grade sheet
C01385 00591 ∂19-Dec-86 1503 HSU@Sushi.Stanford.EDU re: finals and grade sheet
C01386 00592 ∂20-Dec-86 1705 OKUNO@SUMEX-AIM.STANFORD.EDU Re: junet
C01388 00593 ∂21-Dec-86 1919 HSU@Sushi.Stanford.EDU final
C01390 00594 ∂21-Dec-86 1927 HSU@Sushi.Stanford.EDU handouts
C01392 00595 ∂21-Dec-86 2319 HSU@Sushi.Stanford.EDU re: final
C01394 00596 ∂21-Dec-86 2337 AMEHTA@Sushi.Stanford.EDU Final
C01395 00597 ∂22-Dec-86 0934 BJORK@Score.Stanford.EDU Broken terminal
C01396 00598 ∂22-Dec-86 1000 JMC
C01397 00599 ∂22-Dec-86 1100 PALLAS@Sushi.Stanford.EDU Re: morals and ethics
C01399 00600 ∂22-Dec-86 1124 pratt@navajo.stanford.edu mailing list
C01401 00601 ∂22-Dec-86 1141 HSU@Sushi.Stanford.EDU grade sheet
C01402 00602 ∂22-Dec-86 1405 RA finals
C01403 00603 ∂22-Dec-86 1405 HSU@Sushi.Stanford.EDU
C01404 00604 ∂22-Dec-86 1501 JMC
C01405 00605 ∂22-Dec-86 1606 shashank@su-whitney.arpa Re: ethics and restraint of trade
C01408 00606 ∂22-Dec-86 1628 PHYSICSLIB@Sierra.Stanford.EDU Re: For Henry Lowood
C01411 00607 ∂22-Dec-86 1707 GLB
C01412 00608 ∂22-Dec-86 1731 SCHMIDT@SUMEX-AIM.STANFORD.EDU Re: ethics and restraint of trade
C01415 00609 ∂22-Dec-86 1815 @SUMEX-AIM.STANFORD.EDU:MRC@PANDA fundamentalist textbook case
C01422 00610 ∂22-Dec-86 1840 HSU@Sushi.Stanford.EDU grades
C01423 00611 ∂22-Dec-86 1957 DEK affirming Landau
C01424 00612 ∂22-Dec-86 2143 CLT visit
C01425 00613 ∂22-Dec-86 2044 DEK excuse me,...
C01426 00614 ∂22-Dec-86 2349 HSU@Sushi.Stanford.EDU re: grades
C01431 00615 ∂22-Dec-86 2354 HSU@Sushi.Stanford.EDU Tom Lawson
C01433 00616 ∂23-Dec-86 0011 HSU@Sushi.Stanford.EDU re: Tom Lawson
C01434 00617 ∂23-Dec-86 0900 JMC
C01435 00618 ∂23-Dec-86 0902 SB Terminal update
C01436 00619 ∂23-Dec-86 1046 RA Dr. Flaherty
C01437 00620 ∂23-Dec-86 1126 BJORK@Score.Stanford.EDU A thought on the terminal problem.
C01439 00621 ∂23-Dec-86 1644 RA leaving
C01440 00622 ∂24-Dec-86 0929 CLT dinner
C01441 00623 ∂24-Dec-86 1459 RA vacation
C01442 00624 ∂24-Dec-86 1545 PHYSICSLIB@Sierra.Stanford.EDU Keys for Math Library?
C01446 00625 ∂24-Dec-86 1828 LES
C01450 00626 ∂25-Dec-86 2129 ME psych and bboards
C01451 00627 ∂25-Dec-86 2217 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU star wars
C01490 00628 ∂25-Dec-86 2218 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU star wars
C01493 00629 ∂25-Dec-86 2217 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU disregard previous message
C01495 00630 ∂26-Dec-86 0637 @REAGAN.AI.MIT.EDU:Hewitt@XX.LCS.MIT.EDU response to McDermott
C01497 00631 ∂26-Dec-86 1149 VAL reply to message
C01512 00632 ∂26-Dec-86 1213 VAL CS326
C01521 00633 ∂28-Dec-86 1522 netlibd@anl-mcs.ARPA index
C01543 00634 ∂28-Dec-86 1533 netlibd@anl-mcs.ARPA index from fnlib
C01560 00635 ∂28-Dec-86 1540 netlibd@anl-mcs.ARPA sin from fnlib
C01588 00636 ∂28-Dec-86 1603 ANDY@Sushi.Stanford.EDU Forbs education cover story
C01589 00637 ∂28-Dec-86 1612 PHYSICSLIB@Sierra.Stanford.EDU re: Keys for Math Library?
C01590 00638 ∂28-Dec-86 1615 CLT shopping list
C01591 00639 ∂28-Dec-86 1755 FEIGENBAUM@SUMEX-AIM.STANFORD.EDU Re: argument against logic in AI
C01593 00640 ∂28-Dec-86 2307 ME DMWAITS?
C01594 00641 ∂29-Dec-86 0844 ito%aoba.tohoku.junet%utokyo-relay.csnet@RELAY.CS.NET visit
C01596 00642 ∂29-Dec-86 0922 TALEEN@Score.Stanford.EDU nils would like to speak w/you
C01597 00643 ∂29-Dec-86 1005 CLT
C01598 00644 ∂29-Dec-86 1458 Bobrow.pa@Xerox.COM A personal favor
C01600 00645 ∂29-Dec-86 1625 CLT
C01601 00646 ∂29-Dec-86 1855 LES Gosper account
C01602 00647 ∂29-Dec-86 2220 Bobrow.pa@Xerox.COM re: A personal favor
C01604 00648 ∂30-Dec-86 0840 PERLAKI@Score.Stanford.EDU message
C01605 00649 ∂30-Dec-86 0956 RICHARDSON@Score.Stanford.EDU AI Faculty Mtg.
C01606 00650 ∂30-Dec-86 1033 CLT shopping list
C01607 00651 ∂30-Dec-86 1034 CERF@A.ISI.EDU Digital Library Meeting Agenda
C01621 00652 ∂30-Dec-86 1110 TALEEN@Score.Stanford.EDU msg. from Vint Cerf
C01622 00653 ∂30-Dec-86 1115 TALEEN@Score.Stanford.EDU express letter
C01624 00654 ∂30-Dec-86 1313 STEINBERGER@SRI-KL.ARPA Re: Evil Empire
C01634 00655 ∂30-Dec-86 1355 KOHEN@Sushi.Stanford.EDU Re: Evil Empire
C01636 00656 ∂30-Dec-86 1438 TALEEN@Score.Stanford.EDU book held for you at Green library
C01638 00657 ∂30-Dec-86 1441 ANDY@Sushi.Stanford.EDU Re: Evil Empire
C01642 00658 ∂30-Dec-86 1500 CERF@A.ISI.EDU Cancellation of Meeting
C01644 00659 ∂30-Dec-86 1743 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU star wars continued
C01684 00660 ∂30-Dec-86 1759 VAL
C01685 00661 ∂31-Dec-86 0834 FEIGENBAUM@SUMEX-AIM.STANFORD.EDU Play about Turing's life
C01689 00662 ∂31-Dec-86 0900 CLT
C01690 00663 ∂31-Dec-86 0900 JMC
C01691 00664 ∂31-Dec-86 0900 JMC
C01695 00665 ∂31-Dec-86 1446 TALEEN@Score.Stanford.EDU 3:00
C01697 00666 ∂31-Dec-86 1529 Bobrow.pa@Xerox.COM re: A personal favor
C01698 00667 ∂31-Dec-86 1530 STEINBERGER@SRI-KL.ARPA Evil empire, etc.
C01707 00668 ∂31-Dec-86 1529 oshea.pa@Xerox.COM Video lectures for new Xerox President
C01716 ENDMK
C⊗;
∂01-Oct-86 1641 RA new phone
My new telephone number is 723 6321.
∂01-Oct-86 1713 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 1 October 1986
Previous Balance 6.30
Payment(s) 6.30 (check 9/18/86)
-------
Current Charges 4.00 (bicycle lockers)
1.15 (vending machine)
-------
TOTAL AMOUNT DUE 5.15
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 1.0% per month will be
charged on balances remaining unpaid 25 days after bill date above.
An account with a credit balance earns interest of .33% per month,
based on the average daily balance.
∂01-Oct-86 1701 VAL section for proposal
See darpa[1,val]
∂01-Oct-86 1720 VAL Commonsense Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
My talk on the Yale shooting is postponed until October 16 (4pm, MJH252).
A reminder will be sent out a few days before the meeting. There will be no
meeting next week.
Vladimir
∂02-Oct-86 0900 JMC
Cuthbert and Chuck.
∂02-Oct-86 1042 RA ofind command
Can you arrange it so that I too can have a command similar to your OFIND which
will read names from my PHON file?
Thanks,
∂02-Oct-86 1123 VAL reply to message
[In reply to message rcvd 02-Oct-86 11:10-PT.]
I'll be out on Monday, Oct. 13 (Yom Kippur). I believe other Jewish holidays
fall mostly on weekends this year. Also, I'm going to MCC on Oct. 7,8 (next
Tuesday and Wednesday), and Reiter invited me to give a colloquium talk in
Toronto on Oct. 21.
∂02-Oct-86 1234 EVE@CSLI.STANFORD.EDU Symbolic Systems Faculty Roster
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 2 Oct 86 12:34:50 PDT
Date: Thu 2 Oct 86 12:32:39-PDT
From: Eve Wasmer <EVE@CSLI.STANFORD.EDU>
Subject: Symbolic Systems Faculty Roster
To: jmc@SU-AI.ARPA
Professor McCarthy: I'm preparing a roster of SSP faculty office locations,
phone numbers, and EM addresses. The information I have on you is: 460-356, 3-4430. Is this information correct? Thanks.
-------
∂02-Oct-86 1241 VAL
Fernando seems very good, let's try to catch him.
∂02-Oct-86 1446 SJG cryptograms with multiple decodings
John:
In "AI: The Very Idea," Haugeland gives a short (about 20 or so
characters) cryptogram with three possible meanings; I seem to
remember this as bearing on a discussion we had some time ago.
Do you want me to find the thing for you?
Matt
∂02-Oct-86 1454 SJG re: cryptograms with multiple decodings
[In reply to message rcvd 02-Oct-86 14:53-PT.]
OK. I'll try to remember to bring it in tomorrow.
Matt
∂02-Oct-86 1554 RA leaving early
It is Thursday and I will be leaving early.
∂02-Oct-86 1605 RA ofind command
If I put
find:in phon.tex[1,ra]
in my option.txt, I can access my phon file but I can't access yours. It picks
up whatever is listed first. Is there a way to access both files?
Thanks,
∂03-Oct-86 0318 rwsh%computer-lab.cambridge.ac.uk@Cs.Ucl.AC.UK Forwarding: The ELEPHANT Language
Received: from CS.UCL.AC.UK by SAIL.STANFORD.EDU with TCP; 3 Oct 86 03:18:25 PDT
Received: from computer-lab.cambridge.ac.uk by mv1.Cs.Ucl.AC.UK
via Janet with NIFTP id aa01263; 3 Oct 86 10:49 WET
Received: from cl.cam.ac.uk by Jenny.CL.Cam.AC.UK with SMTP id a010549;
3 Oct 86 10:32 BST
Date: Fri, 3 Oct 86 10:28:20 BST
From: Roger Hale <rwsh%computer-lab.cambridge.ac.uk@Cs.Ucl.AC.UK>
To: jmc@su-ai.arpa, clt@su-ai.arpa
Subject: Forwarding: The ELEPHANT Language
Message-Id: <UK.AC.Cam.CL (Steve) 86275.680.2670>
Perhaps my initial message failed to get through....
-------------- forwarded message ---------------
> Date: Wed, 10 Sep 86 14:04:56 BST
> From: Roger Hale <rwsh@CL>
> To: jmc@arpa.su-ai
> Cc: clt@arpa.su-ai
> Subject: The ELEPHANT Language
> Message-Id: <UK.AC.Cam.CL (Steve) 86252.640.5500>
> MsgSet: :active:
I am a PhD student at Cambridge University, where I am working with Ben
Moszkowski on Temporal Logic Programming. Your work on the Elephant language
seems to be quite relevant to this research, and I would be grateful if you
could point me to some (accessible) references on the subject.
Thanks,
Roger Hale
Arpa: rwsh%uk.ac.cam.cl@uk.ac.ucl.cs
Post: Computer Laboratory
University of Cambridge
Corn Exchange Street
Cambridge CB2 3QG
ENGLAND
---------- end of forwarded message ------------
∂03-Oct-86 0939 RA audit a course
If you have no objections, I would like to sit in on Winograd's Phenomenological
Foundation of Cognition... course, MWF 10:00.
Thanks,
∂03-Oct-86 0945 VAL milestones
[In reply to message rcvd 02-Oct-86 14:57-PT.]
Maybe you'll find this chart of possible "milestones" useful for the proposal.
| Objectives
Stage |--------------------------------------------------------------------
| Theory | Implementation
-----------------------------------------------------------------------------
| Define the general structure | Implement algorithms for computing
| of the database and the | the result of circumscription.
1 | syntactic form of facts to be |
| included in it. Study general |
| properties of databases with |
| this structure. |
-----------------------------------------------------------------------------
| Construct the first version | Integrate the circumcsription
2 | of the database. | program with a general purpose
| | proof checker. Use the system for
| | debugging the database.
-----------------------------------------------------------------------------
| Finalize the database. | Investigate the capabilities of
| Investigate its capabilities | heuristics for automated reasoning
3 | and limitations. Determine | based on circumscription.
| directions for further |
| research. |
∂03-Oct-86 0946 CLT jjw,air,AMEHTA@Sushi.Stanford.EDU
To: JMC, RPG, LES, RWW
;avizur
reminder
Jim Boyle from Argonne will be here on Monday Oct. 6th
I have arranged for a demo of the TAO/ELIS system at 11:30 in
S-101 (Sumex-Aim machine room) in the Medical School.
The demo will be given by Gitchang Okuno one of the designers
and implementors of this system. There will be a group
going over from Mjax -- leaving about 11:15.
There will be informal discussion of parallel lisp implementation
at 2pm in Mjax 252.
Boyle will give a seminar at 4pm also in Mjax 252.
∂03-Oct-86 0947 RA [Reply to message recvd: 23 Sep 86 19:14 Pacific Time]
Mary Mainland, now Ash, work number, San Mateo DA offfice, 363 4000. Her
address: 943 Casanueva, Stanford.
∂03-Oct-86 0950 SJG cryptogram
Dumb dogs dig wells
Cold cats cut riffs
Rife rats rot hulls
How feeble it is is up to you.
Matt
∂03-Oct-86 1111 SJG re: cryptogram
[In reply to message rcvd 03-Oct-86 10:45-PT.]
Very little. I'm currently reading through the book, and will
probably be ready to lend it to you in about a week. But if you'd
like to just have a quick look, I brought it in today.
Matt
∂03-Oct-86 1152 RWF Eric Drexler
To: JMC
CC: RWF
will be giving the CS colloquium on Oct 21. Would you like to spend some time
with him? Are there other people he should meet? Would you like to be his
principal contact? Do lunch?
-Bob Floyd
∂03-Oct-86 1159 RA new schedule
To: JMC, ZM
Right now I work 9:00 - 5:00. I leave early on Thursdays and I make it up
either by staying late or coming in early during the week.
My new schedule:
Monday: 8:00 - 5:00 (10:00-11:00 time taken off for CS378)
Tuesday: 9:00 - 6:00 (extra hour to make up for leaving early Thursday)
Wed. : 8:00 - 5:00 (10:00-11:00 time taken off for CS378)
Thurs.: 9:00 - 4:15
Friday: 8:00 - 5:00 (10:00-11:00 time taken off for CS378)
In cases where I don't make it to the office by 8:00, I will work late
to make up for it.
∂03-Oct-86 1314 RA ijcai.tex
IJCAI is now in ijcai.tex[e77,jmc]
∂03-Oct-86 1434 RA would you mind?
Because of the holiday, I would like to leave around 3:00, if it's ok with you.
Thanks,
∂03-Oct-86 1451 RA elephant paper
Since there are no more copies in the file, I tried to pub it but couldn't
because of an error msg. Please let me know how to fix it.
Thanks,
∂03-Oct-86 1458 RA elephant paper
I found a copy of the paper and will xerox and send it.
∂03-Oct-86 2043 shoham@YALE.ARPA two things
Received: from YALE-BULLDOG.ARPA by SAIL.STANFORD.EDU with TCP; 3 Oct 86 20:42:28 PDT
Received: by Yale-Bulldog.YALE.ARPA; 3 Oct 86 22:49:09 EDT (Fri)
Date: 3 Oct 86 22:49:09 EDT (Fri)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8610040249.AA00220@Yale-Bulldog.YALE.ARPA>
Subject: two things
To: jmc@su-ai.ARPA
Dear John,
Instead of only a copy of the new chapter 3, I'm about to send you a draft of the
first five chapters. Since we're talking about 180 pages, I thought I'd warn you in
advance. I'd be a fool to expect you to read all that, but if any of it catches
your fancy, I'll be delighted to hear your comments (including critical ones as
those on ch. 3; hope you like the new version better). I wouldn't like this
draft to be overly distributed, but I'll trust your discretion in passing it on to
others. In particular, I'd love to get reactions from Vladimir, Nils, and anyone in
Mike's group (I'll send them a note about the existence of your copy).
On another topic, I'm really excited about the proposed frame-problem workshop.
The only problem is that timing is awkward, since at that time I'll still be in Israel.
I've asked around informally, and no one seemed to object to postponing it by one or
two weeks. How do you feel about it? If you're inclined to agree, could I asked you
to mention the possibility to the organizers? I have already, but your whisper rings
louder than all my shouts. While we're on the topic of changes, Dave Smith mentioned
the fact Kansas is a less than ideal location, and that a place with more redeeming
features (Utah? Colorado? Somewhere close to the Bay area?) would be more appropriate.
I personally don't feel very strongly about it, but I promised to bring it up.
Regards,
Yoav.
-------
∂04-Oct-86 1800 JMC
Call Gerry about China.
∂04-Oct-86 1820 RESTIVO@Score.Stanford.EDU picture request
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 4 Oct 86 18:20:07 PDT
Date: Sat 4 Oct 86 17:49:24-PDT
From: Chuck Restivo <Restivo@Score.Stanford.EDU>
Subject: picture request
To: jmc@Sail.Stanford.EDU
Message-ID: <12244230944.9.RESTIVO@Score.Stanford.EDU>
[cwr]
I want to obtain two autographed reproductions of your photograph
as it appeared in the 21 September 1986 issue of the San Jose
Mercury's 'West' magazine section. Please advise on the best
way to round this stuff up.
Regards,
- cw
-------
∂04-Oct-86 1926 BRINK@Sushi.Stanford.EDU Tenure track
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 4 Oct 86 19:26:48 PDT
Date: Sat 4 Oct 86 19:25:34-PDT
From: Ed Brink <brink@Sushi.Stanford.EDU>
Subject: Tenure track
To: jmc@Sail.Stanford.EDU
Message-ID: <12244248453.9.BRINK@Sushi.Stanford.EDU>
Food for thought, publishable if you think it's worth it:
I represent a small but nonempty set of folks with enough time in grade (mine
is at IBM) to be able to afford to retire early if there is something worth
doing and enough money in the offing to make the (in my case $20k/yr) pension
reasonable. I'm only 52, and I'm trying to get into the AI PhD program so I
can contribute something while I still can.
Quite different from the tenure-track approach, and if properly handled, maybe
a pool of quite flexible talent.
What do you think?
..Ed
(PS: This is emphatically only what it says it is, and not a seamy attempt to
curry favor or whatever. I wish it weren't so hard to say something without
folks thinking one is trying to DO something by it!)
-------
∂05-Oct-86 0456 @DIAMOND.S4CC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM Knotted Doughnuts
Received: from [128.81.51.3] by SAIL.STANFORD.EDU with TCP; 5 Oct 86 04:56:42 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by DIAMOND.S4CC.Symbolics.COM via CHAOS with CHAOS-MAIL id 16222; Sun 5-Oct-86 07:55:49 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 1216; Sun 5-Oct-86 04:54:26 PDT
Date: Sun, 5 Oct 86 04:48 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: Knotted Doughnuts
To: jmc@SU-AI.ARPA
Message-ID: <861005044855.1.RWG@TSUNAMI.SPA.Symbolics.COM>
is the title of M. Gardner's latest collection. This reminds me that
you once mentioned a colleague who had determined the length to diameter
ratio of certain "knotted cylinders". Can you re-adumbrate the reference?
∂06-Oct-86 1255 BRINK@Sushi.Stanford.EDU re: Tenure track
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Oct 86 12:53:24 PDT
Date: Mon 6 Oct 86 12:50:19-PDT
From: Ed Brink <brink@Sushi.Stanford.EDU>
Subject: re: Tenure track
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 4 Oct 86 22:17:00-PDT
Message-ID: <12244700787.53.BRINK@Sushi.Stanford.EDU>
Thanks for taking 15 minutes. I expected more like 15 seconds.
New PhD's are not, I hope, all 20<x<30 years old. I hope, because I intend
to become one ASAP. But I intend to use my PhDness to invent goodies for
IBM. That's not heretical, is it?
So much for homogeneity.
..Ed
-------
∂06-Oct-86 1346 ANDY@Sushi.Stanford.EDU Re: Einstein, etc.
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Oct 86 13:46:22 PDT
Date: Mon 6 Oct 86 13:45:00-PDT
From: Andy Freeman <ANDY@Sushi.Stanford.EDU>
Subject: Re: Einstein, etc.
To: JMC@Sail.Stanford.EDU
cc: su-etc@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 6 Oct 86 13:09:00-PDT
Message-ID: <12244710740.22.ANDY@Sushi.Stanford.EDU>
Has CS had a Newton, Darwin/Mendel, or Franklin. What would be
comparable to the discovery of oxygen or measuring the speed of light?
-andy
-------
∂06-Oct-86 1418 RA going to the bookstore
I am going to the bookstore
∂06-Oct-86 1500 JMC
Lowood
∂06-Oct-86 1529 REGES@Score.Stanford.EDU Machines for CS306
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Oct 86 15:29:22 PDT
Date: Mon 6 Oct 86 15:28:03-PDT
From: Stuart Reges <REGES@Score.Stanford.EDU>
Subject: Machines for CS306
To: jmc@Sail.Stanford.EDU, clt@Sail.Stanford.EDU
Office: Margaret Jacks 030C, 723-9798
Message-ID: <12244729501.37.REGES@Score.Stanford.EDU>
Do you want to use the TI Explorer workstations for your CS306 class this
quarter rather than using LOTS? The only class making use of the TIs is the
CS022 class, Programming in LISP. That class has 30 students, and we have 10
machines, so we could easily handle more traffic.
-------
∂06-Oct-86 1631 REGES@Score.Stanford.EDU re: Machines for CS306
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Oct 86 16:31:12 PDT
Date: Mon 6 Oct 86 16:05:24-PDT
From: Stuart Reges <REGES@Score.Stanford.EDU>
Subject: re: Machines for CS306
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 6 Oct 86 15:44:00-PDT
Office: Margaret Jacks 030C, 723-9798
Message-ID: <12244736299.37.REGES@Score.Stanford.EDU>
The machines each have their own disks, so commonly used files can be stored on
all machines. They are also on SUNet, so one can use FTP to retrieve files from
other places. We haven't had enough of a crowd on the machines that we have
found the need to limit access. We have given students the combination to the
room and told them when TAs will be available. If you don't need to use the
TIs, you probably shouldn't. But if you think the students would benefit from
being exposed to the programming environment available on the workstation, then
I think it would be appropriate to use them.
-------
∂06-Oct-86 1653 PANDREW@Score.Stanford.EDU CS306
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Oct 86 16:53:50 PDT
Date: Mon 6 Oct 86 16:52:32-PDT
From: Phil Andrew <PAndrew@Score.Stanford.EDU>
Subject: CS306
To: mccarthy@Score.Stanford.EDU
cc: fl@Othello.Stanford.EDU
Message-ID: <12244744881.24.PANDREW@Score.Stanford.EDU>
Prof. McCarthy --
We've been getting many complaints from students wanting to sign up for
CS306 to receive allocation on LOTS. In order for us to register the class,
we need to receive a Class Registration form from you, giving us the details
for the class, such as name, title, who the contact is, number of hours
of console allocation requested, TAs and instructors.
If you need any assistance, you can contact either myself or John Reuling
for more details.
Thank you.
Philip Andrew
LOTS Faculty Liaison
-------
∂06-Oct-86 1804 CLT shopping list
huggies
kerns apricot nectar 6pack small cans
gerbers high meat dinners
strained chicken with veg (3)
strained beef with veg (3)
draino
paper towels
supper - meat and veg
∂06-Oct-86 1851 R.REULING@LEAR.STANFORD.EDU re: Classes using LOTS
Received: from LEAR.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Oct 86 18:51:38 PDT
Date: Mon 6 Oct 86 18:51:08-PDT
From: John Reuling <R.Reuling@LEAR.STANFORD.EDU>
Subject: re: Classes using LOTS
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sun 5 Oct 86 18:38:00-PDT
Office: 246 Margaret Jacks Hall, Stanford; 415/725-5555
Message-ID: <12244766470.18.R.REULING@LEAR.STANFORD.EDU>
As far as I can tell, your class has not yet been registered. I've left a
registration form in your MJH mailbox. If you leave it in my mailbox once
it's completed, I'll see that it gets to Sweet Hall.
Thanks.
-J
-------
∂06-Oct-86 1939 LES Arkady
To: JMC, CLT
Gives tentative acceptance. I'll get the paper rolling.
Or does someone have his biography in text file form?
∂06-Oct-86 2013 PHYSICSLIB@Sierra.Stanford.EDU Meeting next Monday
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Oct 86 20:13:52 PDT
Date: Mon 6 Oct 86 20:13:14-PDT
From: Henry E. Lowood <PHYSICSLIB@Sierra.Stanford.EDU>
Subject: Meeting next Monday
To: jmc@Sail.Stanford.EDU
cc: physicslib@Sierra.Stanford.EDU, cn.arc@Forsythe.Stanford.EDU
Message-ID: <12244781417.10.PHYSICSLIB@Sierra.Stanford.EDU>
TO: John McCarthy
FR: Henry Lowood
Could we change our meeting to 3:15 on Monday, the 13th (from 2:00)? This
would make it possible for Roxanne Nilan to join us.
Thanks,
Henry
-------
∂06-Oct-86 2225 cramer@Sun.COM Paly International Festival
Received: from Sun.COM by SAIL.STANFORD.EDU with TCP; 6 Oct 86 22:24:54 PDT
Received: from snail.sun.com by sun.com (3.2/SMI-3.0)
id AA04618; Mon, 6 Oct 86 22:24:12 PDT
Received: from clem.sun.uucp by snail.sun.com (3.2/SMI-3.2)
id AA13346; Mon, 6 Oct 86 22:25:00 PDT
Received: by clem.sun.uucp (3.2/SMI-3.2)
id AA00456; Mon, 6 Oct 86 22:24:10 PDT
Date: Mon, 6 Oct 86 22:24:10 PDT
From: cramer@Sun.COM (Sam Cramer)
Message-Id: <8610070524.AA00456@clem.sun.uucp>
To: JMC@SAIL.STANFORD.EDU
Subject: Paly International Festival
John, I have just learned that Palo Alto High School will be sponsoring an
"International Festival" in late January. After looking at a list of the
speakers, I have a bad feeling about this event - I suspect it will be a
pro-Third World, anti-western indoctrination session. There is to be one
talk on the "roots of the crisis in Central America" and others on the
"Palestine Question - Israel and PLO Response" (note the symmetry) and
"A personal view of Quaddaffi's Libya - based on five years of living in
Libya".
I am trying to act through the local Jewish Community Relations Council (of
which I am a member) to get more facts and hopefully steer the event back in
the right direction. Do you know anyone in the Stanford/Hoover community that
could be of assistance? I would really like to form an alliance between the
pro-Israel community and local conservatives on this issue. Any ideas?
In particular, I could use specifics on the the following speakers:
Kathleen Namphy (evidently related to the current president/maximum leader of
Haiti), who is listed as a "lecturer in Underdgraduate studies" at Stanford,
and Mohammed Abo-Rabbo, a Phd candidate in Linguistics at Stanford. Do you
know anything about them? I'd like to do a search of the campus paper for
them - is there any way to "grep" through back issues electronically?
Sam
∂06-Oct-86 2250 cheriton@pescadero.stanford.edu Postdocs
Received: from PESCADERO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Oct 86 22:50:16 PDT
Received: by pescadero.stanford.edu with Sendmail; Mon, 6 Oct 86 22:49:05 pdt
Date: Mon, 6 Oct 86 22:49:05 pdt
From: David Cheriton <cheriton@pescadero.stanford.edu>
Subject: Postdocs
To: jmc@su-ai
Reading your letter in SIGPLAN about 100 postdosc for AI,
I heartily agree, especially if that could also be expanded to cover
other fields of CS (in concept, not sharing the 100).
There does need to be more of a chance for good people to contain
in independent research directions right after they graduate,
before they get hit with funding, teaching and other assist. prof. issues.
David C.
∂07-Oct-86 0900 JMC
Claudia re Fredkin check.
∂07-Oct-86 1041 AMEHTA@Sushi.Stanford.EDU Appearance in class
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 10:41:32 PDT
Date: Tue 7 Oct 86 10:40:01-PDT
From: Abhay Mehta <AMEHTA@Sushi.Stanford.EDU>
Subject: Appearance in class
To: jmc@Sail.Stanford.EDU
cc: hsu@Sushi.Stanford.EDU
Message-ID: <12244939211.9.AMEHTA@Sushi.Stanford.EDU>
I will arrange to be present today (10/7) at the end of the class so that
the students can see me.
Abhay
-------
∂07-Oct-86 1226 JOSHI@cis.upenn.edu Workshop
Received: from LINC.CIS.UPENN.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 12:26:13 PDT
Posted-Date: Tue, 7 Oct 86 15:11 EDT
Message-Id: <8610071924.AA13257@linc.cis.upenn.edu>
From: Aravind Joshi <Joshi@cis.upenn.edu>
Subject: Workshop
To: McCarthy@su-ai
Date: Tue, 7 Oct 86 15:11 EDT
Professor John McCarthy
Department of Computer Science
Stanford University
Stanford, CA
Dear John:
During July 87 (from July 13 to August 7, to be precise), Stan Peters,
Bill Rounds (U. Michigan), and I are planning a workshop tentatively
entitled " Mathematical Theories of Language" at the Stanford Unviversity
in conjunction with the LSA Institute. The idea of this workshop came
up very recently and,therefore, was not included in
the eraly plans of the institute and the proposal to AAAI by LSA.
I am enclosing below a short paragraph which provides a tentative description
of the scope of the workshop. The workshop will be largely made up of
several people who are likely to be at the Institute for one reason
or another. However, in addition to these people, we would like to invite
about 5 more people to participate in this workshop. All of these
people (the 5 or so we want to invite plus the ones who are going to be
there anyway), with the exception of possibly one or two are directly
concerned with the natural language processing aspects of AI. I would like to
include at least one person who could report on circumscriptive inference
as it relates to inferences in language (in this connection, I would welcome
your suggestions, in fact, if you would like talk on this topic, we would
be very delighted). I would like to know whether you would entertain a small
proposal (to AAAI) for support of this workshop. The main items of support
are really for the additional people and support for the two
non-resident organizers. The total amount will be in the range of about
$6000 to $7000.
I hope you will be interested in considering this proposal. Of course, we
will submit a formal proposal. Since Stan Peters is right there, I will ask
him to help me with the preparation. I look forward to hearing form you soon.
Best wishes,
Sincerely,
Aravind
Aravind K. Joshi
joshi@cis.upenn.edu
Workshop on Mathematical Theories of Language (July 13 - August 7)
The workshop will meet twice a week for presentation and
discussion of mathematically formulated models of language
systems about which results can be proved that have interesting
interpretations for language. Topics are likely to include
formal languages, automata theory, strong and weak equivalence of
grammatical theories, rigorously defined frameworks for semantic
analysis and their mathematical properties, mathematical
properties of artificial languages as they relate to natural
languages, and mathematical theories of language learning. The
workshop is being organized by Aravind Joshi, Stanley Peters and
William Rounds and is expected to include specially invited
participants as well as participants who will be at the
Linguistic Institute for other activities.
-------
PS: Bonnie Webber and I are considering sending in a proposal later
(for a late 87 or early 88) on the topic "Parallel processing
in natural language processing aspects of AI" (tentative title). This will,
of course, more squarely in AI. Bonnie and I will be touch with you o
about this later.
∂07-Oct-86 1325 UDI@Score.Stanford.EDU Re: error
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 13:25:07 PDT
Date: Tue 7 Oct 86 13:23:49-PDT
From: Ehud Shapiro <UDI@Score.Stanford.EDU>
Subject: Re: error
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Tue 7 Oct 86 12:57:00-PDT
Message-ID: <12244969028.11.UDI@Score.Stanford.EDU>
Thanks, I hope there aren't too many left...
Udi
-------
∂07-Oct-86 1442 CLT Qlisp
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
We are going to start having regular group meetings.
We will try Friday noon - beginning a week from Friday (Oct 17).
(This seems to be slot available in rpg's travel schedule
for the next few weeks.)
Joe will report (briefly) on his visit to Argonne
and the Alliant Users Meeting.
∂07-Oct-86 1533 AMEHTA@Sushi.Stanford.EDU TA Offices
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 15:33:36 PDT
Date: Tue 7 Oct 86 15:22:39-PDT
From: Abhay Mehta <AMEHTA@Sushi.Stanford.EDU>
Subject: TA Offices
To: cs306-distribution: ;
Message-ID: <12244990662.34.AMEHTA@Sushi.Stanford.EDU>
Handout #2 gave the TA office hours but didnt have the location, so here's
the complete information:
Yung-jen Hsu: Tuesdays 2:30p - 4:30p MJH450
Abhay Mehta : Wednesdays 10:00a - 12:00p MJH004 (Basement of MJH)
-------
∂07-Oct-86 1556 PHYSICSLIB@Sierra.Stanford.EDU re: Meeting next Monday
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 15:56:37 PDT
Date: Tue 7 Oct 86 15:55:57-PDT
From: Henry E. Lowood <PHYSICSLIB@Sierra.Stanford.EDU>
Subject: re: Meeting next Monday
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 6 Oct 86 20:32:00-PDT
Message-ID: <12244996724.9.PHYSICSLIB@Sierra.Stanford.EDU>
TO: John McCarthy
FR: Henry Lowood
Just to confirm: we will come by your office at 3:15 on Monday.
Thanks,
Henry
-------
∂07-Oct-86 1626 AMEHTA@Sushi.Stanford.EDU Console time on LOTS
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 16:25:52 PDT
Date: Tue 7 Oct 86 15:34:45-PDT
From: Abhay Mehta <AMEHTA@Sushi.Stanford.EDU>
Subject: Console time on LOTS
To: cs306-distribution: ;
Message-ID: <12244992866.34.AMEHTA@Sushi.Stanford.EDU>
The CS306 class has now been activated on LOTS. Each CS306 student will be
alloted 6 hrs/week on LOTS. If you have any problems with this, please let
me know. Dont forget to 'update' your accounts to include CS306.
Abhay Mehta
-------
∂07-Oct-86 1634 HSU@Sushi.Stanford.EDU cs306
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 16:31:11 PDT
Date: Tue 7 Oct 86 15:55:58-PDT
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: cs306
To: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12244996726.52.HSU@Sushi.Stanford.EDU>
Prof. McCarthy,
I've taken care of setting up the console allocation for CS306 on LOTS,
and sending e-mails as well as hardcopies of classnotes to TV land.
If there is anything else needs to be done, please let me know.
Yung-jen
-------
∂07-Oct-86 1738 @CSNET-RELAY.ARPA,@utterly.ai.toronto.edu:utterly.ai!hector McDermott again
Received: from RELAY.CS.NET by SAIL.STANFORD.EDU with TCP; 7 Oct 86 17:37:55 PDT
Received: from toronto.edu by csnet-relay.csnet id bf00793; 7 Oct 86 12:41 EDT
Received: from utterly.ai.toronto.edu by ai.toronto.edu id AA07796; Tue, 7 Oct 86 12:07:20 edt
Received: by utterly.ai.toronto.edu id AA02279; Tue, 7 Oct 86 12:06:50 EDT
Message-Id: <8610071606.AA02279@utterly.ai.toronto.edu>
Date: 07 Oct 86 12:06:41 EDT (Tue)
From: Hector Levesque <hector%ai.toronto.edu@CSNET-RELAY.ARPA>
To: phayes@SRI-KL.ARPA, reiter%utai.toronto.edu@CSNET-RELAY.ARPA,
stan@SRI-STRIPE.ARPA, forbus%uiuc.arpa@CSNET-RELAY.ARPA,
genesereth@SUMEX-AIM.ARPA, rjb%allegra%ai.toronto.edu@CSNET-RELAY.ARPA,
hewitt@MC.LCS.MIT.EDU, alberta!lksc%ubc-vision.uucp@CSNET-RELAY.ARPA,
jmc@SU-AI.ARPA, val@SU-AI.ARPA, stickel@SRI-STRIPE.ARPA,
tyson@SRI-STRIPE.ARPA, bmoore@SRI-STRIPE.ARPA, bobrow@XEROX.COM,
stefik@XEROX.COM, dekleer.pa@XEROX.COM
Subject: McDermott again
(Ahem!) (This is the kind of message that isn't fun to send or to receive.)
Look: I'm as bad as the next guy with deadlines, so far be it for me to try to
make you feel *guilty* about this, but... it's about time to send in your
commentary. I would very much like to know how things are going. If you just
need a day or two to send it, tell me it's in the mail. If your piece somehow
isn't clicking, and you would like some editorial advice, just say so. I'll
have a look and send you back quickie comments. If you haven't exactly
started per se, please try to make an honest assessment about when you will
(if ever). I can take bad news. Really.
Hector
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
∂07-Oct-86 1851 SINGH@Sierra.Stanford.EDU Many thanks for your response.
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 18:51:49 PDT
Date: Tue 7 Oct 86 18:51:08-PDT
From: Harinder Singh <SINGH@Sierra.Stanford.EDU>
Subject: Many thanks for your response.
To: jmc@Sail.Stanford.EDU
Message-ID: <12245028616.20.SINGH@Sierra.Stanford.EDU>
Prof. McCarthy,
Thank you for taking the time to respond. Sure appreciate it.
- Inder
-------
∂07-Oct-86 2342 FERNANDO@Sushi.Stanford.EDU RAship
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Oct 86 23:42:24 PDT
Date: Tue 7 Oct 86 23:41:00-PDT
From: Tim Fernando <FERNANDO@Sushi.Stanford.EDU>
Subject: RAship
To: jmc@Sail.Stanford.EDU, val@Sail.Stanford.EDU
cc: fernando@Sushi.Stanford.EDU
Message-ID: <12245081383.25.FERNANDO@Sushi.Stanford.EDU>
I have been giving your RAship offer a lot of thought, and this has
in turn led me to think about what, in a large sense, my interests
are and how they might fit in with the group's.
I am very much interested in understanding recursion, self-reference,
induction and self-application, in the context of a 'unified'
-- I apologize for using that term, but I can't think of a better
word -- mathematical theory of computation and deduction. I suspect
that the role self-reference/reflection play in deduction is analogous to that
played by recursion in computation. (In this regard, I find the research
of people like Craig Smorynski and Brian Smith intriguing. ) I would like
to study induction and in particular what people like Dana Scott, Solomon
Feferman and Yiannis Moschovakis have to say about it. Furthermore, I am quite
curious about how recursion/self-reference might look in a theory of non-well-
founded sets -- the sort of work Barwise, Aczel, etc. have been up to. And, of
course, I am interested in McCarhty's own ideas about a mathematical theory of
computation.
With regard to non-monotonic logic, I am interested in it not so much
as an attack on the problem of common sense reasoning (which I personally feel
quite uncomfortable studying) but as a form of non-monotonic induction, in
which there is (perhaps?) a measure of self-reference.
Having said all this, I must confess that I do not think that I can make much
progress in research this year, as I will have to devote most of my time to
the comps (because of my very weak background in the areas of systems
and applications). The idea of actually thinking about a research problem
in circumscription is very appealing, but I cannot promise to spend a lot
of time in it, as I would also like to pursue the other interests I mentioned
above -- in particular, Barwise's work on non-well-founded sets. As far
as funding, I have been told that the department will fund first year
PhD students who are not supported by the research groups they become
associated with. It is not clear to me what you expect from an RA, so
if what I have written above has not completely turned you off and if you
are still willing to support me, please let me know more or less how many
hours a week you would require from me.
-Tim
-------
∂08-Oct-86 0723 OTA-CIT@A.ISI.EDU Followup--Comments on Conference
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 8 Oct 86 07:22:55 PDT
Date: 8 Oct 1986 09:30:37 EDT
Subject: Followup--Comments on Conference
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@CSLI.STANFORD.EDU, waltz@GODOT.THINK.COM,
brown.pa@XEROX.COM, kay@MIT-MULTICS.ARPA, jmc@SU-AI.ARPA,
nilsson@SU-SCORE.ARPA, hector%utai.toronto.edu@RELAY.CS.NET, lenat@MCC.COM,
hollister%ti-csl@RELAY.CS.NET, herb.simon@A.CS.CMU.EDU,
bundy%aiva.edinburgh.ac.uk@CS.UCL.AC.UK, lehnert%umass@RELAY.CS.NET,
forbus@P.CS.UIUC.EDU, wilensky%larch.berkeley.edu@UCBVAX.BERKELEY.EDU,
jlk%gatech.csnet@RELAY.CS.NET, feldman@ROCHESTER.ARPA, norman@NPRDC.ARPA,
lederberg@ROCKEFELLER.ARPA, bkph%mit-oz@MC.LCS.MIT.EDU, 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, chapman@RUSSELL.STANFORD.EDU,
dts%cstvax.edinburgh.ac.uk@CS.UCL.AC.UK
cc: ota@A.ISI.EDU, OTA-CIT@A.ISI.EDU
As a brief followup, we'd appreciate any comments you have on the
computer conference itself. Was it an appropriate and comfortable medium
for this kind of discussion? Do you have suggestions about how it
ought to be run differently, if done again? How does it compare to
face-to-face or telephone discussions?
Thanks again,
Jim Dray, OTA
-------
∂08-Oct-86 0809 shoham@YALE.ARPA
Received: from YALE-BULLDOG.ARPA by SAIL.STANFORD.EDU with TCP; 8 Oct 86 08:08:38 PDT
Received: by Yale-Bulldog.YALE.ARPA; 8 Oct 86 10:23:40 EDT (Wed)
Date: 8 Oct 86 10:23:40 EDT (Wed)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8610081423.AA01254@Yale-Bulldog.YALE.ARPA>
To: Nilsson@su-score.ARPA, Genesereth@su-score.ARPA, lifschitz@su-ai.ARPA,
de2smith@sumex-aim.ARPA, sjg@sumex-aim.ARPA, Grosof@sumex-aim.ARPA
Cc: jmc@su-ai.ARPA
Hello all,
I'm sending John a draft of the first five chapters of my thesis. I'm sure none
of you are interested in reading all 180 pages, so I won't bombard you with
copies. You are of course more than welcome to read any part that interests
you, and your comments will be cherished.
Regards,
Yoav
-------
∂08-Oct-86 0847 RA [Reply to message recvd: 07 Oct 86 23:02 Pacific Time]
Re reference 2:
--------------
There is a 1974 publication:
AUTHOR: Shortliffe, Edward Hance.
TITLE: MYCIN: a rule-based computer program for advising physicians
regarding antimicrobial therapy selection.
IMPRINT: [Stanford, Calif.] 1974.
xvi,395 leaves. illus.
which is his Ph.D. thesis
and there is a 1984 publication with Buchanan:
TITLE: Rule-based expert systems : the MYCIN experiments of the Stanford
Heuristic Programming Project / edited by Bruce G. Buchanan, Edward
H. Shortliffe.
which one should be number 2 in the reference?
Re Reference 23:
---------------
It is the same as Reference 10 (below).
If I am right, 23 should be eliminated
{\bf Laird, John E., Allen Newell and Paul S. Rosenbloom (1986)}:
``Soar: An Architecture for General Intelligence'', to be published.
Re reference 21:
---------------
It seems to me that the beginning of the 2nd line in the reference--
Yovits, M.C. and--should be eliminated, so that the reference should read:
Newell, A., J. C. Shaw and H. A. Simon (1960): ``A variety
of intelligent learningin a General Probelm Solver'', in
{\it Self-Organizing Systems}, Cameron, S. eds., Pergammon, pp 153-189.
∂08-Oct-86 1015 JMC
Staggs.
∂08-Oct-86 1046 porter%wldwst.DEC@decwrl.DEC.COM Student Advising Hours
Received: from [128.45.0.240] by SAIL.STANFORD.EDU with TCP; 8 Oct 86 10:46:14 PDT
Received: by sonora.dec.com (5.54.2/4.7.34)
id AA00982; Wed, 8 Oct 86 10:46:20 PDT
Date: Wed, 8 Oct 86 10:46:20 PDT
Message-Id: <8610081746.AA00982@sonora.dec.com>
From: porter%wldwst.DEC@decwrl.DEC.COM
To: "jmc@sail.stanford.edu"%sonora.DEC@decwrl.DEC.COM, porter@decwrl.DEC.COM
Subject: Student Advising Hours
From: Mark Porter Date: 10/8/86
EMAIL: TRILOGY@SUSHI
Dr. McCarthy,
I have been told that you were assigned as my MSCS advisor for this quarter,
and I was wondering if I could spend a few minutes discussing some of the
degree requirements and various other questions about graduate programs
at Stanford.
Have you set aside certain hours for student advising? If not, are there
any times that are best for you? I am on campus Monday, Wednesday annd Friday
afternoons, but any other time would be fine also.
The easiest way to reach me is at my account on Sushi.
Many Thanks,
Mark Porter
∂08-Oct-86 1058 AMEHTA@Sushi.Stanford.EDU Balance
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 8 Oct 86 10:58:10 PDT
Date: Wed 8 Oct 86 10:55:02-PDT
From: Abhay Mehta <AMEHTA@Sushi.Stanford.EDU>
Subject: Balance
To: cs306-distribution: ;
Message-ID: <12245204089.21.AMEHTA@Sushi.Stanford.EDU>
A lot of you are having difficulty understanding problem 13:
Balance is a function you write that has the following properties:
1. The sexp that it returns has the same fringe as that of the input
list.
2. The returned sexp is balanced.
So (balance '((a . (b . c)) . d)) should be '((a . b) . (c . d))
and (balance '((a . (b . (c . (d . e)))) . f)) will return
'((a . (b . c)) . (d . (e . f)))
Also note that in the clisp environment, sexpressions will be written
as lists wherever possible i.e. '((a . b) . (c . d)) will be written
as '((a . b) c . d)
Abhay
-------
∂08-Oct-86 1132 RA TEXing
What is the next file you would like me to tex?
∂08-Oct-86 1156 SANKAR@Score.Stanford.EDU Compass, Gyros etc.
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 8 Oct 86 11:56:40 PDT
Date: Wed 8 Oct 86 10:30:06-PDT
From: Sriram Sankar <SANKAR@Score.Stanford.EDU>
Subject: Compass, Gyros etc.
To: jmc@Sail.Stanford.EDU
Message-ID: <12245199548.43.SANKAR@Score.Stanford.EDU>
While on the topic, I just thought I'd mention that Gyros have another serious
problem (other than drift which may not be that significant with todays state
of the art). Gyros have to be corrected once in a while on long distance
flights. This is because things change with position on the earth. An extreme
situation is when one flies to the other side of the earth. The gyro will now
erroneously show that up is down, and down is up. I used to know the full
details, but I remember that there is a rule that states how often the
correction has to be made, and also how the correction has to be made.
Since commercial pilots depend on gyros (artificial horizon) to determine a
straight and level flight, they cannot correct their gyros (they have many of
them) by first flying straight and level at constant speed. I believe that
the correction is made using the combined readings of the magnetic compass,
the artificial horizon (a gyro with two degrees of freedom) and the turn and
slip indicator (a gyro with one degree of freedom, the one that the artificial
horizon does not have). However, I am not sure how this is done.
Sriram.
-------
∂08-Oct-86 1224 SJG so long sucker?
Hi John:
I've got a couple of students who I may supervise in getting a
bunch of Symbolics to play Diplomacy against each other. I seem
to remember your telling me about a stylized version of the game
that you invented in your past some time; would you be prepared
to spend half an hour telling the three of us how it works?
I'm getting together with them tomorrow afternoon at 2; we could
see you then, at 1.30, or some other time. What say?
Thanks --
Matt
∂08-Oct-86 1550 RA Nancy Staggs
Nancy Staggs called re office furniture. Her tel. 326 7177.
∂08-Oct-86 1745 1F1BROWN%UKANVAX.BITNET@Forsythe.Stanford.Edu postponement
Received: from LINDY.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 8 Oct 86 17:40:14 PDT
Received: by lindy.stanford.edu with Sendmail; Wed, 8 Oct 86 17:40:55 pdt
Date: Wed, 8 Oct 86 17:39:36 PDT
From: <1F1BROWN@UKANVAX.BITNET>
Reply-To: 1F1BROWN%UKANVAX.BITNET@Forsythe.Stanford.Edu
To: JMC@SAIL.STANFORD.EDU
Subject: postponement
Date: Wed, 8 Oct 86 17:24 CDT
From: <1F1BROWN@UKANVAX.BITNET> (Glenn O. Veach)
Subject: postponement
To: jmc@sail.stanford.edu
X-Original-To: "jmc@sail.stanford.edu", 1F1BROWN
We will postpone the workshop. We will try for the first of April.
As soon as we get the date settled and the arragements made, we will
send out a new announcement.
Frank
∂08-Oct-86 1945 BRINK@Sushi.Stanford.EDU CS306
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 8 Oct 86 19:45:40 PDT
Date: Wed 8 Oct 86 19:44:11-PDT
From: Ed Brink <brink@Sushi.Stanford.EDU>
Subject: CS306
To: mehta@Sushi.Stanford.EDU, hsu@Sushi.Stanford.EDU, jmc@Sail.Stanford.EDU
Message-ID: <12245300418.15.BRINK@Sushi.Stanford.EDU>
On the advice of my advisor (I guess that's what he is for) I am dropping
CS306 in favor of CS242. I just wanted to say how much I have enjoyed 306, and
that I am sorry the realities of trying to make my studies fit my degree program
have made this necessary.
On the bittersweet side, I really didn't have a clue how I was going to get all
that reading and homework done by tomorrow, but that has NO influence on my
decision!
I'm keeping the class notes and EKL notes on purpose. I think they're worth
keeping.
Professor, in case you wonder, I'm the tall guy on your right in front who
remembers the 704.
..Ed
-------
∂08-Oct-86 2331 GLB
I have put the a copy of the tutorial in your mailbox.
∂09-Oct-86 0924 BRINK@Sushi.Stanford.EDU re: CS306
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 9 Oct 86 09:24:02 PDT
Date: Thu 9 Oct 86 09:21:53-PDT
From: Ed Brink <brink@Sushi.Stanford.EDU>
Subject: re: CS306
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 8 Oct 86 21:02:00-PDT
Message-ID: <12245449275.13.BRINK@Sushi.Stanford.EDU>
(Laughter...)
Could be...
But I would have toughed it out if I could have done. I'm supposed to be just
half time, and 13 units is not half time by SITN's standards. Maybe I can
work it in later; I hope so.
..Ed
-------
∂09-Oct-86 0932 VAL McDermott again
Levesque reminds us that it's time to write a reply. How will we go about it?
∂09-Oct-86 0941 VAL re: RAship
To: FERNANDO@SUSHI.STANFORD.EDU, JMC@SAIL.STANFORD.EDU
[In reply to message from FERNANDO@Sushi.Stanford.EDU sent Tue 7 Oct 86 23:41:00-PDT.]
Tim,
Our interests are primarily in commonsense reasoning. My impression is that
you must be able to find people here whose interests are closer to yours.
But, of course, you are always welcome at our seminars.
Vladimir
∂09-Oct-86 1448 RA Nancy Skaggs
Skaggs, 326 7177, called re: office furniture.
∂09-Oct-86 1602 THEODOROU@Sushi.Stanford.EDU Grade for CS206/306
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 9 Oct 86 16:02:48 PDT
Date: Thu 9 Oct 86 16:00:30-PDT
From: Dimitris Theodorou <THEODOROU@Sushi.Stanford.EDU>
Subject: Grade for CS206/306
To: jmc@Sail.Stanford.EDU
Message-ID: <12245521840.41.THEODOROU@Sushi.Stanford.EDU>
I took CS206 from you Fall '84-'85 and then went for a leave of absence
during '85-'86. My transcript still shows a '*' as my grade for that
class. I believe that my grade came out to be a B+. Could you please
submit a grade change for me because I will be graduating this June.
My student number is 3253730.
Thank you,
Dimitris Theodorou
-------
∂10-Oct-86 0059 CLT arpa proposal
To: JMC, LES
arp86[1,clt] contains a draft of my contribution.
I will probably make some further small improvements
and clean up references. However it will do for
filling the blank and getting on with putting the
whole thing together.
∂10-Oct-86 1213 RA CS306 tutors
Edith Gilbertson from SITN called re the tutors' applications, She wants to
know whether you ok Jim Bratnober from HP in Boise and Dirk Hubregs.
She would like me to call her and let her know. If you want to call her,
her number is 5-6963.
∂10-Oct-86 1610 LES Robotics Exhibit Loan
To: JMC@SAIL.STANFORD.EDU, TOB@SAIL.STANFORD.EDU,
BXR@SAIL.STANFORD.EDU
The Computer Museum in Boston has been after us for some time to loan them
some of our old robotics equipment for an exhibit they are planning.
I propose that we do so under the terms stated below. This will enable
us to recall the equipment if we later decide that it should go elsewhere --
e.g. to a local museum. Alternatively, after a few years we might end up
telling them that they can keep some or all of this junk.
I invite comments on this draft letter.
Les
.require "let.pub" source_file;
.csl 1986 October 10
Oliver Strimpel
The Computer Museum
300 Congress Street
Boston, MA 02210
Dear Mr. Strimpel:
.fi
This is a slightly slow response to your letter of July 31 1985, requesting
the loan of some of our earlier experimental equipment for use in your
exhibits, as follows:
.begin indent 10,10; preface 0; crbreak
Rancho Arm
Hydraulic Arm
Stanford Arm
Orm
Cart (probably without camera, alas -- it seems to be lost)
Piano-mover
.end continue
We will agree to make any or all of these artifacts available under the following
terms.
1. Packing and shipping costs to get these items to your facility will
be borne by the Computer Museum.
2. This is to be a renewable five-year loan subject to the following conditions:
.continue
(a) you may terminate it at any time by offering to return these items with the
shipping costs prepaid;
.continue
(b) Stanford may terminate it with three months advance notice, in which case
Stanford will bear the return shipping costs;
.continue
(c) at the end of five years, Stanford may either offer to extend this loan
under the same or different terms or may exercise option (b).
As we discussed, I plan to have a number of our documentary films
transferred to video tape and will have copies made available to you at
copying cost in VHS format.
My understanding is that you are interested in the following films:
.begin indent 10,10; preface 0; crbreak
Butterfinger
Hear Here
Instant Insanity
Motion and Vision
Computer Interactive Picture Processing (if I can find a copy)
Display Simulations of 6-legged Walking
MIT Logo Project (if I can find a copy)
.end
If you agree to these terms, please sign and return a copy of this letter.
Please cross out any films you are not interested in.
.indent 0,0;
∂10-Oct-86 1627 THEODOROU@Sushi.Stanford.EDU re: Grade for CS206/306
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 10 Oct 86 16:27:48 PDT
Date: Fri 10 Oct 86 16:18:00-PDT
From: Dimitris Theodorou <THEODOROU@Sushi.Stanford.EDU>
Subject: re: Grade for CS206/306
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Thu 9 Oct 86 20:17:00-PDT
Message-ID: <12245787170.24.THEODOROU@Sushi.Stanford.EDU>
thank you very much
-------
∂10-Oct-86 1647 RA Monday
To: JMC@SAIL.STANFORD.EDU
CC: ZM@SAIL.STANFORD.EDU, eppley@SCORE.STANFORD.EDU
Because of the holiday I will take Monday off.
∂12-Oct-86 2128 binford@su-whitney.arpa Robotics Exhibit Loan
Received: from WHITNEY.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 12 Oct 86 21:28:27 PDT
Received: by su-whitney.arpa with Sendmail; Sun, 12 Oct 86 21:27:18 pdt
Date: Sun, 12 Oct 86 21:27:18 pdt
From: Tom Binford <binford@su-whitney.ARPA>
To: LES@SAIL.STANFORD.EDU
Cc: JMC@SAIL.STANFORD.EDU, TOB@SAIL.STANFORD.EDU, BXR@SAIL.STANFORD.EDU
In-Reply-To: Les Earnest's message of 10 Oct 86 1610 PDT
Subject: Robotics Exhibit Loan
Les
The letter seems fine to me. I will check about the camera.
I wonder whether the translating camera mount still is there.
Tom
∂13-Oct-86 1010 RA CS 306
Dirk Hebregs the tutor from Lake Stevens, Washington wanted to talk to you
about hardware they can use for the programming in the classs. He would
like you to call him (206) 335 2491.
∂13-Oct-86 1100 JMC
Swedish bill.
∂13-Oct-86 1139 RA gps[f86,jmc]
This file is now on line.
∂13-Oct-86 1212 RA going out
I am going to the bookstore.
∂13-Oct-86 1357 RA Academician Ershov
Jay Powell called to let you know that Ershov will be here in November. She
wanted to know whether you would like to host him at your house. Her tel.
today is 924 6194 and from tomorrow 332 6471.
∂13-Oct-86 1429 GLB Re: cs 306
∂13-Oct-86 1352 B.BELGIUM@HAMLET.STANFORD.EDU Re: cs 306
Received: from HAMLET.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 13 Oct 86 13:52:07 PDT
Date: Mon 13 Oct 86 13:51:09-PDT
From: Phil Andrew <B.Belgium@HAMLET.STANFORD.EDU>
Subject: Re: cs 306
To: GLB@SU-AI.ARPA
cc: fl@OTHELLO.STANFORD.EDU
In-Reply-To: Message from "Gianluigi Bellin <GLB@SAIL.STANFORD.EDU>" of Mon 13 Oct 86 11:39:00-PDT
Message-ID: <12246546868.107.B.BELGIUM@HAMLET.STANFORD.EDU>
This is quite an unusual request due to its size. I would like to have more
information about this program and its space requirements before I increase
the class directory allocation five-fold. Also, what are the reasons that
the students may require more disk allocation once they start running the
program?
Thanks for your cooperation.
Philip Andrew
LOTS Faculty Liaison
-------
∂13-Oct-86 1624 LIBRARY@Score.Stanford.EDU Tech Reports
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 13 Oct 86 16:24:32 PDT
Date: Mon 13 Oct 86 16:22:03-PDT
From: C.S./Math Library <LIBRARY@Score.Stanford.EDU>
Subject: Tech Reports
To: jmc@Sail.Stanford.EDU
Message-ID: <12246574340.35.LIBRARY@Score.Stanford.EDU>
TO: John McCarthy
I sent you the references to what looked like the old technical report
you were seeking for a colleague in Poland. These references would
have been addressed to you from a generic SOCRATES terminal, since I
just walked up the Math-CS Library to do the search there.
There should be at least one copy of the report in the Math-CS Library.
Henry Lowood
-------
∂13-Oct-86 1833 HSU@Sushi.Stanford.EDU cs306
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 13 Oct 86 18:33:53 PDT
Date: Mon 13 Oct 86 18:32:12-PDT
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: cs306
To: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12246598032.13.HSU@Sushi.Stanford.EDU>
Professor McCarthy,
I have several questions regarding the course:
Some TV student asked about the possibility of porting EKL to the
machines on their site, is it possible?
Do you have the handouts from last year? Or is there any place where I
could get hold of one? It has been 4 years since I took your course,
therefore I'd like to take a look at the more "current" organization of
the course/handouts/homeworks so that I can be better prepared to TA the
course.
I'm preparing another handout about some more administrative issues,
do you have any announcements that you'd like to make? I will
distribute them in class tomorrow.
Thanks.
Yung-jen
-------
∂14-Oct-86 0715 AI.PETRIE@MCC.COM Emycin Criticism
Received: from MCC.COM by SAIL.STANFORD.EDU with TCP; 14 Oct 86 07:15:33 PDT
Date: Tue 14 Oct 86 09:15:35-CDT
From: Charles Petrie <AI.PETRIE@MCC.COM>
Subject: Emycin Criticism
To: jmc@SU-AI.ARPA
Reply-To: Petrie@MCC
Message-ID: <12246737002.13.AI.PETRIE@MCC.COM>
I should like to reference one of your criticisms of Emycin. The
example I remember is that of a system for administering a hospital
that continues to follow routine procedures even during a fire.
Could you please give me a reference for such a criticism, if one
exists?
Thanks,
Charles Petrie
-------
∂14-Oct-86 0800 EC.MA1@Forsythe.Stanford.Edu Math/CS": Socrates display
Received: from LINDY.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 14 Oct 86 07:59:58 PDT
Received: by lindy.stanford.edu with Sendmail; Fri, 14 Oct 83 07:59:57 pdt
Date: Tue, 14 Oct 86 08:00:05 PDT
From: Socrate <EC.MA1@Forsythe.Stanford.Edu>
To: JMC@SAIL.STANFORD.EDU
Subject: Math/CS": Socrates display
*
Technical Reports / Search: Find AUTHOR MCCARTHY, J and TITLE MANIFOLDS
Result: 2 citations
Citation 1
AUTHOR: Stanford University. Applied Mathematics and Statistics Laboratory.
(STANAMSL) 36.
TITLE: The stability of invariant manifolds.
February 21, 1955. 25p.
LOCATION: Math & Comp Sci (205036)
SERIES: Technical report.
OTHER: United States. Office of Naval Research. (ONR)
McCarthy, J.
NUMBER: 36. (Report Number)
Nonr-225(11) (NR-041-086). (Contract/Grant Number)
NOTES: Item TECM-27075 (Technical Repts) Language: eng
Citation 2
AUTHOR: Stanford University. Applied Mathematics and Statistics Laboratories.
T.R. 36.
TITLE: THE STABILITY OF INVARIANT MANIFOLDS.
1955.
LOCATION: 000491: Math & Comp Sci
OTHER: McCarthy, J.
NUMBER: T.R. 36. (Report Number)
NOTES: Item TECM-14287 (Technical Repts) Language: unk
∂14-Oct-86 0803 AI.PETRIE@MCC.COM re: Emycin Criticism
Received: from MCC.COM by SAIL.STANFORD.EDU with TCP; 14 Oct 86 08:03:10 PDT
Date: Tue 14 Oct 86 10:02:46-CDT
From: Charles Petrie <AI.PETRIE@MCC.COM>
Subject: re: Emycin Criticism
To: JMC@SU-AI.ARPA
cc: AI.PETRIE@MCC.COM
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Tue 14 Oct 86 07:54:00-CDT
Reply-To: Petrie@MCC
Message-ID: <12246745593.13.AI.PETRIE@MCC.COM>
Thanks for clearing up my confused memory and for the reference.
Charles
-------
∂14-Oct-86 0849 RA be in late
To: JMC@SAIL.STANFORD.EDU, ZM@SAIL.STANFORD.EDU,
eppley@SCORE.STANFORD.EDU
I don't feel well and will be coming in later probably between
11:00 and noon.
Rutie
------
∂14-Oct-86 0907 VAL re: McDermott
[In reply to message rcvd 09-Oct-86 11:50-PT.]
When can we meet to discuss replying to McDermott and the bug in the
Sterling/Shapiro book?
∂14-Oct-86 0917 SJG what is the qualification problem?
Hi John:
I've been arguing with Mike (and the rest of the Logic Group) about
this. The question is, is the qualification problem:
1. Figuring out how to incorporate knowledge implicit in one's
knowledge of a domain into a planning system? I guess this might
be called the *inferred precondition* problem. I know that a potato
in my tailpipe will stop my car's running, but still don't bother
to check it out before turning the key in the ignition.
Or, is it
2. Getting a satisfactory domain description in the first place?
My mother probably doesn't realize that a potato in the tailpipe would
stop her car working, so she doesn't check it out because it makes no
sense for her to under *any* circumstances. But she is really in
trouble if there *is* a potato in the tailpipe.
Now, I argue that only the first of these is really any sort of a
planning problem; the second has much more to do with general issues
in AI and formalization of a (common sense?) domain. So which of these
did you intend to call the "qualification problem"?
Thanks. And thanks again for lunch -- I enjoyed it.
Matt
∂14-Oct-86 1127 RINDFLEISCH@SUMEX-AIM.ARPA Radford AI Mkt Survey Input
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 14 Oct 86 11:24:32 PDT
Date: Tue 14 Oct 86 11:23:06-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Radford AI Mkt Survey Input
To: Roeder@SUMEX-AIM.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA, Buchanan@SUMEX-AIM.ARPA,
Feigenbaum@SUMEX-AIM.ARPA, Genesereth@SUMEX-AIM.ARPA,
Shortliffe@SUMEX-AIM.ARPA, Nilsson@SU-SCORE.ARPA, JMC@SU-AI.ARPA,
Les@SU-AI.ARPA, TOB@SU-AI.ARPA
Message-ID: <12246782061.40.RINDFLEISCH@SUMEX-AIM.ARPA>
Doug, This is our input for the Radford Associates AI market survey, including
a list of CSD/KSL professionals classified according to our best interpretation
of the Radford job descriptions applied to the Stanford setting. Because of
the time pressure in doing this since getting the material last Thursday, I've
only been able to review the list with Bruce Buchanan, Les Earnest, and Ed
Feigenbaum.
It was difficult in some cases to map the Radford industrial classification
system onto our university scientific people for the same sorts of reasons we
have talked about w/r to distinguishing Stanford's CSS and University Manager
classifications. That is, the "Scientist" series (which we feel best describes
our (Sr.) Research Associates) merges scientific and managerial criteria, with
the higher classifications emphasizing predominantly functions of managing and
directing rather than individual scientific achievement. Thus, because they
were not manager/directors, some people who are outstanding, widely-respected
scientists were placed in lower classifications than might be appropriate in
terms of their achievements and real value to the university. The Radford
"Technical Manager" series does not help in resolving this distinction.
Note also the following changes to the list of people since we spoke last:
Delete
Jussi Ketonen Will not be returning from leave at Lucid, Inc.
Add
Daniel Kolkowitz Omitted earlier, my oversight
Igor Rivin New hire
N. Shankar New hire
William Westfield Omitted earlier, my oversight
Doug: I'd be glad to talk to you about the specifics of why we classified
various people as we did if you have any questions.
Other cc's: If I haven't been able to talk to you about this and you want to
know more about the Radford survey, let me know.
Tom R.
-------------
CSD/KSL Technical Personnel Classifications
Radford Associates AI Market Survey
Scientific Director
Earnest, Lester (Sr Res Assoc)
Rindfleisch, Thomas (Sr Res Assoc)
Scientist 4
Engelmore, Robert (Sr Res Assoc)
Fagan, Lawrence (Sr Res Assoc)
Scientist 3
Blum, Robert (Res Assoc)
Brown, Harold (Sr Res Assoc)
Clancey, Bill (Sr Res Assoc)
Friedland, Peter (Sr Res Assoc)
Ginsberg, Matthew (Res Assoc) (Promotion to Sr Res Assoc pending)
Hayes-Roth, Barbara (Sr Res Assoc)
Khatib, Oussama (Sr Res Assoc)
Lifschitz, Vladimir (Sr Res Assoc)
Nii, H. Penny (Sr Res Assoc)
Weyhrauch, Richard (Res Assoc)
Scientist 2
Brinkley, James (Res Assoc)
Cornelius, Craig (Res Assoc)
Ponce, Jean (Res Assoc)
Rivin, Igor (Res Assoc) (New hire about 1/1/87)
Shankar, N (Res Assoc) (New hire 10/1/86)
Singh, Narinder (Res Assoc)
Smith, David (Res Assoc)
Talcott, Carolyn (Res Assoc)
Scientist 1
Walker, Michael (Sci Pgmr IV)
Software Systems Engr 4
Frost, Martin (Sys Pgmr IV)
Gilmurray, Frank (Sys Pgmr IV)
Sweer, Andrew (Sys Pgmr IV)
Yeager, William (CSS)
Software Systems Engr 3
Kolkowitz, Daniel (Sys Pgmr III)
Lane, Christopher (Sci Pgmr IV)
Schmidt, Christopher (Sys Pgmr IV)
Software Systems Engr 2
Acuff, Richard (Sys Pgmr III)
Aiello, Nelleke (Sci Pgmr III)
Barnhouse, Steve (Sci Pgmr III)
Combs, David (Sci Pgmr III)
Crispin, Mark (Sys Pgmr III)
Gelman, Andrew (Sci Pgmr III) (new hire)
Hewett, Michael (Sci Pgmr III)
Nishimura, Sayuri (Sci Pgmr III)
Oliphant, Stephen (Sci Pgmr III) (no BS)
Rappaport, Steven (Sci Pgmr III)
Rice, James (Sci Pgmr III) (promotion to IV pending)
Saraiya, Nakul (Sci Pgmr III) (pending visa)
Tu, Samson (Sci Pgmr III)
Tucker, Robert (Sys Pgmr III)
Westfield, William (Sys Pgmr III)
Wulfman, Cliff (Sci Pgmr III)
Software Systems Engr 1
Reuling, John (Sci Pgmr II) (promotion to III pending)
Knowledge Engineer 1
Rohn, Janice (Sci Pgmr I)
-------
∂14-Oct-86 1130 RA Nancy Staggs
Please call Staggs, re office furniture, 326 7177
∂14-Oct-86 1216 gref@nrl-aic AAAI support
Received: from NRL-AIC.ARPA by SAIL.STANFORD.EDU with TCP; 14 Oct 86 12:15:56 PDT
Date: Tue, 14 Oct 86 15:03:40 edt
From: John Grefenstette <gref@nrl-aic>
Message-Id: <8610141903.AA18516@nrl-aic>
To: jmc@su-ai
Subject: AAAI support
Dear Prof. McCarthy:
I am serving as program chairman for the 2nd International
Conference on Genetic Algorithms and Their Applications.
The Conference is tentatively scheduled for July 28-31, 1987
at MIT. We are seeking nominal support ($2K) from AAAI
in order to subsidize student registration fees.
Additional support is being provided by the Navy Center for
Applied Research in AI at NRL.
Conference topics will include: foundations of genetic algorithms,
machine learning using genetic algorithms, classifier systems,
apportionment of credit algorithms, relationships to other search
and learning paradigms.
Conference participation is open to all researchers interested in
the Conference topics. The first Conference (July 1985) attracted
about 100 participants. We expect about 150 next year.
A call for papers will be issued shortly. Submissions will be
reviewed by the Program Committee:
John H. Holland University of Michigan (Conference Chair)
Lashon B. Booker Navy Center for Applied Research in AI
Dave Davis Bolt Beranek and Newman
Kenneth A. De Jong Navy Center for Applied Research in AI
John J. Grefenstette Navy Center for Applied Research in AI
David E. Goldberg University of Alabama
Stephen F. Smith Carnegie-Mellon Robotics Institute
Stewart W. Wilson Rowland Institute for Science
I would appreciate any information concerning AAAI support.
Thank you.
Dr. John J. Grefenstette
Navy Center for Applied Research in AI
Code 5510
Naval Research Laboratory
Washington, DC 20375-5000
gref@nrl-aic.arpa
(202) 767-2685
∂14-Oct-86 1220 VAL semi-annual context talk
I've scheduled Mike Georgeff to talk about the frame problem next week. How about
Oct. 30 for your context talk?
∂14-Oct-86 1226 RA Chris Rogers
Rogers says that you told him to make an appointment with you for this afternoon.
He will come in at 2:45.
∂14-Oct-86 1443 AMEHTA@Sushi.Stanford.EDU New TA hours
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 14 Oct 86 14:43:51 PDT
Date: Tue 14 Oct 86 14:40:38-PDT
From: Abhay Mehta <AMEHTA@Sushi.Stanford.EDU>
Subject: New TA hours
To: cs306-distribution: ;
Message-ID: <12246818021.8.AMEHTA@Sushi.Stanford.EDU>
My TA hours will now be 9:00 - 11:00 Wednesdays instead of 10 - 12.
Same location.
Abhay Mehta
-------
∂14-Oct-86 1451 DEK Ershov visit
He (and his doctor!) will be staying at my house Nov 4 to 11.
∂14-Oct-86 1517 RA re AAAI workshop
Dr. Larry Fagin needs information re writing a proposal for AAAI workshop.
He would like you to call him at 723 6979. I told his secretary to call
Claudia.
∂14-Oct-86 1710 RA coming in late
To: JMC@SAIL.STANFORD.EDU, ZM@SAIL.STANFORD.EDU,
eppley@SCORE.STANFORD.EDU
I have a doctor appointment tomorrow morning and therefore I will
come in late. I will be here around 11:00.
Rutie
-----
∂14-Oct-86 2015 LES CS300 lecture
To: JMC, VAL, CLT
You are scheduled to describe current research in formal reasoning to the
CS300 class this Thursday (10/16) from 2:45 to 4:00pm in Building 320,
Room 334 (Geology Corner). Does anyone need an overhead projector or
other special equipment?
∂15-Oct-86 0900 JMC
Halmos 4:15 380W
∂15-Oct-86 0900 JMC
prescription, Ash
∂15-Oct-86 0956 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
The Yale Shooting: Non-Monotonic Formalisms Not Guilty
Vladimir Lifschitz (VAL@SU-AI)
Stanford University
Thursday, October 16, 4pm
MJH 252
John McCarthy has proposed to use circumscription for solving two
problems encountered in attempts to formalize reasoning about action, the
"qualification problem" and the "frame problem". His formulation was
found to be inadequate. A simple illustration was given by Steve Hanks
and Drew McDermott, who considered the sequence of 3 actions: LOAD (a gun),
WAIT, and SHOOT (an individual named Fred). They wrote axioms which
describe properties of these actions in the style of McCarthy's approach
and showed that the expected result, Fred's death, cannot be derived from
the axioms using circumscription. This difficulty is mentioned by
McDermott, among others, in his paper, "Critique of Pure Reason", as an
argument against the "logicist" approach to AI.
Three papers presented at AAAI-86 (Kautz, Shoham and Lifschitz)
met this challenge by formalizing, in different contexts, the idea of
"chronological minimization": selecting models in which abnormal events
occur as late as possible. All these formulations had to modify or
generalize McCarthy's definition of circumscription.
In this talk I propose a solution not based on chronological
minimization. With a somewhat different choice of primitive predicates,
the effects of actions can be characterized by simple axioms in the
language of situation calculus plus traditional circumscription. Moreover,
in these applications circumscription is usually "tractable": its result
can be determined by predicate completion and similar methods.
∂15-Oct-86 1000 B.BELGIUM@OTHELLO.STANFORD.EDU Re: LOTS allocation for EKL
Received: from OTHELLO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Oct 86 09:59:11 PDT
Date: Wed 15 Oct 86 09:58:30-PDT
From: Phil Andrew <B.Belgium@OTHELLO.STANFORD.EDU>
Subject: Re: LOTS allocation for EKL
To: JMC@SU-AI.ARPA
cc: fl@OTHELLO.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 13 Oct 86 14:28:00-PDT
Message-ID: <12247028804.208.B.BELGIUM@OTHELLO.STANFORD.EDU>
Ok, I just wanted to make sure -- sometimes people overreact to what they
perceive their needs WILL be ...
The clas directory size is now 1000 pages.
-Phil
-------
∂15-Oct-86 1022 SJG qualification problem problem
John:
I have a nickel riding on your answer to my msg of yesterday.
What say?
Thanks.
Matt
∂15-Oct-86 1124 CLT Qlisp reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
We meet this Friday noon, Oct 17 in MJ352.
Joe will report (briefly) on his visit to Argonne
and the Alliant Users Meeting.
∂15-Oct-86 1257 SJG re: what is the qualification problem?
[In reply to message rcvd 15-Oct-86 12:06-PT.]
I'm not sure who gets it, yet. My position is that if you start
with a domain description simply in terms of domain constraints,
and from these domain constraints are able to come up with a list
of the "inferred" preconditions to any particular action, then you
have effectively solved the qualification problem.
Karen Myers (and others in the Logic Group) argued that no, in order
to solve the qualification problem, you had to come up with a domain
description (in terms of explicit preconditions, domain constraints,
or whatever) rich enough to include *all* of the hidden preconditions.
So they would argue that my mother has NOT solved the qualification
problem, while I would argue that she HAS, and that her potential
difficulty if her car has a potato in its tailpipe is a problem of
quite a different sort.
So who gets the nickel?
Matt
∂15-Oct-86 1448 SJG qualification problem
Unless you object, I'll circulate your opinion to the other
people in the logic group. My fee for promoting your beliefs
is $50.
Matt
∂15-Oct-86 1600 JMC
Halmos
∂16-Oct-86 0934 ACT
10/16/86 TPH Milk 5.75 4.50 2.00 5.00
∂16-Oct-86 1725 RA Joshi, Univ. of Pennsylvania
Joshi called you re a msg. he had sent you; his tel. (215) 898 8540. He will
call you again tomorrow.
∂17-Oct-86 0022 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM difference of squair routes
Received: from [128.81.2.8] by SAIL.STANFORD.EDU with TCP; 17 Oct 86 00:22:03 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by MIRO.SNETC.Symbolics.COM via CHAOS with CHAOS-MAIL id 17329; Fri 17-Oct-86 03:21:54 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 2665; Fri 17-Oct-86 00:19:21 PDT
Date: Fri, 17 Oct 86 00:18 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: difference of squair routes
To: jmc@SU-AI.ARPA
Message-ID: <861017001824.6.RWG@TSUNAMI.SPA.Symbolics.COM>
I just got back from San Diego. Will probably not stay long tonight (Thurs), but should
be in ~2300 to 0400 Friday/Sat.
∂17-Oct-86 0207 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM bleary-eyed guess:
Received: from [128.81.2.8] by SAIL.STANFORD.EDU with TCP; 17 Oct 86 02:07:03 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by MIRO.SNETC.Symbolics.COM via CHAOS with CHAOS-MAIL id 17331; Fri 17-Oct-86 05:07:05 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 2669; Fri 17-Oct-86 02:04:24 PDT
Date: Fri, 17 Oct 86 02:03 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: bleary-eyed guess:
To: jmc@SU-AI.ARPA
Message-ID: <861017020344.8.RWG@TSUNAMI.SPA.Symbolics.COM>
All but a finite number of "best" approximations to, e.g., pi will be of the form
sqrt(q↑2+p)-sqrt(q↑2-p), where p/q is an approximant to pi.
∂17-Oct-86 0821 PHY Janos Komlos of UCSD
To: "@SEARCH.DIS[1,PHY]"@SAIL.STANFORD.EDU
To: Search Committee
From: Leo Guibas
Subject: Janos Komlos of UCSD
Janos is a well-known probabilist/combinatorialist who in recent years
(like many of his hungarian compatriots) has turned his attention to
problems in theoretical computer science, with impressive results. He is
being considered for a joint Math-CS appointment. The Math department is
anxious to act on his case, because Komlos is holding on to a lucrative
offer from Rutgers (current home of Chvatal and Szemeredi as well, for
those who remember...). I have left with Phyllis a folder with his vita
and references, as collected by the Math search committee, in consultation
with Don and myself. Letters from Lovasz and Tarjan are still expected.
Please take a careful look at the folder and try to assess how Janos would
fit into our department. I am personally extremely positive about the
appointment -- so is Don. If there is enough sentiment for it, we can invite
him to come and give a talk. In the interest of time, the Math department
would prefer if we acted without such a visit.
I'd like to be able to respond to Math with our position by sometime next
week. Please let me know what your feelings are, especially if they are
negative. Tentatively, save the lunch hour of next Tuesday 10/21 (sorry Nils)
for a possible meeting, in case we cannot agree by e-mail beforehand.
∂17-Oct-86 0900 JMC
Elliott,prescription
∂17-Oct-86 1322 VAL draft available
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
My talk yesterday was based on the paper on formal theories of action I'm
writing now. Those interested in seeing the first 13 pages of the paper can
pick up copies in my office, MJH362.
Vladimir
∂17-Oct-86 1454 VAL non-monotonic seminar
What do you think about this plan: to have two more talks on the frame problem
(Georgeff and Ginsberg), and then a discussion session? (Then we'll postpone
your context talk by 2 weeks.) Would you be willing to start and lead the
discussion?
∂17-Oct-86 1627 AIR
To: JMC, LES
I'd like to start on October 27. - Arkady
∂17-Oct-86 1810 UDI@Score.Stanford.EDU program
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Oct 86 18:10:38 PDT
Date: Fri 17 Oct 86 16:50:33-PDT
From: Ehud Shapiro <UDI@Score.Stanford.EDU>
Subject: program
To: jmc@Sail.Stanford.EDU
Message-ID: <12247628105.22.UDI@Score.Stanford.EDU>
THis is less then what I thought and promised. The enclosed
program verifies the internal consistency of a circuit (i.e.
undesired internal connections), but does not verify that
there are no other components in the circuit that are connected
to internal nodes of the verified component.
Udi
p.s. What is Lfschitz e-mail address?
/*
Specifying circuits
Extending Program 2.2, Page 23, in the Art of Prolog
*/
% Base relations remain the same
/*
Compound elements:
The basic approach is that every compound elements
be defined as follows:
element(Ports,Internal) :-
subelement1(Ports1,Internal1),
subelement2(Ports2,Internal2),
....
subelementn(Portsn,Internaln),
Ports+Internal=Ports1+Internal1+...+Portsn+Internaln,
distinct(Ports+Internal),
Where + is multiset union.
Note: this approach guarantees only internal consistency
of circuits. If there other other elements connected
to internal ports in the circuit, this fact will not be
detected by the current program, unless the component verified
is the entire circuit.
*/
inverter([Input,Output],[]) :-
transitor(Input,ground,Output),
resistor(power,Output),
distinct([Input,Output]).
nand←gate([Input1,Input2,Output],[X]) :-
transistor(Input1,X,Output),
transistor(Input2,ground,X),
resistor(power,Output),
distinct([Input1,Input2,Output,X]).
and←gate([Input1,Input2,Output],[X1,X2]) :-
nand←gate([Input1,Input2,X1],[X2]),
inverter(X1,Output),
distinct([Input1,Input2,Output,X1,X2]).
distinct([X|Xs]) :- not member(X,Xs), distinct(Xs).
distinct([]).
/*
Note: In these simple examples the lists were constructed "manually",
but in more complex cases they can be appended, to construct
the list of Internal ports of the compound element, and the union
of all ports in the element.
*/
-------
∂18-Oct-86 0417 @DIAMOND.S4CC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM sqrt-sqrt, I'm wet.
Received: from [128.81.51.3] by SAIL.STANFORD.EDU with TCP; 18 Oct 86 04:17:36 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by DIAMOND.S4CC.Symbolics.COM via CHAOS with CHAOS-MAIL id 20130; Sat 18-Oct-86 07:16:26 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 2811; Sat 18-Oct-86 04:14:46 PDT
Date: Sat, 18 Oct 86 04:13 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: sqrt-sqrt, I'm wet.
To: jmc@SU-AI.ARPA
Message-ID: <861018041340.4.RWG@TSUNAMI.SPA.Symbolics.COM>
Empirically, my conjecture about sqrt(q↑2+p)-sqrt(q↑2-p) is a complete
crock. I've only looked at sqrt(n)-sqrt(m) ~ pi, but evidently things
are much more complicated than with rational approximations. But, as
you might suspect, the number of digits of agreement for "best" (n,m)
pairs is roughly digitsin(n) + digitsin(n-m).
On the bright side, there seems to be a lot of exploitable smoothness.
If (n,m) is a good pair, then the next possible pair is (n+k+1,m+k),
and if sqrt(n+k+1)-sqrt(m+k)=x, then k is a rational function of n, m,
and x↑2. Furthermore, k seems to increase fairly smoothly. The analog
of continued fraction terms is the gaps between new best ks. In the case
of continued fractions, these can be guessed from the approximation error.
The case at hand, however, looks completely crazy.
Also, it may be just a fluke, but at around n=10↑6, the distance between
new best pairs seems to be stretching dramatically. I'm going to let
a bfloat version run over"night".
∂18-Oct-86 1217 CLT rivin
To: JMC, LES
I talked to him today. He is still thinking.
I couldn't tell from the conversation which way he is leaning.
∂18-Oct-86 1345 shoham-yoav@YALE.ARPA Re: inaccurate history
Received: from YALE-CELRAY.ARPA by SAIL.STANFORD.EDU with TCP; 18 Oct 86 13:44:52 PDT
Received: by yale-celray.YALE.ARPA; Sat, 18 Oct 86 16:31:45 edt
Date: Sat, 18 Oct 86 16:31:45 edt
From: Yoav Shoham <shoham-yoav@YALE.ARPA>
Message-Id: <8610182031.AA12388@yale-celray.YALE.ARPA>
Subject: Re: inaccurate history
To: John McCarthy <JMC@SAIL.STANFORD.EDU>
In-Reply-To: John McCarthy <JMC@SAIL.STANFORD.EDU>, 18 Oct 86 1258 PDT
the 1969 paper that introduced the situation calculus had a treatment
of the law of falling bodies that included continuous time ...
Oops. I'll have to go over the paper again. Thanks a lot for these comments,
by the way. I hope more will be coming. With enough of them, I might end
up with something halfway decent.
Yoav.
-------
-------
∂18-Oct-86 1525 CERF@A.ISI.EDU [CERF@A.ISI.EDU: Next Digital Library Meeting]
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 18 Oct 86 15:25:31 PDT
Date: 18 Oct 1986 18:24-EDT
Sender: CERF@A.ISI.EDU
Subject: [CERF@A.ISI.EDU: Next Digital Library Meeting]
From: CERF@A.ISI.EDU
To: jmc@SU-AI.ARPA
Message-ID: <[A.ISI.EDU]18-Oct-86 18:24:52.CERF>
John,
I hope this reaches you - by this time your mailbox may well have migrated to
a private workstation or other system...please let me know if, despite the
rather short notice, it would be possible for you to join us in DC for the next
Digital Library discussion.
If the dates are not compatible with your currentplans, could you suggest some
alternatives which would work for; this is all assuming, of course, that you
continue to entertain an interest in this area.
Many thanks,
Vint Cerf
Begin forwarded message
Received: By A.ISI.EDU via direct-append with Hermes; 10 Oct 86 08:55:09-EDT
Date: 10 Oct 1986 08:55-EDT
From: CERF@A.ISI.EDU
To: Bibliotects:
Subject: Next Digital Library Meeting
Message-ID: <[A.ISI.EDU]10-Oct-86 08:55:07.CERF>
Sender: CERF@A.ISI.EDU
Fcc: PS:<CERF>FCC.TXT
Warning: This document may contain naked TAB characters. TAB stops
are at every 8 columns (8, 16, 24, ...). - VGC
Dear Fellow Bibliotects:
After the August inaugural meeting of the Digital Library Working Group (DLWG),
Bob Kahn and I made a first pass at a possible outline for the DLWG report. We
also made a stab at assigning specific sections to working group members as
an aid to focusing attention, responsibility, and opportunity for public
shaming should you not have anything to contribute at the next meeting...
Herewith, the outline:
0) Executive Summary
Kahn 1) Introduction
Essential concepts: univorm access to digital information;
idea of transitive copyrights; leyered interfaces; new
retrieval techniques and systems (e.g. connection machines);
importance of scale; technology limitations and goals; regular
capture of information in manipulable (standard) form...
Kahn 2) Architecture (Bibliotecture?)
repository concept; communications; multilevel client/access
services; accounting/control.
Popek 3) Repositories (is a system of speculative reasoning a supposeitory?)
interfaces; distributed; accommadation of heterogeneity.
Cerf 4) Communication Environment
networking; public; private; ISDN; packet switched; other;
bandwidths; costs
Cerf 5) Information collection and authenticity (trademark)
What information? Public domain? copyrighted?
Carbonell 6) Information access, search,retrieval and validation
Hillis Natural Language Aspects. Special searching
facilities
Hillis 7) Information Manipulation Opportunities
Reddy 8) Information Presentation Concepts
Reddy 9) Research Topics, Technology Limits and Goals
Cataloging, Indexing, Representation, ...
Lyons 10) Legal and Adminstrative Aspects
Content liability, copyright concepts, relicensing idea,
accounting/collection requirements, role of copyright coord corp
11) Program Plan
Who participates? What initial Services?
12) Summary
13) References
14) Business Plan
Site license/sponsor/year. Unlimited access to information
(lump annual royalty or per piece royalty? Copy Control?)
*************
It is proposed to try to convene a second meeting during the period
7-8-9 November which are a Friday-Sat-Sun. In Washington DC if possible.
Please acknowledge receipt of this message and respond to CERF@A.ISI.EDU
regarding your availability on the dates suggested. If you have alternate
dates or preferences as to including or avoiding weekends, please let me know.
If you have material prepared based on your notes from the last meeting, please
feel free to send it to me for further distribution.
I can be reached at (703) 620-8990 or at my ARPANET Mail address.
Invitations will be extended to others such as John McCarthy at Stanford and
Don Lindberg at the National Library of Medicine (Director) who are part of the
working group but could not attend last time.
I will begin putting material on-line to make more concrete the Digital Library
concept and to provide a target for you all to shoot at (please shoot at the
concepts and not the author!).
Looking forward to hearing from you soon.
Vint
--------------------
End forwarded message
∂19-Oct-86 0334 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM would you believe just damp?
Received: from [128.81.2.8] by SAIL.STANFORD.EDU with TCP; 19 Oct 86 03:34:17 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by MIRO.SNETC.Symbolics.COM via CHAOS with CHAOS-MAIL id 18037; Sun 19-Oct-86 06:33:19 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 2850; Sun 19-Oct-86 03:31:33 PDT
Date: Sun, 19 Oct 86 03:30 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: would you believe just damp?
To: jmc@SU-AI.ARPA
cc: rwg@RUSSIAN.SPA.Symbolics.COM
Message-ID: <861019033031.2.RWG@TSUNAMI.SPA.Symbolics.COM>
For 1/2, the successive best approximations seem indeed of the form
sqrt(q↑2+p)-sqrt(q↑2-p), where p/q = 1/2, 2/4, 3/6, 4/8, ..., excepting
the first two (a fairly finite number), which are sqrt(0)-sqrt(0), and
sqrt(2)-sqrt(1). (Actually, the first one is 0*1/(0*2) !)
1/3 and 2/3 settle into the analogous patterns, the latter having two
sporadic intrusions at (31,24) and (134,119). 3/2 is fairly strange:
except for four sporadics following (2,0), the general (ai,bi) lie in
two interlaced quadratic progressions, but with (ai+bi)/2 not square.
1/sqrt(2) is amazingly simple: sqrt(n (n+1)/2)-sqrt(n (n-1)/2).
(= 1/sqrt(2) + 1/(8 sqrt(2) n↑2) + O(n↑-4)).
You might think this could generalize to n (n+or-a)/2 for a/sqrt(2),
but 3/sqrt(2) comes out as two interlaced quadratics (plus two sporadics),
neither of which is n (n+or-3)/2.
sqrt(a n↑2 + b n) - sqrt(a n↑2 - b n) correctly predicts b/sqrt(a) for b=1,
a = 3, 4, 5, 7, 8, 9, ... but skips every other "best" for
a = 2, 6, 10, 14, ... .
For b = 2, the formula catches only every fourth "best" for a = 8 (i.e. the
case 1/sqrt(2) again), wins for a = 9, 10, 11, 13, 14, misses every other
for a = 12, and fails utterly on a = 15, which eventually interlaces
sqrt(60*n↑2+49*n+10)-sqrt(60*n↑2+41*n+7) and
sqrt(60*n↑2+79*n+26)-sqrt(60*n↑2+71*n+21) ! Both of these expand as
2/sqrt(15) + 1/(14400 sqrt(15) n↑2) + ...). Something hairy is going on
here.
Note that all the above progressions of "bests" are infinitely denser than
those for pi, 1/pi, 2↑(1/3) and 2↑(-1/3), which are roughly exponential
rather than quadratic. (This has to do with smoothness.)
∂19-Oct-86 0413 CERF@A.ISI.EDU re: [CERF@A.ISI.EDU: Next Digital Library Meeting]
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 19 Oct 86 04:13:36 PDT
Date: 19 Oct 1986 07:13-EDT
Sender: CERF@A.ISI.EDU
Subject: re: [CERF@A.ISI.EDU: Next Digital Library Meeting]
From: CERF@A.ISI.EDU
To: JMC@SU-AI.ARPA
Message-ID: <[A.ISI.EDU]19-Oct-86 07:13:01.CERF>
In-Reply-To: The message of 18 Oct 86 1548 PDT from John McCarthy <JMC@SAIL.STANFORD.EDU>
John,
Many thanks for such a prompt response! I have to agree that red eyes do
not a refreshed mind make. Suppose, for the sake of argument that we started
on Friday evening, in effect - some of us, at any rate, and that you arrived
after traveling on Friday, some time Friday evening also. You could choose
to start then or wait until Saturday, as you saw fit.
We would make Saturday the major effort and half of Sunday, ending up
after lunch.
If that works for you, I am reasonably confident we can make a similar
arrangement with everyone else.
It will be a great pleasure to see you again after such a long time.
Vint
∂20-Oct-86 1050 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu Sorry, can't make it on Friday
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Oct 86 10:50:50 PDT
Received: by lindy.stanford.edu; Mon, 20 Oct 86 10:51:30 PDT
Date: Mon, 20 Oct 86 10:51:04 PDT
From: <ELLIOTT@slacvm.bitnet>
Reply-To: ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu
To: JMC@sail.stanford.edu
Subject: Sorry, can't make it on Friday
Date: 20 October 86 10:49-PST
From: ELLIOTT@SLACVM
To: JMC@SAIL
Subject: Sorry, can't make it on Friday
Date: 20 October 1986, 10:47:19 PST
From: Bloom, Elliott ELLIOTT at SLACVM
To: JMC at SAIL.STANFORD
Subject: Sorry, can't make it on Friday
Dear John,
I forgot that I had a meeting on Friday which includes a working
lunch. Can we make it the following week some day around 12pm.
Greetings,
Elliott
∂20-Oct-86 1141 VAL Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
MORE AGENTS ARE BETTER THAN ONE
Michael Georgeff
Artificial Intelligence Center
SRI International
Thursday, October 23, 4pm
MJH 252
A recent paper by Steve Hanks and Drew Mcdermott shows how some
previous "solutions" to the frame problem turn out to be inadequate,
despite appearances otherwise. They use a simple example -- come to
be called the "Yale Shooting Problem" -- for which it is impossible to
derive some expected results -- in this case, that the target of a
shooting event ceases living. Such difficulties, they suggest, call
into question the utility of nonmonotonic logics for solving the frame
problem.
In this talk, we describe a theory of action suited to multiagent
domains, and show how this formulation avoids the problems raised by
Hanks and McDermott. In particular, we show how the Yale Shooting
Problem can be solved using a generalized form of the situation
calculus for multiagent domains, together with notions of causality
and independence. The solution does not rely on complex
generalizations of nonmonotonic logics or circumscription, but instead
uses traditional circumscription. We will also argue that most
problems traditionally viewed as involving a single agent are better
formulated as multiagent problems, and that the frame problem, as
usually posed, is not what we should be attempting to solve.
∂20-Oct-86 1456 guibas@decwrl.DEC.COM Komlos
Received: from [128.45.0.240] by SAIL.STANFORD.EDU with TCP; 20 Oct 86 14:56:06 PDT
Received: from src.dec.com by sonora.dec.com (5.54.3/4.7.34)
id AA23479; Mon, 20 Oct 86 14:56:07 PDT
Received: by src.dec.com (4.22.05/4.7.34)
id AA11423; Mon, 20 Oct 86 14:56:04 pdt
Date: Mon, 20 Oct 86 14:56:04 pdt
From: guibas@decwrl.DEC.COM (Leonidas Guibas)
Message-Id: <8610202156.AA11423@src.dec.com>
To: pratt@navajo.stanford.edu, jmc@sail.stanford.edu
Cc: guibas@decwrl.DEC.COM
Subject: Komlos
Will you please let me know what your reaction to Komlos is by the end of
the afternoon? I am hoping to NOT have to call a meeting for noon tomorrow.
Thanks,
L.
∂20-Oct-86 1608 Mailer failed mail returned
To: JMC@SAIL.STANFORD.EDU
In processing the following command:
MAIL
The following message was unsent because of a command error:
------- Begin undelivered message: -------
∂20-Oct-86 1608 JMC re: Janos Komlos of UCSD
To: guibas@DECWRL.DEC.COM, Nilsson@SCORE.STANFORD.EDU
[In reply to message from Leo rcvd 17-Oct-86 08:21-PT.]
I have looked at the Komlos folder. He seems to be a first class
combinatorist, and I have no objections. I leave to others the
following questions. (1) To what extent will he participate in
the work of this department as well as mathematics. (2) Will the
numerical analysts be disappointed that the joint appointment is
not in the area of "scientific computation".
Or is this a proposed additional joint appointment? If so, does it
use up part of the theory slot?
------- End undelivered message -------
∂20-Oct-86 1730 NORDMARK@Sushi.Stanford.EDU Re: Gene Golub said you might help
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Oct 86 17:29:16 PDT
Date: Mon 20 Oct 86 17:23:39-PDT
From: Erik Nordmark <NORDMARK@Sushi.Stanford.EDU>
Subject: Re: Gene Golub said you might help
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 20 Oct 86 15:38:00-PDT
Message-ID: <12248420560.32.NORDMARK@Sushi.Stanford.EDU>
Well, that's how US banks earn their money - charging 250% fee!
I guess I can help to get that bill payed, but it will take the
2 weeks it takes to get mail to Sweden.
Are you going to be in your office tomorrow around 9am or 11am-- ?
Erik
-------
∂20-Oct-86 1748 shoham-yoav@YALE.ARPA Re: inclusion in proposal
Received: from YALE-CELRAY.ARPA by SAIL.STANFORD.EDU with TCP; 20 Oct 86 17:47:57 PDT
Received: by yale-celray.YALE.ARPA; Mon, 20 Oct 86 20:07:42 edt
Date: Mon, 20 Oct 86 20:07:42 edt
From: Yoav Shoham <shoham-yoav@YALE.ARPA>
Message-Id: <8610210007.AA02533@yale-celray.YALE.ARPA>
Subject: Re: inclusion in proposal
To: John McCarthy <JMC@SAIL.STANFORD.EDU>
Cc: VAL@SAIL.STANFORD.EDU, LES@SAIL.STANFORD.EDU
In-Reply-To: John McCarthy <JMC@SAIL.STANFORD.EDU>, 20 Oct 86 1444 PDT
We are currently writing the DARPA proposal for continuation of the work of
the Formal Reasoning Group. Actually we've finished the content part
and Les Earnest is doing the budget. It just occurred to me that it
might be convenient for you to be included. We will be asking for
support for the next three years, and we could put in a request to
support your work summers and for the non-teaching part of the
academic year + one student research assistant. Nils thinks it's a
good idea if you have no other source of research support in mind.
If you want to do it, we would need a sketch of what you
propose to do for the next three years. Les, LES@SAIL, would make
suggestions about DARPAizing it on receipt of the draft.
I appreciate and accept the offer. It seems like that would cover a substantial
part of my "abstract" work, which I'm sure is the hardest to fund. By when
do you want my draft? Also, I could use some guidance in writing it. How long
should it be? What have you written and how can I make my proposal compatible
with it? Should I stress possible applications, or remain creatively theoretical?
Should I include a CV? etcetera.
Yoav
-------
-------
∂20-Oct-86 1803 GLB Re: CS 306
∂20-Oct-86 1052 B.BELGIUM@OTHELLO.STANFORD.EDU Re: CS 306
Received: from OTHELLO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Oct 86 10:52:03 PDT
Date: Mon 20 Oct 86 10:45:38-PDT
From: Phil Andrew <B.Belgium@OTHELLO.STANFORD.EDU>
Subject: Re: CS 306
To: Reuling@SU-SCORE.ARPA
cc: FL@LEAR.STANFORD.EDU, GLB@SU-AI.ARPA
In-Reply-To: <12248339526.20.REULING@Score.Stanford.EDU>
Message-ID: <12248348105.122.B.BELGIUM@OTHELLO.STANFORD.EDU>
The disk space request for CS306 has been approved (I got several pieces of
mail from people in the know, which had me convinced of the request's
validity very quickly).
-Phil
∂20-Oct-86 1845 GLB
Now EKL is running at LOTS, with the files in the tutorial preloaded.
Students at LOTS need only type CLASS:<CS306>EKL306 and start their own proof.
I did a couple of exercises myself, and everything seems to work.
I would suggest to give preliminary sets of exercises for the students
to get familiar with the system, before assigning any demanding problem.
Most exercises in Section 2 of the tutorial are elementary. It would also
be useful to assign some pencil-and-paper proofs in Natural Deduction,
just to get the idea. I'll be glad to help the TA.
∂21-Oct-86 0003 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM Halmos
Received: from [128.81.2.8] by SAIL.STANFORD.EDU with TCP; 21 Oct 86 00:03:17 PDT
Received: from RUSSIAN.SPA.Symbolics.COM by MIRO.SNETC.Symbolics.COM via CHAOS with CHAOS-MAIL id 18455; Tue 21-Oct-86 03:02:02 EDT
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 3108; Tue 21-Oct-86 00:00:33 PDT
Date: Mon, 20 Oct 86 23:59 PDT
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: Halmos
To: jmc@SU-AI.ARPA
Message-ID: <861020235930.7.RWG@TSUNAMI.SPA.Symbolics.COM>
Salamin's been pretty mellow lately. Shall I invite him?
(I assume Halmos doesn't smoke!)
∂21-Oct-86 0813 BRONSTEIN@Score.Stanford.EDU re: Airline REGULATORY organization?
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Oct 86 08:13:51 PDT
Date: Tue 21 Oct 86 08:12:20-PDT
From: Alexandre Bronstein <BRONSTEIN@Score.Stanford.EDU>
Subject: re: Airline REGULATORY organization?
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 20 Oct 86 20:39:00-PDT
Message-ID: <12248582341.22.BRONSTEIN@Score.Stanford.EDU>
Thanks..
Alex
-------
∂21-Oct-86 0900 JMC
dentist
∂21-Oct-86 1111 RA 11:00 appointment
You have an 11:00 appointment with Nordmark and he is here waiting for you.
∂21-Oct-86 1122 NORDMARK@Sushi.Stanford.EDU re: Gene Golub said you might help
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Oct 86 11:22:02 PDT
Date: Tue 21 Oct 86 11:20:19-PDT
From: Erik Nordmark <NORDMARK@Sushi.Stanford.EDU>
Subject: re: Gene Golub said you might help
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 20 Oct 86 17:31:00-PDT
Message-ID: <12248616563.41.NORDMARK@Sushi.Stanford.EDU>
I see your at home right now, can we set another time?
Erik
-------
∂21-Oct-86 1717 roy@navajo.stanford.edu Your copy of "The New Republic".
Received: from NAVAJO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Oct 86 17:16:55 PDT
Received: by navajo.stanford.edu; Tue, 21 Oct 86 17:16:04 PDT
Date: Tue, 21 Oct 86 17:16:04 PDT
From: Shaibal Roy <roy@navajo.stanford.edu>
Subject: Your copy of "The New Republic".
To: jmc@sail.stanford.edu
Your copy of "The New Republic", Sept 15 & 22, 1986 has somehow
ended up on a public desk in MJH450. If you have no use for it,
can I keep it? The cover page has a photograph that I'd like to
have.
-shaibal
∂22-Oct-86 0810 shoham@YALE.ARPA Re: addendum
Received: from YALE-BULLDOG.ARPA by SAIL.STANFORD.EDU with TCP; 22 Oct 86 08:09:55 PDT
Received: by Yale-Bulldog.YALE.ARPA; 22 Oct 86 10:22:06 EDT (Wed)
Date: 22 Oct 86 10:22:06 EDT (Wed)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8610221422.AA09484@Yale-Bulldog.YALE.ARPA>
Subject: Re: addendum
To: John McCarthy <JMC@SAIL.STANFORD.EDU>
In-Reply-To: John McCarthy <JMC@SAIL.STANFORD.EDU>, 22 Oct 86 0026 PDT
I was imagining that we would ask for money to support you for your
non-teaching time during the academic year and full time summer. You
mention that this would be for the theoretical part of your work. If
you have some other support, perhaps we should ask for less.
I have none, although I'm hoping that in time I'll manage to secure some
industry support for robotics applications. Frankly, I don't expect to to
any hardcore applications for two years, and I'd like to have at least one
student doing purely theoretical stuff at all times. The bottom line is
that I think the support you mentioned would come in handy, but of course
I'll be grateful for anything. I'll try to get the two pages + cv out today.
By the way, I've been bitten by a new bug lately. I call it the "from pretzels
to predicates" problem, but that's just because I like the name. Essentially,
I'm looking for a mechanism to escape the brittleness of logic. I want to
encode logical operations in a behavior of a machine, with slightly deviant
behaviors corresponding to slightly-less-than-logical operations, which degrade
gracefully into stream-of-consciousness thinking. The idea is that even the
notion of a "concept" will not be not well defined, but more or less well defined,
with a certain behavior corresponding to the "ideal" or "propototypical"
concept. I realize that before I embark on this I will have to make it more
precise (otherwise, God forbid, I might degrade into a connectionist), but it's
at least something to think about when my lack of progress on chapter 6
overwhelms me.
Yoav.
-------
∂22-Oct-86 0900 JMC
Halmos
∂22-Oct-86 0939 CLT Qlisp reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
We meet this Friday noon, Oct 17 in MJ352.
(actually the room is reserve for 12:15-1:15)
We will write and discuss some simple Qlisp programs.
Bring programs and problems to program!
∂22-Oct-86 1028 CLT Qlisp reminder
no
∂22-Oct-86 1251 shoham@YALE.ARPA proposal
Received: from YALE-BULLDOG.ARPA by SAIL.STANFORD.EDU with TCP; 22 Oct 86 12:42:53 PDT
Received: by Yale-Bulldog.YALE.ARPA; 22 Oct 86 14:45:55 EDT (Wed)
Date: 22 Oct 86 14:45:55 EDT (Wed)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8610221845.AA12266@Yale-Bulldog.YALE.ARPA>
Subject: proposal
To: McCarthy@sail.stanford.edu
Cc: Les@sail.stanford.edu
Well, I've come up with something. I'm not pleased with it, and am willing to
make many changes to it. Given more time I'd sit on it for a while, and
try to feel less glib and mercenary about things. As things stand, I'm
shipping it to you now `as is'. In the interest of speed I'm sending the
LATEX source, which you can either read on line, or else Les can just run it
through LATEX. The first two pages are my resume. The last two pages the
proposal. Please 1. feel free to change the proposal (although I'd like to
see what I'm committed to before it goes off), 2. let me know what you'd
like me to change myself, and 3. let me know if you'd like me to send you a hard
copy as well.
Regards,
Yoav.
%----------------------------------------------------------------------------------
\documentstyle{article}
\setlength{\parskip}{\medskipamount}
\addtolength{\textheight}{2.0in}
\addtolength{\textwidth}{1.5in}
\addtolength{\oddsidemargin}{-0.75in}
\pagestyle{empty}
\addtolength{\topmargin}{-1.0in}
\begin{document}
\begin{center}
{\large YOAV SHOHAM}\\
Yale University \\
Computer Science Department \\
10 Hillhouse Avenue \\
New Haven, Ct. 06520 \\
\end{center}
{\large \sc Employment}
\begin{description}
\item
Assistant Professor in Computer Science \\
Stanford University \\
Starting in April 1987
\item
Visiting Scientist \\
Applied Mathematics Department \\
Weizmann Institute of Science, Rehovot, Israel \\
January-March 1987
\item
Research Assistant in Computer Science \\
Yale University \\
1982-1986
\item
Teaching Assistant in Computer Science \\
Yale University \\
1982-1983
\end{description}
{\large \sc Education}
\begin{description}
\item Yale University, New Haven, Connecticut \\
Ph.D., expected, 1986 \\
Dissertation area: Computer Science / Artificial Intelligence \\
Dissertation title: Reasoning about Change
\item
Technion, Haifa, Israel \\
B.A. Computer Science, 1981, with distinction
\end{description}
{\large \sc Visits, grants and awards}\\
\begin{description}
\item
1985: Visited the AI Center at SRI
International for about six weeks, working with Dr. Stanley Rosenschein.
\item
1984: Invited to Japan to participate in the FGCS conference, visit ICOT
and the Science University of Tokyo for a total duration of one month.
\item
1984: Participated in Commonsense Summer, a project organized by Dr.
Jerry Hobbs at the
AI Center of SRI International.
\item
1982: Received a Fulbright travel grant for travel to the U.S.
\item
1980,1981: Received the Technion President Award
\end{description}
{\large \sc Publications}
\begin{enumerate}
\item %1
{\it FAME: A Prolog Program That Solves Problems
in Combinatorics}, Y. Shoham,
Proc. Second International Logic Programming Conference,
Uppsala, Sweden, 1984.
\item %2
{\it Knowledge Inversion},
Y. Shoham and D.V. McDermott,
Proc. AAAI,
Austin, Texas, 1984.
\item %3
{\it Prolog Predicates as Denoting Directed
Relations}, Y. Shoham and D.V. McDermott,
Proc. FGCS, Tokyo, Japan, 1984.
\item %4
{\it Temporal Notation and Causal Terminology},
Y. Shoham and T. Dean,
Proc. Seventh Annual Conference of the Cognitive Science Society,
Irvine, California, 1985.
\item %5
{\it Naive Kinematics: One Aspect of Shape},
Y. Shoham, Proc. 9th IJCAI, Los Angeles, California, 1985.
\item %6
{\it Reasoning about Causation in Knowledge-Based Systems},
Y. Shoham, Proc. IEEE
Conference on Artificial Intelligence Applications, Miami,
Florida, 1985.
\item %7
{\it Ten Requirements from a Theory of Change}, Y. Shoham,
Journal of New Generation Computing 3(4), special issue on
knowledge representation, 1985.
\item %8
{\it Temporal Reasoning}, Y. Shoham and D.V. McDermott,
in {\it The Encyclopedia of Artificial Intelligence}, Shapiro, S.C.
(Ed.), Wiley-Interscience, New York (to appear).
\item %9
{\it A Propositional Modal Logic of Time Intervals\/} (short version), J.Y. Halpern and Y. Shoham,
Proc. IEEE Symp. on Logic in Computer Science, Boston, MA, June 1986.
\item %10
{\it Reified Temporal Logics: semantical and ontological considerations}, Y. Shoham,
Proc. 7th ECAI, Brighton, U.K., July 1986 (best-paper award). Revised version
to appear in the {\it Journal of Artificial Intelligence}, under the title
{\it Temporal Logics in AI}.
\item %11
{\it Chronological Ignorance: time, knowledge, nonmonotonicity and causal theories,}
Y. Shoham, Proc. AAAI, Philadelphia, PA, 1986.
\end{enumerate}
\newpage
{\large \sc Research Goals}
Stated broadly, my goal is to produce a system for reasoning about time that
has two properties:
\begin{enumerate}
\item
The system is required to be both rigorous and natural: it needs
precise and yet intuitive semantics.
\item
The rigor is not the solution
to any problem, only a prerequisite for giving one.
The system must support specific sorts of inferences;
Those should be
efficient, and hopefully also natural.
\end{enumerate}
The above requirements are not to come at the expense of realistic models of the world.
The rules according to which the world operates are complex, and formulations which
require one to overidealize it
are of no interest.
The proposed project is a continuation of my Ph.D. dissertation.
In the following I expand on these goals, and explain how I plan to achieve them.
The rigor of the system will derive from its logical basis. Without going into too
many details, there are quite a few ways in which to represent temporal
information in a logical system (see publication 8 for an overview of such
systems in AI, philosophy, and theoretical computer science, and publication
10 for a close scrutiny of temporal logics used in AI).
Careful selection of a temporal formalism results also in a very intuitive system.
The popularity of McCarthy and Hayes' situation calculus,
McDermott's temporal logic, and Allen's interval calculus
This first goal has to a large extent already been met (see `timetable' below).
The second one has not, although considerable progress has been made towards it,
and that is where the research effort will be concentrated.
One can identify in AI several classes of tasks which require reasoning about time,
and they include
prediction, explanation of observed phenomena, planning, and learning of
new rules of lawful change.
Although these classes of tasks are obviously related, in AI they have given rise to
by and large disjoint fields of research. Furthermore, most of the work in those areas
has been rather informal (to the detriment of its eventual effectiveness, it
is claimed here). I intend to first concentrate on the task of prediction,
and then expand to reasoning about planning (which includes the intermediate task
of explanation).
Several problems arise already in the context of prediction. The best-known one
is the so-called {\it frame problem}. In my dissertation I argue that the
frame problem actually embodies two distinct problems,
which I have called respectively the {\it qualification problem\/} and the
{\it extended prediction problem}. Briefly, they are the problems of
making statements of the form ``if this is true now then that is true later,''
without having the `this' part grow too large, and without having to make too
many such inferences.
We are now in a position to give solutions to these problems. Beside my own solution
(which appears in part in publication 11 and in full in my dissertation), there
have been related solutions proposed by Lifschitz, Kautz, Kowalski, and several
others. Most of these solutions rely on some form of nonmonotonic reasoning,
or reasoning in frameworks which allow making ``defeasible'' inferences. This area,
pioneered by McCarthy, Reiter and
McDermott, has seen many recent advances, due to McCarthy, Lifschitz,
Etherington, Perlis, myself, and many others.
As I have said, despite the recent progress, there is much to do before we can claim
to have laid the problem of temporal prediction to its final rest. The two main
issues that I see are extending current solutions to handle
continuous and concurrent change, and applying our theories to more than toy test
cases. The following is a proposed timetable for tackling these issues.
\begin{description}
\item{I. } (Completed). Develop a temporal formalism, define its precise syntax and
semantics, determine is complexity and expressiveness properties.
\item{II. } (Completed). Develop a framework for nonmonotonic reasoning appropriate
for temporal reasoning.
\item{III. } (Months 1-6) Debug the theory of causal theories,
which is defined in terms of the nonmonotonic temporal logic,
and apply it to a substantial test case. In that connection we will have a
seminar on causal reasoning, and I expect a student to write a program peforming
causal projection, either in the domain of medical diagnosis or in the domain
of rigid physical objects.
\item{IV .} (Months 1-12) Extend the theory of causal theories handle simultaneous
cause and effect, and especially continuous time. I have explored these research
directions at some length, but more work is needed here in order to devise
efficient algorithms for these cases.
\item{V. } (Months 13-18) The work so far will have concentrated on representing
and reasoning about forward projection in time. During this time period we will
start worrying about backwards projection, which in its simpler form is the
`explanation' task, and its more complex form is the `planning' task. These
tasks are inherently harder than the prediction task, because of the special form
of causal rules (which is reflected in the fact that for people too planning
is harder than predicting). This task will amount to asking metalogical questions of the
part of the underlying nonmonotonic temporal logic). Our work at this stage will
stress the need for completely clear
semantics, which most existing nontrivial planning systems lack.
\item{VI. } (Months 13-24) Demonstrate the effectiveness of the planner in
the domain of route planning and resource management. This domain, which
features continuous time, partial knowledge at the time of planning, and the
effect of actions in time, will also interface nicely with the next stage of the
project.
\item{VII. } (Months 18-36)
It has to be admitted that so far logic has played a limited role in ``real world''
applications. At the same time, real applications tend to suffer from
unclarity, unmaintainability and inextensibility. I would like to bridge this
gap between anaemic logical systems and theory-less working programs. This
is similar to Rosenschein's effort in applying formal theories of action to the
SRI robot. In our case too I plan to reason about a mobile robot, but with
the aim of escaping the bottleness of the sensing problem. For this reason
I am thinking about airborne robots, which on the one hand do not have the immediate
problem of running into obstacles (and in those periods when they do I plan to
take manual control over them), and on the other hand are very suitable to reasoning
about time, deadlines, resource management, technical malfunctions, and so on.
By the time this stage is reached I hope to have a parallel research effort going on which will
investigate the hardware constraints and which will involve avionics experts, but this
part of the project will be independent. It will concentrate on issues of knowledge
representation and reasoning: Which of the real-world features require representation?
What alogorithms are needed? Are causal theories sufficiently expressive for this
task? And so on. By the end of this period I aim to have a prototype
``airborne reasoner'' which will run in a simulated environment. Shortly afterwards
I hope to have a physical airborne device (construction of which will not
be part of this proposed research), at which time I will be able to
experiment with the ``airborne reasoner'' in real situations.
\end{description}
\end{document}
-------
∂22-Oct-86 1400 M.MMMIKEY@HAMLET.STANFORD.EDU re: Phony Status Symbols
Received: from HAMLET.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Oct 86 14:00:22 PDT
Date: Wed 22 Oct 86 13:58:51-PDT
From: MMMike Yang <M.MMMIKEY@HAMLET.STANFORD.EDU>
Subject: re: Phony Status Symbols
To: JMC@SU-AI.ARPA
cc: ILAN@SU-SCORE.ARPA, su-etc@SU-SCORE.ARPA, mis@PSYCH.STANFORD.EDU,
helen@PSYCH.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 22 Oct 86 10:44:00-PDT
Message-ID: <12248907566.205.M.MMMIKEY@HAMLET.STANFORD.EDU>
If you want to know if you view a car phone as a phony status symbol,
think if you've every considered just installing a car phone antenna
on your car, so everyone *thinks* you have a car phone. I know I have.
- Mike
-------
∂22-Oct-86 1434 ILAN@Score.Stanford.EDU Automatic Status
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Oct 86 14:34:18 PDT
Date: Wed 22 Oct 86 14:22:54-PDT
From: Ilan Vardi <ILAN@Score.Stanford.EDU>
Subject: Automatic Status
To: su-etc@Score.Stanford.EDU
cc: jmc@Sail.Stanford.EDU, ilan@Score.Stanford.EDU, helen@Psych.Stanford.EDU
Message-ID: <12248911945.22.ILAN@Score.Stanford.EDU>
It used to be that driving an automatic was a status symbol, now it
seems the reverse is true. Cars used to have metal emblems saying
``Automatic'', now they say ``5 speed''. I suppose that standard
tranmissions are lower on the pecking order than car phones.
Another question: Now that car phones are affordable, are people going to
get special plates with their phone numbers, making the freeway the pickup
bar of the 1990's?
-------
∂22-Oct-86 1542 CRISPIN@SUMEX-AIM.ARPA re: Phony Status Symbols
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Oct 86 15:42:14 PDT
Date: Wed 22 Oct 86 15:42:36-PDT
From: Mark Crispin <Crispin@SUMEX-AIM.ARPA>
Subject: re: Phony Status Symbols
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 22 Oct 86 10:44:00-PDT
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12248926455.57.CRISPIN@SUMEX-AIM.ARPA>
Hi. If you would like to talk with me about car phones I'd be happy to.
I have been part of a clique of local hacker/car phone owners for over a
year now and know *quite* a bit about the topic. I may be able to give
you some advice that could save you a lot of money/headaches!
-- Mark --
-------
∂22-Oct-86 1600 JMC
Halmos
∂22-Oct-86 1800 JMC
Gloria
∂23-Oct-86 0100 JMC
Tandberg
∂23-Oct-86 0949 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
MORE AGENTS ARE BETTER THAN ONE
Michael Georgeff
Artificial Intelligence Center
SRI International
Thursday, October 23, 4pm
MJH 252
A recent paper by Steve Hanks and Drew Mcdermott shows how some
previous "solutions" to the frame problem turn out to be inadequate,
despite appearances otherwise. They use a simple example -- come to
be called the "Yale Shooting Problem" -- for which it is impossible to
derive some expected results -- in this case, that the target of a
shooting event ceases living. Such difficulties, they suggest, call
into question the utility of nonmonotonic logics for solving the frame
problem.
In this talk, we describe a theory of action suited to multiagent
domains, and show how this formulation avoids the problems raised by
Hanks and McDermott. In particular, we show how the Yale Shooting
Problem can be solved using a generalized form of the situation
calculus for multiagent domains, together with notions of causality
and independence. The solution does not rely on complex
generalizations of nonmonotonic logics or circumscription, but instead
uses traditional circumscription. We will also argue that most
problems traditionally viewed as involving a single agent are better
formulated as multiagent problems, and that the frame problem, as
usually posed, is not what we should be attempting to solve.
∂23-Oct-86 1208 PERLAKI@Score.Stanford.EDU message
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 23 Oct 86 12:08:27 PDT
Date: Thu 23 Oct 86 11:40:39-PDT
From: Agnes M. Perlaki <PERLAKI@Score.Stanford.EDU>
Subject: message
To: McCarthy@Score.Stanford.EDU
Message-ID: <12249144552.33.PERLAKI@Score.Stanford.EDU>
Adry Marshall called. Re: child photos. Please call 408 294-9493.
She called earlier today, but the system was down, so I am just now sending
you the message.
-Agi
-------
∂23-Oct-86 1214 RA Invitation to speak
Bill Park from Manicopa County College, Phoenix, Arizone called to invite
you to give a talk on the history of AI, either in mid-Feb or mid-March.
He will call back around 12:30.
∂23-Oct-86 1232 UDI@Score.Stanford.EDU RE: partial solution
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 23 Oct 86 12:32:28 PDT
Date: Thu 23 Oct 86 12:26:34-PDT
From: Ehud Shapiro <UDI@Score.Stanford.EDU>
Subject: RE: partial solution
To: jmc@Sail.Stanford.EDU
cc: val@Sail.Stanford.EDU
Message-ID: <12249152912.11.UDI@Score.Stanford.EDU>
Vladimir seems to think that this is actually a full
solution, since you were interested only in uniqueness
of ports; is that so?
Udi
p.s. If this actually solves the problem that you have posed,
then perhaps I would write a memo to the Prolog Digest, noting
the problem and the solution.
-------
∂23-Oct-86 1649 MATHY@Score.Stanford.EDU People count
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 23 Oct 86 16:49:52 PDT
Date: Thu 23 Oct 86 16:26:10-PDT
From: Rosemary Mathy <MATHY@Score.Stanford.EDU>
Subject: People count
To: jmc@Sail.Stanford.EDU
cc: RA@Sail.Stanford.EDU
Message-ID: <12249196530.9.MATHY@Score.Stanford.EDU>
There were 25 people in 352 at 4:30 on Thursday, October 23rd.
-------
∂23-Oct-86 2000 JMC
pattern matching in McC and T, unification, pseudo lists
∂23-Oct-86 2102 YOM Thanks
To: JMC@SAIL.STANFORD.EDU, nilsson@SCORE.STANFORD.EDU
I'm leaving for Boston tomorrow,
on my way to Israel (arriving
there sometime in December).
Thank you for your hospitality.
Yoram
∂24-Oct-86 0153 NILSSON@Score.Stanford.EDU Binford
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Oct 86 01:53:26 PDT
Date: Wed 22 Oct 86 19:18:59-PDT
From: Nils Nilsson <NILSSON@Score.Stanford.EDU>
Subject: Binford
To: jmc@Sail.Stanford.EDU
cc: nilsson@Score.Stanford.EDU
Message-ID: <12248965846.12.NILSSON@Score.Stanford.EDU>
John, would you consent to chair the "Binford Reappointment Committee?"
Apparently we do need to initiate a formal reappointment process although
I understand the paper work does not need to be extensive. We need to
get about 6 evaluation letters. I have also asked Bob Cannon and myself
to serve on the committee. (So far, I have agreed to serve; I'd chair it
except that I need to chair the Winograd promotion committee---unless
you want to do that instead.) One reason for asking you is that you recall
the circumstances under which Binford was appointed. I don't think much
work is involved. We will need to meet once to decide on to whom to send
letters and then once again to decide what to recommend to the CSD faculty.
-Nils
-------
∂24-Oct-86 1037 RA Amir Naqwi's orals
Naqwi called to remind you that you are on his oral committee and that
his orals are on Monday, 10/27, 4:00pm, at Building 500, room 501. He
said that if you want to talk to him, his number is 3-4503.
∂24-Oct-86 1158 CRISPIN@SUMEX-AIM.ARPA cretinism
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Oct 86 11:58:05 PDT
Date: Fri 24 Oct 86 11:54:53-PDT
From: Mark Crispin <Crispin@SUMEX-AIM.ARPA>
Subject: cretinism
To: JMC@SU-AI.ARPA
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12249409287.57.CRISPIN@SUMEX-AIM.ARPA>
I find this story very upsetting:
a077 0736 24 Oct 86
PM-Textbook Flap,0147
URGENT
Textbook Ruling Decides in favor of Fundamentalists
By KRISTI UMBREIT
Associated Press Writer
GREENEVILLE, Tenn. (AP) - Seven families who filed suit contending
their children's textbooks contained themes offensive to their strict
Christian beliefs won their case today.
A federal judge said the families are entitled to damages from the
Hawkins County school board because the district's insistance in
using the Holt, Rinehart, Winston series of reading books violated
their First Amendment rights to religious freedom.
''In forcing the plaintiff-students to read from the Holt series or
to forfeit a free public education, the defendants have burdened the
plaintiffs' right of free exercise of their religion,'' said U.S.
District Judge Thomas G. Hull in a 27-page opinion.
Hull set a tentative Dec. 15 date to convene a jury, which will
decide how much the families are entitled to in damages.
''I can say I'm very grateful and I'm very thankful,'' said Vicki
Frost, one of the parents and the leading witness in the trial.
''It's a wonderful victory. We're ecstatic,'' said Michael Farris,
who represented the parents.
Timothy Dyk, a lawyer for the school system who was paid by
television producer Norman Lear's People for the American Way, said
he would file an appeal today or Monday.
''This is not the end of the road,'' he said at a news conference in
Washington. ''We intend to take this case as far as is necessary to
get this decision reversed.''
Anthony Podesta, president of People for the American Way, called
the ruling a ''recipe for disaster for public education that will
disrupt education for all children. It will invite every sect in the
country to pick and choose which parts of the public school
curriculum it will accept.''
Podesta also said the ruling could have a chilling effect on
textbook publishers. He said they would begin to revise their books
''to accommodate the lowest common denominator philosophy, to try to
make textbooks as inoffensive as possible, and in doing so continue
to dumb-down textbooks.''
The families sued the Hawkins County schools in 1983 after objecting
to hundreds of passages in the readers, including stories by Margaret
Meade, Isaac Asimov and Danish fairy-tale writer Hans Christian
Andersen.
The parents said the books contained references to witchcraft,
exotic religions, one-world government and relative ethics, all of
which violate their Christian beliefs.
''The Diary of Anne Frank'' was objected to because it tolerated all
religions. ''The Wizard of Oz'' was deemed objectionable because it
told children that traits such as courage, intelligence and
compassion are personally developed rather than God-given, and it
depicted a witch as good.
Hull, in his decision, said Hawkins County school officials must
teach the reading program without jeopardizing the students' rights
to a free public education or religious freedom. He did not specify
how the district was to do that.
''The plaintiffs have sincerely held religious beliefs which are
entitled to protection under the free exercise clause of the First
Amendment and which are offended by certain recurring themes in the
Holt series,'' Hull said.
''The state of Tennessee and the Hawkins County Board of Education
have a legitimate, compelling, and overriding interest in the
education of public school students; that this interest does not
necessitate uniform use of the Holt series and can be achieved by
less restrictive means,'' he said.
Hull said that since Tennessee laws require education while allowing
youngsters to be taught at home or in private schools, the state is
acknowledging that there are ways to accommodate children to achieve
its goals.
During testimony in the non-jury trial, Vicki Frost, one of the
parents, listed 16 areas where she said the books jeopardized her
children's religious beliefs.
Textbook analysts also took the stand, saying the 1983 readers
contain few references to Christianity and favor Hinduism and Eastern
religions.
The school board, backed by state attorneys, argued that the changes
demanded by the parents would defeat public education's purpose of
exposing students to varying viewpoints.
The demands also would create disruption in the curricula because of
the need for separate classrooms, examinations and lesson plans, the
schools' attorneys said.
The families were represented by Michael Farris, whose services were
paid for by the conservative group Concerned Women For America. Dyk
was paid by television producer Norman Lear's People for the American
Way.
AP-NY-10-24-86 1306EDT
-------
∂24-Oct-86 1330 VAL Judea Pearl on causal reasoning
Here is my reply to Judea Pearl concerning his paper on causal reasoning.
Thank you for sending me your note on causal reasoning in default logic. I
enjoyed reading it very much, and I hope you are planning to publish it.
Your idea of distinguishing between expectation-evoking and explanation-evoking
rules was new to me, and I agree with the analysis of the examples in your paper
on the basis of that distinction.
Your paper was especially interesting to me because I am studying now some
formalizations of reasoning about action which explicitly represent causality.
The form of causality I am looking at is somewhat different from what you
discuss in your paper; to me, causality is a relation between an action and a
property of the world caused by it (e.g., teaching Joe how to read and write
is the cause of his being literate). It seems to be close to what you call
expectation-evoking rules. Maybe the reason why the distinction you
introduce has not found its way into the literature is that the usual examples
do not involve default rules of both kinds simultaneously.
I would like to ask your opinion about an interesting example of invalid
default reasoning, which I first learned from Benjamin Grosof (I do not know
whether the example is actually his). Given that (i) Joe is a high school
dropout, (ii) high school dropouts are normally adults, and (iii) adults
are normally employed, -- we do not want to conclude by default that Joe is
employed. I tried to analyze this example in the style of your paper, but
it somehow did not quite work. I classified rule (ii) as a V-def, and rule
(iii) as an X-def, but then your X-V system apparently sanctions the
inference we want to block.
It is not clear to me whether your classification of rules applies to default
reasoning in general, or only to a special (though important) class of
applications. In many cases, default reasoning seems to deal neither with
expectations nor explanations. For instance, it can be required by a
knowledge representation convention: in a closed-world database we store
only positive facts and view negative facts as defaults.
As far as the solution outlined in Section 3 is concerned, it is difficult
for me to judge whether it can lead to a satisfactory system without
seeing a more detailed description of its syntax and semantics. I guess
I would prefer that your ideas be formalized, if possible, in the framework
of one of the existing theories of default reasoning, such as Reiter's
default logic or circumscription. Perhaps it is possible to express the same
distinctions in a more economical way, without introducing two forms of
implication.
∂24-Oct-86 1355 BSCOTT@Score.Stanford.EDU Binford's file
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Oct 86 13:55:42 PDT
Date: Fri 24 Oct 86 13:53:53-PDT
From: Betty Scott <BSCOTT@Score.Stanford.EDU>
Subject: Binford's file
To: JMC@Sail.Stanford.EDU
cc: BScott@Score.Stanford.EDU
Message-ID: <12249430949.31.BSCOTT@Score.Stanford.EDU>
John, Nils said you needed to see Binford's file. It is rather extensive and
I am reluctant to have it floating around. Would you consider looking at it
in Nils' conference room, and I can then arrange to have copies made of any-
thing which you need for your committee.
Betty
-------
∂24-Oct-86 1426 BSCOTT@Score.Stanford.EDU re: Binford's file
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Oct 86 14:26:18 PDT
Date: Fri 24 Oct 86 14:22:06-PDT
From: Betty Scott <BSCOTT@Score.Stanford.EDU>
Subject: re: Binford's file
To: JMC@Sail.Stanford.EDU
cc: BSCOTT@Score.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 24 Oct 86 14:22:00-PDT
Message-ID: <12249436086.31.BSCOTT@Score.Stanford.EDU>
O.K., just let me know when you want to look at it.
Betty
-------
∂24-Oct-86 1428 RICHARDSON@Score.Stanford.EDU Binford
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Oct 86 14:28:41 PDT
Date: Fri 24 Oct 86 14:27:00-PDT
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: Binford
To: jmc@Sail.Stanford.EDU
Message-ID: <12249436978.24.RICHARDSON@Score.Stanford.EDU>
John,
Nils has asked that I forward the following schedule to you regarding
the Binford reappointment.
-Anne
-----
Tom Binford (reappointment as Professor, Research in CSD)
<<<continuous or term?>>>
John McCarthy (chair)
Bob Cannon <<<must be confirmed>>>
Nils Nilsson
1) Now: send out request for evaluation letters. We need 12 letters.
Check format of draft with Shari Austin-Kitzmiller in SOE to avoid
bureaucratic problems later.
2) by December 5, 1986: committee makes preliminary recommendation to CSD
faculty
3) by December 12, 1986: preliminary papers ready for SOE EXCOM meeting
4) by January 6, 1986: CSD faculty approves or disapproves final committee
recommendation
5) by January 21, 1987: final papers ready for SOE EXCOM meeting
6) March 10: Board of Trustee approval
-------
-------
∂24-Oct-86 1429 RICHARDSON@Score.Stanford.EDU Binford
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Oct 86 14:29:33 PDT
Date: Fri 24 Oct 86 14:27:51-PDT
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: Binford
To: jmc@Sail.Stanford.EDU
Message-ID: <12249437135.24.RICHARDSON@Score.Stanford.EDU>
John,
Bob Cannon has just confirmed that he will be on the committee.
-Anne
-------
∂24-Oct-86 1501 RICHARDSON@Score.Stanford.EDU re: Binford
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Oct 86 15:01:04 PDT
Date: Fri 24 Oct 86 14:34:12-PDT
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: re: Binford
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 24 Oct 86 14:31:00-PDT
Message-ID: <12249438291.24.RICHARDSON@Score.Stanford.EDU>
John,
My mistake - I sent this message out at the same time as the other
appointments. Six it is.
Sorry,
Anne
-------
∂24-Oct-86 1704 VAL Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
RAMIFICATION AND QUALIFICATION IN THE BLOCKS WORLD
Matt Ginsberg
David Smith
Thursday, October 30, 4pm
MJH 252
In this talk, we discuss the need to infer properties of actions
from general domain information. Specifically, we discuss the
need to deduce the indirect consequences of actions (the
ramification problem), and the need to determine inferentially
under what circumstances a particular action will be blocked
because its successful execution would involve the violation of
a domain constraint (the qualification problem).
We present a formal description of action that addresses these
problems by considering a single model of the domain, and updating
it to reflect the successful execution of actions. The bulk of the
talk will involve the investigation of simple blocks world problems
that existing formalisms have difficulty dealing with, including
the Hanks-McDermott problem, and two new problems that we describe
as "the dumbbell and the pulley".
∂24-Oct-86 1740 CERF@A.ISI.EDU Nov 7-9 meeting - update
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 24 Oct 86 17:36:27 PDT
Date: 24 Oct 1986 15:03-EDT
Sender: CERF@A.ISI.EDU
Subject: Nov 7-9 meeting - update
From: CERF@A.ISI.EDU
To: Bibliotects:
Message-ID: <[A.ISI.EDU]24-Oct-86 15:03:06.CERF>
Dear Bibliotects:
The Digital Library Working Group meeting scheduled for Nov 7-9
in Reston, VA now looks something like this:
1. Meeting starts with dinner at 7:00 PM on Friday, 7 November.
This will probably be a catered affair at NRI in Reston.
2. We would hold informal discussions about the agenda for the
next day over and after dinner.
3. We would plan a full day of activities for Saturday, Nov 8
and would have meals at NRI, except for dinner which would
be out (at some pleasant oasis).
4. We would wrap up the meeting by Sunday noon.
Hotel accommodations are available at the Reston Sheraton
(not to be confused with the McLean/Tyson's Sheraton). NRI
can make arrangements for you if you would like.
Participants expected thus far:
John McCarthy, Raj Reddy (Saturday only?), Bob Kahn, Butler
Lampson, Vint Cerf, Patrice Lyons.
I have yet to hear from Jaime Carbonell, Jerry Popek and Bob
Sproull.
Danny Hillis and Don Lindberg cannot join us owing to prior
committments.
The next meeting is proposed for January 5-6, 1987 (Mon/Tues)
in Reston, unless there is a strong sentiment for warmer climes,
perhaps on the west coast.
I will send out an agenda soon as well as any on-line material
we have available.
Please confirm your attendance (times/days or arrival/departure),
whether you need travel assistance or assistance with hotel
reservations.
Looking forward to another stimulating session.
Vint
∂25-Oct-86 0452 HSU@Sushi.Stanford.EDU Re: question
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 25 Oct 86 04:52:14 PDT
Date: Sat 25 Oct 86 04:50:28-PDT
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: Re: question
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 24 Oct 86 21:08:00-PDT
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12249594169.8.HSU@Sushi.Stanford.EDU>
Thanks for your message. I had to leave for an appointment at 5pm, and
I didn't find you in the office when I went there the second time.
Sorry for my sudden "disappearance".
The questions I wanted to ask were quite trivial -- basically to clarify
some of the homework problems. However, I don't have the text with me
now. I will send you another message with the questions soon.
Yung-jen
-------
∂25-Oct-86 1025 ASHOK@Sushi.Stanford.EDU [John McCarthy <JMC@SAIL.STANFORD.EDU>: Conway problem]
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 25 Oct 86 10:25:11 PDT
Date: Sat 25 Oct 86 10:23:22-PDT
From: Ashok Subramanian <ashok@Sushi.Stanford.EDU>
Subject: [John McCarthy <JMC@SAIL.STANFORD.EDU>: Conway problem]
To: jmc@Sail.Stanford.EDU
cc: ashok@Sushi.Stanford.EDU
Message-ID: <12249654770.16.ASHOK@Sushi.Stanford.EDU>
>> How do you understand the Conway problem about the angel and the devil?
>> Must the devil destroy a planet while the angel is on it? Is it that
>> the angel can't see what planets have been destroyed and may land on
>> one? Is it that the devil attempts to surround the angel by a barrier
>> of destroyed planets?
I support the third interpretation. The way I see it, the devil and
the angel take turns. The devil destroys a planet, and then the angel
takes a leap into space. If the devil destroys the planet the angel is
sitting on, then the angel is not killed, but must necessarily seek a
new planet on the next move. Both parties can see the current state of
the world, and have no finite bound on computing power. The devil also
knows where the angel is at all times.
The only way the devil can win is for him to surround the angel by a
barrier of destroyed planets, and then home in on the angel.
--ashok
-------
∂25-Oct-86 1203 CRISPIN@SUMEX-AIM.ARPA Re: textbooks
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 25 Oct 86 12:03:09 PDT
Date: Sat 25 Oct 86 12:03:30-PDT
From: Mark Crispin <Crispin@SUMEX-AIM.ARPA>
Subject: Re: textbooks
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 24 Oct 86 21:36:00-PDT
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12249673000.18.CRISPIN@SUMEX-AIM.ARPA>
All I know is the AP and TV reports, which said basically the same thing.
I hope this decision will be reversed. Otherwise, it would be yet another
blow against US education. I am becoming increasingly convinced that it is
difficult if not impossible to get a reasonable pre-college education in
the US. My elememtary school education was in the glory days of the early
'60s where it was considered important to catch up with the Russians in
space.
-------
∂25-Oct-86 1637 CRISPIN@SUMEX-AIM.ARPA re: textbooks
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 25 Oct 86 16:37:20 PDT
Date: Sat 25 Oct 86 16:37:42-PDT
From: Mark Crispin <Crispin@SUMEX-AIM.ARPA>
Subject: re: textbooks
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 25 Oct 86 13:21:00-PDT
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12249722915.11.CRISPIN@SUMEX-AIM.ARPA>
I believe that both deserve the blame, along with the American perception
that *everybody* deserves a high school diploma (much less a college one)
as a *right* not a hard-earned privilege. I had absolutely *no* homework
to do throughout junior high school and high school (I did all the assignments
in the 10 minutes between classes). The bad habits I acquired then haunted me
in college, where my grades were quite mediocre, and to today -- I can't
study my Japanese lessons effectively even though I am strongly motivated to!
What was the point? I got beaten up for being too smart and bringing up the
class average when I did no work -- if I did serious studying I would probably
be dead!
I hate the American education system for what it did to me, and even more so
when I found out *real* educations are possible in other countries. Foo, it
is insane that colleges have remedial English composition classes for
American students!!!
You're right, the hippie-dippie movement was to blame in part, as typified by
the "New Math". I remember having an argument with a teacher since I didn't
"show my work" in a single-digit division. To me, the answer was obvious by
inspection, and to go through the laborious diagrams required by "New Math"
were tedium. But let's not forget the fundamentalists -- MY science textbooks
had creationism in them and insisted that species were static. It took years
for me to comprehend why I was being confused by this as opposed to everything
about dinosaurs I saw in the museums -- I was being lied to.
-------
∂26-Oct-86 0900 JMC
call about picture
∂26-Oct-86 1221 vijay@ernie.Berkeley.EDU AI DISC: Searle replies to Hofstadter
Received: from ERNIE.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 26 Oct 86 12:21:30 PST
Received: by ernie.Berkeley.EDU (5.53/1.16)
id AA13141; Sun, 26 Oct 86 12:22:12 PST
Date: Sun, 26 Oct 86 12:22:12 PST
From: vijay@ernie.Berkeley.EDU (Vijay Ramamoorthy)
Message-Id: <8610262022.AA13141@ernie.Berkeley.EDU>
To: jmc@sail.stanford.edu
Subject: AI DISC: Searle replies to Hofstadter
Reply to Hofstadter
John R. Searle
In an earlier published discussion of the Chinese Room,
Hofstadter (along with his coauthor, Dan Dennett) misquoted
me five separate times. Hofstadter and Dennett were forced
to admit these misquotations in "The New York Review of
Books". ("The Myth of the Computer: An Exchange," "The New
York Review of Books", XXIX, June 24, 1982) Now given this
rather unfortunate scholarly history, one might suppose that
Hofstadter would be more careful on future occasions, but
alas, such is not the case.
In the very first paragraph of his current offering he
writes,
Only BRAINS can experience anything, says Searle.
To me, that is not only mystical, but also incred-
ibly chauvinistic and incomprehensible. It is a
religion, pure and simple. (Hofstadter, Reply to
Searle, p. 1)
But if he had bothered to look at the actual texts I wrote
and to which he is supposed to be replying, he would read
the following:
Among other things, I am accused of holding the
preposterous 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
- 2 -
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
substance in advance. (Searle, Turing the Chinese
Room, p. 1)
This simple blunder by Hofstadter (I assume it is not
deliberate on his part) is sufficient to render most of the
rest of what he says irrelevant. In particular his discus-
sion of antimatter is simply irrelevant. What I do in fact
argue is "not" that as a matter of principle only brains
can have experiences. That is a view I characterize as
"preposterous". My view rather is that since brains do
cause experiences, it follows that any other system capable
of causing experiences would have to have causal powers at
least equivalent to those of the brain. Whether that is
possible for any given system is an empirical question like
any other, not to be settled by "a priori" logical specula-
tion. So when I say that you can't do it with beer cans and
ping-pong balls, that is an emprical claim on my part. The
evidence for it is overwhelming, because we know that the
biochemistry of neurons is totally unlike the chemistry of
beer cans and ping pong balls.
If we try to imagine a plausibility argument for the
actual existence of nonneuronal conscious systems I think
the best is the following: It is not only possible, but it
seems likely, that forms of life may have evolved elsewhere
in the universe; there is no reason to suppose that life is
confined to our little solar system. But now given the
- 3 -
probable existence of a fairly large number of living forms
elsewhere in the universe it is surely not out of the ques-
tion that some of these may have evolved consciousness in
chemical forms different from ours.
Such a possibility is not only consistent with my posi-
tion, it is in fact a natural consequence of it; for notice
what we are imagining is a system or set of systems that
have a different chemical structure from that of our brains
but have equivalent causal powers. But Hofstadter is com-
mitted to a much more extreme position, a position which I
have attacked and which he has failed to defend. He is
committed not just to the view that "some" non-neuronal
based system "might" be capable of causing and sustaining
consciousness, but rather, that "any" system whatever, made
out of any material whatever, "must" (and this is not just,
"might for all we know", but "must") have consciousness and
other sorts of experiences, provided only that it instan-
tiates the right computer program. This has the conse-
quence, to repeat, that a system made out of beer cans and
ping-pong balls would have to have every conscious experi-
ence that human beings have, provided only that it instan-
tiated the right programs. He claims this is a "deceptive"
and trick example. But there is nothing "deceptive" or
tricky about it: It is simply an attempt to take seriously
his claim that the biochemistry of the brain is totally
irrelevant to the mind, that any old substance will do if it
has the right "pattern". He suggests that the problem with
- 4 -
the beer cans and the ping-pong balls is that they are not
"active, connected, and interactive " enough. Fine. Make
them as active, connected, and interactive as you like.
Equip each beer can with wind-up motors that let it run
around on wheels in all directions. Equip each ping-pong
ball with wind machines that will blow it about in any
direction you like. Connect the ping pong balls and the
beer cans with fishlines until the whole system is as
"active, connected and interactive" as you want. All the
same, it is still empirically absurd to suppose that the
beer can/ping-pong ball system must necessarily produce the
same experiences as my present experiences of thirst pro-
duced by the impact of angiotensin 2 on my hypothalamus or
the cocaine sensation that the drug taker feels when the
cocaine competitively inhibits the capacity of the synaptic
receptors to reabsorb norepinephrine.
It is important to state exactly what is at issue here.
The point is not that for all we know the system of beer
cans and ping pong balls might have thoughts and feelings.
(Who knows what it feels like to be a system of beer cans
bombarded by ping pong balls?) Hofstadter's claim is that if
the system has the right syntax it MUST have thoughts and
feeling, because that is all there is to having thoughts
and feelings. So any system whatever -- made of beer cans,
old rusting cars, tinkertoys (his "reductio ad absurdum",
not mine), anything at all -- provided only that it has the
right syntax, i.e. it runs on the right computer program,
- 5 -
must have all the thoughts and feelings that any human being
has. And it is important to be clear about the sense of
"syntax" in question since many of Hofstadter's remarks make
it clear that he is still confusing the formal syntax of a
system with the physical powers of the system that can be
"represented" in a formal sysntactical system. (This is the
confusion I was trying to correct at the end of my original
reply in the section entitled, "Models and Reality").
Remember, where computers are concerned the syntax can be
entirely a matter of 0's and 1's. It is independent of any
physical, chemical, biochemical features of the system in
question.
In all of his mistakes and misquotations, he has simply
not touched the two central propositions at issue. First,
syntax in the computational sense simply is not sufficient
by itself to guarantee the presence of consciousness, inten-
tionality, etc. Syntax is not sufficient for semantics.
Second, since any system is syntactical in this sense, to
produce mental phenomena a system must have something more.
What is it? The answer is the obvious one. Since brains do
it, any non-neuronal system would have to be able to dupli-
cate (in whatever medium it uses) the specific causal powers
of brains.
Hofstadter claims that there are
large numbers of philosophers, psychologists, AI
researchers, physicists, and cognitive scientists
whose views are substantially the same as mine.
- 6 -
I think that at one time that may have been true, but right
now strong AI is very much dwindling and on the defensive.
And the reason is clear: The preposterous consequences of
strong AI are becoming apparent to more and more people. To
repeat, it is not a serious scientific hypothesis to suppose
that we could duplicate the specific experiences which
cocaine users have, and which are caused by the action of
cocaine on norepinephrine, just by mimicking the formal
structure of the biochemical process in a system of beer
cans, however "active, connected, and interactive". Ask any
neuropharmacologist if he or she thinks they might just as
well quit studying the catecholamines (i.e. norepinephrine,
epinephrine, dopamine, and seratonin) and concentrate on the
beer cans. Ask them if they think it is really all done
with syntax, with 0's and 1's. (And by the way, ask your-
self. Suppose it really mattered, suppose your brain was in
deep trouble and you needed help. Would you rather that the
researchers stuck to the the biochemistry or would you just
as soon that they investigated beer cans? Having read Hofs-
tadter would you tell them that the biochemistry doesn't
matter since it is all syntax anyway, and syntax can be done
in any medium whatever provided only that it is "active,
connected, and interactive"?)
Hofstadter seems obsessed by something he calls "reli-
gion." My suggestion that neurobiology really matters in the
study of the mind/brain is characterized by Hofstadter as
"religious" or "religion" on four separate occasions in this
- 7 -
one short reply. I wish he had told us what religion he
thinks he is talking about. But there is indeed one "reli-
gious" precept I do adhere to: If you are going to quote
your opponent in order to criticize his views you had better
get the quotations right. Next time around I suggest that
Hofstadter would do better to try to adhere to that "reli-
gion".
∂26-Oct-86 1606 CLT
(setq a (cons nil nil))
(null (rplaca a a))
(eq a (car a))
(eq a (cdr a))
∂26-Oct-86 1609 CLT
(setq l (list 0 1 2 3))
(null (rplacd (last l) l))
(car l)
(car(cdr l))
(car(cdr(cdr l)))
(car(cdr (cdr(cdr l))))
(car(cdr(cdr (cdr(cdr l)))))
∂26-Oct-86 2137 CLT
(setq prinlevel 6) ;init nil
(setq prinlength 6) ;init nil
∂26-Oct-86 2336 vijay@ernie.Berkeley.EDU AI DISC: A restatement of the Dreyfus Position
Received: from ERNIE.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 26 Oct 86 23:32:27 PST
Received: by ernie.Berkeley.EDU (5.53/1.16)
id AA28278; Sun, 26 Oct 86 22:50:48 PST
Date: Sun, 26 Oct 86 22:50:48 PST
From: vijay@ernie.Berkeley.EDU (Vijay Ramamoorthy)
Message-Id: <8610270650.AA28278@ernie.Berkeley.EDU>
To: jmc@sail.stanford.edu
Subject: AI DISC: A restatement of the Dreyfus Position
Following is a restatement of the Dreyfus position-- the beginning
sections may seem familiar, but the ending sections describe/imply
the "new" Dreyfus position (as they have termed it).
-----------
TOWARDS A RECONCILIATION OF PHENOMENOLOGY AND AI
I. The Traditional Account of Skills
The rationalist philosophical tradition which descends
from Socrates, to Plato, to Leibniz, to Kant, to conven-
tional AI and knowledge engineering takes it for granted
that understanding a domain consists in having a theory
about that domain. A theory formulates the relationships
between objective, context-free features (attributes, fac-
tors, data points, cues, etc.) in terms of abstract princi-
ples (covering laws, rules, programs, etc.) As this tradi-
tion develops, even everyday practice is assumed to be based
on unconscious theory.
In one of his earliest dialogues, "The Euthyphro", Plato
tells us of an encounter between Socrates and Euthyphro, a
religious prophet and so an expert on pious behavior.
Socrates asks Euthyphro to tell him how to recognize piety:
"I want to know what is characteristic of piety ... to use
as a standard whereby to judge your actions and those of
other men." But instead of revealing his piety-recognizing
principles, Euthyphro does just what every expert does when
cornered by Socrates. He gives him examples from his field
of expertise; in this case situations in the past in which
- 2 -
men and gods have done things which everyone considers
pious. Socrates persists throughout the dialogue in demand-
ing that Euthyphro tell him his rules, but although Euthy-
phro claims he knows how to tell pious acts from impious
ones, he cannot state the rules which generate his judg-
ments. Socrates ran into the same problem with craftsmen,
poets and even statesmen. None could articulate the theory
underlying his behavior.
Plato admired Socrates and sympathized with his prob-
lem. So he developed a partial account of what caused the
difficulty. In theoretical domains such as mathematics, at
least, experts had once known the rules they use but then
they had forgotten them. In these domains the rules are
there functioning in the expert's mind whether he is cons-
cious of them or not. How else could we account for the
fact that he can perform the task? The role of the philoso-
pher is to help such experts remember the principles on
which they act.
Plato's account did not apply to everyday skills but
only to theoretical domains in which there is `a priori`
knowledge. It took another thousand years before Leibniz
boldly generalized the rationalist account to all forms of
intelligent activity.
[T]he most important observations and turns of skill
in all sorts of trades and professions are as yet
unwritten. This fact is proved by experience when
- 3 -
passing from theory to practice we desire to accom-
plish something. Of course, we can also write up this
practice, since it is at bottom just another theory more
complex and particular....[1]
Now, three centuries after Leibniz, knowledge engineers
such as Edward Feigenbaum say with confidence that the rules
the expert uses have been put in a part of their mental com-
puters where they work automatically.
When we learned how to tie our shoes, we had to
think very hard about the steps involved ... Now
that we've tied many shoes over our lifetime, that
knowledge is "compiled," to use the computing term
for it; it no longer needs our conscious atten-
tion.[2] We also have a new name for what Socrates
and Plato were doing: "knowledge acquisition
research".[3]
But although philosophers and even the man in the
street have become convinced that expertise consists in
applying sophisticated heuristics to masses of facts, there
are few available rules. As Feigenbaum explains:
[A]n expert's knowledge is often ill-specified or
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [1] Leibniz, ←λS←λe←λl←λe←λc←λt←λi←λo←λn←λs, ed. Philip Wiener (New York:
Scribner, 1951), p. 48 (Our italics.)
$9 [2] Edward Feigenbaum and Pamela McCorduck, ←λT←λh←λe ←λF←λi←λf←λt←λh ←λG←λe←λn-
←λe←λr←λa←λt←λi←λo←λn, Addison-Wesley Pub. Co., 1983, p. 55
$9 [3] Ibid., p. 79
$9
- 4 -
incomplete because the expert himself doesn't always
know exactly what it is he knows about his
domain.[4]
So the knowledge engineer has to help him recollect what he
once knew.
[An expert's] knowledge is currently acquired in a
very painstaking way; individual computer scientists
work with individual experts to explicate the
expert's heuristics -- the problem of knowledge
acquisition is the critical bottleneck in artificial
intelligence.[5]
When Feigenbaum suggests to an expert the rules the expert
seems to be using he gets a Euthyphro-like response.
"That's true, but if you see enough patients/rocks/chip
designs/instruments readings, you see that it isn't true
after all." [6] and Feigenbaum comments with Socratic annoy-
ance: "At this point, knowledge threatens to become ten
thousand special cases."[7]
There are also other hints of trouble. Ever since the
inception of Artificial Intelligence, researchers have been
trying to produce artificial experts by programming the com-
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [4] Ibid., p. 85.
$9 [5] Ibid., pp. 79-80.
$9 [6] Ibid., p. 82.
$9 [7] Ibid., p. 82.
- 5 -
puter to follow the rules used by masters in various
domains. Yet, although computers are faster and more accu-
rate than people in applying rules, master-level performance
has remained out of reach. A recent evaluation of expert
systems lists only 9 as being in commercial use out of some
180 that have been developed.[8] Of all the expert systems
funded by the military according to an authority in DARPA,
only one has gone beyond the prototype stage.
The same story is repeated in every area of expertise.
In each area where there are experts with years of experi-
ence the computer can do better than the beginner, and can
even exhibit useful competence, but it cannot rival the very
experts whose facts and supposed heuristics it is processing
with incredible speed and unerring accuracy.
II. A Phenomenological Account of Skill Acquisition
In the face of this impasse it is necessary, in spite
of the authority and influence of Plato and 2000 years of
philosophy, for us to take a fresh look at what a skill is
and what the expert acquires when he achieves expertise.
One must be prepared to abandon the traditional view that a
beginner starts with specific cases and, as he becomes more
proficient, abstracts and interiorizes more and more sophis-
ticated rules. It might turn out that skill acquisition
moves in just the opposite direction: from abstract rules to
←←←←←←←←←←←←←←←←←←←←←←←←←
[8] Donald A. Waterman, "A Guide to Expert Systems",
Addison- Wesley Publishing Company, 1986
- 6 -
particular cases. Since we all have many areas in which we
are experts, we have the necessary data, so let's look and
see how adults learn new skills.
Stage 1: Novice
Normally, the instruction process begins with the
instructor decomposing the task environment into context-
free features which the beginner can recognize without bene-
fit of experience. The beginner is then given rules for
determining actions on the basis of these features, like a
computer following a program. The beginning student wants
to do a good job, but lacking any coherent sense of the
overall task, he judges his performance mainly by how well
he follows his learned rules. After he has acquired more
than just a few rules, so much concentration is required
during the exercise of his skill that his capacity to talk
or listen to advice is severely limited.
For purposes of illustration, we shall consider two
variations: a bodily or motor skill and an intellectual
skill. The student automobile driver learns to recognize
such interpretation-free features as speed (indicated by his
speedometer) and distance (as estimated by a previously
acquired skill). Safe following distances are defined in
terms of speed; conditions that allow safe entry into
traffic are defined in terms of speed and distance of oncom-
ing traffic; timing of shifts of gear is specified in terms
of speed, etc. These rules ignore context. They do not
- 7 -
refer to traffic density or anticipated stops.
The novice chess player learns a numerical value for
each type of piece regardless of its position, and the rule:
"always exchange if the total value of pieces captured
exceeds the value of pieces lost." He also learns that when
no advantageous exchanges can be found center control should
be sought, and he is given a rule defining center squares
and one for calculating extent of control. Most beginners
are notoriously slow players, as they attempt to remember
all these rules and their priorities.
Stage 2: Advanced beginner
As the novice gains experience actually coping with
real situations, he begins to note, or an instructor points
out, perspicuous examples of meaningful additional com-
ponents of the situation. After seeing a sufficient number
of examples, the student learns to recognize them. Instruc-
tional maxims now can refer to these new `situational aspects`
recognized on the basis of experience, as well as to the
objectively defined `non-situational features` recognizable by
the novice. The advanced beginner confronts his environ-
ment, seeks out features and aspects, and determines his
actions by applying rules. He shares the novice's minimal
concern with quality of performance, instead focusing on
quality of rule following. The advanced beginner's perfor-
mance, while improved, remains slow, uncoordinated, and
laborious.
- 8 -
The advanced beginner driver uses (situational) engine
sounds as well as (non-situational) speed in his gear-
shifting rules, and observes demeanor as well as position
and velocity to anticipate behavior of pedestrians or other
drivers. He learns to distinguish the behavior of the dis-
tracted or drunken driver from that of the impatient but
alert one. No number of words can serve the function of a
few choice examples in learning this distinction. Engine
sounds cannot be adequately captured by words, and no list
of objective facts about a particular pedestrian enables one
to predict his behavior in a crosswalk as well as can the
driver who has observed many pedestrians crossing streets
under a variety of conditions.
With experience, the chess beginner learns to recognize
over-extended positions and how to avoid them. Similarly,
he begins to recognize such situational aspects of positions
as a weakened king's side or a strong pawn structure despite
the lack of precise and universally valid definitional
rules.
Stage 3: Competence
With increasing experience, the number of features and
aspects to be taken account of becomes overwhelming. To
cope with this information explosion, the performer learns,
or is taught, to adopt a hierarchical view of decision-
making. By first choosing a plan, goal or perspective which
organizes the situation and by then examining only the small
- 9 -
set of features and aspects that he has learned are the most
important given that plan, the performer can simplify and
improve his performance.
Choosing a plan, a goal or perspective, is no simple
matter for the competent performer. He has expectations
about how things will turn out but he is often surprised
that his expectations are not fulfilled. No wonder. Nobody
gives him any rules for how to choose a perspective, so the
competent performer has to make up various rules which he
then adopts or discards in various situations depending on
how they work out. This procedure is frustrating, however,
since each rule works on some occasion and fails on others,
and no set of objective features and aspects correlate
strongly with these successes and failures. Nonetheless the
choice is unavoidable. While the advanced beginner can get
along without recognizing and using a particular situational
aspect until a sufficient number of examples makes identifi-
cation easy and sure, to perform competently `requires` choos-
ing an organizing goal or perspective. Furthermore, the
choice of perspective crucially affects behavior in a way
that one particular aspect rarely does.
This combination of necessity and uncertainty intro-
duces an important new type of relationship between the per-
former and his environment. The novice and the advanced
beginner applying rules and maxims feel little or no respon-
sibility for the outcome of their acts. If they have made
- 10 -
no mistakes, an unfortunate outcome is viewed as the result
of inadequately specified elements or rules. The competent
performer, on the other hand, after wrestling with the ques-
tion of a choice of perspective or goal, feels responsible
for, and thus emotionally involved in, the result of his
choice. An outcome that is clearly successful is deeply
satisfying and leaves a vivid memory of the situation
encountered as seen from the goal or perspective finally
chosen. Disasters, likewise, are not easily forgotten.
Remembered whole situations differ in one important
respect from remembered aspects. The mental image of an
aspect is flat in the sense that no parts stand out as
salient. A whole situation, on the other hand, since it is
the result of a chosen plan or perspective, has a "three-
dimensional" quality. Certain elements stand out as more or
less important with respect to the plan, while other
irrelevant elements are forgotten. Moreover, the competent
performer, gripped by the situation that his decision has
produced, experiences the situation not only in terms of
foreground and background elements but also in terms of
senses of opportunity, risk, expectation, threat, etc. As
we shall soon see, if he stops reflecting on problematic
situations as a detached observer, and stops thinking of
himself as a computer following better and better rules,
these gripping, holistic experiences become the basis of the
competent performer's next advance in skill.
- 11 -
A competent driver beginning a trip decides, perhaps,
that he is in a hurry. He then selects a route with atten-
tion to distance and time, ignores scenic beauty, and as he
drives, he chooses his maneuvers with little concern for
passenger comfort or for courtesy. He follows more closely
than normal, enters traffic more daringly, occasionally
violates a law. He feels elated when decisions work out and
no police car appears, and shaken by near accidents and
traffic tickets.
The class A chess player, here classed as competent,
may decide after studying a position that his opponent has
weakened his king's defenses so that an attack against the
king is a viable goal. If the attack is chosen, features
involving weaknesses in his own position created by his
attack are ignored as are losses of pieces inessential to
the attack. Removal of pieces defending the enemy king
becomes salient. Successful plans induce euphoria and mis-
takes are felt in the pit of the stomach.
In both of these cases, we find a common pattern:
detached planning, conscious assessment of elements that are
salient with respect to the plan, and analytical rule-guided
choice of action, followed by an emotionally involved
experience of the outcome.
Stage 4: Proficiency
Considerable experience at the level of competency sets
- 12 -
the stage for yet further skill enhancement. Having experi-
enced many situations, chosen plans in each, and having
obtained vivid, involved demonstrations of the adequacy or
inadequacy of the plan, the performer involved in the world
of the skill, "notices," or "is struck by" a certain plan,
goal or perspective. No longer is the spell of involvement
broken by detached conscious planning.
Since there are generally far fewer "ways of seeing"
than "ways of acting," after understanding without conscious
effort what is going on, the proficient performer will still
have to think about what to do. During this thinking, ele-
ments that present themselves as salient are assessed and
combined by rule to produce decisions about how best to
manipulate the environment. The spell of involvement in the
world of the activity will thus temporarily be broken.
On the basis of prior experience, a proficient driver
approaching a curve on a rainy day may sense that he is
traveling too fast. Then, on the basis of such salient ele-
ments as visibility, angle of road bank, criticalness of
time, etc., he decides whether to take his foot off the gas
or to step on the brake. (These factors would be used by the
`competent` driver consciously to `decide that` he is speeding.)
The proficient chess player, who is classed a master,
can recognize a large repertoire of types of positions.
Recognizing almost immediately and without conscious effort
the sense of a position, he sets about calculating the move
- 13 -
that best achieves his goal. He may, for example, know that
he should attack, but he must deliberate about how best to
do so.
Stage 5: Expertise
The proficient performer, immersed in the world of his
skillful activity, `sees` what needs to be done, but `decides`
how to do it. With enough experience with a variety of
situations, all seen from the same perspective (with the
same goal, significant, issue), but requiring different tac-
tical decisions, the proficient performer gradually decom-
poses this class of situations into subclasses, each of
which share the same decision, single action, or tactic.
Thus allows an immediate intuitive response to each situa-
tion which is characteristic of expertise.
The number of classes of discriminable situations,
built up on the basis of experience, must be immense. It
has been estimated that a master chess player can discrim-
inate roughly 50,000 types of positions. Automobile driving
probably requires the ability to discriminate a similar
number of situations. These discriminable classes of stimu-
lae, unlike the situational elements learned by the advanced
beginner, bear no names and, in fact, defy complete verbal
description.
While the proficient performer must be aware of the
meaning of the situation in order to decide what to do in
- 14 -
it, the expert is only sometimes aware of the significance
of the situation. In those areas of activity such as mar-
tial arts, sports such as basketball and tennis, as well as
the give and take of everyday social situations, an expert
has no time to deliberate. The quality of expertise then
depends on how many subtlety different types of situations
the brain can appropriately discriminate. Each type of
situation calls forth an appropriate type of response. What
transparently ←λm←λu←λs←λt be done ←λi←λs done.
In other types of activity such as business management,
sensitive press conferences, sports like golf and bowling,
and tournament chess, there is time to deliberate. Deli-
beration occurs in either of two types of cases. In one
type, the same stimulae present themselves compellingly in
several different gestalts, i.e., under several different
perspectives. Here the expert must decide what to see as
the current issue. More commonly, but less crucially, the
issue is clear but more than one response presents itself as
attractive. Here the expert must decide what to do. In
both cases the expert must deliberate. He uses such tech-
niques as counting out (in formal games) or getting more
information to decide between almost equally plausible
interpretations or courses of action (in the everyday
world). If the expert has had sufficient experience, simply
chosing whatever looks best no matter how slight the differ-
ence will normally produce high quality performance. The
above types of deliberative decision, however, determine who
- 15 -
will have the edge.
The expert chess player, classed as an international
master or grandmaster, in most situations experiences a com-
pelling sense of the issue and the best move. Excellent
chess players can play at the rate of 5-10 seconds a move
and even faster without any serious degradation in perfor-
mance. At this speed they must depend almost entirely on
intuition and hardly at all on analysis and comparison of
alternatives. We recently performed an experiment in which
an international master, Julio Kaplan, was required rapidly
to add numbers presented to him audibly at the rate of about
one number per second while at the same time playing five-
second-a-move chess against a slightly weaker, but master
level, player. Even with his analytical mind almost com-
pletely occupied by adding numbers, Kaplan more than held
his own against the master in a series of games. Deprived
during most of the game[9] of the time necessary to see
problems or construct plans, Kaplan still produced fluid and
coordinated, long-range strategic play.
Julio is only rarely aware of making plans and having
expectations at all. But here the question arises: How can
the expert initiate and carry through long-range strategies
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [9] At a few points in the game when things do not
seem obvious and deliberation of the type described
above is called for, Julio sneaks in a bit of short
range look-ahead. His addition slows down a fraction
of a second and he sometimes makes mistakes in his ad-
dition.
$9
- 16 -
without having assessed the situation, chosen a perspective,
made a plan, and formed expectation about how the situation
will work out? To answer this question the tradition has
assumed that goal directed action must be based on conscious
or unconscious `planning` involving beliefs, desires, and
goals. If, however, the expert responds to each situation
as it comes along in a way which has proven to be appropri-
ate in the past, his behavior will achieve the past objec-
tives without his having to have these objectives as goals
in his conscious or unconscious mind. Thus the expert is
moving into the future, and although he does not consciously
entertain expectations, he is set to respond to some
developments rather than others. If events take a turn that
is new to him, he will be startled, and, at best, fall back
to competence.
Kaplan's performance seems somewhat less amazing when
one realizes that a chess position is as meaningful,
interesting, and important to a professional chess player as
a face in a receiving line is to a professional politician.
Bobby Fischer, perhaps history's greatest chess player, once
said that for him "chess is life." Almost anyone can add
numbers and simultaneously recognize and respond to faces,
even though the face will never exactly match the same face
seen previously, and politicians can recognize thousands of
faces just as Julio Kaplan can recognize thousands of chess
positions similar to ones previously encountered.
- 17 -
Herbert Simon has studied the chess master's almost
instantaneous understanding of chess positions and accom-
panying compelling sense of the best move. While it is in
principle possible, in a formal game like chess, to associ-
ate a move with each different board position, there are far
too many positions for such a method to be applied in prac-
tice. To cope with this difficulty he claims that chess
masters are familiar with thousands of patterns, which he
calls chunks. Each chunk is a remembered description of a
small group of pieces in a certain relationship to each
other. The master supposedly associates an appropriate move
with each chunk. Each position thus causes a small number
of moves to spring to the master's mind and he chooses among
these without need for rule-like calculations.[10]
There are at least two problems with Simon's specula-
tion. Because most chess positions are comprised of several
chunks at each move, more than one move would come to mind
and need to be evaluated before the player would have a
sense of which was best. Yet, Julio Kaplan rarely seems to
require such evaluation when he plays rapidly while simul-
taneously adding numbers. Hence Simon's conceptualization of
chess in terms of chunk recognition, while providing a
theory about why moves spring to mind, still seems to fall
far short of the actual phenomenon of masterful play.
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [10] Herbert A. Simon, "Models of Thought", Yale
University Press, New Haven and London, 1979, pp. 386-
403.
$9
- 18 -
Furthermore, for Simon chunks such as a standard castled
king's formation are defined independently of the rest of
the position. A configuration that didn't quite fit the
description of a chunk, but in a real chess position played
the same role as the chunk, would not count as such. But
chess players can recognize the functional equivalence of
configurations which don't fall under a single definition.
For example, in some cases a configuration would count as a
standard castled king's formation even if one pawn were
advanced, but in other cases it would not.
Simon's model can easily be programmed as a standard AI
production system but it cannot account for the expert's
normal ability intuitively to see one clearly preferable
move. An elaboration of the production rule approach could
in principle solve this problem without needing a production
rule for every possible board position. What would be
required would be a production rule for a large number of
board positions and a rule for computing similarity so that
every position could be treated as identical to the one to
which it was most similar. Such a system would map all
positions directly into moves. The practical drawback is
that no chess player has introspective knowledge of any
rules for computing similarity. As we shall see in the next
section, neural net models of the brain can accomplish this
mapping without being programmed with explicit rules.
The expert driver, generally without any awareness, not
- 19 -
only knows by feel and familiarity when an action such as
slowing down is required, but he knows how to perform the
action without calculating and comparing alternatives. He
shifts gears when appropriate with no conscious awareness of
his acts. Most drivers have experienced the disconcerting
breakdown that occurs when suddenly one reflects on the gear
shifting process and tries to decide what to do. Suddenly
the smooth, almost automatic, sequence of actions that
results from the performer's involved immersion in the world
of his skill is disrupted, and the performer sees himself,
just as does the competent performer, as the manipulator of
a complex mechanism. He detachedly calculates his actions
even more poorly than does the competent performer since he
has forgotten many of the guiding rules that he knew and
used when competent, and his performance suddenly becomes
halting, uncertain, and even inappropriate.
It seems that a beginner makes inferences using rules
and facts just like a heuristically programmed computer, but
that with talent and a great deal of involved experience the
beginner develops into an expert who intuitively sees what
to do without applying rules.
Given our account of the five stages of skill acquisi-
tion, we can understand why the knowledge engineers from
Socrates to Feigenbaum have had such trouble getting the
expert to articulate the rules he is using. The expert is
simply not following any rules! He is doing just what
- 20 -
Feigenbaum feared he might be doing -- discriminating
thousands of special cases. This in turn explains why
expert systems are never as good as experts. If one asks
the experts for rules one will, in effect, force the expert
to regress to the level of a beginner and state the rules he
still remembers but no longer uses. If one programs these
rules on a computer one can use the speed and accuracy of
the computer and its ability to store and access millions of
facts to outdo a human beginner using the same rules. But
no amount of rules and facts can capture the knowledge an
expert has when he has stored his experience of the actual
outcomes of tens of thousands of situations.
On the basis of our skill model we predict that in any
domain in which people exhibit holistic understanding, no
system based upon heuristics will consistently do as well as
experienced experts, even if they were the informants who
provided the heuristic rules. There is, however, at least
one expert system that is both expert and heuristic and per-
forms as well as anyone in its field. It is the very
impressive XCON developed at Digital Equipment Corporation
to check the way components of VAX computers can be combined
to meet customers' needs. This, however, is not a case of
intuitive expertise. Even the experienced "technical edi-
tors" who perform the job at DEC depend on heuristic-based
problem solving and take about ten minutes to work out even
a simple case, so it is no surprise that an expert system
can rival the best experts. In domains like the above where
- 21 -
there are no intuitive experts -- loading cargo vehicles
seems to be another case -- one can expect expert systems to
be of great value.
III. The Connectionist Model
The knowledge engineer might still say that in spite of
appearances the mind/brain of the expert `must` be reasoning
-- making millions of rapid and accurate inferences like a
computer. After all the brain is not "wonder tissue" and how
else could it work? But there `are` models for what might be
going on in the hardware that make no use of the sort of
symbols and principles presupposed in traditional philosophy
and AI. That is, they do not use symbols that correspond to
recognizable features of the world and rules that represent
these features' relationships.
Such models are called connectionist models and use
parallel distributed processing. Specifically we have in
mind the most sophisticated models which not only have a set
of elements that encodes the stimulae from the real world
and a set that encodes an associated response, but which
also have intermediate (hidden) nodes which play a role in
translating the input into the output. A recent example is
Hinton's network for learning concepts by means of distri-
buted representations.[11] Hinton's network can be trained
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [11] Geoffrey Hinton, "Learning Distributed Represen-
tations of Concepts," to appear in Proc. 8th Annual
Conference Cognitive Science Soc., Amherst, Mass., Aug.
1986
$9
- 22 -
to encode relationships in a domain, which human beings con-
ceptualize in terms of features, without the network being
given the features that human beings use. Hinton produces
examples of cases in which in the trained network some nodes
can be interpreted as corresponding to the features that
human beings pick out, although they only roughly correspond
to these features. Most nodes, however, cannot be semanti-
cally interpreted at all. A feature used in a concept is
either present or not. In the net, however, although cer-
tain nodes are more active when a certain feature is present
in the domain, the amount of activity varies not just with
the presence or absence of this feature, but is affected by
the presence or absence of other features as well.
Given the state of the art in conventional AI and what
connectionism can do, how might one explain how the brain
produces skilled behavior? We will sketch the mechanisms
underlying the first and last stages of skill acquisition.
As we have already noted the novice uses rules and
features like a conventional AI program. Obviously, this
processing is performed by interconnected neurons, but the
question is open which if any of the connectionist models
proposed this far can instantiate procedures which allow the
appropriate level of abstraction. As far as we can see it
makes no philosophical difference how the formal operations
employed by conventional AI in producing novice performance
are instantiated.
- 23 -
While it is easy to see how conventional AI explains
the feature detecting and inference making of the novice,
and hard to see how a connectionist network would implement
the required step-wise processing, when we turn to the
expert, things are reversed. A connectionist account of
learning by examples and recognition of similarity seems
much more natural than any conventional AI account.
A neural net such as Hinton's is demonstrably capable
of adjusting its connection strengths and thresholds so as
to produce a desired output for each member of a specified
set of inputs. (Hinton's program for learning family rela-
tionships learns 100 such input/output pairs.)
Similarity recognition, while crucial to intuitive
expertise, plays no role in Hinton's concept learning model.
It could, however, easily be incorporated in a connectionist
architecture. Consider the case where a Hinton-like network
is used to map inputs representing chess positions into out-
puts which are representations of the issue and one or more
associated moves. Any new chess position which is not
identical with any previous learned input will produce some
particular output. If that output is similar to one of the
learned issue/action outputs which has been associated with
some given input position P, one can say that the system has
recognized the new input position as similar to position P.
Such a net can be said to recognize similarity without using
a predefined similarity measure -- without asking and
- 24 -
answering the question: "Similar with respect to what?".
Similarity simply means in this case whatever the net takes
to be similar based on its particular past training.
Sometimes the outputs will not be interpretable as
representations of any issue and move. Then the system can
be said to recognize that the current input is not similar
to any input to which it has been exposed. In the above
way, a large enough net should be able to discriminate the
approximately 50,000 different situations which a grandmas-
ter needs to distinguish, and to respond to a new situation
as similar to one of these or as outside its intuitive
expertise.
IV. Conclusion
What are the implications of all this for conventional
AI? That, of course, depends on what you take to be the
essential commitments of conventional AI, and also on what
one means by a distributed representation. We shall attempt
to distinguish three progressively non-conventional senses
of distribution. Although we describe each kind of distri-
bution abstractly, we shall also use Hinton's network for
learning family relationships to illustrate what we have in
mind by each type. In so doing, we shall discuss possible
interpretations of Hinton's first level of hidden nodes.
Conventional AI generally takes as the features in
terms of which to encode the relevant structure of a domain
- 25 -
the sort of features which human beings can pick out and
use. Thus from Newell and Simon's protocols, through
frames, scripts, procedures and productions, to recent
knowledge engineering, AI has used introspection to deter-
mine what features to use. As we have seen, even in a net
as simple as Hinton's, after learning a set of associations
some nodes can sometimes be interpreted as simple feature
detectors, i.e., as detecting the sort of features human
beings use in conceptualizing a domain. This would be a
case of non-distributed representation.
These everyday features (for Hinton, generation,
nationality, etc.) could also be represented in a distri-
buted way, i.e., when one of these features is present a
certain `group` of nodes would always be active. However,
representations can be distributed in a second, stronger
sense. Single nodes or even sets of nodes need not detect
the sorts of features people pick out. They could equally
well detect combinations of these features. To take the
simplest case: If there are two hidden nodes and three
features A, B, and C, and only one of these features can be
present in any given input, then if one hidden node detects
the presence of either A or B, one detects the presence of B
or C, the behavior of these two will uniquely determine
which feature is present. So it is not necessary that one
node detect A, one detect B, and both on, say, detects C.
In Hinton's model one node or group of nodes might, for
- 26 -
example, detect both second generation Italians and English
people on the left side of their family tree. Once one
admits that the features are distributed this second way,
there is no reason to believe that in large networks the
distributed features are the sort that the mind can recog-
nize or even unconsciously use. Moreover, in the third and
most extreme case there is no reason to suppose that the
features being detected are any combination of the features
people ordinarily detect. In Hinton's model it is not obvi-
ous that nationality or generation need play any role in
learning to map the input information (person and relation-
ship) into the output (another person). In this case the
Hinton model could be said to be using features which are
abstract and non-mental.
Hinton has picked a domain, family relationships, which
is constructed by human beings precisely in terms of the
features human beings normally notice, such as generation
and nationality. Furthermore, though some people can be
better than others in figuring out these family relation-
ships, the domain is so small and combinatorial it does not
lend itself to intuitive expertise. Hinton's paper analyzes
those cases in which, for certain random initial connection
strengths, some nodes can be interpreted as representing the
everyday features people use. Our experience with Hinton's
model shows, however, that even his model seems, for some
random initial connection strengths, to learn its associa-
tions without any obvious use of these everyday features.
- 27 -
If in domains that people analyze in terms of everyday
features the network uses other features, one would expect
the situation to be even worse for interpreting nodes of
neural nets that have been trained to capture intuitive
expertise. For example, if it turns out that the brain
processes information using distributions of the third kind
in domains such as chess, then if one could train a neural
net to capture the mapping of situations to issues and moves
that a given player has seen, there would be no reason to
believe that the individual nodes or combinations of nodes
could be interpreted as detecting the sort of features that
people talk about in chess, or even the 50,000 context-free
chunks that Simon would like to isolate.
It seems that the above combination of phenomenology
and connectionism may well be devastating to conventional AI
as a `practical` endeavor. Could anything be salvaged of the
cognitivist-rationalist intuition that for any domain that
can be mastered, there exists a set of features and rules
which explains that mastery? If the connectionist hierarch-
ical account turns out to be correct[12], then, given any
particular domain and the experience of a particular person
in that domain, there would, indeed, always exist an account
←←←←←←←←←←←←←←←←←←←←←←←←←
$9 [12] Obviously no one thinks that the Hinton type
model captures in detail what it going on in the brain.
For example, the brain seems to use lateral connections
with feedback at each level and neurons respond in com-
plex ways to their neighbors. However, these con-
siderations seem to only complicate but not to negate
the philosophical issue we wish to raise.
$9
- 28 -
of his expertise in terms of higher-order abstract features.
To construct this account from the network that has learned
his associations, each node one level above the input nodes
could be interpreted as detecting when one of a certain set
of input patterns is present. If the set of input patterns
which that particular node detects is given an invented name
(it almost certainly won't have a name in our vocabulary),
the nodes could be interpreted as detecting the feature so
named. Hence, every node one level above the input level
can be characterized as a feature detector. Similarly,
every node a level above these nodes can be interpreted as
detecting a higher-order feature which is defined as the
presence of one of a specified set of patterns among the
first level features detectors. And so on up the hierarchy.
The fact that a given expert's expertise can always be
accounted for in terms of relations among a number of highly
abstract features of his skill domain does not, however,
preserve the rationalist intuition that these explanatory
features must capture the essential structure of the domain.
If the net is taught one more association of an input/output
pair (where the input prior to training produces an output
different from the one to be learned), the interpretation of
at least some of the nodes will have to be changed. So the
features which some of the nodes picked out before the last
instance of training could turn out not to have been invari-
ant structural features of the domain.[13]
←←←←←←←←←←←←←←←←←←←←←←←←←
[13] As long as the behavior of each pattern recog-
- 29 -
Stuart E. Dreyfus
Hubert L. Dreyfus
University of California,
Berkeley
←←←←←←←←←←←←←←←←←←←←←←←←←
nizing net at each point in its training must be ex-
plained in terms of some different features, there is
no guarantee that the similarities picked out by a
given net, in the way described in the previous sec-
tion, are the ones that would produce high quality per-
formance across the domain. Unless the system detects
as similar the class of inputs which require the same
response in the real world, there will not be high
quality performance. The system will respond to
overall patterns without analysis like an expert but
will not produce masterly performance. Class A and ex-
pert level chess players, when playing intuitively, do
not share a sense of similarity with each other, i.e.,
they often disagree with each other as to whether two
board positions present the same issue or require the
same type of move.
Chess grandmasters, on the contrary, exhibit remarkable
unanimity when evaluating a position. One can speculate
that they have had so much chess experience that the number
of neurons in their brain dedicated to the chess net has ap-
proached the minimum necessary for encoding their experi-
ence, (or, conversely, their number of chess experiences is
approaching the largest number that a fixed number of neu-
rons can correctly associate), and that therefore the
abstract, higher-order, non-mental features they detect are
the minimal necessary features for getting the job done.
This seems to be the only remnant of the cognitivist-
rationalist intuition that expertise can always be explained
in terms of stable features that depend only on the domain
and not on the particular set of experiences of the expert.
Such abstract, higher-order features, however, would not be
useful to conventional AI and would only be domain indepen-
dent relative to the size of the net of neurons dedicated to
this domain.
∂27-Oct-86 0100 JMC
tandberg
∂27-Oct-86 0900 JMC
reservations for Reston meeting
∂27-Oct-86 1042 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu lunch this week?
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 27 Oct 86 10:39:11 PST
Received: by lindy.stanford.edu; Mon, 27 Oct 86 10:38:54 PST
Date: Mon, 27 Oct 86 10:37:05 PST
From: <ELLIOTT@slacvm.bitnet>
Reply-To: ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu
To: JMC@sail.stanford.edu
Subject: lunch this week?
Date: 27 October 86 10:06-PST
From: ELLIOTT@SLACVM
To: JMC@SAIL
Subject: lunch this week?
Date: 27 October 1986, 10:05:01 PST
From: Bloom, Elliott ELLIOTT at SLACVM
To: JMC at SAIL.STANFORD
Subject: lunch this week?
I am free Tues, Wed. for lunch. Can you make it one of those
two days? Thurs and Frid. are difficult for me this week.
∂27-Oct-86 1353 AIR lectures
Would it be OK, if I attend your lectures?
∂27-Oct-86 1428 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu Re: re: lunch this week?
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 27 Oct 86 14:27:57 PST
Received: by lindy.stanford.edu; Mon, 27 Oct 86 14:27:38 PST
Date: Mon, 27 Oct 86 14:27:34 PST
From: <ELLIOTT@slacvm.bitnet>
Reply-To: ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu
To: JMC@sail.stanford.edu
Subject: Re: re: lunch this week?
Date: 27 October 86 14:22-PST
From: ELLIOTT@SLACVM
To: JMC@SAIL.STANFORD.EDU
Subject: Re: re: lunch this week?
Date: 27 October 1986, 14:21:54 PST
From: Bloom, Elliott ELLIOTT at SLACVM
To: JMC at SAIL.STANFORD.EDU.STANFORD
Subject: Re: re: lunch this week?
In-Reply-To: JMC AT SAIL.STANFORD.EDU.STANFORD -- 10/27/86, 11:08
Meet you at 12:05 at the Faculty Club.
∂27-Oct-86 1424 MRC cellular
To answer your last question, there are two ways to set up a system:
. the cheap way is to set up a few very powerful tranmitters.
. the right way is to set up lots and lots of small, less powerful
transmitters
The right way also gives you more capacity. There are 333 channels
per company for a total of 666. You can't have the same channels used
by adjacent transmitters for obvious reasons. If a cell was small
(e.g. a block) the channels can be reused perhaps three or four blocks
away for another call...
I *strongly* recommend an organization known as Priority Communications
in Santa Clara. Keith, the owner, charges a bit more than the other
guys do but he does *quality* work and he *supports* his work. Dealer's
Car Stereo et al will butcher your car in the process of installing the
phone and will make mistakes such as cutting your antenna coax to fit
instead of the required 8'. Keith's a Cellular One dealer. I think
that for the nonce Cellular One looks to be a better bet than GTE. I
am a GTE customer, but I am not a happy customer...
∂27-Oct-86 1604 TEICH@Sushi.Stanford.EDU [David Teich <TEICH@Sushi.Stanford.EDU>: degree plan]
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 27 Oct 86 16:04:48 PST
Date: Mon 27 Oct 86 16:02:45-PST
From: David Teich <TEICH@Sushi.Stanford.EDU>
Subject: [David Teich <TEICH@Sushi.Stanford.EDU>: degree plan]
To: jmc@Sail.Stanford.EDU
Message-ID: <12250251765.20.TEICH@Sushi.Stanford.EDU>
Mail-From: TEICH created at 26-Oct-86 18:51:12
Date: Sun 26 Oct 86 18:51:10-PST
From: David Teich <TEICH@Sushi.Stanford.EDU>
Subject: degree plan
To: mccarthy@Sushi.Stanford.EDU
cc: teich@Sushi.Stanford.EDU
Message-ID: <12250020280.16.TEICH@Sushi.Stanford.EDU>
I'd like to get together with you sometime this week to discuss my
degree plan and then to get the paperwork signed. My only class on
tues-thurs is from 2:30-4, so if you can do that on one of those days,
any other time is fine.
One thing I'd like to discuss with you is the possibility of substituting
an extra AI course for CS212. I had a 360/370 assembler class in 1979, and
I spent 2 & 1/2 years as a senior computer operator on a large computer
floor (~13 IBM & DEC computers). Since my interest lay in the area of
artificial intelligence applications for human-reasoning and business
applications, I feel that I have as good an exposure to systems
architecture as I will need. The extra course in AI would allow me
the freedom to further explore the cognitive sciences aspect of the
field.
For my regular 12 units of AI electives I was planning on taking:
CS275 - Syntax of NL
276 - Semantics of NL
306 - Recursive Programming
326 - Epistemological Problems of AI
But I would also like to take CS328B - Applying Cognitive Psychology to
Computer Systems, and also possibly take a CS329 topics course if one
came along in this area. If I could substitute CS328B, or some other
AI related course in place of CS212, I feel that I would gain a much
better idea of what is possible in my area of interest, and how I would
best be able to enter an appropriate area of research.
David Teich
-------
-------
∂28-Oct-86 0651 JOSHI@cis.upenn.edu supporting your proposal
Received: from LINC.CIS.UPENN.EDU by SAIL.STANFORD.EDU with TCP; 28 Oct 86 06:51:13 PST
Received: by linc.cis.upenn.edu
id AA00922; Tue, 28 Oct 86 09:50:09 EST
Posted-Date: Tue, 28 Oct 86 09:37 EST
Message-Id: <8610281450.AA00922@linc.cis.upenn.edu>
From: Aravind Joshi <Joshi@cis.upenn.edu>
Subject: supporting your proposal
To: John McCarthy <JMC@sail.stanford.edu>
Date: Tue, 28 Oct 86 09:37 EST
John:
Thank you very much for message informing me about your support for
the proposed workshop. I understand your concerns but I want to assure
you that I will use my best judgement in making sure that a very substantial
part of the workshop will deal with issues in natural language processing
that are relevant to AI. Having walked on both sides of the fence (dividing
linguitics and AI) I feel I can use my judgement effectively. Even then
I think I know what worries you and I share some of your concerns too.
I hope to make it a successful workshop.
I appreciate your comments on my contribution to AAAI and I hope to continue
my contributions to AAAI in the future.
Thanks again.
With best wishes,
Aravind
∂28-Oct-86 0800 JMC
408 294-9493 Audrey Marshall
∂28-Oct-86 1027 VAL reply to message
[In reply to message rcvd 27-Oct-86 19:21-PT.]
I am sorry, who is Tandberg?
∂28-Oct-86 1045 VAL re: reply to message
[In reply to message rcvd 28-Oct-86 10:35-PT.]
What do we want Tannberg to do at this point? Will you write to him, or should I
do that?
I am sending an application form to Moscow, in which I'm saying that I want to
spend there a week before the Congress and want to take my wife with me. (I
realize that those things definitely won't be permitted, but I think I should
ask).
Would you be able and willing to come if I invite Kheifets over this Saturday
or Sunday night?
∂28-Oct-86 1100 JMC
Call about picture.
∂28-Oct-86 1137 TAJNAI@Score.Stanford.EDU Inference Corporation
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 28 Oct 86 11:34:47 PST
Date: Tue 28 Oct 86 11:32:59-PST
From: Carolyn Tajnai <TAJNAI@Score.Stanford.EDU>
Subject: Inference Corporation
To: JMC@Sail.Stanford.EDU
cc: Diaz@Score.Stanford.EDU, Macmilk@Score.Stanford.EDU
Message-ID: <12250464798.15.TAJNAI@Score.Stanford.EDU>
I just had a call from Karen Farr, Chuck Williams' secretary.
Inference is dropping their Forum membership.
They will probably ask you to stay over after your next SAB meeting
and make a Forum visit the next day. I'm still holding the $2,000
liaision fee that I set aside for your visit.
I'm disappointed that they are dropping. Is there anything you can
do?
Carolyn
-------
∂28-Oct-86 1151 AI.ELLIE@MCC.COM group picture
Received: from MCC.COM by SAIL.STANFORD.EDU with TCP; 28 Oct 86 11:51:45 PST
Date: Tue 28 Oct 86 13:51:46-CST
From: Ellie Huck <AI.ELLIE@MCC.COM>
Subject: group picture
To: jmc@SU-AI.ARPA
cc: ai.buchberger@MCC.COM
Message-ID: <12250468220.23.AI.ELLIE@MCC.COM>
On the afternoon of November 15, 1986, the MCC AI Program is
having a group picture taken. If your schedule permits, we
would like to have you be included in the picture.
Hope you can make it, we look forward to seeing you then.
-------
∂28-Oct-86 1422 AI.ELLIE@MCC.COM ERROR
Received: from MCC.COM by SAIL.STANFORD.EDU with TCP; 28 Oct 86 14:22:08 PST
Date: Tue 28 Oct 86 16:16:38-CST
From: Ellie Huck <AI.ELLIE@MCC.COM>
Subject: ERROR
To: jmc@SU-AI.ARPA
cc: ai.buchberger@MCC.COM
Message-ID: <12250494591.23.AI.ELLIE@MCC.COM>
Dr. McCarthy -- in my message of today regarding a Group Picture, I
gave the wrong address. It should have read November 25 NOT November
15. Very sorry for the inconvenience. Ellie
-------
∂28-Oct-86 1500 JMC
Oliger
∂28-Oct-86 1857 RWF
To: JMC@SAIL.STANFORD.EDU, su-etc@SAIL.STANFORD.EDU
I find myself in (rare) disagreement with JMC on an issue of personal liberties.
There is no plausible case for requiring any one book in an education, but the
systematic exclusion of a class of books may be constitutionally and morally
impermissible , by analogy with the systematic exclusion of an ethnic group
from juries. There is an overwhelming state interest in education of the young,
which justifies compulsory education. If you grant that much, you must grant
accreditation to prevent the frustration of that interest, e.g. by fraudulent
schools. You can reasonably require the teaching of history, physical ed.,
arithmetic, etc. I thiink there would be a presumption of the right of the
state to specify curriculum in general, and require private schools to adhere
to the specs. This argument is of course an amateur legal argument, and does
not address the moral issues.
The moral argument is that our children are not our chattels. We recognize that
parents do not have the righ to abuse their children, though practical reasons
make us reluctant to intervene. To prevent children from reading fantasy, myth,
non-Christian religious doctrine, non-biblical science, etc., is morally wrong.
It is not practical for the state to intervene in the home to permit free
use of libraries, but it is monstrous for the state to collude in depriving
children of the liberty to read. Clearly people that will teach their own
children just to avoid the witches of Oz do not intend to let their children
read freely under any circumstances.
Ask yourselves: if you knew a child whose parents imposed the reading limitations
suggested by this case, would you, if need be, even violate law to help the
child evade the restrictions? I would, perceiving the right to read everything
as anatural right that precedes law (I am a Lockean, at least on this matter; I
can't prove that people have natural rights, but allstates that don't act as if
there were natural rights are tyrranies). The state is not compelled to actively
protect all natural rights, but when the state does not cater to strongly felt
and legitimate needs, there is the risk of public disorder when people try to
enforce their own rights.his is analogous to the need for a criminal law, to
prevent vendettas and blood feuds. Here, if there is a compelling moral case for
helping the child, and the likelihood of disorder when private parties do so,
I think the state has an overwhelming interest in acting.
What say y'all?
-Bob Floyd
∂28-Oct-86 1856 POSER@CSLI.STANFORD.EDU Re: parent suit
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 28 Oct 86 18:56:19 PST
Date: Tue 28 Oct 86 19:00:05-PST
From: Bill Poser <POSER@CSLI.STANFORD.EDU>
Subject: Re: parent suit
To: JMC@SU-AI.ARPA
cc: su-etc@SU-AI.ARPA, crispin@SUMEX-AIM.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Tue 28 Oct 86 15:46:00-PST
I think that I agree with JMC. Parents should be able to
educate their kids at home if they want to, provided that they
maintain some minimum educational standards. I doubt that Shakespeare
is essential to the sort of minimum education that ought to be required,
but these people are so far out that I wonder whether anything approaching
a decent education is possible by their standards. An education consisting
solely of Bible-reading is not adequate.
Actually, I don't understand exactly what these parents objected to.
From what I first read I thought it would be some sort of relatively
strong material (sex, people eating each other to survive,
comparative religion) but one of the things I read (the NYT article?)
contained two excerpts from materials that they found objectionable.
I am at a loss to explain what they objected to. These excerpts seemed
totally innocuous to me. Does anyone know what their real gripe is
about?
-------
∂29-Oct-86 0549 JUSTESON@Sushi.Stanford.EDU Core courses possibly to waive
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 29 Oct 86 05:49:18 PST
Date: Wed 29 Oct 86 05:47:30-PST
From: John S. Justeson <JUSTESON@Sushi.Stanford.EDU>
Subject: Core courses possibly to waive
To: jmc@Sail.Stanford.EDU
Message-ID: <12250664050.9.JUSTESON@Sushi.Stanford.EDU>
The core courses that I think fall into the category you had in mind as
hardware-oriented and not as useful are listed below. Use your own
discretion; if you feel the course is relevant then you needn't inquire
about having it waived.
212: Computer Architecture and Organization
240A,B: Operating Systems
243: Advanced Compiling Techniques
In place of these, more courses would be taken in logic, AI, and perhaps
mathematically oriented courses; also language-oriented CS and perhaps
CS-oriented linguistics courses.
Thanks...
-- John
-------
∂29-Oct-86 0929 CLT Qlisp meeting reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
Time: Friday, Oct 31 12:15-1:15
Place: MJ352
Topic: Theme and variations on Qlisp
Byron Davies will talk about his parallel lisp implementations
on the lisp machine.
∂29-Oct-86 1155 RA Dr. Bloom
A reminder that you have lunch today at noon at the faculty club with Dr. Bloom.
∂29-Oct-86 1623 CLT arpa proposal
There is a copy of the current state on your desk.
Please look at your section to see if it is ok.
Could you check Shoham's section to see if it
fits and supply a title.
Rutie has the online file.
∂29-Oct-86 1708 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
RAMIFICATION AND QUALIFICATION IN THE BLOCKS WORLD
Matt Ginsberg
David Smith
Thursday, October 30, 4pm
MJH 252
In this talk, we discuss the need to infer properties of actions
from general domain information. Specifically, we discuss the
need to deduce the indirect consequences of actions (the
ramification problem), and the need to determine inferentially
under what circumstances a particular action will be blocked
because its successful execution would involve the violation of
a domain constraint (the qualification problem).
We present a formal description of action that addresses these
problems by considering a single model of the domain, and updating
it to reflect the successful execution of actions. The bulk of the
talk will involve the investigation of simple blocks world problems
that existing formalisms have difficulty dealing with, including
the Hanks-McDermott problem, and two new problems that we describe
as "the dumbbell and the pulley".
∂29-Oct-86 1730 cramer@Sun.COM Paly International Festival
Received: from Sun.COM by SAIL.STANFORD.EDU with TCP; 29 Oct 86 17:28:36 PST
Received: from snail.sun.com by sun.com (3.2/SMI-3.0)
id AA00141; Wed, 29 Oct 86 17:27:59 PST
Received: from clem.sun.uucp by snail.sun.com (3.2/SMI-3.2)
id AA28183; Wed, 29 Oct 86 17:27:09 PST
Received: by clem.sun.uucp (3.2/SMI-3.2)
id AA00582; Wed, 29 Oct 86 17:27:44 PST
Date: Wed, 29 Oct 86 17:27:44 PST
From: cramer@Sun.COM (Sam Cramer)
Message-Id: <8610300127.AA00582@clem.sun.uucp>
To: jmc@sail.stanford.edu
Subject: Paly International Festival
I sent you a brief note about a month ago regarding this event. I've found
out a bit more about it, and wanted to let you know.
The International Festival is a 3 day event taking place at the end of January.
The intent is to educate Paly students about things international. I have a
list of speakers are planned for this event; below I'm listing the ones that
seem questionable. I'm curious if you know anything about these lecturers.
Nancy Moss, Phd., sociologist and health care researcher, Stanford and UCSF.
"Why the US is involved in Central America" - geographical, historical,
and economic bases [sic] for US interest in Central America
Kathleen Namphy, Phd., Lecturer in undergraduate studies, Stanford
"The Palestine Question - Israel and PLO response" - OR -
"February rebellion against Duvalier" - Dr. Namphy is the sister-in-law
of the current President of Haiti.
Mohammed Abo-Rabbo, Phd. candidate in linguitsics, Stanford (Palestinian)
"The Dark Ages or the Golden Age?" - what Europe calls the Dark Ages
the Islamic world sees as their Golden Age - OR -
"A Personal View of Quaddafi's Libya" - based on five years living in
Libya
Rev K. Gilliam, Graduate Student and Poet, Stanford
"'Grenadian Requiem' and other poetry on Haiti and Grenada" - OR -
"Spiritual Motifs reflected in the photography and poetry of Vincent
Gilliam" - on the religions of the Mediterranean region.
William Gould, J.D., Professor of Law, Stanford
"South Africa" - has visit and talked on S. Africa before
Joel Samoff, Phd., Professor of African Politics, Stanford School of Education.
"Perspectives on the struggle for Southern Africa" - regional and
African continental points of view on southern Africa's contested
terrain.
Paul Strassberg, Executive Director - Volunteers in Asia, Stanford.
"Third World Poverty: can I do anything about it?" - ways for
students to get involved.
Terrence Emmons, Phd., Professor of History, Stanford.
"The Soviet Union today" by a Russian specialist.
Martin and Dorthea Hellman - lived in the USSR.
"The Soviet Union: Myth and Reality" - Dr. Hellman is at Stanford.
[elswhere mentions that he is an EE prof.]
Incidentally, there are two people from Hoover on the list (Larry Diamond
and Alvin Rabushka), so the situation isn't totally bleak.
Do any of these names ring a bell? Some look mighty fishy, others may be
innocuous. Please pass the list along to anyone else who might know about the
speakers.
Thanks,
Sam Cramer
∂29-Oct-86 1852 VAL Kheifets
We meet for dinner at our place (1050 Miller) on Saturday, 7pm. Carolyn and
Timothy are, of course, invited too.
∂29-Oct-86 2100 ASHOK@Sushi.Stanford.EDU Re: Parent suit
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 29 Oct 86 21:00:30 PST
Date: Wed 29 Oct 86 20:59:44-PST
From: Ashok Subramanian <ashok@Sushi.Stanford.EDU>
Subject: Re: Parent suit
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 29 Oct 86 00:11:00-PST
Message-ID: <12250830117.14.ASHOK@Sushi.Stanford.EDU>
The ad for the movie "28 UP" mentions something similar "Give me a child
until he is seven, and I will give you the man," and attributes it to
Saint Francis Xavier.
--ashok
-------
∂29-Oct-86 2131 RWF
To: su-etc@SAIL.STANFORD.EDU, JMC@SAIL.STANFORD.EDU,
stone@PSYCH.STANFORD.EDU, r.roland@LEAR.STANFORD.EDU
Some of you apparently did not read my
argument carefully. What I wrote was, I think,
closely reasoned.
I did not say that the constitution requires
compulsory education or a particular curriculum.
We can all agree it does not. I said there is an
OVERWHWHELMING STATE INTEREST in education, which
JUSTIFIES compulsory education. That is,
legislators are wise to reqquire it, and the
constitution permits it. If you grant that, you
grant SOME control of content and quality. That
states do not in factspecify more than the names
of required courses does not contradict their
having a presumptive right to do so (JMC note).
Fraudulentcourses like creation science may and do
require exercise of that right, though usually by
courts.
Valerie Stone not only gets my initials wrong,
she stoops to a personal dig. Valerie, ALL legal
arguments by non-lawyers are amateur. Your legal
argument is not a refutation of mine. In fact, I
grant you the constitution guarantees religious
freedom (though not an unlimited right to harm
your children for religious reasons; this has been
addjudicated in blood transfusion cases, and most
pertinently in denying the Amish the right to
limit the time span of their children's
educaation). I grant you as above that the
constitution neither guarantees the right to, nor
mandates any particular type of, education. Now
what?
The legal argument I made, in a little more
detail, is that first amendment freedom of the
press rights prohibit systematic exclusion of a
class of books from a public or state-mandated
education in the same way that fourteenth
amendment equal protection of the laws rights
prohibit systematic exclusion of a class of
citizens from juries. (If religious rights are in
conflict with this, they are not the children's
religious rights anyway).
John McCarthy (more or less) asserts the legal
riight of parents to educate their own children.
I have no quarrel with that, but my arguments
about content are not thereby refuted. For
example, parents who believe the hollow earth
theory may well have the right to indoctrinate
their children or teach them at home but not both.
Those I have named flunk the midterm, and must
reread my original argument, It was, I think,
narrowly drawn, non-doctrinaire, and logically
strong. If you find REAL weaknesses in it you
still have a chance at an A for the course.
Valerie Stone's scenario of fundamentalists
taking over the public school system has in fact
been tried, and thee federal courts have overruled
"creation science" in Arkansas about five years
ago. Louisiana is trying it again, because it is a
vote-getter, and will fail. That's why you don't
need the right to teach evolutiion at home (unless
you want it taught competently). Hooray for the
constitution--badly treated as it is, it
repeatedly saves us from our worst excesses.
Don't let your worst nightmares get to you, O
reader. I went to school in Alabama, Arkansas,
and Louisiana, an instinctive atheist in a crowd
of Baptists. It wasn'tt a very good education,
but the teachers were there in good faith, and
they let me sit in the back and read the
encyclopedia, which was all right with me. The
main thing wrong was wrong elsewhere: so much of
it is a waste of time. Like liberals,
fundamentalists are often in public view in
adversarial situations; forget not that most of us
(bboard types excepted) don't spend much of our
time arguing.
-Bob Floyd
∂30-Oct-86 0134 CERF@A.ISI.EDU Digital Library Meeting
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 30 Oct 86 01:33:59 PST
Date: 30 Oct 1986 04:33-EST
Sender: CERF@A.ISI.EDU
Subject: Digital Library Meeting
From: CERF@A.ISI.EDU
To: jmc@SU-AI.ARPA, lampson@SRC.DEC.COM
To: reddy@A.CS.CMU.EDU
Message-ID: <[A.ISI.EDU]30-Oct-86 04:33:47.CERF>
John, Butler and Raj,
Could you advise as to your arrival/departure plans for the Digital
Library meeting so we can plan for meals and transportation?
We need to know which of you will arrive on Friday, Nov 7 and will
want dinner. Similarly, if you are unable to stay for Sunday,
Nov 9, we need to know this. I am assuming that all three of you
will be able to stay at least through dinner on Saturday, Nov 8.
Thanks.
Vint
∂30-Oct-86 0140 CERF@A.ISI.EDU January meeting
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 30 Oct 86 01:39:51 PST
Date: 30 Oct 1986 04:37-EST
Sender: CERF@A.ISI.EDU
Subject: January meeting
From: CERF@A.ISI.EDU
To: Bibliotects:
Message-ID: <[A.ISI.EDU]30-Oct-86 04:37:16.CERF>
Gents,
could I get a straw vote on the January 5-6 dates? I'd like to make sure that
90% of you can make it then. It is crucial that all of us meet at once by then
to assure concensus on the system design and report contents.
Many thanks,
Vint
∂30-Oct-86 0224 JUSTESON@Sushi.Stanford.EDU Courses to waive
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Oct 86 02:24:00 PST
Date: Thu 30 Oct 86 02:23:17-PST
From: John S. Justeson <JUSTESON@Sushi.Stanford.EDU>
Subject: Courses to waive
To: jmc@Sail.Stanford.EDU
Message-ID: <12250889016.14.JUSTESON@Sushi.Stanford.EDU>
The two-course sequence 240AB has apparently been replaced by just the
first, 240A, in the core requirements; so the 4 courses I sent you
should be reduced to three.
-- John
-------
∂30-Oct-86 1356 UDI@Score.Stanford.EDU CFL
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Oct 86 13:56:41 PST
Date: Thu 30 Oct 86 13:26:51-PST
From: Ehud Shapiro <UDI@Score.Stanford.EDU>
Subject: CFL
To: jjw@Sail.Stanford.EDU
cc: jmc@Sail.Stanford.EDU
Message-ID: <12251009816.29.UDI@Score.Stanford.EDU>
Joe, I have given jmc a paper by Jacob Levi, a student of mine,
a concurrent functional language, and its compilation to FCP.
It might be of interest to you too. Will be glad to hear comments
on it.
Udi
-------
∂30-Oct-86 1615 RA leaving
It is Thursday and I am leaving early.
∂30-Oct-86 2154 NILSSON@Score.Stanford.EDU Binford Letter
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Oct 86 21:54:08 PST
Date: Thu 30 Oct 86 08:10:39-PST
From: Nils Nilsson <NILSSON@Score.Stanford.EDU>
Subject: Binford Letter
To: jmc@Sail.Stanford.EDU
cc: nilsson@Score.Stanford.EDU
Message-ID: <12250952253.9.NILSSON@Score.Stanford.EDU>
John, Here is a letter that has the blessing of the School of
Engrg (adapted to Tom's case). I made some top-of-the-head suggestions
about to whom to write the letter and with whom to compare Tom. (We
ought to ask Tom for more names in both categories. Can you do that?)
Dear xxxx:
[[[Candidate People: Horn, Kanade, Rosenfeld, Brady, Fischler, ...]]]
I am writing you in regard to Thomas Binford. It is our custom at
Stanford to review and evaluate the work of our research faculty
periodically through inquiries of people of eminence outside of the
University. This particular review of Professor Binford is concerned
with his reappointment as Professor (Research).
A key criterion for continued appointment includes the requirement that a
professor continue to perform research of high quality and impact leading
to clear visibility of our institution and recognition of the faculty
member. The accomplishments and continued promise of faculty members are
evaluated by comparison and ranking with outstanding contemporaries. This
comparison and ranking is the keystone of our decision process.
We request then, not a recommendation from you regarding our decision,
but rather an assessment of the stature, general reputation and quality
of work. We hope that your assessment can be responsive to our criteria
and in the context of comparison of him to his contemporaries whose work
may be familiar to you; e.g. B. K. P. Horn, M. Brady, A. Rosenfeld, T.
Kanade, and <others?> [[[making sure that we delete from this list the
person to whom the letter is addressed!]]]. For your use, I enclose a
recent vitae of Thomas Binford.
In developing your assessment, the following questions may provide a useful
framework:
1. How well and in what capacity do you know Tom?
2. How would you characterize his ability to carry out creative and
significant research?
3. What basis is there for expecting him to make significant
contributions in the future?
4. Can you cite any specific contributions that have had a significant
effect on Tom's reputation?
5. How does Tom rank in comparison with his peers?
I am deeply appreciative of your taking time from a busy schedule to
respond to this inquiry. I also wish to assure you that your response
will be kept confidential within the University administration. We would
appreciate a response from you by 15 December, 1986; if this presents a
problem could you please contact me.
Sincerely,
John McCarthy
-------
∂30-Oct-86 2203 cheriton@pescadero.stanford.edu Room switching
Received: from PESCADERO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Oct 86 22:03:21 PST
Received: by pescadero.stanford.edu with Sendmail; Thu, 30 Oct 86 21:45:26 pst
Date: Thu, 30 Oct 86 21:45:26 pst
From: David Cheriton <cheriton@pescadero.stanford.edu>
Subject: Room switching
To: jmc@su-ai
We had 39 people at the seminar today. I really dont see it feasible to
switch to a smaller room with this level of attendance.
Are there not other rooms available at this time?
David C.
∂30-Oct-86 2234 FULLERTON@Sierra.Stanford.EDU Interview
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Oct 86 22:33:12 PST
Date: Thu 30 Oct 86 22:08:49-PST
From: Dwain Fullerton <FULLERTON@Sierra.Stanford.EDU>
Subject: Interview
To: jmc@Sail.Stanford.EDU
cc: nilsson@Score.Stanford.EDU
Message-ID: <12251104838.20.FULLERTON@Sierra.Stanford.EDU>
Professor McCarthy,
A couple of days ago a representative from SRA, an
educational subsidiary of IBM, called and asked if I would try to
arrange for some interviews with Stanford faculty working in AI
and Expert Systems. SRA's purpose, as far as I can determine, is
to make a videotape featuring representatives from Stanford, MIT,
and Carnegie-Mellon talking about developments and directions in
these fields. The tape is intended for IBM internally plus
senior executives of other major companies. At present SRA plans
to be on campus Tuesday, November 4 and Wednesday, November 5.
Are you interested in talking with them?
Best regards,
Dwain Fullerton
-------
∂30-Oct-86 2236 UDI@Score.Stanford.EDU Re: another bug
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Oct 86 22:36:33 PST
Date: Thu 30 Oct 86 11:56:51-PST
From: Ehud Shapiro <UDI@Score.Stanford.EDU>
Subject: Re: another bug
To: JMC@Sail.Stanford.EDU
cc: leon%case.RELAY.CS.NET@Score.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 25 Oct 86 15:52:00-PDT
Message-ID: <12250993432.29.UDI@Score.Stanford.EDU>
+, -, *, and / are defined coorectly, but not ↑.
In any event, we need to makie explicit in the text that polynomial
are parsed correctly, due to the operator declarations.
Thanks for the comment.
Udi
-------
∂31-Oct-86 0127 CERF@A.ISI.EDU re: Digital Library Meeting
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 01:27:29 PST
Date: 31 Oct 1986 04:26-EST
Sender: CERF@A.ISI.EDU
Subject: re: Digital Library Meeting
From: CERF@A.ISI.EDU
To: JMC@SU-AI.ARPA
Message-ID: <[A.ISI.EDU]31-Oct-86 04:26:48.CERF>
In-Reply-To: The message of 30 Oct 86 1247 PST from John McCarthy <JMC@SAIL.STANFORD.EDU>
John,
thanks for your prompt reply! We will make reservations for you at the
Sheraton Reston for the evenings of Friday Nov 7 and Saturday Nov 8.
I would suggest a cab from the airport to the hotel (it is about 8 miles)
and we can pick you up at the hotel around 6:45 PM.
The meeting will be at our offices (NRI is short for Corporation for
national Research Initiatives) in Reston [actually, we share space
with AFIPS in their building].
Looking forward to seeing you again after a rather long hiatus.
Vint
∂31-Oct-86 0836 FULLERTON@Sierra.Stanford.EDU re: Interview
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 08:36:17 PST
Date: Fri 31 Oct 86 08:34:38-PST
From: Dwain Fullerton <FULLERTON@Sierra.Stanford.EDU>
Subject: re: Interview
To: JMC@Sail.Stanford.EDU
cc: FULLERTON@Sierra.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Thu 30 Oct 86 22:37:00-PST
Message-ID: <12251218762.15.FULLERTON@Sierra.Stanford.EDU>
As far as I can tell there's no money in it for us, and the only benefit would
be visibility before an industrial audience--the value of which is always
difficult to measure. I'll count you as unavailable unless you advise me to
the contrary.
Best,
Dwain
-------
∂31-Oct-86 0900 JMC
psa
∂31-Oct-86 0921 alliant!alliant.Alliant.COM!jat%mit-eddie.UUCP@harvard.harvard.edu.HARVARD.EDU visit
Received: from HARVARD.HARVARD.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 09:19:57 PST
Received: by harvard.harvard.edu.HARVARD.EDU; Fri, 31 Oct 86 12:19:46 EST
Received: by EDDIE (5.31/4.7) id AA27984; Fri, 31 Oct 86 10:59:12 EST
Received: by alliant.Alliant.COM (4.12/UUCP-Project/rel-1.0/10-29-86)
id AA09465; Fri, 31 Oct 86 10:54:45 est
Date: Fri, 31 Oct 86 10:54:45 est
From: mit-eddie!jat@alliant.Alliant.COM (Jack Test)
Message-Id: <8610311554.AA09465@alliant.Alliant.COM>
To: clt@sail.stanford.edu, jjw@sail.stanford.edu, jmc@sail.stanford.edu,
les@sail.stanford.edu
Subject: visit
I'm going to be visiting the Bay-Area November 18-19 and would like to
meet with you to exchange some information about what we've been doing
at Alliant and about what you've been doing at Stanford on your Alliant.
How about Tuesday afternoon (the 18th)? Please let me know what time
would be convenient for you.
-Jack
∂31-Oct-86 1018 NILSSON@Score.Stanford.EDU re: Binford Letter
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 10:18:50 PST
Date: Fri 31 Oct 86 10:11:12-PST
From: Nils Nilsson <NILSSON@Score.Stanford.EDU>
Subject: re: Binford Letter
To: JMC@Sail.Stanford.EDU
cc: NILSSON@Score.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Thu 30 Oct 86 22:30:00-PST
Message-ID: <12251236344.47.NILSSON@Score.Stanford.EDU>
Strike out any tonal vulgarities you don't like.
-------
∂31-Oct-86 1111 VAL Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
On Thursday, November 6, at 4pm we meet in MJH 301 (not 252, as usual), for
discussion of the last three talks and of reasoning about action in general.
Vladimir
∂31-Oct-86 1335 CERF@A.ISI.EDU Reservations
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 13:35:03 PST
Date: 31 Oct 1986 16:34-EST
Sender: CERF@A.ISI.EDU
Subject: Reservations
From: CERF@A.ISI.EDU
To: jmc@SU-AI.ARPA, reddy@A.CS.CMU.EDU
Message-ID: <[A.ISI.EDU]31-Oct-86 16:34:10.CERF>
Reservations at the Reston Sheraton have been made for Friday night
for Raj and for Friday and Saturday night for John. Specifics of
the Friday night guarantees will be forthcoming Monday.
Logistics for getting you from the hotel to NRI will be finalized
next week once we know who else we need to ferry from A to B.
Looking forward to seeing you both.
Vint
∂31-Oct-86 1435 POSER@CSLI.STANFORD.EDU Re: Is this stupid propaganda?
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 14:31:59 PST
Date: Fri 31 Oct 86 14:28:18-PST
From: Bill Poser <POSER@CSLI.STANFORD.EDU>
Subject: Re: Is this stupid propaganda?
To: JMC@SU-AI.ARPA
cc: su-etc@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 31 Oct 86 11:59:00-PST
What is the evidence that the hypothesis that KAL007 ws ordered
to fly over the Soviet Union by the US government was put forward by
the Soviet Union?
-------
∂31-Oct-86 1442 CLT levy
The paper seems rather ho hum (it has a good reference section).
It seems to me he is being a little premature in job hunting,
and I don't see much overlap in his research interests and
our current group interests.
∂31-Oct-86 1454 CLT tonight
I am going to need to go shopping after Timothy's supper,
so perhaps we should eat independently.
∂31-Oct-86 1530 RA Inference meeting
Richard Schroeppel called to remind you of the meeting Monday at 10:30.
∂31-Oct-86 1732 POSER@CSLI.STANFORD.EDU re: Is this stupid propaganda?
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 17:32:14 PST
Date: Fri 31 Oct 86 17:31:29-PST
From: Bill Poser <POSER@CSLI.STANFORD.EDU>
Subject: re: Is this stupid propaganda?
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 31 Oct 86 16:06:00-PST
I see. I certainly wouldn't put it past them, and I don't know
what really happened, but I am also quite capable of believing
that the US government would do such a thing. It does seem to me
that people can make conjectures like this without any need for
suggestions from the Soviet Union. Most of the Kennedy conspiracy
theories, for example, are probably not of Soviet origin.
-------
∂31-Oct-86 2026 GOLDBERG@CSLI.STANFORD.EDU re: Is this stupid propaganda?
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 20:26:32 PST
Date: Fri 31 Oct 86 20:07:27-PST
From: Jeffrey Goldberg <GOLDBERG@CSLI.STANFORD.EDU>
Subject: re: Is this stupid propaganda?
To: JMC@SU-AI.ARPA
cc: GOLDBERG@CSLI.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 31 Oct 86 18:04:00-PST
My apologies for bringing up something that was thoroughly gone over before.
I do not think that I changed the subject inappropriately, however. You
made a rather sweeping allegation (those who suspected US culpability in
the KAL007 matter were dupes).
Anyway, to address your question, I think that cartoon you describe is
disgusting. It is worse than stupid propaganda, it is insulting in a very
fundamental way. There are people who seem to like conspiracy theories and
will believe anything as long as it is evil enough. But people who are prone
to such paranoia are not always on the left. The country is full of people
like Gary Richard Arnold (Congressional candidate from the 16th district a
few years ago), the more extreme members of the John Birch Society,
etc. So, I think it is rather presumptuous of you to claim that it is
Moscow propaganda that is behind either the claim that the US is
partially responsible for the 007 tragedy (or even those who might
believe that the US military created AIDS). While you might disagree
with (or even think that some of these positions are crazy, I find it
frightening that you would associate particular opinions as coming
(even indirectly) from agents of Moscow.
You are probably right about military jets. It is extremely
provocative to shoot down a jet that is not actually attacking.
-Jeff Goldberg
Feel free to post this if you think it would be helpful to the
discussion.
-------
jmc - In neither case am I simply presuming that Moscow is behind it.
However, in the KAL007 case I don't remember the facts very well, except
that the Soviets started saying that the U.S. was responsible for the
South Korean flight just as soon as they began admitting that they shot it
down - a matter of two or three days. In the present case, Pravda cites
an unnamed Western newspaper, quoting unnamed Western scientists, but no
news story so far has named any. It occurs to me, however, that there
may be no deeper motivation than getting revenge for Western stories
about bacterial and chemical warfare in Laos and Afghanistan and the
story of several years ago about the anthrax outbreak in Sverdlovsk
being at a bacteriological warfare laboratory. The evidence on both
matters is quite strong but not overwhelming.
What about the stupidity question?
∂31-Oct-86 2139 ASHOK@Sushi.Stanford.EDU Irrational Punch
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 21:39:13 PST
Date: Fri 31 Oct 86 21:37:47-PST
From: Ashok Subramanian <ashok@Sushi.Stanford.EDU>
Subject: Irrational Punch
To: rwf@Sail.Stanford.EDU, jmc@Sail.Stanford.EDU, pchen@Sushi.Stanford.EDU,
ganGOLLI@Sushi.Stanford.EDU, jjw@Sail.Stanford.EDU,
bhayes@Cascade.Stanford.EDU
cc: ashok@Sushi.Stanford.EDU
Message-ID: <12251361331.9.ASHOK@Sushi.Stanford.EDU>
SLICK PROOFS AND THE POWER OF HANDWAVING
In class this Wednesday, I presented a "solution" to the Irrational
Punch problem. I am embarrassed that I did not spot the flaw in the
proof; I am also shocked at the ease with which I managed to pull the
wool over your eyes ....
Here is the original solution ----
------------------------------------------------------------------------------
Consider three points in a straight line ---
A ----------- B --------------------- C
1 sqrt(2)
The claim is that three punches centered at these points will wipe out
all points. For, if (x,y) were to survive, then (all coordinates relative to
point B) we have
x↑2 + y↑2 \in Q,
(x+1)↑2 + y↑2 \in Q, and
(x-sqrt(2))↑2 + y↑2 \in Q,
The first two equations imply that x is rational, and the first and
the third equations imply that x*sqrt(2) is rational. So taking
ratios, we get that sqrt(2) is irrational, which is false.
The same solution generalises to irrational punch in n-dimensional
Euclidean space, for any n.
------------------------------------------------------------------------------
COMMENT: The above solution is flawed, because there is no
contradiction if x=0. Our equations then become
y \in Q,
sqrt(y↑2 +1) \in Q, and
sqrt(y↑2 +2) \in Q,
and it is NOT obvious that there is no solution.
In fact, if we put y=p/q, p,q relatively prime integers, we get
sqrt(p↑2 + q↑2) \in Q, and
sqrt(p↑2 + 2*q↑2) \in Q.
In other words, we are trying to find three squares in arithmetic
progression, where the difference between terms is also a square. This
is a problem due to Fermat (see Dickson chapter XIV footnote 4), and I
do not know whether there is a solution. I suspect there is one;
however, I haven't toyed with the problem or the literature enough to
answer this question.
There are ways out of this difficulty, however. One is to try to see
what happens if we change the numbers 1 and sqrt(2) to something more
appropriate. We get other Diophantine problems, and we might try to
prove that one of these problems has no solutions. However, here are
two better approaches ---
a) Fermat showed that there cannot be FOUR squares in arithmetic
progression, using the method of descent. (See for example Theorem 3,
Chapter 4 of Mordell's book.) Hence if we choose four collinear
points as shown below, the original proof goes through.
A ----------------- B ---- C ------- D
AB = 1; AC = sqrt(2); AD = sqrt(3)
Notice that the generalization to n dimensions goes through as before;
it will always go through if all our punch-centers are collinear.
b) Solve the harder "Transcendental Punch" problem, i.e., "weaken" the
punch so that it knocks out only those points that are at a
transcendental distance away from the center. This allows us to
construct a very simple solution ---
A ----------- B --------------------- C
1 e
An inspection of the same equations as before shows that there cannot
be a point at an algebraic distance from all three, for if there were
such a point, e would be algebraic. Since the weaker punch succeeds in
eliminating every point, so will the stronger punch.
------------------------------------------------------------------------------
ANYBODY SEE ANY BUGS ?
--ashok
-------
∂31-Oct-86 2204 ASHOK@Sushi.Stanford.EDU Disjoint triangles in the plane
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Oct 86 22:04:23 PST
Date: Fri 31 Oct 86 22:03:01-PST
From: Ashok Subramanian <ashok@Sushi.Stanford.EDU>
Subject: Disjoint triangles in the plane
To: jmc@Sail.Stanford.EDU
cc: ashok@Sushi.Stanford.EDU
Message-ID: <12251365926.9.ASHOK@Sushi.Stanford.EDU>
>> You remarked that the solution given in Halmos's class for the
>> triangle problem was computationally inefficient. It seems to me that
>> it is very efficient, because it involves simply a change of
>> co-ordinates and then sorting the x-co-ordinates of the points.
Yes, but how does one figure out which coordinate system to change to?
I haven't thought much about this problem, but it seems hard to do
this in an efficient fashion, in the WORST CASE.
ashok
-------
∂01-Nov-86 0716 DEK
∂01-Nov-86 0132 JMC
Do you know if Ershov has arrived in the U.S.?
*** No; he was supposed to have arrived in D.C. last week, and I am
scheduled to call the arrangement people on Monday to learn about
next week. The plan is for him to stay at our house beginning Tuesday night,
as you probably know. I've just got back from France, so I haven't been
in touch with the people making arrangements.
∂01-Nov-86 1044 NILSSON@Score.Stanford.EDU AI Quals
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 1 Nov 86 10:43:55 PST
Date: Sat 1 Nov 86 10:41:31-PST
From: Nils Nilsson <NILSSON@Score.Stanford.EDU>
Subject: AI Quals
To: Feigenbaum@SUMEX-AIM.Stanford.EDU, Genesereth@SUMEX-AIM.Stanford.EDU,
WINOGRAD@CSLI.Stanford.EDU, JMC@Sail.Stanford.EDU,
Rosenbloom@SUMEX-AIM.Stanford.EDU, Buchanan@SUMEX-AIM.Stanford.EDU,
TOB@Sail.Stanford.EDU, Shortliffe@SUMEX-AIM.Stanford.EDU
cc: nilsson@Score.Stanford.EDU
Message-ID: <12251504007.15.NILSSON@Score.Stanford.EDU>
Victoria has sent a note around asking faculty when they would
be establishing the various qual dates. What do we AI-ers want to
do about the Quals this year (I'm not speaking as the chairman now,
just a fellow AI-er)? Victoria also reminded us that we are
being encouraged (by the new PhD committee) to have two quals a year.
How do we want to react to that? I'll volunteer to collect opinions.
Anyone want to volunteer to organize the AI quals this year?
(Tom, does the fact that there will be a robotics qual mean that you
will mainly be dealing with that instead of the AI qual. I presume we
still want vision to be a subject area for the AI quals and that you,
as our visionary, will still be on the AI qual panel.) -Nils
-------
∂01-Nov-86 1134 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu Lunch and more
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 1 Nov 86 11:34:15 PST
Received: by lindy.stanford.edu; Sat, 1 Nov 86 11:32:31 PST
Date: Sat, 1 Nov 86 11:34:01 PST
From: <ELLIOTT@slacvm.bitnet>
Reply-To: ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu
To: JMC@sail.stanford.edu
Subject: Lunch and more
Date: 1 November 86 11:07-PST
From: ELLIOTT@SLACVM
To: JMC@SAIL
Subject: Lunch and more
Date: 1 November 1986, 11:03:39 PST
From: Bloom, Elliott ELLIOTT at SLACVM
To: JMC at SAIL.STANFORD
Subject: Lunch and more
Thursday may be possible for me. I have to check my calender at SLAC. I will
finalize on Sunday. You promised me some information on judges to vote for
on Tuesday. Please don't forget.
∂01-Nov-86 1817 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu lunch
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 1 Nov 86 18:17:33 PST
Received: by lindy.stanford.edu; Sat, 1 Nov 86 18:15:48 PST
Date: Sat, 1 Nov 86 18:17:19 PST
From: <ELLIOTT@slacvm.bitnet>
Reply-To: ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu
To: JMC@sail.stanford.edu
Subject: lunch
Date: 1 November 86 18:16-PST
From: ELLIOTT@SLACVM
To: JMC@SAIL
Subject: lunch
Date: 1 November 1986, 18:14:56 PST
From: Bloom, Elliott ELLIOTT at SLACVM
To: JMC at SAIL.STANFORD
Subject: lunch
Unfortunately, I have meetings scheduled both on Tues and Thurs of
next week which don't allow an early lunch. Let's try for the week
after. Thanks for the info on the judges.
∂01-Nov-86 2330 CLT rivin
To: JMC, LES
here is his reply
∂01-Nov-86 2221 rivin@ALDERAAN.SCRC.Symbolics.COM qlisp&offer
Received: from [192.10.41.109] by SAIL.STANFORD.EDU with TCP; 1 Nov 86 22:21:44 PST
Received: from GANNET.SCRC.Symbolics.COM by ALDERAAN.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 15992; Fri 31-Oct-86 15:37:00 EST
Date: Fri, 31 Oct 86 15:37 EST
From: igor rivin <rivin@ALLEGHENY.SCRC.Symbolics.COM>
Subject: qlisp&offer
To: clt@SU-AI.ARPA
cc: rivin@ALLEGHENY.SCRC.Symbolics.COM
Message-ID: <861031153742.7.RIVIN@GANNET.SCRC.Symbolics.COM>
Hi again. I have mulled the issue over and over, pondering all the advantages of
the position you guys are offering, and finally came to the conclusion that the
external and internal state around here at this moment does not permit me to
honestly accept a two year commitment. The situation will very likely
crystallize within the next month, but whether or not you are willing to give me
an extension is, of course, up to you.
Whatever happens in this respect, I am very interested in the qlisp project and
would like to maintain some kind of link.
Igor.
∂01-Nov-86 2340 CLT
ron burback is the potential qlisper from colorado
∂02-Nov-86 1043 CLT
wil you take care of timothy this afternoon
∂02-Nov-86 1927 binford@su-whitney.arpa AI Quals
Received: from WHITNEY.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 2 Nov 86 19:26:56 PST
Received: by su-whitney.arpa with Sendmail; Sun, 2 Nov 86 19:29:44 pst
Date: Sun, 2 Nov 86 19:29:44 pst
From: Tom Binford <binford@su-whitney.ARPA>
To: NILSSON@Score.Stanford.EDU
Cc: Feigenbaum@SUMEX-AIM.Stanford.EDU, Genesereth@SUMEX-AIM.Stanford.EDU,
WINOGRAD@CSLI.Stanford.EDU, JMC@Sail.Stanford.EDU,
Rosenbloom@SUMEX-AIM.Stanford.EDU, Buchanan@SUMEX-AIM.Stanford.EDU,
TOB@Sail.Stanford.EDU, Shortliffe@SUMEX-AIM.Stanford.EDU,
nilsson@Score.Stanford.EDU
In-Reply-To: Nils Nilsson's message of Sat 1 Nov 86 10:41:31-PST
Subject: AI Quals
Nils
Although there will be a robotics qual, I will still participate on the
AI qual.
I prefer having a single qual. Is there a convincing need for a second?
Tom
∂03-Nov-86 1129 cramer@Sun.COM re: Paly International Festival
Received: from Sun.COM by SAIL.STANFORD.EDU with TCP; 3 Nov 86 11:29:36 PST
Received: from snail.sun.com by sun.com (3.2/SMI-3.0)
id AA00349; Mon, 3 Nov 86 11:28:40 PST
Received: from clem.sun.uucp by snail.sun.com (3.2/SMI-3.2)
id AA24978; Mon, 3 Nov 86 11:27:51 PST
Received: by clem.sun.uucp (3.2/SMI-3.2)
id AA00646; Mon, 3 Nov 86 11:28:53 PST
Date: Mon, 3 Nov 86 11:28:53 PST
From: cramer@Sun.COM (Sam Cramer)
Message-Id: <8611031928.AA00646@clem.sun.uucp>
To: jmc@sail.stanford.edu
Subject: re: Paly International Festival
Did you get the list of speakers that I sent you?
Do you know anything about any of them?
Thanks,
Sam -- cramer@sun.com
∂03-Nov-86 1219 coraki!pratt@Sun.COM Programming Languages
Received: from Sun.COM by SAIL.STANFORD.EDU with TCP; 3 Nov 86 12:16:52 PST
Received: from sun.uucp (sun.sun.com) by sun.com (3.2/SMI-3.0)
id AA00611; Mon, 3 Nov 86 12:14:52 PST
Received: by sun.uucp (1.1/SMI-3.2)
id AA09556; Mon, 3 Nov 86 12:12:56 PST
Received: by coraki.uucp (3.2/SMI-1.2)
id AA02475; Mon, 3 Nov 86 11:49:54 PST
Date: Mon, 3 Nov 86 11:49:54 PST
From: coraki!pratt@Sun.COM (Vaughan Pratt)
Message-Id: <8611031949.AA02475@coraki.uucp>
To: jmc@su-ai.arpa, zm@su-ai.arpa
Subject: Programming Languages
Cc: coraki!pratt@Sun.COM
The way the department is presently going about hiring people in the
area of programming languages is in my view not well thought out. I
consider us (JMC, ZM, VRP) the three faculty members in the department
who can fairly claim to represent the programming language research
community. Yet I know that I for one am not even being consulted on
some of the PL people being brought up for serious consideration.
An additional problem is that PL is falling between the cracks.
There is no committee that is adequately staffed to deal with PL
applicants; as a result we as a department are simply blowing away
the wheat along with the chaff in this area.
Do either of you share, at least in some degree, my perception of this
problem, and if so should we attempt to get organized to deal with it?
-v
∂03-Nov-86 1319 YM
do you have Vicotr Kuo mail address? (US Mail or EMAIL). Thanks. -Yoni
∂03-Nov-86 1401 WINOGRAD@CSLI.STANFORD.EDU Quals
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 3 Nov 86 13:56:55 PST
Date: Mon 3 Nov 86 13:55:02-PST
From: Terry Winograd <WINOGRAD@CSLI.STANFORD.EDU>
Subject: Quals
To: ai-faculty: ;
The reason for having two quals is for the benefit of the students.
First, it gives them something finer-grained for fitting into their
overall schedule, and second it potentially gives them more opportunities
to pass it, given that the limit is in years, not tries. Of copurse,
neither of these is an absolute, but if we can do it without a lot
of difficulties, I think we should be accommodating.
Since it is an oral, scheduling it more than once isn't that big a deal
(not like making up two separate written exams). The issue of
comparability is a relative one -- the fact that having a single qual
give us more cross-comparative data misses the fact that it changes
from year to year anyway. Also, my experience is that our decisions
aren't terribly swayed by comparisons, but that we really do have more
stable evaluation functions.
Also, having it twice a year might make it easier to get more faculty
focus (I'm more likely to say yes to a day twice than two a two-day
session once). Overall it seems like there is very little additional
cost.
p.s., this is spoken in my role as PhD committee chair, NOT as
a volunteer to run the qual.
--t
-------
∂03-Nov-86 1353 CLT calendar item
To: JMC, LES, JJW, AIR, RPG, ARG
Tuesday Nov 18 at 2:45pm Jack Test will be here to talk about what they
have been doing at Alliant and to find out what we are doing.
I will arrange a room and send a reminder.
∂03-Nov-86 1512 G.BEESON@CSLI.STANFORD.EDU regrets
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 3 Nov 86 15:12:16 PST
Date: Mon 3 Nov 86 15:10:30-PST
From: Michael Beeson <G.BEESON@CSLI.STANFORD.EDU>
Subject: regrets
To: jmc@SU-AI.ARPA
I regret missing your talk at San Jose recently. I had planned not only
to attend it but to go out to dinner with you. However, in the event I
was too busy with last-minute preparations for the trip to Italy which
I left on two days later. Too busy, that is, even to
REMEMBER your lecture--not too busy to attend. Second-hand reports
are that it was a success, pitched just right for the audience.
-------
∂03-Nov-86 1515 GLB homework
I have prepared the suggested homework for EKL and gave a copy to Hsu for comments.
∂03-Nov-86 1616 RA leaving at 4:30
To: ZM, JMC
I need to leave at 4:30 today.
∂04-Nov-86 0640 AI.BOYER@MCC.COM visit
Received: from MCC.COM by SAIL.STANFORD.EDU with TCP; 4 Nov 86 06:40:00 PST
Date: Tue, 4 Nov 1986 08:38 CST
Message-ID: <AI.BOYER.12252246151.BABYL@MCC.COM>
From: AI.BOYER@MCC.COM
To: jmc@su-ai
Subject: visit
The 24th and 25th of November would be fine days for you to
consult here. Look forward to your visit.
∂04-Nov-86 0801 ZM Re: Programming Languages
To: JMC@SAIL.STANFORD.EDU, coraki!pratt@Sun.COM
[Reply to message recvd: 03 Nov 86 22:48 Pacific Time]
A late lunch please (1pm) Zohar
∂04-Nov-86 1107 AIR EBOS
Do you need IBM letter you gave me few days ago?
Should I keep it?
Arkady
∂04-Nov-86 1144 RA vacation
To: JMC, ZM
I will go on a short vacation begin. Fri. 11/24 and will be back after
Thanksgiving. Total 4 workdays.
∂04-Nov-86 1155 JUTTA@Score.Stanford.EDU advisor signature
Received: from [36.36.0.195] by SAIL.STANFORD.EDU with TCP; 4 Nov 86 11:55:15 PST
Date: Tue 4 Nov 86 11:52:04-PST
From: Jutta McCormick <JUTTA@Score.Stanford.EDU>
Subject: advisor signature
To: ZM@Sail.Stanford.EDU, jmc@Sail.Stanford.EDU, ejm@Shasta.Stanford.EDU,
oliger@Navajo.Stanford.EDU, wiederhold@SUMEX-AIM.Stanford.EDU
cc: jutta@Score.Stanford.EDU
Stanford-Phone: (415) 723-0572
Message-ID: <12252303281.28.JUTTA@Score.Stanford.EDU>
In the past I have signed for MS students' advisors on Applications for
Candidacy if they followed advisor-approved checksheets and on pre-approved
program changes. The procedure has been for students to submit these forms
to me for double-checking and forwarding to the advisor if there were
questions. Please let me know if you would like me to continue this
practice or how you would like your advisees' documents handled.
--Jutta McCormick
--------
-------
∂04-Nov-86 1403 GLB homework
The point of the logic exercises was to make sure that students know what
a formal proof is. Is it your view program verification should be done as
much as possible in intuitive reasoning, so that it is better to leave
entirely aside the notion of formal proof (or rather, leave it to EKL)?
I've made another draft of the assignment, consisting entirely of EKL proofs
and left it on your desk.
∂04-Nov-86 1417 RA correction on vacation dates
To: ZM, JMC
The date for my vacation was wrong (looked at the wrong month), my vacation
will start Fri. 11/21. I will be back after Thanksgiving.
∂04-Nov-86 1423 RA Trip to Virginia
Dottie Smith from Dr. Cerf office called to let you know that she made
reservations for you for Fri. and Sat. at the Sheraton International in
Reston VA. The confirmation no. is 65568701169. Fri. dinner is at 7:00pm
Dr. Cerf will probably pick you up. They will contact you when you get there.
Have you decided which flight you are going to take on your way back?
Dottie's tel. is (703) 620 8990.
∂04-Nov-86 1515 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 4 November 1986
Previous Balance 5.15
Payment(s) 5.15 (check 10/21/86)
-------
Current Charges 4.00 (bicycle lockers)
0.60 (vending machine)
-------
TOTAL AMOUNT DUE 4.60
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 1.0% per month will be
charged on balances remaining unpaid 25 days after bill date above.
An account with a credit balance earns interest of .33% per month,
based on the average daily balance.
∂04-Nov-86 1526 CERF@A.ISI.EDU Confirmation Numbers for DC trip
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 4 Nov 86 15:15:03 PST
Date: 4 Nov 1986 17:11-EST
Sender: CERF@A.ISI.EDU
Subject: Confirmation Numbers for DC trip
From: CERF@A.ISI.EDU
To: jmc@SU-AI.ARPA
Message-ID: <[A.ISI.EDU] 4-Nov-86 17:11:00.CERF>
John,
Your secretary should be getting this information from our Dottie Smith,
but just to make sure you have it, the confirmation number for guarantee
of your Friday night stay at the Sheraton Reston is 65568701169.
Our plan is to pick you up from the Sheraton just a bit before 7 PM on
Friday night to bring you to our NRI offices nearby where you will be
fed in the company of others working on the digital library. Raj Reddy
will be staying at the same hotel and most likely we will pick him up
at the same time.
Looking forward to seeing you soon.
Vint
∂04-Nov-86 1629 CLT Qlisp meeting reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
Time: Friday, Nov 7 12:15-1:15
Place: MJ352
Topic: Qlisp programming
bring programs and programming problems to discuss
∂04-Nov-86 1741 cramer@Sun.COM re: Paly International Festival
Received: from Sun.COM by SAIL.STANFORD.EDU with TCP; 4 Nov 86 17:41:07 PST
Received: from snail.sun.com by sun.com (3.2/SMI-3.0)
id AA00281; Tue, 4 Nov 86 17:33:00 PST
Received: from clem.sun.uucp by snail.sun.com (3.2/SMI-3.2)
id AA01101; Tue, 4 Nov 86 17:32:10 PST
Received: by clem.sun.uucp (3.2/SMI-3.2)
id AA02228; Tue, 4 Nov 86 17:33:10 PST
Date: Tue, 4 Nov 86 17:33:10 PST
From: cramer@Sun.COM (Sam Cramer)
Message-Id: <8611050133.AA02228@clem.sun.uucp>
To: JMC@SAIL.STANFORD.EDU
Subject: re: Paly International Festival
I think that Beyond War is linked to the Werner Erhard est scam. Personal
growth through scaring the wits out of people.
Sam
∂04-Nov-86 1849 vijay@ernie.Berkeley.EDU Statements...
Received: from ERNIE.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 4 Nov 86 18:48:02 PST
Received: by ernie.Berkeley.EDU (5.53/1.16)
id AA20543; Tue, 4 Nov 86 18:48:56 PST
Date: Tue, 4 Nov 86 18:48:56 PST
From: vijay@ernie.Berkeley.EDU (Vijay Ramamoorthy)
Message-Id: <8611050248.AA20543@ernie.Berkeley.EDU>
To: jmc@sail.stanford.edu
Subject: Statements...
Hi,
Have you received Searle's, Hofstadter's and Dreyfus's latest
statements? I have received "returned mail" from your electronic
address.
-VR
∂04-Nov-86 2133 CLT friday's meeting
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
It has been pointed out to me that Friday's
Siglunch features Ken Kahn talking about
Vulcan and Object-Oriented system for Concurrent Prolog.
If there is sufficient interest we can postpone the
qlisp programming so people can go to Ken's talk.
Assume the Qlisp meeting is on unless you hear otherwise.
(Which will only happen if several people tell me they
want to go to Ken's talk.)
∂04-Nov-86 2359 coraki!pratt@Sun.COM lunch
Received: from Sun.COM by SAIL.STANFORD.EDU with TCP; 4 Nov 86 23:58:57 PST
Received: from sun.uucp (sun.sun.com) by sun.com (3.2/SMI-3.0)
id AA00442; Tue, 4 Nov 86 23:36:38 PST
Received: by sun.uucp (1.1/SMI-3.2)
id AA29557; Tue, 4 Nov 86 22:28:15 PST
Received: by coraki.uucp (3.2/SMI-1.2)
id AA04632; Tue, 4 Nov 86 21:29:08 PST
Date: Tue, 4 Nov 86 21:29:08 PST
From: coraki!pratt@Sun.COM (Vaughan Pratt)
Message-Id: <8611050529.AA04632@coraki.uucp>
To: jmc@su-ai.arpa, zm@su-ai.arpa
Subject: lunch
I'm teaching at 1 on Tu&Th and consulting Fri, leaving Mon & Wed, both
of which are presently free.
-v
∂05-Nov-86 0602 DEK Ershov
I'm free on Thursday between 2 and 3 in my office here (and I'm tied
up from 3 to 5). Dinner Thursday will be splendid.
Next Monday looks bad, as I'm busy from 8am till midnight and flying
to NY Tues morning! Therefore, let us try to focus on Thursday.
∂05-Nov-86 0900 UDI@Score.Stanford.EDU re: game of life
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 5 Nov 86 09:00:48 PST
Date: Wed 5 Nov 86 08:59:29-PST
From: Ehud Shapiro <UDI@Score.Stanford.EDU>
Subject: re: game of life
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Tue 4 Nov 86 21:04:00-PST
Message-ID: <12252534006.38.UDI@Score.Stanford.EDU>
Thanks, Udi
-------
∂05-Nov-86 1112 RA Maddalana
I reserved the private room for you for tomorrow at 7:00. It will seat 8. I told
her that we have 6 and will call her back if we have more. It is reserved under
your name.
∂05-Nov-86 1339 DEK right
namely, I get Ershov tomorrow from 2 til 3;
Jill and I join you and the others for dinner in the evening.
∂05-Nov-86 1742 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
On Thursday, November 6, at 4pm we meet in MJH 301 (not 252, as usual), for
discussion of the last three talks and of reasoning about action in general.
Vladimir
∂05-Nov-86 1750 VAL Ershov
It looks like I won't be able to pick him up after lunch. But we'll try to
join you for dinner.
∂05-Nov-86 1940 LES DARPA budget
Here is a budget to go with the DARPA proposal. -Les
Proposal to DARPA
for
Research in Formal Reasoning
Budget for Three Years beginning 1 February 1986
Personnel Annual Cost
Prof. John McCarthy (25% acad. yr., 50% Sum.) 29,085
Yoav Shoham, Asst. Professor (50% acad. yr., 100% Sum.) 33,330
Vladimir Lifschitz, Sr. Research Assoc. (100%) 60,504
Carolyn Talcott, Research Assoc. (75%) 36,000
N. Shankar, Research Assoc. (100%) 45,996
Lester Earnest, Sr. Research Assoc. (20%) 15,593
--------, Research Assoc. (100%) 48,000
--------, Student Res. Assist. (50% acad. yr., 100% Sum.) 14,325
--------, Student Res. Assist. (50% acad. yr., 100% Sum.) 14,325
--------, Student Res. Assist. (50% acad. yr., 100% Sum.) 14,325
--------, Student Res. Assist. (50% acad. yr., 100% Sum.) 14,325
--------, Student Res. Assist. (50% acad. yr., 100% Sum.) 14,325
Rutie Adler, Secretary (50%) 11,958
---------
Annual salary subtotal 352,091
Annual budget begin dates 2/1/87- 2/1/88- 2/1/89-
End dates 1/31/88 1/31/89 1/31/90
Annual salary subtotal 352,091 352,091 352,091
Allowance for salary increases 8,802 29,928 51,053
(6% beginning 9/1/87,
12% beginning 9/1/88,
18% beginning 9/1/89)
--------- --------- ---------
Salary total by year 360,893 382,019 403,144
Staff benefits (24.7% till 9/1/87, 91,095 100,439 108,647
26.0% till 9/1/88, 26.7% till 9/1/89,
27.3% thereafter)
Travel (4 East Coast trips/year 8,000 8,000 8,000
@ $1000, 8 Western trips/yr.
@ $500)
Computer time costs 85,000 90,000 95,000
Other direct costs (publications, 25,000 25,000 25,000
supplies, telephones and
other services)
--------- --------- ---------
Subtotal 569,988 605,458 639,791
Indirect Costs (73%) 416,091 441,984 467,047
--------- --------- ---------
Total by year 986,079 1,047,442 1,106,838
--------- --------- ---------
Cumulative totals 986,079 2,033,521 3,140,359
∂06-Nov-86 1155 vijay@ernie.Berkeley.EDU AI DISC: HOFSTADTER REPLY TO SEARLE
Received: from ERNIE.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 6 Nov 86 11:55:34 PST
Received: by ernie.Berkeley.EDU (5.53/1.16)
id AA12211; Thu, 6 Nov 86 11:44:01 PST
Date: Thu, 6 Nov 86 11:44:01 PST
From: vijay@ernie.Berkeley.EDU (Vijay Ramamoorthy)
Message-Id: <8611061944.AA12211@ernie.Berkeley.EDU>
To: jmc@sail.stanford.edu
Subject: AI DISC: HOFSTADTER REPLY TO SEARLE
Searlian Circularity Revealed!
A Final Commentary on Searlianism
Douglas R. Hofstadter
Since the argument with Searle is getting rather tiresome, I think that
I will do something that even he would approve of: I will simply quote him at
length, thereby letting him have the last word, and simultaneously satisfying
his desire to be quoted accurately. I shall cite from his original paper,
"Minds, Brains, and Programs", in the section entitled "The combination reply".
We start in the middle of his second paragraph, jump to the end of the third
paragraph, then to the middle of the fourth paragraph, and finish up quoting
the beginning and the end of the fifth paragraph of that section. At that
point, I quit, letting Searle's curiously circular position stand up, to the
extent that it can, for itself.
.... If we could build a robot whose behavior was
indistinguishable over a large range from human behavior,
we would attribute intentionality to it, pending some reason
not to. ....
.... But.... If we knew independently how to account for
its behavior without such assumptions, we would not attribute
intentionality to it, especially if we knew it had a formal
program. ....
.... Suppose we knew that the robot's behavior was entirely
accounted for by the fact that a man inside it was receiving
uninterpreted formal symbols from the robot's sensory receptors
and sending out uninterpreted formal symbols to its motor
mechanisms, and the man was doing this symbol manipulation
in accordance with a bunch of rules. Furthermore, suppose
the man knows none of these facts about the robot, all he
knows is which operations to perform on which meaningless
symbols. In such a case we would regard the robot as an
ingenious mechanical dummy. The hypothesis that the robot
has a mind would now be unwarranted and unnecessary....
.... [W]e find it completely natural to ascribe intentionality
to members of certain other primate species such as apes and
monkeys and to domestic animals such as dogs. .... We would
certainly make similar assumptions about the robot unless we
had some reason not to, but as soon as we knew that the behavior
was the result of a formal program, and that the actual causal
properties of the physical substance were irrelevant we would
abandon the assumption of intentionality.
∂06-Nov-86 1230 CLT igor
Here is my message to him decorated with his reply.
We should talk about it, and I will try agin to call him.
∂06-Nov-86 0931 rivin@YUKON.SCRC.Symbolics.COM qlisp&offer
Received: from SCRC-YUKON.ARPA by SAIL.STANFORD.EDU with TCP; 6 Nov 86 09:31:30 PST
Received: from GANNET.SCRC.Symbolics.COM by YUKON.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 114168; Wed 5-Nov-86 17:44:15 EST
Date: Wed, 5 Nov 86 17:44 EST
From: igor rivin <rivin@ALLEGHENY.SCRC.Symbolics.COM>
Subject: qlisp&offer
To: CLT@SAIL.STANFORD.EDU, rivin@PEGASUS.SCRC.Symbolics.COM,
rivin@ALLEGHENY.SCRC.Symbolics.COM
In-Reply-To: The message of 5 Nov 86 00:42 EST from Carolyn Talcott <CLT@SAIL.STANFORD.EDU>
Message-ID: <861105174422.3.RIVIN@GANNET.SCRC.Symbolics.COM>
Received: from STONY-BROOK.SCRC.Symbolics.COM by ALLEGHENY.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 9139; Wed 5-Nov-86 01:45:37 EST
Received: from PEGASUS.SCRC.Symbolics.COM by STONY-BROOK.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 150554; Wed 5-Nov-86 01:45:43 EST
Received: from SAIL.STANFORD.EDU (su-ai.arpa.ARPA) by scrc-pegasus id AA17979; Wed, 5 Nov 86 01:42:30 est
Date: 04 Nov 86 2142 PST
From: Carolyn Talcott <CLT@SAIL.STANFORD.EDU>
Subject: qlisp&offer
To: rivin@SCRC-PEGASUS.ARPA
Hi,
I've been trying to call you but without success, so I'll send a msg.
I have been out of town the last couple of days... Just returned last night...
I gather one thing you are waiting for the is the long term plan
for MACSYMA at Symbolics?
That is true to a certain extent.
Would you be more comfortable with a one year commitment.
We could easily live with anything between one and two years.
That would be much better. I would like having the option.
We can give you a while longer to decide.
Great!
If there is some alternative you would like to suggest
What kind of thing do you have in mind? I am not really familiar with
the environment enough to suggest intelligent things, I am afraid...
One question that does leap to mind, though, is what are the prospects
of this mutating into a faculty position at some point?
An assurance of such a position would override all the considerations that are
making me hesitate at the moment.
, we
are flexible within the bounds of getting the job done
in a reasonable length of time.
I certainly understand.
What is a good time to call you (at work or at home)?
I am around (at Symbolics) from 1pm to about 5pm and then from 6pm to ~2-3am.
The second block is intermittent, though -- I can be reached with a probability
of about 50% at any time between 6pm and 3am. (All the times above are EST).
I am looking forward to hearing from you.
Igor.
∂06-Nov-86 1430 RA proposal
Les thought we should have your bio in the proposal too. I will also add his
and Carolyn's. Is this ok with you?
Do you know whether Carolyn is coming back today?
thanks
∂06-Nov-86 1608 RA trip back
Frank, Dina Bolla, called to let you know that he could not get you a
window seat for the way back, only an aisle seat. He said you should ask
for one when you get to the airport, they may have one then.
∂06-Nov-86 1617 RA leaving
It's Thursday and I'm leaving early.
Have a nice trip.
∂06-Nov-86 1626 HSU@Sushi.Stanford.EDU discussion
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Nov 86 16:26:35 PST
Date: Thu 6 Nov 86 04:32:04-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: discussion
To: jmc@Sail.Stanford.EDU
cc: hsu@Sushi.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12252747471.13.HSU@Sushi.Stanford.EDU>
Professor McCarthy:
Is there anything needs to be done before the midterm session today?
I could meet you in your office if you want.
Also, I would like to talk to you about my thesis work. Could I discuss
with you while we wait for the exam?
Thank you.
Yung-jen
-------
∂06-Nov-86 2021 HSU@Sushi.Stanford.EDU re: discussion
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Nov 86 20:21:46 PST
Date: Thu 6 Nov 86 20:20:06-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: re: discussion
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Thu 6 Nov 86 17:03:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12252920053.25.HSU@Sushi.Stanford.EDU>
When will be appropriate for you? I'll be available most of the mornings.
I'm looking forward to talking to you, thanks.
Yung-jen
-------
∂06-Nov-86 2204 HSU@Sushi.Stanford.EDU re: discussion
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 6 Nov 86 22:04:49 PST
Date: Thu 6 Nov 86 22:02:45-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: re: discussion
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Thu 6 Nov 86 21:59:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12252938742.25.HSU@Sushi.Stanford.EDU>
Tuesday (11/11) 11am sounds fine with me. I'll meet you in your office
then. Thanks for your prompt response.
Yung-jen
-------
∂07-Nov-86 0026 CERF@A.ISI.EDU Meeting This weekend
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 7 Nov 86 00:26:21 PST
Date: 7 Nov 1986 03:25-EST
Sender: CERF@A.ISI.EDU
Subject: Meeting This weekend
From: CERF@A.ISI.EDU
To: jmc@SU-AI.ARPA
Message-ID: <[A.ISI.EDU] 7-Nov-86 03:25:34.CERF>
Good Morning, John.
Here's how the meeting seems to be shaping up. Bob Kahn, Patrice and
I will participate along with you and Raj Reddy as well as John Bowles
who is acting as a business consultant for NRI. There is a chance that
Butler Lampson will make it, but we haven't got confirmation.
There are a number of other library working group members, including
Jerry Popek, Jaime Carbonnell, Bob Sproull, Danny Hillis (Connection
Machine designer), Dave Clark (MIT), who were not able to make it to
this November session but will join us in January (5-6).
Raj is probably going to fly back to Pittsburgh on Saturday night;
you may want to alter your plans and fly back on a 5:30 PM flight
from Dulles on Saturday or take the morning flight Sunday. Of course,
if we seem to have gotten into some deep territory late Saturday,
all of us would be delighted to continue the discussion Sunday, but
I thought you should be aware of options open.
Although the agenda is quite flexible, I suspect we ought to review
the results of the August meeting on Friday night in preparation for
Saturday. If you have particular topics you want to explore, please
don't hesitate to bring them up (I know you will!).
See you soon.
Vint
∂07-Nov-86 0600 JMC
Susie re trip
∂07-Nov-86 0936 JJW EKL on Sushi
To: GLB@SAIL.STANFORD.EDU, hsu@SUSHI.STANFORD.EDU
CC: JMC@SAIL.STANFORD.EDU
I've updated <EKL> on Sushi so that students can use it there.
Also, John Reuling set up a directory group so that Jane and I
can both change any files on the <EKL> directory on Sushi.
Joe
∂08-Nov-86 1355 CLT calendar item
mon 10-nov 18:30 Dinner chez Hayes for Ershov
∂08-Nov-86 1750 SJG Frame problem workshop
To: "@NMI.DIS[1,SJG]"@SAIL.STANFORD.EDU
Hi!
You remember me: I was organizing a non-monotonic workshop a while
back. Well, I stopped when I saw that a workshop on logical solutions
to the frame problem was being arranged for the spring, and it seemed
likely that that workshop would serve the same purpose.
I now find -- to my dismay -- that said workshop is going to be held
in KANSAS. Do you guys feel about Kansas the way I do? Is it worth
leaning on the organizer to move it?
Anyway, if you'd like it moved (like I would), please send me a message
to that effect. I'll collect them all and forward them to Frank Brown.
Thanks.
Matt
∂08-Nov-86 1836 JJW Qlisp questions (long message)
To: QLISP@SAIL.STANFORD.EDU
At our meeting this past Friday we decided to look at the definition of
Qlisp to see if we should add or change anything, and to come up with a
more well-defined semantics for the language.
I'd like to begin with a summary of Qlisp as I understand it, along with
some questions that come up or points that aren't clear to me. Let's
start with a question:
Question 1: Is Qlisp based on Common Lisp, or Scheme, or neither (i.e.,
whatever we want it to be)? If Common Lisp, do we strictly adhere to
the separate value and function namespaces? (The Qlisp paper by JMC and
RPG is ambiguous in this regard.)
Qlisp adds the following primitives to the base language: QLET, QLAMBDA,
QCATCH, SUSPEND-PROCESS and RESUME-PROCESS; and changes the definitions
of CATCH, THROW and UNWIND-PROTECT. It also implicitly changes several
features of the base language. The Qlisp paper suggests that other new
primitive forms may be defined if there is a need for them.
QLET is like ordinary LET, except that it can spawn processes to compute
values in parallel. The general form (QLET pred (bindings) body) first
evaluates PRED, and then there are three cases:
(a) The value of PRED is NIL. QLET then behaves like LET, doing the
bindings sequentially and executing the body. (The base language
determines whether the bindings must be evaluated left-to-right, or
in an arbitrary order. Any use of QLET with a predicate that might
not be NIL would, of course, be stupid to depend on such an order.)
(b) The value of PRED is neither NIL nor EAGER. QLET spawns a new
process for each of the bindings and waits for all of these
processes to finish. (An implementation could save one process
creation by letting the parent process handle the last of the
bindings.) It then proceeds to evaluate the body.
(c) The value of PRED is EAGER. QLET spawns a new process for each of
the bindings, and proceeds to evaluate the body with the variables
bound to placeholders for the values being computed by these
processes. (These are "futures" in Multilisp terminology, and this
term will be used from here on.) The evaluation may proceed until
the body requires one of the values, or the body may terminate
without requiring any of the values.
Examples of forms that require values: predicates, conditionals,
arithmetic, accessor functions (car, cdr, etc.). Examples of forms
that don't require values: function application, variable binding,
assignment, constructors (cons, etc.).
Question 2: Can we be more exact about this? It seems that for every
primitive operation in the base language we will have to state which
arguments are required and which are not.
If a value is required, then if the process computing it has
finished and stored the value in the future, the evaluation of the
body can proceed without delay. If the process has not completed,
the process computing the body must wait until it completes. It may
then proceed, possibly waiting again if it needs another value that
has not been determined.
The value of the last form in the body is returned as the value of
the QLET form. Processes created by the QLET may still be running
if their values were not required to evaluate the body. Their
values may have been ignored entirely, or may have been used in
forms that don't require values. For example, the QLET could return
a cons-cell containing futures for values still being determined, or
in the simplest case could return a future itself [e.g.,
(qlet 'eager ((v (foo))) v)]. Therefore we continue running such
processes even after exiting the QLET. It is then possible that any
process may encounter an undetermined future and have to wait, and
that more than one process may be waiting on the same future.
QLAMBDA introduces a new object called a "process closure". A process
closure is similar to an ordinary closure in that it returns a value
when it is applied to arguments, and maintains an environment, but
differs in that only one application of a process closure is ever
active. Thus it provides a primitive with which one can define critical
regions and synchronization.
The general form (QLAMBDA pred (variables) body) evaluates PRED, and has
three cases:
(a) The value of PRED is NIL. QLAMBDA behaves like LAMBDA, returning an
ordinary closure.
(b) The value of PRED is neither NIL nor EAGER. QLAMBDA spawns a new
process, and returns a process closure. The process closure is a
Lisp data structure similar to an ordinary closure, but it contains
a pointer to the new process, which we will call a "QLAMBDA
process". We will call any process that applies a process closure
to a set of arguments a "calling process". The QLAMBDA process
waits for a message. When the calling process applies the process
closure, it sends the arguments to the QLAMBDA process. The QLAMBDA
process computes the body of the QLAMBDA form with the arguments
bound to the variables, and returns the resulting value. The
calling process waits for the value to be returned unless it is
ignoring the return value. Applications of the process closure are
synchronized, so that no two are in progress at the same time.
Question 3: When is a value considered to be ignored? The only example
given in the Qlisp paper is that a top-level form in an explicit or
implicit PROGN (except the last top-level form, which is the value of
the PROGN), ignores its value.
Question 4: If a process closure is applied in a value-ignoring
position, but the QLAMBDA process is busy, does the calling process wait
until its message has been read by the QLAMBDA process, or does it just
enqueue the message and continue? Remarks in the Qlisp paper tend to
imply the latter.
(c) The value of PRED is EAGER. A process closure and an QLAMBDA
process are created just as above, but the QLAMBDA process
immediately starts computing the body of the QLAMBDA form. To do
this, the variables are bound to futures. When the process closure
is applied, the values of the futures are set to the arguments and
the QLAMBDA process may proceed if it had to wait on one of the
futures. If the QLAMBDA process happens to compute its value
without waiting on any of the futures, it will wait at that point
until it is called. In either case, once it returns a value to its
caller, it again begins to evaluate the QLAMBDA body with a new set
of futures bound to the variables.
CATCH, QCATCH and THROW provide non-local exits and allow the killing of
processes. The value returned by (CATCH tag form) or (QCATCH tag form)
is the value of FORM, if no THROW is executed before FORM returns a
value. CATCH kills any processes still executing when FORM returns,
while QCATCH waits for them to finish. If (THROW tag value) is
executed, we search upwards through the stack of function calls and the
tree of process creations until we find a CATCH or QCATCH with a
matching tag. We then kill all processes created as a result of the
CATCH or QCATCH and return the value in the THROW as the value of the
CATCH or QCATCH.
Question 5: There is a third possibility: CATCH could return a value
normally (no THROW was executed), but some processes could still be
computing part of that value, as a result of eager QLETs. We will want
these processes to continue, but might not want to wait for their
values. Shouldn't there be a way to exit the CATCH but leave processes
running? I would argue that ordinary return from a CATCH shouldn't kill
any processes; that THROW should be the only way to kill something.
Question 6: What does it mean to kill a QLAMBDA process? Since the
process closure still exists and may be applied after we leave the CATCH
or QCATCH, we don't want to prevent the QLAMBDA process from ever
running again. But if, at the time we kill processes, a QLAMBDA process
is working on computing some value that is no longer needed, we will
want to stop it and unwind it back to the point where it is ready to
accept a new set of arguments.
Question 7: If a THROW occurs inside a QLAMBDA process, and there is no
corresponding CATCH in that process, what is the "parent" process that
we should examine to continue searching for the tag? Is it the process
that created the process closure? That one might no longer exist. Is
it the process currently calling the QLAMBDA process? That one might be
gone too, if the call was in a value-ignoring position.
(to be continued ...)
If anyone has read this far, I'd be glad to hear your opinions.
∂08-Nov-86 2127 kuo@eniac.seas.upenn.edu Happy Birthday to Timothy!
Received: from LINC.CIS.UPENN.EDU by SAIL.STANFORD.EDU with TCP; 8 Nov 86 21:26:58 PST
Received: by linc.cis.upenn.edu
id AA14583; Sun, 9 Nov 86 00:26:16 EST
Date: Sun, 9 Nov 86 00:25:27 est
From: kuo@eniac.seas.upenn.edu (Victor Kuo)
Posted-Date: Sun, 9 Nov 86 00:25:27 est
Message-Id: <8611090525.AA19248@eniac.seas.upenn.edu>
To: CLT@su-ai.arpa, JMC@su-ai.arpa
Subject: Happy Birthday to Timothy!
On the occasion of Timothy's one year's birthday, the one traditionally
the Chinese celebrate the most, please accept our sincere congratulations
to you, the parents. My family and I will share your joys with you today
since today is also my birthday. -- Victor
∂08-Nov-86 2315 BAUDINET@Sushi.Stanford.EDU re: death valley: question
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 8 Nov 86 23:14:50 PST
Date: Sat 8 Nov 86 23:12:15-PST
From: Marianne Baudinet <BAUDINET@Sushi.Stanford.EDU>
Subject: re: death valley: question
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 8 Nov 86 22:05:00-PST
Message-ID: <12253475682.7.BAUDINET@Sushi.Stanford.EDU>
Thanks for the information,
Marianne
-------
∂09-Nov-86 0919 VAL
Please send me the title and abstract of your talk on contexts.
∂09-Nov-86 1234 vijay@ernie.Berkeley.EDU AI DISC: DEADLINE
Received: from ERNIE.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 9 Nov 86 12:34:47 PST
Received: by ernie.Berkeley.EDU (5.57/1.16)
id AA02557; Sun, 9 Nov 86 12:34:38 PST
Date: Sun, 9 Nov 86 12:34:38 PST
From: vijay@ernie.Berkeley.EDU (Vijay Ramamoorthy)
Message-Id: <8611092034.AA02557@ernie.Berkeley.EDU>
To: jmc@sail.stanford.edu
Subject: AI DISC: DEADLINE
Hi,
Well, we have a tremendous amount of information. Very soon
we'll be sending out final copies of all correspondences.
We are announcing now that the end of this "discussion" will
be on Nov. 30. Anything that comes in after Nov. 20, however,
and "implicates" any of the participants, the "implicated"
parties will have a chance to respond - even if it runs after
the deadline...
Soon after that, we will have a more compact, edited version
ready to be published. More about this "edited" version later.
-VR RA JS DB
∂09-Nov-86 1633 GLB
(wipe-out)
(get-proofs lispax prf ekl sys)
(get-proofs natnum prf ekl sys)
(get-proofs normal prf ekl sys)
(proof proddd)
;we need to delete the definition of append, otherwise * would be ambiguous
(deletel lispax#46 lispax#47) y
(axiom |∀n.sexp n|)
(label simpinfo)
(define prodd |∀u x.prodd(nil)=1 ∧
prodd(x.u)=times(x,prodd(u))| listinductiondef)
(label prodd_def)
(ue (phi |λu.allp(natnum,u)⊃natnum prodd(u)|) listinduction
(open allp prodd) )
(label proddsort)
(ue (phi |λu.allp(natnum,u)∧member(0,u)⊃prodd(u)=0|) listinduction
(open allp member prodd) proddsort (use normal timesdef mode: always))
;∀U.ALLP(NATNUM,U)∧MEMBER(0,U)⊃PRODD(U)=0
∂09-Nov-86 1705 coraki!pratt@sun.com fndsch@navajo = Foundations search
Received: from NAVAJO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 9 Nov 86 17:05:15 PST
Received: from Sun.COM by navajo.stanford.edu with TCP; Sun, 9 Nov 86 17:01:50 PST
Received: from sun.uucp (sun.sun.com) by sun.com (3.2/SMI-3.0)
id AA12914; Sun, 9 Nov 86 17:02:16 PST
Received: by sun.uucp (1.1/SMI-3.2)
id AA20755; Sun, 9 Nov 86 17:02:40 PST
Received: by coraki.uucp (3.2/SMI-1.2)
id AA00610; Sun, 9 Nov 86 17:02:06 PST
Date: Sun, 9 Nov 86 17:02:06 PST
From: coraki!pratt@sun.com (Vaughan Pratt)
Message-Id: <8611100102.AA00610@coraki.uucp>
To: fndsch@navajo.stanford.edu
Subject: fndsch@navajo = Foundations search
Mail to fndsch@navajo will go to the Foundations search committee.
The list is maintained in /csd2/pratt/lists/fndsch on navajo, and
was taken from SEARCH.DIS[1,PHY] on sail. It reads as follows.
guibas@src.dec.com
guibas@su-score.arpa
phy@su-ai.arpa
dek@su-ai.arpa
pratt@navajo.stanford.edu
mayr@su-score.arpa
jmc@su-ai.arpa
papa@su-score.arpa
ullman@navajo.stanford.edu
nilsson@su-score.arpa
haddad@sushi.stanford.edu
schaffer@sushi.stanford.edu
∂10-Nov-86 1027 CLT message for Ershov
Chuck Sword, a Geophysics grad student
(advisor = Jon Claerbout)
offers to show you the computing facilities in Geophysics.
They have a Convex and lots of nice graphics facilities for
displaying geophysical data.
They are interested in Russia, have visited Novosibirsk and
and corresponded with Alekseev at the computation center there.
They would very much like to talk to you.
Chuck can be reached at 3-1319 or chuck@hanauma
∂10-Nov-86 1108 VAL Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
FORMALIZING THE NOTION OF CONTEXT
John McCarthy
Thursday, November 13, 4pm
MJH 252
Getting a general database of common sense knowledge and
expressing it in logic requires formalizing the notion of context.
Since no context is absolutely general, any context must be elaboration
tolerant and we discuss this notion. Another formalism that seems
useful involves entering and leaving contexts; this is a generalization
of natural deduction.
∂10-Nov-86 1141 coraki!pratt@sun.com John Mitchell
Received: from NAVAJO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 10 Nov 86 11:40:57 PST
Received: from Sun.COM by navajo.stanford.edu with TCP; Mon, 10 Nov 86 11:38:16 PST
Received: from sun.uucp (sun.sun.com) by sun.com (3.2/SMI-3.0)
id AA25754; Mon, 10 Nov 86 11:38:40 PST
Received: by sun.uucp (1.1/SMI-3.2)
id AA00761; Mon, 10 Nov 86 11:38:40 PST
Received: by coraki.uucp (3.2/SMI-1.2)
id AA00153; Mon, 10 Nov 86 10:52:19 PST
Date: Mon, 10 Nov 86 10:52:19 PST
From: coraki!pratt@sun.com (Vaughan Pratt)
Message-Id: <8611101852.AA00153@coraki.uucp>
To: fndsch@navajo.stanford.edu
Subject: John Mitchell
I have a vitae and a stack of papers from John Mitchell, which I have
passed on to Leo. John should be returning to Bell from Europe about
now. He is interested in a position for the coming year. I believe
time is of the essence in this case.
-v
∂10-Nov-86 1420 RA Heftler letter
Re the list of past presidents of AAAI, do you have the list or shall
I contact Claudia for that?
Thanks,
∂10-Nov-86 1617 UDI@Score.Stanford.EDU israeli companies
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 10 Nov 86 16:17:03 PST
Date: Mon 10 Nov 86 16:14:31-PST
From: Ehud Shapiro <UDI@Score.Stanford.EDU>
Subject: israeli companies
To: jmc@Sail.Stanford.EDU
cc: boaz@CSLI.Stanford.EDU
Message-ID: <12253923924.23.UDI@Score.Stanford.EDU>
WE have a local expert in Israeli AI companies, Boaz Lezinger,
who is visiting CSLI. I have asked him to contact you about this.
Let me know if somthing comes out of it.
Regards, Udi
-------
∂10-Nov-86 1822 G.GORIN@HAMLET.STANFORD.EDU invitation to lunch
Received: from HAMLET.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 10 Nov 86 18:22:26 PST
Date: Mon 10 Nov 86 18:13:00-PST
From: Ralph Gorin <G.GORIN@HAMLET.STANFORD.EDU>
Subject: invitation to lunch
To: jmc@SU-AI.ARPA
Message-ID: <12253945493.105.G.GORIN@HAMLET.STANFORD.EDU>
Dear John,
Please be my guest at lunch on 5 December at the Faculty Club.
I'm hosting lunch for Hal Raveche, Dean of the School of Science at
Rensselaer. Dr. Raveche is leading a major program to upgrade the
sciences at RPI. Naturally, the recruitment of faculty for an
expansion of Computer Science at RPI is included in his Science
Initiatives program.
I'm sure he'd appreciate this opportunity to meet you,
to discuss our new Ph.Ds, directions for Computer Science, etc.
As you may know, Tom Spencer and Rob Schreiber are now in Computer
Science at RPI. And, Jim Meindl is newly installed as their
Provost.
Ralph
-------
∂10-Nov-86 1833 HSU@Sushi.Stanford.EDU EKL demo
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 10 Nov 86 18:33:50 PST
Date: Mon 10 Nov 86 18:24:05-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: EKL demo
To: glb@Sail.Stanford.EDU, jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12253947509.7.HSU@Sushi.Stanford.EDU>
CS 306 --- EKL Demo
Time: 6-9pm, Wednesday, November 12
Place: Rm.303, Sweet Hall
-------
∂10-Nov-86 2226 JJW Qlisp questions (continued)
To: QLISP@SAIL.STANFORD.EDU
RPG's answer to Question 1 in the previous message is that Qlisp must be
(as promised to DARPA) an extension of full Common Lisp, including the
separate function and value namespaces and all of the other features of
the language. So that answers that.
Here's a continuation of the descriptions of current Qlisp features along
with some more questions:
UNWIND-PROTECT is used to prevent terminated processes from leaving things
in an undesirable state. The form (UNWIND-PROTECT form cleanup) evaluates
FORM and guarantees that CLEANUP will be executed whether FORM returns
normally or is terminated by a THROW.
Question 8: The Qlisp paper only discusses the case where a THROW takes
place because of a computation done inside FORM. Another important case
is when processes containing UNWIND-PROTECT are killed because of a THROW
in another process. When this happens, do we execute the cleanup code for
all UNWIND-PROTECTs in processes being killed?
SUSPEND-PROCESS and RESUME-PROCESS are used (not surprisingly) to suspend
and resume processes. They take as an argument a process closure and
suspend or resume the associated process. SUSPEND-PROCESS can also take
no arguments, in which case it causes a process to suspend itself.
Question 9: What is the effect of SUSPEND-PROCESS to an already suspended
process, or RESUME-PROCESS to a process that is not suspended?
Question 10: How does suspension interact with killing of processes? If a
suspended process is killed, does it wake up to execute UNWIND-PROTECT
cleanup forms? Can a process be suspended while it is executing a cleanup
form? [If the answer to this is no, then (UNWIND-PROTECT NIL (FOO)) will
be a concise way to ensure that (FOO) is executed without possibility of
being killed or suspended.]
Question 11: If a process is not associated with a process closure (i.e.,
it was created by QLET), can we suspend or resume it?
∂10-Nov-86 2306 @MIRO.SNETC.Symbolics.COM:rwg@RUSSIAN.SPA.Symbolics.COM stanford indians
Received: from [128.81.2.8] by SAIL.STANFORD.EDU with TCP; 10 Nov 86 23:05:53 PST
Received: from RUSSIAN.SPA.Symbolics.COM by MIRO.SNETC.Symbolics.COM via CHAOS with CHAOS-MAIL id 25835; Tue 11-Nov-86 02:04:57 EST
Received: from TSUNAMI.SPA.Symbolics.COM by RUSSIAN.SPA.Symbolics.COM via CHAOS with CHAOS-MAIL id 5230; Mon 10-Nov-86 23:01:53 PST
Date: Mon, 10 Nov 86 23:01 PST
From: Bill Gosper <rwg@RUSSIAN.SPA.Symbolics.COM>
Subject: stanford indians
To: jmc@SU-AI.ARPA
Message-ID: <861110230149.0.RWG@TSUNAMI.SPA.Symbolics.COM>
Venkata Vemireddy, the main waiter and manager's son at Pasand, wants your
advice on how to get into Stanford as an undergraduate in business
administration. He is getting his associate degree from Foothill (bad gpa
1st year, but 3.57/4.0 second year.) Math SAT 580, English 300(!). Not
much extracurricular, due to restaurant demands, but was hospital
volunteer.
Is there anything you can think to add to what probably comes with
his application materials, e.g.? If you deliver such advice in person,
you get a free dinner.
∂11-Nov-86 0908 VAL re: Who is Frolov?
[In reply to message rcvd 10-Nov-86 22:36-PT.]
I am ashamed I didn't recognize the name of our distinguished colleague
Comrade Frolov.
∂11-Nov-86 0910 CLT igor
\+Office Phone: (617)-577-7741&\cr
rivin@SCRC-PEGASUS.ARPA
∂11-Nov-86 0927 BOAZ@CSLI.STANFORD.EDU
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 11 Nov 86 09:27:49 PST
Date: Tue 11 Nov 86 09:26:13-PST
From: Boaz Lazinger <BOAZ@CSLI.STANFORD.EDU>
To: jmc@SU-AI.ARPA
Hello John,
Udi asked me to contact you regarding Israeli AI companies.
I headed the CS&AI Dept. at the National Council for R&D , prior to comming here to CSLI.
My phone #:
723-2030 (w)
8559061 (h)
Boaz.
-------
∂11-Nov-86 1153 CLT
∂11-Nov-86 1146 THEORYNT%YKTVMX.bitnet@WISCVM.WISC.EDU Military Funding of Mathematics
Received: from WISCVM.WISC.EDU by SAIL.STANFORD.EDU with TCP; 11 Nov 86 11:46:11 PST
Received: from (THEORYNT)YKTVMX.BITNET by WISCVM.WISC.EDU on 11/11/86
at 13:24:17 CST
Date: Tues, 11 Nov 1986, 02:27:12 EST
From: Michael Shub <shub@ibm.com>
Subject: Military Funding of Mathematics
Resent-date: 11 Nov 1986 13:55:28-EST (Tuesday)
Resent-From: THEORYNT%YKTVMX.bitnet@WISCVM.WISC.EDU
Reply-To: THEORYNT%YKTVMX.bitnet@WISCVM.WISC.EDU
Resent-to: Theory-List@ibm.com
November 4,1986
This fall, a group of mathematicians met to discuss our concern with
military funding in mathematics. This group includes Lipman Bers, Lucy
Garnett, Linda Keen, Irwin Kra, Lee Mosher, Barbara Simons, Mike Shub,
Jean Taylor and Bill Thurston. We are in touch with many more. As a result
of these meetings we have written two resolutions and plan to introduce
them at the AMS Council and Business Meeting in San Antonio in January.
To generate discussion and thought we hope to have them published in the
January issue of the Notices. One of us, Bill Thurston, has written a
position paper on the subject which has been submitted for publication in
the same journal.
Enclosed are copies of the two resolutions along with brief
supporting comments written by Moe Hirsch. In order to generate support
for these motions, we need as many cosponsors as possible. We ask that
you cosponsor those motions which feel that you can, even though you may
have different reasons for supporting them than the reasons given in
Hirsch's paper. In addition, we ask you to circulate the motions among
your colleagues and return the signatures to:
Lucy Garnett, 180 Park Row Apt. 24B , New York, N.Y. 10038
or by electronic mail usingbitnet address
DPSBB@CUNYVM.
Also, please send copies of the motions to other possible signators.
For a copy of Thurston's position paper entitled "Military funding in
mathematics" write to him directly :
Bill Thurston, Mathematics Department, Washington Road, Princeton,
N.J. 08544.
MOTION1
Many scientists consider SDI (commonly referred to as Star Wars)
incapable of achieving its stated goals and dangerously destabilizing.
Participation by universities and professional organizations lends a
spurious scientific legitimacy to it. Therefore the AMS will lend no
support to the Star Wars program. In particular, no one acting as a
representative of the AMS shall participate in efforts to obtain
funding for Star Wars research or to mediate between agencies
granting Star Wars research and those seeking to apply for it.
MOTION 2
The AMS is concerned about the increasing militarization of support for
mathematics research. There is a tendency to distribute this support
through narrowly focussed (mission oriented) programs which circumvent
normal peer review procedures. This tendency, unless checked, may skew
and ultimately injure mathematics in the United States. Therefore those
representing the AMS are requested to direct their efforts towards
increasing the fraction of non-military funding for mathematics research,
as well as towards increasing total research support.
Comments from Moe Hirsch
Thanks to the efforts of many individuals and organizations, the
need for increased governmental funding of mathematical research has been
widely recognized, and such funding has recently grown significantly.
Recently, the proportion of funding coming from military agencies has
been increasing. While recognizing that much valuable research has been
supported by military branches of the government, we think there are
larger issues involved than just the amount of support. We deplore the
increasing imbalance between civilian and military support for
mathematics for the following reasons:
1. As military agencies become a major source of support the direction of
mathematical research will inevitably be deformed to suit military rather
than scientific needs.
Military agencies are increasingly mission oriented in their research
programs. Mathematicians and their graduate students will unavoidably
feel pressure -- both internally, and from their departments-- to design
their research projects so as to increase the likelihood of grants.
Unorthodox lines of research will be neglected in favor of whatever fad
has current popularity.
2. Increased dependence on military funding will have a chilling effect on
free public discussion of issues related to research.
Already some government officials have openly proclaimed that critics of
their missions need not apply. More insidious is the natural tendency not
to bite the hand that feeds us. What grant applicant would not consider
carefully, before criticising the agency's mission, whether this might not
endanger the grant?
3. Participation by universities, individuals and professional
organizations in such dubious military projects as SDI (Star Wars) lends a
spurious legitimacy to these projects.
We are often advised, in effect, to "take the Star Wars money even if you
are opposed to it -- especially if you are opposed." Whatever the morality
of such an attitude may be, it dangerously misleads the public into
believing that the scientific community supports enterprises such as SDI,
when in fact the overwhelming number of experts believe it to be deeply
flawed on scientific grounds, and very dangerous politically.
4. Military funding of mathematics is increasingly departing from
the vital traditions of peer review, open access to conferences, and no
restraint on publication.
DARPA has a publicly announced its reliance on its own consultants rather
than peer review of grant applications. It is true and commendable that
most military support of mathematical research has followed traditional
practices. But we see ominous signs that some agencies are sharply
departing from a tradition that, however vital, has no force of law.
When cosponsoring these motions, please be clear as to which ones you are
cosponsoring, also please indicate if you are a member of the AMS. Once
again the addresses are DPSBB@CUNY on the bitnet to reach Lucy Garnett
by electronic mail and Lucy Garnett, 180 Park Row, Apt.24b, New York,NY
10038 for ordinary mail. Please send this message on to several other
people who might be interested. Our mailing lists and capabilities are
rather limited.
∂11-Nov-86 1239 RLG job
can we meet Thursday, at time of your choice?
if job needs doing before then, i can meet lunch tomorrow (12-1pm), or eve today.
Bob.
∂11-Nov-86 1445 RA leaving?
If you don't have anything urgent for me to do today, I would like to leave
early, about 3:00.
Thanks,
∂12-Nov-86 0957 RA Marylin Salmansohn, ACM
Would like you to call her (212) 869 7440 ext. 255 re your paper in the ACM
anthology.
∂12-Nov-86 1004 JIMENEZ@Score.Stanford.EDU message
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 12 Nov 86 10:04:43 PST
Date: Wed 12 Nov 86 10:02:41-PST
From: Tina Jimenez <JIMENEZ@Score.Stanford.EDU>
Subject: message
To: JMC@Sail.Stanford.EDU
Message-ID: <12254380519.32.JIMENEZ@Score.Stanford.EDU>
Marilyn Silmansohn of ACM New York phoned, please call. 212 869-7440
EXT 255.
Tina
-------
∂12-Nov-86 1054 PERLAKI@Score.Stanford.EDU message
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 12 Nov 86 10:54:10 PST
Date: Wed 12 Nov 86 10:31:09-PST
From: Agnes M. Perlaki <PERLAKI@Score.Stanford.EDU>
Subject: message
To: mccarthy@Score.Stanford.EDU
Message-ID: <12254385703.55.PERLAKI@Score.Stanford.EDU>
Please call Peter Hirsch at IBM Palo Alto at 855-3117.
-Agi
-------
∂12-Nov-86 1100 JMC
expenses to Kahn
∂12-Nov-86 1117 RA leaving
To: JMC@SAIL.STANFORD.EDU, ZM@SAIL.STANFORD.EDU,
eppley@SCORE.STANFORD.EDU
I am leaving now for an 11:30 doctor appointment. I will be back in about
two hours.
Rutie
------
∂12-Nov-86 1307 VAL re: Here's a draft.
[In reply to message rcvd 12-Nov-86 13:01-PT.]
Great. Just one correction: I emigrated in 1976.
∂12-Nov-86 1627 CLT Qlisp meeting reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
Time: Friday, Nov 14 12:15-1:15
Place: MJ352
Topic: Joe Weening will tell us about Multilisp
∂12-Nov-86 1650 6058598%PUCC.PRINCETON.EDU@WISCVM.WISC.EDU SDI debate
Received: from WISCVM.WISC.EDU by SAIL.STANFORD.EDU with TCP; 12 Nov 86 16:50:39 PST
Received: from (VMMAIL)PUCC.BITNET by WISCVM.WISC.EDU on 11/12/86 at
18:49:59 CST
Received: by PUCC (Mailer X1.23b) id 2738; Wed, 12 Nov 86 19:48:01 EST
Date: Wed, 12 Nov 86 19:47:45 EST
From: Bill Crandall <6058598%PUCC.BITNET@WISCVM.WISC.EDU>
Subject: SDI debate
To: John McCarthy <jmc@sail.stanford.edu>
I am writing my thesis at Princeton on software development
for the Strategic Defense Initiative. As you know, there has been
a tremendous debate within the software community on this subject
and there has been no clear resolution to the problem.
I would like to know why you believe that software for a complex
strategic battle management system could be built. Specifically,
* how could a robust software system be developed so that
the defensive system would not fail during an attack?
* what would this software system's architecture be like?
* what types of advances in software engineering (for example,
automated programming, rule-based programmer's assistants, etc.)
do you think would be necessary, useful and possible in
SDI's development?
Finally, regardless of your opinion in the software debate, what
do you think of the Strategic Defense Initiative overall? Does it
make sound strategic sense? Is it possible to create a perfect
defense or would a leaky one suffice?
Thank you in advance for taking the time to answer these questions.
Your reply will be very useful to me in my research.
Sincerely,
Bill Crandall
Princeton University
∂12-Nov-86 1711 VAL Reminder: Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
FORMALIZING THE NOTION OF CONTEXT
John McCarthy
Thursday, November 13, 4pm
MJH 252
Getting a general database of common sense knowledge and
expressing it in logic requires formalizing the notion of context.
Since no context is absolutely general, any context must be elaboration
tolerant and we discuss this notion. Another formalism that seems
useful involves entering and leaving contexts; this is a generalization
of natural deduction.
∂12-Nov-86 1734 HADDAD@sushi.stanford.edu procedural questions about the faculty search
Received: from NAVAJO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 12 Nov 86 17:34:56 PST
Received: from Sushi.Stanford.EDU by navajo.stanford.edu with TCP; Wed, 12 Nov 86 13:53:05 PST
Date: Wed 12 Nov 86 13:50:40-PST
From: Ramsey Haddad <HADDAD@sushi.stanford.edu>
Subject: procedural questions about the faculty search
To: fndsch@navajo.stanford.edu
Message-Id: <12254422024.47.HADDAD@Sushi.Stanford.EDU>
Alex and I are new to this and thus have a bunch of questions. Here's
a representative sample:
1) How many slots are we trying to fill? 2? more?
2) What is confidential? Obviously recommendations are. And, other
committee members specific opinions, too, I assume? What about the
simple fact that person X is applying for a slot?
3) How do recommendations work? Does the applicant give us a list
of (people) references? If we know someone who is familiar with the
applicant can we just ask for their opinion?
4) What is the procedure for getting our hands on the publications
that the applicant submits?
In other words we have a whole bunch of uncertainties about "what
SHOULDN'T be done", "what SHOULD be done", "what should be done -- but
DISCREETLY" and "HOW things are done".
These uncertainties include:
5) How should we go about getting answers to our questions?
Ramsey (& Alex)
-------
∂12-Nov-86 1830 ullman@navajo.stanford.edu Re: procedural questions about the faculty search
Received: from NAVAJO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 12 Nov 86 18:30:50 PST
Received: by navajo.stanford.edu; Wed, 12 Nov 86 18:27:56 PST
Date: Wed, 12 Nov 86 18:27:56 PST
From: Jeff Ullman <ullman@navajo.stanford.edu>
Subject: Re: procedural questions about the faculty search
To: HADDAD@sushi.stanford.edu, fndsch@navajo.stanford.edu
Let me try some answers:
1. up to 2 slots. Personally, I don't think we should worry about
such things. If 3 outstanding people were available, we should
try making cases.
2. I would not reveal ANYTHING about matters of recruiting.
3. Junior people generally give a list of references; Stanford hopes
for 5, but often 3-4 will do. Senior cases generally require 10-15
references, from a list chosen by the committee, possibly after
consultation with the victim.
4. Write to the applicant and ask them.
---jeff
∂13-Nov-86 0907 TAJNAI@Score.Stanford.EDU Nobel Prize
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 13 Nov 86 09:06:55 PST
Date: Thu 13 Nov 86 09:05:23-PST
From: Carolyn Tajnai <TAJNAI@Score.Stanford.EDU>
Subject: Nobel Prize
To: jmc@Sail.Stanford.EDU
Message-ID: <12254632232.16.TAJNAI@Score.Stanford.EDU>
I dreamed last night that you won a Nobel Prize. Hope it is a
prophetic dream.
Carolyn
-------
∂13-Nov-86 1502 WIEDERHOLD@SUMEX-AIM.ARPA John Sowa - Industrial lecturer Fall quarter 87/88
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 13 Nov 86 15:02:11 PST
Date: Thu 13 Nov 86 15:01:23-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: John Sowa - Industrial lecturer Fall quarter 87/88
To: jmc@SU-AI.ARPA
cc: "*PS:<WIEDERHOLD>SABBATICAL.MSGS-87.1"@SUMEX-AIM.ARPA
Message-ID: <12254697042.83.WIEDERHOLD@SUMEX-AIM.ARPA>
I had a chance to talk more with John Sowa from IBM Yorktown
(Conceptual graphs .. ) and he is getting all the required blessings
and committments to do the lectureship the fall quarter from IBM.
The local IBM group will host him for the complement of the
time here. Since I plan to be on sabbatical that quarter
I'd like him also to be available to my students, but that should not
require a financial committment from us.
We should be able to get a course description in time for the bulletin.
Let me know if I should do anything else to make his visit happen.
Gio
-------
∂13-Nov-86 1517 CERF@A.ISI.EDU Thanks
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 13 Nov 86 15:17:01 PST
Date: 13 Nov 1986 18:16-EST
Sender: CERF@A.ISI.EDU
Subject: Thanks
From: CERF@A.ISI.EDU
To: reddy@A.CS.CMU.EDU, jmc@SU-AI.ARPA
Cc: kahn@A.ISI.EDU
Message-ID: <[A.ISI.EDU]13-Nov-86 18:16:00.CERF>
Raj and John,
Just wanted you to know how much it helped to have your thoughts
and ideas in connection with the digital library project. Having a
fresh perspective is always valuable and since you've both obviously
thought a lot about this topic, yu were especially helpful.
My next step is to try to boil down the substance of our discussions
into a document we can share with the rest of the working group in
advance of the Jan meetings.
We expect to meet with representatives from the Library of Congress
soon and will raise many of the issues that John addressed concerning
the utility of having things that go into the LOC also be readily
available in digital form.
I'll keep you posted on events as they unfold.
Vint
∂13-Nov-86 1557 RA leaving
It's Thursday and I am leaving early.
∂13-Nov-86 2230 BRINK@Sushi.Stanford.EDU Thanks
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 13 Nov 86 22:30:25 PST
Date: Thu 13 Nov 86 22:21:15-PST
From: Ed Brink <brink@Sushi.Stanford.EDU>
Subject: Thanks
To: jmc@Sail.Stanford.EDU
Message-ID: <12254777117.12.BRINK@Sushi.Stanford.EDU>
..for letting me butt in on your session this afternoon. It was over my head
in some ways, but I learned a lot. You have a way of speaking English even
when the subject is pretty doggone abstruse, which can make the novice think he
understands until he opens his mouth.
If I decide to take IBM up on its generous offer to put me to pasture early, I
may decide the Farm has as green grass as any, and come looking for a job.
Wish I hadn't let my advisor talk me into 242 instead of 306. There is no
comparison.
Thanks again..
..Ed Brink
-------
∂14-Nov-86 0538 CERF@A.ISI.EDU re: Thanks
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 14 Nov 86 05:38:29 PST
Date: 14 Nov 1986 08:37-EST
Sender: CERF@A.ISI.EDU
Subject: re: Thanks
From: CERF@A.ISI.EDU
To: JMC@SU-AI.ARPA
Cc: reddy@A.CS.CMU.EDU, kahn@A.ISI.EDU
Message-ID: <[A.ISI.EDU]14-Nov-86 08:37:59.CERF>
In-Reply-To: The message of 13 Nov 86 1525 PST from John McCarthy <JMC@SAIL.STANFORD.EDU>
John,
many thanks for taking the trouble to put your thoughts on "paper" -
[how should we say that in the future?].
I quite agree that we must have document stored in a way that makes
them searchable - Danny Hillis would be in complete agreement with
you on that point. We may wind up with multiple representations in
the archive, if necessary, to achieve both the ability to search
on the full text and the ability to transmit copies which can
be accurately reproduced ("protrayed?") at the receiving end,
on paper or on soft-copy display.
I believe there will be opportunities for NRI and for the working
group participants to address legislators and other key people
about the digital library, knowledge bank and related ideas, so
your willingness to participate along those lines is much appreciated.
Looking forward to seeing you in January.
Vint
∂14-Nov-86 1253 RA Walter Rosenblith
You don't have a Walter Rosenblith in your phon file. Where can I find
his address, or what is his affliation?
Thanks,
∂14-Nov-86 1331 shoham@YALE.ARPA update
Received: from YALE-BULLDOG.ARPA by SAIL.STANFORD.EDU with TCP; 14 Nov 86 13:29:56 PST
Received: by Yale-Bulldog.YALE.ARPA; 14 Nov 86 15:44:40 EST (Fri)
Date: 14 Nov 86 15:44:40 EST (Fri)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8611142044.AA01297@Yale-Bulldog.YALE.ARPA>
Subject: update
To: nilsson@su-score.ARPA, jmc@su-ai.ARPA, genesereth@su-score.ARPA
Dear Nils, John and Mike,
Just a short update on my status. I've just completed a full draft of my
thesis and delivered it to my readers (Mike Fischer and Pat Hayes). I hope
they respond fast, so that I can submit the final version to the grad
school in a month. In the meanwhile I plan to spend a week in Jamaica,
trying to convert my mental state to something resembling sanity. I guess
all this means that I won't get to Stanford and April. Please let me know
of anything you think I should be doing in preparation for my job.
Regards,
Yoav
-------
∂14-Nov-86 1421 CLT calendar item
To: JMC, LES, JJW, AIR, RPG, ARG
Time: Tuesday, Nov 18 2:45
Place: Nils conference room
Event: Meeting with Jack Test re Alliant
∂14-Nov-86 1446 AI.NOVAK@R20.UTEXAS.EDU Lisp and Ada
Received: from R20.UTEXAS.EDU by SAIL.STANFORD.EDU with TCP; 14 Nov 86 14:46:22 PST
Date: Fri 14 Nov 86 16:46:07-CST
From: Gordon Novak Jr. <AI.NOVAK@R20.UTEXAS.EDU>
Subject: Lisp and Ada
To: Feigenbaum@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA
Message-ID: <12254956405.28.AI.NOVAK@R20.UTEXAS.EDU>
Dear Ed and John,
We get our major research funding from the US Army Research Office. It is
my understanding that Hon. J. R. Sculley, Assistant Secretary of the Army
for Research, Development, and Acquisition (who handles an annual procurement
budget of some $40 billion) has issued an edict that all embedded computer
software for the Army must be written in Ada, period. Our program manager
at ARO, Dr. Ronald Green, thinks this decision might be turned around with
the proper support. I have sent the message that follows. I think it would
be very valuable if you gentlemen could also send messages (to green@brl)
supporting the use of Lisp when appropriate. It seems to me that it would be
clearly bad for AI and for the military if they prohibited Lisp.
Many thanks, Gordon
Use of Lisp for Embedded Computer Software for the Army
Gordon S. Novak Jr.
Director, Artificial Intelligence Laboratory
Department of Computer Sciences
University of Texas at Austin
It would be most unfortunate if the US Army prohibited the use of the Lisp
programming language for the software of embedded computer systems by
requiring that all such software be written in Ada.
DARPA has funded the development of a Lisp chip set and processor by Texas
Instruments. This project is on schedule; a processor based on the chip set
is running and is expected to be commercially available in the first half of
1987. The TI chip set will provide a ruggedized processor specialized to run
Lisp in a package that fits in a six-inch cube. Such a processor is clearly
suitable for use in a variety of military platforms. It seems clear that
DARPA funded this project so that Lisp could be used in military applications
as an embedded processor and language.
Inability to use Lisp as a language for military applications would hamper
the application of Artificial Intelligence to military problems. Artificial
intelligence is a technology in which the U.S. is the world leader; it has
been identified by military study groups as one of the most important advanced
technology areas for military application. Lisp provides capabilities for
symbolic computation that are not available in Ada. If artificial intelligence
applications had to be recoded in Ada before they could be deployed, it would
be necessary to reinvent much of Lisp and add it to the Ada program, but at
greatly increased development cost and orders of magnitude performance
degradation compared to processors specially designed to run Lisp efficiently.
It is likely that military platforms of the future will involve many on-board
computers that communicate with each other; the Boeing 767 airplane has over
60 computers that perform different tasks. While it is appropriate for many
of the computers in a military system to be programmed in Ada, it should
be possible for some of the computers to be Lisp processors running Lisp when
that is most appropriate to the application. To prohibit such an arrangement
would be to throw away a significant technological advantage of the United
States.
-------
∂14-Nov-86 1624 RA Staff TGIF
I'll be at a staff TGIF for the next half hour.
∂14-Nov-86 1636 shoham@YALE.ARPA re: update
Received: from YALE-BULLDOG.ARPA by SAIL.STANFORD.EDU with TCP; 14 Nov 86 16:36:44 PST
Received: by Yale-Bulldog.YALE.ARPA; 14 Nov 86 19:03:45 EST (Fri)
Date: 14 Nov 86 19:03:45 EST (Fri)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8611150003.AA03187@Yale-Bulldog.YALE.ARPA>
Subject: re: update
To: John McCarthy <JMC@SAIL.STANFORD.EDU>
Cc: nilsson@su-score.ARPA, genesereth@su-ai.ARPA
In-Reply-To: John McCarthy <JMC@SAIL.STANFORD.EDU>, 14 Nov 86 1336 PST
[In reply to message from shoham@YALE.ARPA sent 14 Nov 86 15:44:40 EST.]
The phrase "I won't get to Stanford and April" must be a misprint for
something else. What?
and => until
-------
∂15-Nov-86 0900 JMC
Vicky
∂15-Nov-86 0900 JMC
reservations to L.A.
∂15-Nov-86 1334 ILAN@Score.Stanford.EDU re: saying for today
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Nov 86 13:34:40 PST
Date: Sat 15 Nov 86 13:33:01-PST
From: Ilan Vardi <ILAN@Score.Stanford.EDU>
Subject: re: saying for today
To: su-etc@Score.Stanford.EDU
cc: jmc@Sail.Stanford.EDU
Message-ID: <12255205243.16.ILAN@Score.Stanford.EDU>
I suppose you're against putting money in the kitty?
-------
∂15-Nov-86 1557 GOLDBERG@CSLI.STANFORD.EDU Re: Saying for today
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Nov 86 15:57:08 PST
Date: Sat 15 Nov 86 15:55:51-PST
From: Jeffrey Goldberg <GOLDBERG@CSLI.STANFORD.EDU>
Subject: Re: Saying for today
To: JMC@SU-AI.ARPA
cc: su-etc@SU-AI.ARPA, GOLDBERG@CSLI.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 15 Nov 86 12:40:00-PST
Supporting the Contras and attacking Lybia for state-sponsored terrorism is
like teaching a child patience by having him watch you beat up people who
interrupt you. (Or maybe it's just like asking for Washington D.C. to be
bombed.)
-Jeff
-------
∂15-Nov-86 1702 CLT
shopping item - unscented tide
∂15-Nov-86 1824 WIEDERHOLD@SUMEX-AIM.ARPA Request for sabbatical style Leave-of-absence
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Nov 86 18:24:16 PST
Date: Sat 15 Nov 86 18:23:34-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: Request for sabbatical style Leave-of-absence
To: nilsson@SU-SCORE.ARPA
cc: bscott@SU-SCORE.ARPA, winslett@SU-SCORE.ARPA, Sowa@IBM.COM,
Milton@SRI-STRIPE.ARPA, shortliffe@SUMEX-AIM.ARPA, vian@SUMEX-AIM.ARPA,
jmc@SU-AI.ARPA
Message-ID: <12255258137.32.WIEDERHOLD@SUMEX-AIM.ARPA>
I want to take off for the Fall-quarter of 1987-1988 only.
Since Ted Shortliffe will be back then my MIS teaching and other
responsibilities can be backstopped.
Jack Milton from UC-Davis (KBMS-associate investigator) will
handle the CS545 seminar jointly with Marianne Winslett - who should be a
Research Associate then.
Limiting my absence to a quarter should minimize the effect on my PhD students
and will be more than compensated with a planned stay by John Sowa from IBM
SRI. I expect him to come here for an industrial lectureship (CS309A)
-- topic: Conceptual Graphs --- during that period. John McCarthy expects
that it will be possible from Stanfords' side and John is making
arrangements with IBM for the complementary support.
I hope you will be able to grant the leave.
Gio
-------
∂15-Nov-86 2134 niz@lindy.STANFORD.EDU forsythe connection... crash recovery
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Nov 86 21:34:31 PST
Received: by lindy.STANFORD.EDU; Sat, 15 Nov 86 21:29:42 PST
Date: Sat, 15 Nov 86 21:29:42 PST
From: Jim Nisbet <niz@lindy.STANFORD.EDU>
Subject: forsythe connection... crash recovery
To: jmc@sail.STANFORD.EDU
I just checked your Socrates account (ed.a69); it shows the last logoff
being at 9:27. The system didn't crash then; I don't know why you lost
your connection -- I've been trying to track down a bad port on our Vax,
so please let me know if it happens again. (You should be able to logon
again to your Socrates account -- if you still can't get on then give me
a call at 723-4206.)
Thanks,
/j
∂15-Nov-86 2145 niz@lindy.STANFORD.EDU Ok... I took ttyj5 out of the tty list
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Nov 86 21:45:06 PST
Received: by lindy.STANFORD.EDU; Sat, 15 Nov 86 21:40:16 PST
Date: Sat, 15 Nov 86 21:40:16 PST
From: Jim Nisbet <niz@lindy.STANFORD.EDU>
Subject: Ok... I took ttyj5 out of the tty list
To: jmc@sail.STANFORD.EDU
I took ttyj5 out of the tty list for the moment so I can test it out. Can
you try logging in one more time... the other ports seem to be ok--I just
tried it myself.
/j
∂16-Nov-86 1250 CLT tonight
i forgot to ask if you will be here so I can
play with glb at 9
∂17-Nov-86 0900 CLT
smog check
∂17-Nov-86 0917 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu lunch this week
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Nov 86 09:17:39 PST
Received: by lindy.STANFORD.EDU; Mon, 17 Nov 86 09:12:23 PST
Date: Mon, 17 Nov 86 09:17:40 PST
From: <ELLIOTT@slacvm.bitnet>
Reply-To: ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu
To: JMC@sail.STANFORD.EDU
Subject: lunch this week
Date: 17 November 86 09:13-PST
From: ELLIOTT@SLACVM
To: JMC@SAIL
Subject: lunch this week
Date: 17 November 1986, 09:11:30 PST
From: Bloom, Elliott ELLIOTT at SLACVM
To: JMC at SAIL.STANFORD
Subject: lunch this week
I am free for lunch today, Wed, and Frid; I can also make
it late on Tuesday (about 1pm).
∂17-Nov-86 0926 wbm@su-whitney.arpa Thanks for Reply: integrals
Received: from WHITNEY.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Nov 86 09:26:28 PST
Received: by su-whitney.arpa with Sendmail; Mon, 3 Nov 86 18:54:25 pst
Date: Mon, 3 Nov 86 18:54:25 pst
From: Wally Mann <wbm@su-whitney.arpa>
Subject: Thanks for Reply: integrals
To: jmc@sail
John,
Thanks for your suggestion.
Wally
∂17-Nov-86 1024 ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu Re: re: lunch this week
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Nov 86 10:24:46 PST
Received: by lindy.STANFORD.EDU; Mon, 17 Nov 86 10:19:27 PST
Date: Mon, 17 Nov 86 10:24:39 PST
From: <ELLIOTT@slacvm.bitnet>
Reply-To: ELLIOTT%SLACVM.BITNET@forsythe.stanford.edu
To: JMC@sail.stanford.edu
Subject: Re: re: lunch this week
Date: 17 November 86 10:23-PST
From: ELLIOTT@SLACVM
To: JMC@SAIL.STANFORD.EDU
Subject: Re: re: lunch this week
Date: 17 November 1986, 10:23:08 PST
From: Bloom, Elliott ELLIOTT at SLACVM
To: JMC at SAIL.STANFORD.EDU.STANFORD
Subject: Re: re: lunch this week
In-Reply-To: JMC AT SAIL.STANFORD.EDU.STANFORD -- 11/17/86, 09:54
Fine meet you at the faculty club at 12:10 on Friday.
Have a nice week.
∂17-Nov-86 1024 RA Paula Simons, Communication Dept.
Simons would like to do a profile on you. Her tel. no. is 328 3846. She
will call back later.
∂17-Nov-86 1037 VAL Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
LOGIC PROGRAMMING AND CIRCUMSCRIPTION
Vladimir Lifschitz
Thursday, November 20, 4pm
MJH 252
The talk will be based on my paper "On the declarative semantics of
logic programs with negation". A few copies of the paper are available
in my office, MJH 362.
ABSTRACT. A logic program can be viewed as a predicate formula, and its
declarative meaning can be defined by specifying a certain Herbrand
model of that formula. For programs without negation, this model is
defined either as the Herbrand model with the minimal set of positive
ground atoms, or, equivalently, as the minimal fixed point of a certain
operator associated with the formula (Van Emden and Kowalski). These
solutions do not apply to general logic programs, because a program
with negation may have many minimal Herbrand models, and the corresponding
operator may have many minimal fixed points. Apt, Blair and Walker and,
independently, Van Gelder, introduced a class of general logic programs
which disallow certain combinations of recursion and negation, and showed
how to use the fixed point approach to define a declarative semantics for
such programs. Using the concept of circumscription, we extend the minimal
model approach to stratified programs and show that it leads to the same
semantics.
∂17-Nov-86 1111 RA Re: inclusion
[Reply to message recvd: 17 Nov 86 00:28 Pacific Time]
Where will I find the 1975 IJCAI report?
Thanks,
∂17-Nov-86 1241 RA your trip to Virginia, Nov. 7
Shall I charge this trip to Stanford or is somebody else paying for it?
Thanks,
∂17-Nov-86 1401 RA expenses for Kahn
You did not give me a receipt for SF Airport parking. Do you remember how
much it was?
Thanks,
∂17-Nov-86 1403 RA new room for the seminar
To: VAL
CC: JMC
You have room 050 Bldg. 420 for the rest of this quarter. Let me know
whether you like the room (ASAP) and we'll request the same room for next
quarter.
∂17-Nov-86 1410 RA Ellie Gray
Please call Ellie Gray (312) 984 7416 re: letter she sent you.
∂17-Nov-86 1449 ILAN@Score.Stanford.EDU Psych paper
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Nov 86 14:49:25 PST
Date: Mon 17 Nov 86 14:47:48-PST
From: Ilan Vardi <ILAN@Score.Stanford.EDU>
Subject: Psych paper
To: jmc@Score.Stanford.EDU
Message-ID: <12255743145.43.ILAN@Score.Stanford.EDU>
Just latex this file to see it!
Ilan
\documentstyle[12pt]{article}
\begin{document}
\title{A Differential Equation Model of Visual-Motor Adaptation}
\date{\today}
\author{Helen Cunningham\\Psychology Department\\Stanford University
\and Ilan Vardi\\Mathematics Department\\Stanford University}
\maketitle
\medskip
\begin{center}
\mbox{\Large\bf Abstract} \end{center}
\medskip
In this paper we present a new way of modelling
visual-motor performance under an altered mapping between
visual and motor spaces. We have set up a differential equation model
in which experimental performance is a solution of the
differential equations. The motivation for this
approach is our interest in
finding parameters which mental representations would minimize.
In our experiment,
we have used a computer display screen and digitizing tablet to
simulate distortions of the normal visual-motor map, by rotating
motor space (the digitizing tablet) relative to visual space
(the screen). This paper will focus on visual-motor performance
under a 90-degree rotation.
Experimental subjects perform a simple aiming task in which they are instructed
to move a circular cursor in a straight line so as to ``capture'' a
target that appears in one of eight possible screen locations
positioned radially around the center (the starting position). Movement
trajectories are sampled as $x,y$ screen coordinates at 15-millisecond
intervals and analyzed post-experimentally with respect to their spatial
and temporal characteristics.
\smallskip
This paper is concerned with the form that movement trajectories take
as subjects adapt to the mapping.
Under the initial mapping trajectories
are always straight. Under a 90-degree rotation, trajectories
are at first very distorted, usually consisting of a sequence of straight
segments connected by sharp angles. If the target is acquired,
it is by such a zig-zagging path. Often at the outset, however, the
target is not acquired.
After the first 20 or so trials subjects uniformly begin to adopt
a new strategy. This strategy is apparent in the
movement records and it described by subjects as seeming
``optimal'' or ``easiest''.
The basic form that movement trajectories settle into is that of smooth
arcs.
\smallskip
We believe that the
smooth
trajectory minimizes some set of abstract parameters
on the subject's internal representation of the space, and this
is the subject of our present research. To reach this goal we have
decided first to model the trajectories as solutions of differential
equations.
Note that our approach differs
from classic control theory models which explain performance in terms of
the mechanics of the hand/arm system.
\smallskip
Before going to the details of the model, the
experimental situation will be described.
\smallskip
The subject is instructed to view a display screen on which appears
a moveable cursor and a center spot. Upon centering the cursor,
a trial is initiated by a cueing tone then followed by the
appearance of a target in one of eight possible locations. The subject's
task is to move the cursor to the target in as straight a line as possible.
Speed is not emphasized, but the target remains visible for only 3.7
seconds. Early in the adaptation period the target is seldom acquired
in the allotted time.
The digitizing tablet on which cursor movements are made is
rotated 90 degrees clockwise with respect to its normal orientation.
Under this rotation, a target appearing at the top of the screen corresponds
to a location the right side of the tablet and so the subject must move
the hand rightward in order to capture an upward target. The
location corresponds to upward (forward) movement on the tablet and
the new location corresponds to rightward movement on the tablet.
\smallskip
The model consits of representing the old mapping as an influence on
the new map. We model this
influence by a rotational counterclockwise field
around the origin. We model the influence of
the target as an inward pointing field
centered at the right hand position where the target is on the tablet.
\begin{figure}
\vspace{1in}
\caption{Model representing influences of old map and target}
\end{figure}
Mathematically we see that the vector field representing the influence
of the old map is
$$
\left(
\begin{array}{c}
-f←{rot}(r)\, y\\g←{rot}(r)\,x
\end{array}
\right)
$$
where $r=x↑2 + y↑2$,
and $f←{rot}$ and $g←{rot}$ are
positive functions to be specified later.
Similarily the vector field
corresponding to a target centered at $(a,0)$ is
$$
\left( \begin{array}{c}
f←{target}(r←a)\, (a-x)\\-g←{target}(r←a)\, y
\end{array}\right)
$$
where $r←a =(a-x)↑2 +y↑2$ and
again $f←{target}$ and $g←{target}$ are positive
functions.
We superimpose the two vector fields and get as our governing
differential equation
$$
\left( \begin{array}{c}
\dot{x}\\ \dot{y}
\end{array}\right)
=
\left(
\begin{array}{c}
af←{target}(r←a)\\
0
\end{array}
\right)
+
\left(
\begin{array}{cc}
-f←{target}(r←a) & -f←{rot}(r) \\
g←{rot}(r) & -g←{target}(r←a)
\end{array}
\right)
\left(
\begin{array}{c}
x\\y
\end{array}
\right)
$$
(as usual we have written $\dot{x}$ for $\frac{dx}{dt}$ etc$\ldots$).
\smallskip
The initial condition consists of
a velocity pointing in the $y$ axis (the initial position
is the origin)
$$
\left(\begin{array}{c}
\dot{x}(0)\\ \dot{y}(0)
\end{array}
\right)
=
\left(
\begin{array}{c}
v←{initial}\\0
\end{array}
\right)
$$
This corresponds to the
fact that the old map gave the target on the $y$ axis.
\medskip
The third part of our paper is a computer simulation of
the above differential equation.
We look for natural functions $f←{rot},\;g←{rot},
\;f←{target},\;g←{target}$ along with initial
condition $v←{initial}$
so that the solutions of the differential equation correspond
faithfully to the trajectories found in the experiment.
\end{document}
-------
∂18-Nov-86 0111 rivin@STONY-BROOK.SCRC.Symbolics.COM teaching
Received: from SCRC-STONY-BROOK.ARPA by SAIL.STANFORD.EDU with TCP; 18 Nov 86 01:11:11 PST
Received: from GANNET.SCRC.Symbolics.COM by STONY-BROOK.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 2138; Tue 18-Nov-86 03:44:06 EST
Date: Tue, 18 Nov 86 03:43 EST
From: igor rivin <rivin@ALLEGHENY.SCRC.Symbolics.COM>
Subject: teaching
To: JMC@SAIL.STANFORD.EDU, rivin@PEGASUS.SCRC.Symbolics.COM
In-Reply-To: The message of 17 Nov 86 17:37 EST from John McCarthy <JMC@SAIL.STANFORD.EDU>
Message-ID: <861118034353.4.RIVIN@GANNET.SCRC.Symbolics.COM>
Received: from STONY-BROOK.SCRC.Symbolics.COM by ALLEGHENY.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 11727; Mon 17-Nov-86 16:37:24 EST
Received: from PEGASUS.SCRC.Symbolics.COM by STONY-BROOK.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 1903; Mon 17-Nov-86 17:48:56 EST
Received: from SAIL.STANFORD.EDU (su-ai.arpa.ARPA) by scrc-pegasus id AA01387; Mon, 17 Nov 86 17:38:01 est
Date: 17 Nov 86 1437 PST
From: John McCarthy <JMC@SAIL.STANFORD.EDU>
Subject: teaching
To: rivin@SCRC-PEGASUS.ARPA
If your are interested in teaching a course in mathematical computation,
I'm sure it could be arranged. It might also establish a need here
that no-one previously knew existed.
I would find teaching such a course very interesting.
Igor.
∂18-Nov-86 0925 VAL re: new room for the seminar
To: RA
CC: JMC
[In reply to message from RA rcvd 17-Nov-86 14:03-PT.]
Room 050 Bldg. 420 is fine.
Thank you,
Vladimir
∂18-Nov-86 0943 RA Ellie Gray
Gray called again (312) 984 7416, re: the letter she sent you.
∂18-Nov-86 1030 CLT msg from SF (missing a J)
[My reply also included]
∂17-Nov-86 2211 SF@CSLI.STANFORD.EDU Japan trip
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Nov 86 22:11:21 PST
Date: Mon 17 Nov 86 22:09:16-PST
From: Sol Feferman <SF@CSLI.STANFORD.EDU>
Subject: Japan trip
To: clt@SU-AI.ARPA, mc@SU-AI.ARPA
We just returned from a 2 week trip to Japan, one week each in Tokyo and
Kyoto. It was extremely interesting in all respects, and the people there
were very helpful to us; everything went as planned. Prices are out of
sight, but a watchful eye will keep that within bounds. The major
expense is hotels, which can be twice as much as here. Meals can also
be exorbitant, but one can do very well in judiciously chosen places.
it takes a bit of planning, but I think it is very much worth it.
Sato and others asked me to urge you to come, and I can do that without
reservation.
P.S. I understand I should submit my expenses to John's secretary,
Ruth Adler. Please let me know if that is not correct. I will
be able to get this together within a week. Is any reprt necessary,
or would that go with a final report for the grant?
-------
SF@CSLI.STANFORD.EDU
Japan trip
I'm glad to hear you had a good trip to Japan.
Yes, Rutie will take care of your expenses.
There is no need to send a report to NSF for each trip.
I suppose the final report will consist of individual
reports on trips to Japan plus a summary of the
activities here during the visits of Sato et. al. to Stanford.
If you want to write something now, you could put it in
a file and mail it to me to keep for putting in the final report.
Otherwise I will probably bug you for a page at report time.
∂18-Nov-86 1259 AAAI-OFFICE@SUMEX-AIM.ARPA workshop request
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Nov 86 12:59:55 PST
Date: Tue 18 Nov 86 12:54:42-PST
From: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: workshop request
To: JMC@SU-AI.ARPA
cc: aaai-OFFICE@SUMEX-AIM.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Message-ID: <12255984700.81.AAAI-OFFICE@SUMEX-AIM.ARPA>
John,
I just received a note requesting funding for a Space Station AUtomation
workshop. I'm concerned about this request because it is not a
workshop per se. It is a session at the Space Station Automation
Symposium III. You should be getting a copy of the request.
Claudia
-------
∂18-Nov-86 1510 VAL Przymusinski
Les says we can pay Przymusinski out of the NSF money as a consultant. Is
this ok?
∂18-Nov-86 1713 VAL re: Przymusinski
[In reply to message rcvd 18-Nov-86 16:31-PT.]
I'd like to invite him to spend 3 or 4 days here, give a talk and try to do
some work with me on computing circumscription using logic programming. We
would pay his airfare and a $200 honorarium. (He can stay at my place, so it
won't be necessary to pay for a hotel). If we decide to do that then I can
also ask Ullman if he wants to co-sponsor the visit.
∂18-Nov-86 1748 @REAGAN.AI.MIT.EDU:JCMA@AI.AI.MIT.EDU Seminar - Formalizing the Notion of Context (SU)
Received: from REAGAN.AI.MIT.EDU by SAIL.STANFORD.EDU with TCP; 18 Nov 86 17:48:13 PST
Received: from MORRISON.AI.MIT.EDU by REAGAN.AI.MIT.EDU via CHAOS with CHAOS-MAIL id 11820; Tue 18-Nov-86 20:47:48 EST
Date: Tue, 18 Nov 86 20:48 EST
From: John C. Mallery <JCMA@AI.AI.MIT.EDU>
Subject: Seminar - Formalizing the Notion of Context (SU)
To: jmc@SU-AI.ARPA
In-Reply-To: The message of 10 Nov 86 14:08 EST from Vladimir Lifschitz <VAL@SAIL.STANFORD.EDU>
Message-ID: <861118204803.6.JCMA@MORRISON.AI.MIT.EDU>
Date: 10 Nov 86 1108 PST
From: Vladimir Lifschitz <VAL@SAIL.STANFORD.EDU>
Commonsense and Non-Monotonic Reasoning Seminar
FORMALIZING THE NOTION OF CONTEXT
John McCarthy
Thursday, November 13, 4pm
MJH 252
Getting a general database of common sense knowledge and
expressing it in logic requires formalizing the notion of context.
Since no context is absolutely general, any context must be elaboration
tolerant and we discuss this notion. Another formalism that seems
useful involves entering and leaving contexts; this is a generalization
of natural deduction.
If there is a paper on this, I would certainly like to get a copy.
Thanks.
John Mallery
Artificial Intelligence Laboratory, NE43-797
Massachusetts Institute of Technology
545 Technology Square
Cambridge, Massachusetts 02139
∂18-Nov-86 2115 FULLERTON@Sierra.Stanford.EDU TV
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Nov 86 21:15:09 PST
Date: Tue 18 Nov 86 21:11:49-PST
From: Dwain Fullerton <FULLERTON@Sierra.Stanford.EDU>
Subject: TV
To: buchanan@SUMEX-AIM.Stanford.EDU
cc: jmc@Sail.Stanford.EDU
Message-ID: <12256075195.19.FULLERTON@Sierra.Stanford.EDU>
Bruce,
More on the SRA saga: John McCarthy tells me that SRA sent him a
sample tape, with "a lot of flashy visuals, but extremely feeble content."
He advises anyone thinking of taking part in the project to charge a large
fee.
Best,
Dwain
-------
∂19-Nov-86 1300 RA Binford
Ann Richardson wanted to remind you of the Binford reppointment letter.
∂19-Nov-86 1306 RA Ellie Gray
Please call Ellie Gray (312) 984 7416.
∂19-Nov-86 1634 VAL New Place for the Commonsense Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
We are moving to a bigger room: 050, Bldg. 420 (main quad, next to MJH).
LOGIC PROGRAMMING AND CIRCUMSCRIPTION
Vladimir Lifschitz
Thursday, November 20, 4pm
Room 050, Bldg. 420
The talk will be based on my paper "On the declarative semantics of
logic programs with negation". A few copies of the paper are available
in my office, MJH 362.
ABSTRACT. A logic program can be viewed as a predicate formula, and its
declarative meaning can be defined by specifying a certain Herbrand
model of that formula. For programs without negation, this model is
defined either as the Herbrand model with the minimal set of positive
ground atoms, or, equivalently, as the minimal fixed point of a certain
operator associated with the formula (Van Emden and Kowalski). These
solutions do not apply to general logic programs, because a program
with negation may have many minimal Herbrand models, and the corresponding
operator may have many minimal fixed points. Apt, Blair and Walker and,
independently, Van Gelder, introduced a class of general logic programs
which disallow certain combinations of recursion and negation, and showed
how to use the fixed point approach to define a declarative semantics for
such programs. Using the concept of circumscription, we extend the minimal
model approach to stratified programs and show that it leads to the same
semantics.
∂19-Nov-86 1642 OKUNO@SUMEX-AIM.ARPA Catch and Qcatch
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 19 Nov 86 16:40:34 PST
Date: Wed 19 Nov 86 16:38:53-PST
From: Hiroshi "Gitchang" Okuno <Okuno@SUMEX-AIM.ARPA>
Subject: Catch and Qcatch
To: "@q.dis[1,clt]"@SU-AI.ARPA
Message-ID: <12256287656.48.OKUNO@SUMEX-AIM.ARPA>
Concerning Joe's Question 5:
Question 5: There is a third possibility: CATCH could return a value
normally (no THROW was executed), but some processes could still be
computing part of that value, as a result of eager QLETs. We will want
these processes to continue, but might not want to wait for their
values. Shouldn't there be a way to exit the CATCH but leave processes
running? I would argue that ordinary return from a CATCH shouldn't kill
any processes; that THROW should be the only way to kill something.
I agree with Joe. Moreover, if THROW is the only way to kill
something, QCATCH and CATCH are the same. That is, QCATCH is not
needed. If I'm wrong - QCATCH and CATCH are not the same, could you
give me an example? The example for CATCH in the RPG's paper didn't
explain the difference of QCATCH and CATCH.
(catch 'quit
(qlet t ((x
(do ((l l1 (cdr l)))
((null l) 'neither)
(cond ((p (car l))
(throw 'quit l1) ))))
(y
(do ((l l2 (cdr l)))
((null l) 'neither)
(cond ((p (car l))
(throw 'quit l2) )))))
x ))
[This program is written in TAO's style.]
The Multilisp approach that there is no way for one task to kill
another task (according to Joe) seems sound. However, I cannot escape
from the power of CATCH-THROW under QLISP interpretation. These
constructs will be useful for a concurrent expert system.
- Gitchang -
-------
∂19-Nov-86 2020 FULLERTON@Sierra.Stanford.EDU re: TV
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 19 Nov 86 20:20:43 PST
Date: Wed 19 Nov 86 20:17:09-PST
From: Dwain Fullerton <FULLERTON@Sierra.Stanford.EDU>
Subject: re: TV
To: JMC@Sail.Stanford.EDU
cc: FULLERTON@Sierra.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 19 Nov 86 19:08:00-PST
Message-ID: <12256327388.12.FULLERTON@Sierra.Stanford.EDU>
John,
Interesting on the amount, and at the very least your presence will
guarantee some solid content to the tape, as opposed to the flashy visuals.
Best,
Dwain
-------
∂19-Nov-86 2058 CHANDRA%OSU-20@ohio-state.ARPA Workshop on High Levl Tools
Received: from OHIO-STATE.ARPA by SAIL.STANFORD.EDU with TCP; 19 Nov 86 20:57:59 PST
Return-Path: <CHANDRA@OSU-20>
Received: from OSU-20 (osu-20.ARPA) by ohio-state.ARPA (4.12/6.1.OSU-CIS)
id AA24249; Wed, 19 Nov 86 22:48:11 est
Message-Id: <8611200348.AA24249@ohio-state.ARPA>
Date: Wed 19 Nov 86 22:49:24-EST
From: B. Chandrasekaran <Chandra%OSU-20@ohio-state.ARPA>
Subject: Workshop on High Levl Tools
To: JMC@SU-AI.arpa
Cc: Chandra%OSU-20@ohio-state.ARPA
The workshop, which AAAI supported to the tune of $5,000
(and DARPA and Ohio State supported to pick up the remaining approximately
$8,000 in costs), went off very successfully. About 80 papers
were submitted out of which about 25 were chosen and the two day
workshop was very productive. I expect that a workshop report
will be appearing in AI magazine one of these days. I need to know
what sort of formalities AAAI needs to close the books on this,
and what you, as Chair of the Committee that made the awards, would
personally like to see. I can send you the program and the preprints.
If you'd like me to do that, or if there is anything else you would
like to see, or I should be sending to AAAI, please let me know.
I will be soon requesting support for another workshop to be
held at the University of Maryland: I have been asked to be
Workshop Chairman with several people from a number of
different universities in the program committee. Before I
request any support for that one, I just wanted to make sure
that the requirements for the High Level Tools Workshop are
all deemed to be in order by AAAI.
-------
∂20-Nov-86 0747 JJW Re: Catch and Qcatch
To: Okuno@SUMEX-AIM.STANFORD.EDU
CC: Qlisp@SAIL.STANFORD.EDU
CATCH and QCATCH are still different, because QCATCH will wait until
all its descendant tasks are finished, while CATCH will return without
waiting.
I'm not advocating Multilisp's approach that says a task can't kill
another task (or cause that task to be killed in some way). I think
the ability to halt execution in another task is desirable in the case
of or-parallelism, or "speculative computation" as Halstead calls it
(and which he seems to be against).
Joe
∂20-Nov-86 0902 MAYR@Score.Stanford.EDU Re: industry lecturers and visiting faculty for 1987-88
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Nov 86 09:01:55 PST
Date: Thu 20 Nov 86 08:57:25-PST
From: Ernst W. Mayr <MAYR@Score.Stanford.EDU>
Subject: Re: industry lecturers and visiting faculty for 1987-88
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 19 Nov 86 21:20:00-PST
Message-ID: <12256465792.9.MAYR@Score.Stanford.EDU>
Cynthia Dwork, IBM Almaden research lab. I'll get you a course description
etc. within a few days. It'll be about distributed computing (I guess).
-ernst
-------
∂20-Nov-86 1006 CHANDRA%OSU-20@ohio-state.ARPA re: Workshop on High Levl Tools
Received: from OHIO-STATE.ARPA by SAIL.STANFORD.EDU with TCP; 20 Nov 86 10:06:07 PST
Return-Path: <CHANDRA%OSU-20@ohio-state.ARPA>
Received: from OSU-20 (osu-20.ARPA) by ohio-state.ARPA (4.12/6.1.OSU-CIS)
id AA00399; Thu, 20 Nov 86 13:04:46 est
Message-Id: <8611201804.AA00399@ohio-state.ARPA>
Date: Thu 20 Nov 86 09:13:57-EST
From: B. Chandrasekaran <Chandra%OSU-20@ohio-state.ARPA>
Subject: re: Workshop on High Levl Tools
To: JMC@SAIL.STANFORD.EDU
Cc: Chandra%OSU-20@ohio-state.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 19 Nov 86 21:18:00-EST
Thanks, I'll get in touch w. Claudia.
-------
Nils:
The following is the only proposal for visiting faculty,
apart from industry lecturers, that I have received so far. Terry
expects to interact with Nygaard and so does Dave Unger in CSL.
Nygaard is one of the inventors of Simula, and we haven't had
anyone with that point of view which originated object-oriented
programming. Therefore, I suggest that you negotiate with him
with the aid of co-ordinates to be supplied by Terry.
∂20-Nov-86 1034 WINOGRAD@CSLI.STANFORD.EDU Re: industry lecturers and visiting faculty for 1987-88
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Nov 86 10:34:43 PST
Date: Thu 20 Nov 86 10:33:00-PST
From: Terry Winograd <WINOGRAD@CSLI.STANFORD.EDU>
Subject: Re: industry lecturers and visiting faculty for 1987-88
To: JMC@SAIL.STANFORD.EDU
Cc: Winograd@CSLI.STANFORD.EDU
I have been discussing with Nils the possibility of having Kristen
Nygaard here for one quarter during the 77-78 year. He would teach a
course on object-oriented programming, focussing on his current work on
the Beta language. He would need one quarter's worth of regular faculty
salary support. Are you the right person to negotiate with about this
possibility? --t
∂20-Nov-86 1140 @REAGAN.AI.MIT.EDU:JCMA@AI.AI.MIT.EDU re: Seminar - Formalizing the Notion of Context (SU)
Received: from REAGAN.AI.MIT.EDU by SAIL.STANFORD.EDU with TCP; 20 Nov 86 11:40:20 PST
Received: from MORRISON.AI.MIT.EDU by REAGAN.AI.MIT.EDU via CHAOS with CHAOS-MAIL id 12117; Thu 20-Nov-86 14:39:27 EST
Date: Thu, 20 Nov 86 14:39 EST
From: John C. Mallery <JCMA@AI.AI.MIT.EDU>
Subject: re: Seminar - Formalizing the Notion of Context (SU)
To: JMC@SAIL.STANFORD.EDU
In-Reply-To: The message of 19 Nov 86 22:05 EST from John McCarthy <JMC@SAIL.STANFORD.EDU>
Message-ID: <861120143926.4.JCMA@MORRISON.AI.MIT.EDU>
Date: 19 Nov 86 1905 PST
From: John McCarthy <JMC@SAIL.STANFORD.EDU>
[In reply to message sent Tue, 18 Nov 86 20:48 EST.]
There isn't yet, but I'll keep you on the list.
Thanks.
∂20-Nov-86 1157 VAL re: Przymusinski
[In reply to message rcvd 19-Nov-86 19:04-PT.]
I asked Ullman whether he'd like to "co-sponsor" Teodor's visit, and he replied
that he would be happy to let Teodor "use the NAIL! meeting as a forum". I meant
to ask him whether he would pay a part of Teodor's expenses; is this what I
did? And did Ullman mean to say that he didn't want to pay?
∂20-Nov-86 1346 RLG seminar
i have a conflict with this afternoon's seminar. How important do you
rate this particular seminar? (Logic Programs & Non-mon)
p.s. notes from last week are in progress right now and should be avail
this afternoon if all goes well.
∂20-Nov-86 1408 RICHARDSON@Score.Stanford.EDU CSD Booklet
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Nov 86 14:07:07 PST
Date: Thu 20 Nov 86 14:03:23-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: CSD Booklet
To: jmc@Sail.Stanford.EDU
Message-ID: <12256521491.18.RICHARDSON@Score.Stanford.EDU>
We are currently in the process of updating our CSD booklet and would
appreciate any additions/changes that you may have to the blurb below.
We would like to have these additions/changes from you by January 1,
1987. If I do not hear from you by that time I will assume that the
blurb is alright as it stands.
Thanks,
Anne
-----
@b{John McCarthy},
@i(Professor of Computer Science.)@\
Prof. McCarthy's 1986 research interests are mainly in artificial intelligence,
especially the formalization of common sense knowledge of how to achieve
goals. In the late 1970s McCarthy developed a method of non-monotonic
reasoning called circumscription and published papers on it in 1980 and
1986. He also has a project to develop a version of Lisp for parallel
processors called Qlisp. @comment[updated 1/86]
-------
∂20-Nov-86 1424 RA to do before I leave?
Is there anything you want me to do today before I leave for my vacation?
I called the communication office regarding their recommendation for a
telephone answering machine and they will call me back with information.
I hope they do it today so that I can order it before I leave.
∂20-Nov-86 1456 CHIOU@STAR.STANFORD.EDU request funding for Space Station Automation Workshop
Received: from STAR.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Nov 86 14:54:46 PST
Date: Thu 20 Nov 86 14:53:39-PST
From: Wun Chiou <CHIOU@STAR.STANFORD.EDU>
Subject: request funding for Space Station Automation Workshop
To: jmc@SU-AI.ARPA
Cc: chiou@STAR.STANFORD.EDU
Message-ID: <VAX-MM(194)+TOPSLIB(120)+PONY(0) 20-Nov-86 14:53:39.STAR.STANFORD.EDU>
Professor McCarthy:
Claudia Mazzetti wants me to amend more information on the request letter
I sent you on November 14 concerning co-sponsorship again and funding
for the Space Station Automation workshop next year.
1. The workshop is unique in its kind and we need to have AI researchers
actively participate in the design phase of our Space Station.
2. Researchers from NASA have in the last two workshops enthusiastically
participated. The same group of co-organizers will be jointing me
again to conduct the workshop.
3. This workshop will be a day ahead of the SPIE advances in Intelligent
Robotics Systems conference. The Automation and Robotics Council of
the IEEE will also endorse the workshop. One of the reasons using SPIE
is for the administrative support.
4. The request for the funding (up to 10K) is strictly for the purpose of
reimburse whose researchers from the universities and nonprofit institutes.
No money will be used for honorariums. This is one way to encourage
the participation from the academia.
5. We will like very much to invite you to be our Keynote speaker next year.
This year's keynote speaker is Dr. Aaron Cohen, director of NASA Johnson
Space Center.
Thanks
Wun
-------
∂20-Nov-86 1510 WINOGRAD@CSLI.STANFORD.EDU re: industry lecturers and visiting faculty for 1987-88
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Nov 86 15:08:24 PST
Date: Thu 20 Nov 86 15:06:00-PST
From: Terry Winograd <WINOGRAD@CSLI.STANFORD.EDU>
Subject: re: industry lecturers and visiting faculty for 1987-88
To: JMC@SAIL.STANFORD.EDU
Cc: Winograd@CSLI.STANFORD.EDU
Fine, what's the deadline? --t
∂20-Nov-86 1529 CLT Qlisp meeting reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
Our next meetng will be Friday Dec 5.
Probably topic - semantics of Qlisp primitives
∂20-Nov-86 1549 RLG notes on context seminar
i have about half of the contexts notes typed in. When i have them all
typed in, i will print out a copy for you and one for val and deliver them.
in the meantime, if you wish to benefit from the half i have done, it is
in the file CONTXT.TXT on [1,RLG]
p.s. typing them in is slower than i thought it would be because it is also
more beneficial to my understanding than i expected--i understand more
as i rework the notes.
∂21-Nov-86 0049 ILAN@Score.Stanford.EDU isolde
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Nov 86 00:49:09 PST
Date: Fri 21 Nov 86 00:47:30-PST
From: Ilan Vardi <ILAN@Score.Stanford.EDU>
Subject: isolde
To: jmc@Score.Stanford.EDU
Message-ID: <12256638749.9.ILAN@Score.Stanford.EDU>
\documentstyle[12pt]{article}
\title{Is technical typing becoming a lost art?}
\author{Ilan Vardi}
\maketitle
\begin{document}
\pagestyle{empty}
Here at Stanford the computer is
rapidly replacing the traditional typewriter for technical
typing, but as
often is the case an advance in technology means the loss of a
traditional craft.
A case
in point is Isolde Field, a secretary with the Mathematics Department
for the last 26 years. Isolde has refined her skill to such an extent
that she is able to do complicated mathematical typing as fast
as people using computer typesetting. To do this, Isolde
has ``Hot-Rodded'' her typewriter to her own specifications. She
lays out special keys on her desk and uses them to type
mathematical symbols. Whenever a symbol is needed she
picks up the correct key from her desk and replaces a key from the
typewriter with it. Isolde has become so good at this that she no
longer needs to look down to tell where the keys are and where to put
them.
To speed up the system she has removed the
``hood'' from the typewriter,
and the machine sits on her desk much like a dragster
ready for the quarter mile.
Her typewriter is a heavy duty model and has only broken down once
in the last sixteen years. She prefers this kind
to the Selectric because the other secretaries keep having
to get the typeballs replaced on their IBM's.
Isolde believes that it takes about as much time for her
to change keys as it takes a wordprocessor to write the special commands,
plus she doesn't have to tire herself by looking at a computer screen all day.
Just having finished typing a 112 page paper
including corrections in a week and a
half she seems to prove her point quite well.
Isolde admits that the computer is better
for editing, but she adds that the mathematicians whose papers she types
are pretty good at getting it right the first time.
Time is catching up, however, and except for Isolde and Priscilla Feigen,
the secretaries at the Mathematics Department are using
computer typesetting (Knuth's \TeX and Milgram's TECNO-TYPE, both
written at Stanford) for technical papers. Isolde says that when she
retires, she's going to take her typewriter with her. It seems that
she will also be one of the last with her great skill.
\end{document}
-------
∂21-Nov-86 0741 RICHARDSON@Score.Stanford.EDU re: CSD Booklet
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Nov 86 07:41:04 PST
Date: Fri 21 Nov 86 07:39:26-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: re: CSD Booklet
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Thu 20 Nov 86 19:04:00-PST
Message-ID: <12256713738.12.RICHARDSON@Score.Stanford.EDU>
Thanks!
-Anne
-------
∂21-Nov-86 0923 MAYR@Score.Stanford.EDU industrial lectureship
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Nov 86 09:23:51 PST
Date: Fri 21 Nov 86 09:22:11-PST
From: Ernst W. Mayr <MAYR@Score.Stanford.EDU>
Subject: industrial lectureship
To: jmc@Sail.Stanford.EDU
Message-ID: <12256732443.38.MAYR@Score.Stanford.EDU>
20-Nov-86 17:12:02-PST,916;000000000001
Return-Path: <dwork%ALMVMA.BITNET@WISCVM.WISC.EDU>
Received: from WISCVM.WISC.EDU (IBM-ETHER.WISC.EDU.#Internet) by SU-SCORE.ARPA with TCP; Thu 20 Nov 86 17:11:56-PST
Received: from (DWORK)ALMVMA.BITNET by WISCVM.WISC.EDU on 11/19/86 at
17:58:11 CST
Date: Wed 19 November 1986
From: dwork%almvma.bitnet@wiscvm.arpa
To: mayr@su-score
Hi, Ernst. I have chosen the following keywords for my projected
course.
fault-tolerant distributed computing, cryptographic protocols,
randomness, interactive proof systems, knowledge
I haven't yet prepared a real course description or outline, but
I will be emphasizing applications of randomization and
cryptographic techniques to fault-tolerant distributed computing.
Please let me know if you require more specifics.
I would like to teach in the 2nd quarter (called "winter"?).
Again, thank you for your help arranging this.
Cynthia Dwork
-------
∂21-Nov-86 1044 VAL Counterexample
Given program:
p(1)
q(2) (1)
r(x)←¬q(x)
After your transformation:
natnum(0)
natnum(s(n))←natnum(n)
p'(1,n)←natnum(n)
q'(2,n)←natnum(n) (2)
r'(x,s(n))←¬q'(x,n),natnum(n)
p(x)←p'(x,n),natnum(n)
q(x)←q'(x,n),natnum(n)
r(x)←r'(x,n),natnum(n)
Claim. In every minimal Herbrand model of (2) q(1) is false.
Refutation. The following model is a counterexample:
natnum(n) ≡ n=0 ∨ n=s(0) ∨ n=s(s(0)) ∨ ...,
p'(x,n) ≡ x=1 ∧ natnum(n),
q'(x,n) ≡ natnum(n),
r'(x,n) ≡ false,
p(x) ≡ x=1,
q(x) ≡ true,
r(x) ≡ false.
∂21-Nov-86 1233 AIR RT and ANDREW
Do you have parts of ANDREW manual?
Arkady
∂22-Nov-86 1415 RLG contexts notes
they are done and a copy is on your office door...
∂23-Nov-86 1659 NILSSON@Score.Stanford.EDU AI Quals
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 23 Nov 86 16:59:41 PST
Date: Sun 23 Nov 86 16:57:44-PST
From: Nils Nilsson <NILSSON@Score.Stanford.EDU>
Subject: AI Quals
To: jmc@Sail.Stanford.EDU
cc: nilsson@Score.Stanford.EDU
Message-ID: <12257339662.22.NILSSON@Score.Stanford.EDU>
John, How about you running the AI quals this Spring? (Running them
mainly means getting us all to agree on what the format and content
will be; advertising this to the students; setting a time in May;
and then participating as one of the examiners.) One reason for asking
you is that I know you had ideas about what ought to be asked that differed
from how the quals have been handled during the last few years, and it's
perhaps time for you to have your impact on them. My office can help
with administrative arrangements if you are willing to take the lead
on technical matters. -Nils
-------
∂23-Nov-86 2203 hpm@rover.ri.cmu.edu "Mind Children" book
Received: from ROVER.RI.CMU.EDU by SAIL.STANFORD.EDU with TCP; 23 Nov 86 22:03:25 PST
Date: 24 Nov 1986 00:41-EST
From: Hans.Moravec@rover.ri.cmu.edu
To: jmc@su-ai
Subject: "Mind Children" book
Message-Id: <533194883/hpm@rover.ri.cmu.edu>
Dear John,
I took the liberty of invoking your name. "Mind Children" is
the book that grew out of the essay that more or less finished off my
AI qual ten years ago. I sent you a very partial draft many months
ago. I understand my "animal" argument differs fundamentally from
your "reasoning" approach to building a mind, but I'm more convinced
than ever that they're both necessary parts of the solution.
Anyway, my editor would like get the book on Harvard's
high energy track. Doing so requires approval by an editorial
committee - a few endorsements by notables would assure that -
thus the invokation. You'll be receiving a copy of the draft
from Harvard by next week, probably. This draft, by the way,
is without four chapters of speculative physics that was in there
originally ("Harvard doesn't publish science fiction ...").
Oh well - best wishes and thanks -- Hans
∂24-Nov-86 0742 CLT igor
To: JMC, LES
Here is his current state of mind. Re his question about travel funds
for the AMS meeting I assume they can be provided (if he is actually
on board by January)?
------------------------------------------------------------------------------
∂22-Nov-86 1248 rivin@STONY-BROOK.SCRC.Symbolics.COM life
Received: from SCRC-STONY-BROOK.ARPA by SAIL.STANFORD.EDU with TCP; 22 Nov 86 12:48:21 PST
Received: from GANNET.SCRC.Symbolics.COM by STONY-BROOK.SCRC.Symbolics.COM via CHAOS with CHAOS-MAIL id 7457; Sat 22-Nov-86 15:46:21 EST
Date: Sat, 22 Nov 86 15:46 EST
From: igor rivin <rivin@ALLEGHENY.SCRC.Symbolics.COM>
Subject: life
To: clt@SU-AI.ARPA
Message-ID: <861122154616.3.RIVIN@GANNET.SCRC.Symbolics.COM>
Hi. At this point it looks pretty certain that I will accept your offer.
The only reason I cannot make an even more positive statement is that
there is local political garbage going on, and the MACSYMA group needs my
support, which cannot be provided if I don't work here. Sigh...
This should all be over by Tuesday or so.
I do have one question -- I have been invited to give a talk at the
American Math Society Winter Meeting (San Antonio, Jan 21-24). The
question is how would I go about getting travel funds for this sort of
thing (presumably the timing could be arranged so that I am officially at
Stanford at that time, under the further assumption that I accept your
offer on Tuesday...).
Thanks.
Igor.
∂24-Nov-86 1208 ceci@portia office hours
Received: from PORTIA by SAIL.STANFORD.EDU with TCP; 24 Nov 86 12:08:50 PST
Received: by portia; Mon, 24 Nov 86 12:06:09 PST
Date: Mon, 24 Nov 86 12:06:09 PST
From: Cecilia Castillo <ceci@portia>
Subject: office hours
To: jmc@sail.STANFORD.EDU
24.nov.86
12:00pm
Professor McCarthy,
I would like to come in and see you about getting your initials on
my checksheet. I need your initials so that I will not have to repeat
a class that I have already taken. I would like to waive Statistics 116,
I had 2 quarters of Probability Theory at UCLA using the book -
Introduction to Probability Theory by Hoel, Port and Stone.
When would be a convenient time to come and see you?
Cecilia Castillo
ceci@sushi
∂24-Nov-86 1944 LES DARPA Proposal
To: JMC
CC: CLT
I reach Bob Simpson at DARPA late this afternoon and asked him if he had
received the draft proposal. He said that he had but had not had a chance
to read it yet. While we were talking on the phone, he leafed through
it, looked at the budget, and gave me the quick response that he had been
thinking in terms of a smaller amount, on order of $500k/year instead of
$1 million/year.
He also remarked that getting more funding might be facilitated by carving
the proposal into smaller pieces and selling some of them to other program
managers, such as Steve Squires. He mentioned specifically that Squires
has responsibility for the more theoretical AI work while he (Simpson) is
responsible for the more applied projects. He mentioned that one of the
things that he wants out of most of his programs is a deliverable
prototype (usually software) that can be tied to a military program and
delivered to a government laboratory.
He said that he would look over the proposal more carefully and give us
more complete feedback soon.
Though I didn't discuss this with Simpson, I am led to wonder whether we
are dealing with the right guy -- most of our research appears to mesh
better with Squires' alleged interests than with Simpson's. Would you
like to poke Squires with this question or would you like me to?
∂24-Nov-86 2131 CLT DARPA Proposal
To: LES, JMC
Where does Scherlis fit into the hierarchy of Program managers?
He should be interested in the MTC part anyway.
∂24-Nov-86 2209 OKUNO@SUMEX-AIM.ARPA One more question and Re: Qcatch and Catch
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Nov 86 22:08:42 PST
Date: Mon 24 Nov 86 22:07:45-PST
From: Hiroshi "Gitchang" Okuno <Okuno@SUMEX-AIM.ARPA>
Subject: One more question and Re: Qcatch and Catch
To: JJW@SU-AI.ARPA, qlisp@SU-AI.ARPA
In-Reply-To: Message from "Joe Weening <JJW@SAIL.STANFORD.EDU>" of Thu 20 Nov 86 07:47:00-PST
Message-ID: <12257658243.24.OKUNO@SUMEX-AIM.ARPA>
Joe,
Yes, CATCH and QCATCH are still different. However, I think that the
capability of QLISP to wait until all its descendant task are finished
is somewhat overloaded. That is, QCATCH has a capability of
synchronization. On the other hand, QLET and QLAMBDA under EAGER
evaluation don't specify the synchornization as you pointed out.
What I meant in the previous message is that QCATCH may be thrown away
and new construct for synchornization will be introduced.
----------------
I have one more question to all people of QLISP mailing list:
My Question: Is process switching more expensive than process
creation/initialization?
The Advanced Architectures Project at KSL assumes that "switching
between stack groups is certainly an expensive operation in simulation
and may also be so in a target machine". However, I cannot believe
this assumption. This assumption is true on Symbolics/TI-Explorer.
The TAO/ELIS experience shows that it takes about 40 micro seconds for
process switching and much faster than process creation. (TAO/ELIS
system adopts deep-binding mechanism for variable binding for fast
process switching.) Or, TSS systems seem to me to assume that process
switching is less expensive.
Am I wrong?
Thanks,
- Gitchang -
-------
∂24-Nov-86 2227 @RELAY.CS.NET,@utterly.ai.toronto.edu:hector@AI.AI.MIT.EDU Update on McDermott critique
Received: from RELAY.CS.NET by SAIL.STANFORD.EDU with TCP; 24 Nov 86 22:27:31 PST
Received: from toronto.edu by csnet-relay.csnet id ab04970; 25 Nov 86 1:22 EST
Received: from utterly.ai.toronto.edu by ai.toronto.edu id AA06212; Mon, 24 Nov 86 20:08:54 est
Received: from ai by utterly.ai.toronto.edu via ETHER with SMTP id AA04423; Mon, 24 Nov 86 20:12:18 EST
Message-Id: <8611250112.AA04423@utterly.ai.toronto.edu>
Date: 24 Nov 86 20:12:08 EST (Mon)
From: Hector Levesque <hector%ai.toronto.edu@RELAY.CS.NET>
To: james@ROCHESTER.ARPA, ec%brown.csnet@RELAY.CS.NET,
JON.DOYLE%CMU-CS-A.ARPA@RELAY.CS.NET, jon.doyle%c.cs.cmu.edu@RELAY.CS.NET,
forbus@A.CS.UIUC.EDU, hinton%cmu-cs-a%ai.toronto.edu@RELAY.CS.NET,
hobbs@SRI-WARBUCKS.ARPA, israel@SRI-STRIPE.ARPA, jmc@SU-AI.ARPA,
val@SU-AI.ARPA, nilsson@SU-SCORE.ARPA, pentland@SRI-STRIPE.ARPA,
watdaisy!dlpoole%watmath.uucp@RELAY.CS.NET,
reiter%utai.toronto.edu@RELAY.CS.NET, stickel@SRI-STRIPE.ARPA,
tyson@SRI-STRIPE.ARPA, tw@CSLI.STANFORD.EDU
Subject: Update on McDermott critique
Nothing to report, dammit. I'm still waiting for the contribution of a number
of people (15, at last count). I have promises from most of them, though.
Anyway, I have yours and thanks. I will keep you up to date as (or if) things
develops.
Hector
∂24-Nov-86 2334 HSU@Sushi.Stanford.EDU cs306
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Nov 86 23:34:06 PST
Date: Mon 24 Nov 86 23:32:22-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: cs306
To: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12257673647.17.HSU@Sushi.Stanford.EDU>
Prof. McCarthy:
Is the assignment on EKL proofs the last one for this course? I'm
considering whether I should extend the due date for this assignment
for two days (to 12/4 instead of 12/2).
Yung-jen
-------
∂25-Nov-86 0206 binford@su-whitney.arpa references
Received: from WHITNEY.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 25 Nov 86 01:57:14 PST
Received: by su-whitney.arpa with Sendmail; Tue, 25 Nov 86 01:56:20 pst
Date: Tue, 25 Nov 86 01:56:20 pst
From: Tom Binford <binford@su-whitney.ARPA>
To: jmc@su-ai
Subject: references
John
Here is a list
Prof. Patrick Winston, MIT
Prof. Berthold Horn, MIT
Prof. Mike Brady, Oxford
Prof. Ruzena Bajcsy, chairperson, CSD, Univ of Penn
Prof. Ram Nevatia, USC
Tom
∂25-Nov-86 0538 JJW Process switching
To: Qlisp@SAIL.STANFORD.EDU
I remember this coming up during a discussion at an AAP meeting about two
years ago, and RPG and I were both a bit surprised that this model of
process switching was being adopted. Certainly it's true with
Symbolics/TI stack groups, since the dynamic context needs to be
re-established whenever you switch processes. (They use shallow binding,
I presume.)
As you point out, deep binding gets around this problem, and deep binding
also may be necessary to implement a reasonable semantics for dynamic
variables in a multiprocessor. This is a good topic for us to think about
right now, in fact. Lexically bound variables cause essentially no
overhead during process switching, so they're the best and I think we want
to encourage a lexical style of programming. There's actually some
"hidden" overhead associated with all process switching, including lexical
variables, because of the cache context that gets destroyed and slows down
the new process while it gets built up again.
So I think we're pretty much going on the assumption that process
switching is cheaper than process creation.
Another can of worms to open is the issue of preemptive vs. non-preemptive
scheduling, or "time sharing" between runnable processes. Does each
process stay running until it terminates or suspends as defined by the
language, or can the runtime system arbitarily suspend processes to give
others some time to run? The latter approach is more "fair" to the
processes; it lets some programs terminate that might otherwise not and
might have other properties that help in the analysis of programs; but it
probably has more overhead.
∂25-Nov-86 0837 AI.JMC@MCC.COM Franklin's whale
Received: from MCC.COM by SAIL.STANFORD.EDU with TCP; 25 Nov 86 08:37:33 PST
Date: Tue 25 Nov 86 10:37:18-CST
From: John McCarthy <AI.JMC@MCC.COM>
Subject: Franklin's whale
To: jmc@SU-AI.ARPA
Message-ID: <12257772848.34.AI.JMC@MCC.COM>
It seems that ``Franklin's Whale'', sometimes called the
Niagara whale, was not really a whale at all
but a relative of the salmon. Its large size and the fact that
it was prized for an oil rather like sperm whale oil (whose biological
function is unknown) gave it the name. It seems to have become extinct
around 1780, one of the earliest American victims of the white man's
greed and rapacity. Although larger specimens were sometimes reported,
none of the Massachusetts whalers' account books so far examined
lists any longer than 20 feet. The specimen in the British museum
is 16 feet long and the one in the Smithsonian is 14 feet. Neither
has been exhibited in more than a century, presumably because
they are much deteriorated, as are almost all fish preserved by
eighteenth century metheds of taxidermy.
-------
∂25-Nov-86 1230 shoham@YALE.ARPA Re: reference to chronological minimization
Received: from YALE-BULLDOG.ARPA by SAIL.STANFORD.EDU with TCP; 25 Nov 86 12:30:04 PST
Received: by Yale-Bulldog.YALE.ARPA; 25 Nov 86 10:17:40 EST (Tue)
Date: 25 Nov 86 10:17:40 EST (Tue)
From: Yoav Shoham <shoham@YALE.ARPA>
Message-Id: <8611251517.AA03623@Yale-Bulldog.YALE.ARPA>
Subject: Re: reference to chronological minimization
To: John McCarthy <JMC@SAIL.STANFORD.EDU>
In-Reply-To: John McCarthy <JMC@SAIL.STANFORD.EDU>, 21 Nov 86 1503 PST
I may need to use my syntactic formulation of chronological minimization
in a paper (in a very preliminary state). If what I use still turns out
to resemble chronological minimization when I get done tinkering with it I
would like to give you credit for the concept by referring to your
forthcoming thesis. I hope that's ok.
Of course. And I'd be interested in seeing the paper.
-------
∂25-Nov-86 1403 AAAI-OFFICE@SUMEX-AIM.ARPA Mathilda R. Wilson Foundation
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 25 Nov 86 14:03:23 PST
Date: Tue 25 Nov 86 13:59:01-PST
From: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: Mathilda R. Wilson Foundation
To: MCK%mit-oz@MC.LCS.MIT.EDU, Reddy@A.CS.CMU.EDU, ai.woody@MCC.COM
cc: aaaI-OFFICE@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Message-ID: <12257831416.69.AAAI-OFFICE@SUMEX-AIM.ARPA>
I just received a phone call from Frederick Nash, President, of
Mathilda Wilson Foundation (widow of Dodge) offering financial
support for post-graduate research for young university professors.
Nash's partner, Pierre Hefler, originally contacted McCarthy regarding
their intent and John advised him to contact this office and the
officers to investigate the possibility of the AAAI acting as
the administrators of these monies. (Their IRS status does not
allow them to give monies to individuals.)
Nash informed me that the foundation has about $1M in income each year
and could contribute a large share of its yearly income to an area
which they felt was important. They felt supporting AI basic research
was very important. AN example of the level of their support is the
foundation's contribution to a chair in veternary medicine at Michigan
State for $350,000 a year. The AAAI could conceivably receive the
same amount. Alternatively, the AAAI could select those candidates and
then the foundation could give their appropriate institution the
amount to administer.
I thought I should forwarn you because Mr. Nash (partner in the law
firm of Botman, Langley and Dahling in Detroit) will be calling you
about this proposal. I think it is a worthly proposal.
Cheers,
Claudia
-------
∂26-Nov-86 0629 AI.WOODY@MCC.COM Re: Mathilda R. Wilson Foundation
Received: from MCC.COM by SAIL.STANFORD.EDU with TCP; 26 Nov 86 06:29:18 PST
Date: Wed 26 Nov 86 08:22:29-CST
From: Woody Bledsoe <AI.Woody@MCC.COM>
Subject: Re: Mathilda R. Wilson Foundation
To: AAAI-OFFICE@SUMEX-AIM.ARPA, MCK%mit-oz@MC.LCS.MIT.EDU, Reddy@A.CS.CMU.EDU
cc: jmc@SU-AI.ARPA, AI.Woody@MCC.COM
In-Reply-To: <12257831416.69.AAAI-OFFICE@SUMEX-AIM.ARPA>
Message-ID: <12258010451.76.AI.WOODY@MCC.COM>
This sounds good to me. Of course, it would take some kind of
AAAI committee to administer it but the effort would be worthwhile.
Woody
-------
∂26-Nov-86 0847 M.MRIZEK@LEAR.STANFORD.EDU LOTS 10th Birthday Celebration
Received: from LEAR.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 26 Nov 86 08:47:39 PST
Date: Wed 26 Nov 86 08:44:03-PST
From: Lori Mrizek <M.MRIZEK@LEAR.STANFORD.EDU>
Subject: LOTS 10th Birthday Celebration
To: jmc@SU-AI.ARPA
cc: m.mrizek@LEAR.STANFORD.EDU
Message-ID: <12258036222.19.M.MRIZEK@LEAR.STANFORD.EDU>
Professor McCarthy,
Ralph and Queenette would like me to get some dates from you in January
that you would be available to attend the LOTS 10th Birthday Celebration.
We are considering January 23rd, from 3-5 p.m. Would you be able to
attend on that date? Please give me a few more dates you would be
available in case the 23rd does not work out. Thank you.
Lori Ann Mrizek
3-5279 or m.mrizek@othello
-------
∂26-Nov-86 1139 rms@prep.ai.mit.edu
Received: from PREP.AI.MIT.EDU by SAIL.STANFORD.EDU with TCP; 26 Nov 86 11:39:35 PST
Received: by prep.ai.mit.edu; Wed, 26 Nov 86 14:02:29 EST
Date: Wed, 26 Nov 86 14:02:29 EST
From: rms@prep.ai.mit.edu (Richard M. Stallman)
Message-Id: <8611261902.AA00666@prep.ai.mit.edu>
To: jmc@su-ai.arpa
1. How is your project progressing? Did you decide to work with GNU Emacs?
2. What did the bureaucrats think of GNU?
I sent mail to someone (I forget who) who is working on a version of
GNU Emacs for editing Japanese. I suggested using the representation
I worked out for you, so that it would at the same time become able
to handle any other characters.
∂26-Nov-86 1158 RICHARDSON@Score.Stanford.EDU AI Courses
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 26 Nov 86 11:57:54 PST
Date: Wed 26 Nov 86 11:55:58-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: AI Courses
To: ai.list: ;
Message-ID: <12258071158.11.RICHARDSON@Score.Stanford.EDU>
I am trying to put together a meeting of the AI faculty during the first
or second week of December to discuss courses for next year. Please let
me know your availability for those two weeks.
Thanks,
Anne
-------
∂26-Nov-86 1316 AAAI-OFFICE@SUMEX-AIM.ARPA Re: new Dahl proposal
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 26 Nov 86 13:16:00 PST
Date: Wed 26 Nov 86 13:05:47-PST
From: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: Re: new Dahl proposal
To: JMC@SU-AI.ARPA
cc: AAAI-OFFICE@SUMEX-AIM.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 26 Nov 86 12:58:00-PST
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Message-ID: <12258083869.13.AAAI-OFFICE@SUMEX-AIM.ARPA>
I don't think I have the page about the proposal. Can you send it
to me?
Thanks,
Claudia
-------
∂26-Nov-86 1537 rms@prep.ai.mit.edu reply to message
Received: from PREP.AI.MIT.EDU by SAIL.STANFORD.EDU with TCP; 26 Nov 86 15:36:14 PST
Received: by prep.ai.mit.edu; Wed, 26 Nov 86 18:26:36 EST
Date: Wed, 26 Nov 86 18:26:36 EST
From: rms@prep.ai.mit.edu (Richard M. Stallman)
Message-Id: <8611262326.AA02529@prep.ai.mit.edu>
To: JMC@sail.stanford.edu
In-Reply-To: John McCarthy's message of 26 Nov 86 1353 PST <8611262309.AA02399@prep.ai.mit.edu>
Subject: reply to message
You were going to show the GNU Emacs manual to some bureaucrats and see
if they wanted to support the GNU project.
∂26-Nov-86 1537 rms@prep.ai.mit.edu reply to message
Received: from PREP.AI.MIT.EDU by SAIL.STANFORD.EDU with TCP; 26 Nov 86 15:36:55 PST
Received: by prep.ai.mit.edu; Wed, 26 Nov 86 18:25:41 EST
Date: Wed, 26 Nov 86 18:25:41 EST
From: rms@prep.ai.mit.edu (Richard M. Stallman)
Message-Id: <8611262325.AA02526@prep.ai.mit.edu>
To: JMC@sail.stanford.edu
Cc: AIR@sail.stanford.edu, CLT@sail.stanford.edu
In-Reply-To: John McCarthy's message of 26 Nov 86 1353 PST <8611262309.AA02399@prep.ai.mit.edu>
Subject: reply to message
By "editor part including editor-based Unix commands", do you mean
the Lisp-style syntax for invoking inferior processes and passing
arguments, etc?
I'm not sure whether it is any easier to use GNU Emacs Lisp for this than
to use Franz or some other Lisp that has more documentation available.
Just for demoing the capability of writing shell-like commands in that
way, does it matter whether you do it in an editor?
If it does, so you want to use GNU Emacs Lisp, there is a certain
amount of documentation. You can ask rlk@athena.mit.edu to send you a
copy of the chapter he wrote. I will try to find and send you an
introduction to Emacs extension someone wrote which talks a lot
about how to use the built-in help to find what you need to know
to write an extension.
It also helps to remember that I followed Maclisp and Lisp machine conventions
a lot. Look at a Maclisp manual and use C-h C-f to see if the function
you want is present in GNU Emacs.
The program etc/digest-doc in the Emacs distribution uses the DOC file
to create a printable summary of documentation strings.
For this application you won't have to interface much with Emacs
editing primitives. You can assume that the users command is
provided as list structure. See the definition of the command
eval-expression, probably in simple.el, for the easy way to set
up a user command that will read the list structure and pass it
to the rest.
You will need the subprocess primitives a lot. So read the source
file src/process.c, first just reading the names of the subrs and their
documentation strings, and then reading the C code to see how they
work. I'm not sure whether you will need to add any primitive
facilities. If you want to be able to pas streams to subprocesses
other than in the standard ways, you will have to change the C code in
that file.
What these things won't answer, ask me.
∂26-Nov-86 2000 JMC
Norm Howard
∂26-Nov-86 2100 JMC
gry
∂26-Nov-86 2100 JMC
Searle and Dreyfus
∂26-Nov-86 2314 SINGH@Sierra.Stanford.EDU Re: Lynch mob spirit
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 26 Nov 86 23:14:30 PST
Date: Wed 26 Nov 86 23:10:56-PST
From: Harinder Singh <SINGH@Sierra.Stanford.EDU>
Subject: Re: Lynch mob spirit
To: JMC@Sail.Stanford.EDU
cc: su-etc@Sail.Stanford.EDU, SINGH@Sierra.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 26 Nov 86 14:37:00-PST
Message-ID: <12258194033.11.SINGH@Sierra.Stanford.EDU>
I don't have much to say on the specific issues under discussion
but I do agree wholeheartedly with Prof. McCarthy's general observations
about a number of `intelligent' people and assemblies thereof.
From my own experience I've found that mere intelligence does not
guarantee very much - super smart people are quite abundant, especially
around here, but whether they have even any *intellectual* honesty to offer
is often an open question. Other things like personal warmth, integrity,
compassion, etc are another story altogether, being more subjective
parameters.
Truly, there *is* such a thing as being *over*-educated (sort of
like boiling vegetables like hell to cook them)! Personal values aside,
I'm sure many people have come across objectively verifiable instances
of very intelligent and educated people trying to bluff and con their
way out of commitments and such. (Not to suggest, of course, that one must
be educated in order to be a liar or a crook.)
Two of my favorite `shining examples' of personal honesty, integrity
and un-compromisingly clear thinking are of virtually *illiterate* and poor
individuals. One was an itinerant cobbler's son, maybe about ten years old,
and the other was a spastic man selling leather goods from a wheelchair on a
New Delhi sidewalk.
The detailed stories must wait. Suffice it to say here that, to me,
their examples put the convoluted, super-clever sophistry of many a `well
educated' person to shame and have served me on occasion as standards for
action.
---------------------------------
PS I realize that this message may be going farther afield than
may have been intended by Prof. McCarthy. I'm happy to let it do so!
-------
∂28-Nov-86 1317 WINSLETT@Score.Stanford.EDU Reagan Library
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 28 Nov 86 13:17:52 PST
Date: Fri 28 Nov 86 13:16:06-PST
From: Marianne Winslett <WINSLETT@Score.Stanford.EDU>
Subject: Reagan Library
To: jmc@Sail.Stanford.EDU
Message-ID: <12258610034.15.WINSLETT@Score.Stanford.EDU>
You are right; all the objections that I have read have been political,
and that does not seem sufficient reason to prevent installation of
a library here.
I would like to speak up for the other point of view, that is, that
putting a library there is a terrible waste of a beautiful spot.
I wouldn't mind if they put it down by Sand Hill and 280, but those
lovely hills should be left as they are for now. I feel pretty
strongly about this, and signed a petition against the plan for this
reason. At the time I thought most other signers shared my sentiments,
but perhaps they are mostly politically motivated.
--Marianne
BTW, politically I would describe myself as "Reagan-neutral". He
seems to polarize most people, doesn't he?
-------
∂28-Nov-86 1555 SJG re: closest possible world
[In reply to message rcvd 28-Nov-86 15:52-PT.]
Not necessarily. The closest possible world of "If I move the bookcase to l"
is the closest possible world where the "consequences" of "move(bookcase,l)"
hold. These consequences will be a list of the "direct" consequences of
the move action, and will surely include at(bookcase,l), but maybe more,
such as moved(bookcase).
Is this an answer, or are you asking something else? We haven't thought
about time much, in the belief that the discrete case needs to be
understood first .
∂28-Nov-86 1618 SJG re: closest possible world
[In reply to message rcvd 28-Nov-86 16:07-PT.]
OK, you're probably right. When we discuss the formalism, the
distinction is made clear, although it usually doesn't matter.
(but see p.20, middle and p.31, middle.)
You are surely right that we don't face the bird turd example (or the
potato example) any more squarely than the other default approaches.
Doing so would involve working in a default framework that allowed you
to notice that some conclusion was default true, and then just not
bother thinking about it any more (so that if it was provably false,
you might make a mistake), and no one has any formalism that will
manage this. My multi-valued logics are intended to do this, but
need more formal work if I'm going to make that claim sharp ...)
∂28-Nov-86 1632 SJG re: closest possible world
[In reply to message rcvd 28-Nov-86 16:28-PT.]
I'm not sure I agree with this. When the robot goes to start
the car, we don't want it to check the tailpipe. But when the
car doesn't start, we *do* want it to check the tailpipe. Are
you saying that there is some sort of built-in procedural
distinction? Should the robot check for a bomb under the hood?
(Maybe he should, if he's a Libyan diplomat.)
I think this is a much more subtle thing than can possible be
"hacked around" -- whether or not to check for a qualification
before taking an action seems to be a function of how important
it is if you're wrong, how likely the qualification is anyway, and
how expensive the test is. It would be sad if all of this had
to be procedurally encoded.
∂28-Nov-86 1637 SJG re: closest possible world
[In reply to message rcvd 28-Nov-86 16:31-PT.]
The paper has had an odd history; I've thought for the past two
months that it was done (I put dates on drafts only). But Dave
always found something else that needed rewriting.
[Aha. I think you have the most recent version anyway -- I meant
eq. (31) on p.23 -- sorry. In both cases, we are using an extended
consequence set to record the fact that the action has been carried
out. Do you have a more compelling example -- bear in mind we're
ignoring time -- of a situation where one cannot consider the
result of moving the bookcase to be the nearest possible world in
which it's at its new location? The newest version ends in the
middle of p.28, with reference 14. If that's not what you've
got, I'll be happy to give you one.]
∂28-Nov-86 2239 JSW Scope and Extent rules for Qlisp
To: Qlisp@SAIL.STANFORD.EDU
Looking through the Common Lisp manual [Steele Jr., Guy L., Common Lisp: The
Language, Digital Press, 1984] [CLtL] for potential problems in defining
Qlisp, I was suprised to find that there really aren't very many. The main
issue of interest is scope and extent, discussed in chapter 3 of [CLtL].
Common Lisp scope rules (lexical and indefinite scope) don't appear to
have any particularly bad interactions with multiprocessing. Indefinite
extent (used for lexical binding) also appears to be OK.
There is a problem with dynamic extent, which is discussed on page 38 of
[CLtL]. Namely, it is no longer the case that objects with dynamic extent
maintain a stack-like discipline. Consider for instance the following:
(defvar *a*) ;*a* is pervasively declared special
(defun fun ()
(let ((*a* 17))
(qlet 'eager ((x (crunch)))
(cons 'foo x))))
(defun crunch ()
(let ((*a* 29))
(waste-some-time))
(+ 3 *a*))
(setq *a* 5)
(fun)
A likely sequence of events is the following:
fun binds 17 to *a* (in main process)
crunch binds 29 to *a* (in new process)
fun unbinds *a*
crunch unbinds *a*
crunch adds 3 to *a* (what value should it get?)
The "reasonable" answer is that the *a* added to 3 at the end of crunch has
the value 17. This would have been the value in a serial execution, because
it would not yet have been unbound.
To define this precisely, we have to talk about each binding in the
context of a process. At any given time, some bindings may be established
in some processes but not in others. Establishing a binding in a process
does not affect any bindings in any other processes. When a new process
is created, it shares the (lexical and special) bindings of its parent.
Disestablishing a binding in a process does not effect other processes
that may share that binding.
We can get this semantics if we use an association list to hold dynamic
bindings; inherit the parent process's binding list when starting a
process; and allow garbage collection to reclaim entries of the list. Of
course such an implementation is too inefficient.
The same thing can be achieved with deep binding, if we keep track of possible
references to a binding by multiple processes, and only reclaim a binding cell
when no more processes refer to it. This requires keeping careful track of
process creation and termination.
It isn't clear whether shallow binding can achieve this semantics at all in a
reasonable way.
This takes care of special variable bindings. Other things that have dynamic
extent are:
a. CATCH/UNWIND-PROTECT: these are given a major overhaul in Qlisp already.
b. BLOCK exit points: this is like a lexical version of CATCH and THROW, but
has not yet been altered for Qlisp. Consider a case like the following:
(block here
(qlet t ((x (if (simplep a)
(return-from here (quick-answer a))
(foo a)))
(y (foo b)))
(work-on x y)))
Should the process computing Y be killed when the RETURN-FROM form is
evaluated in the process computing X?
If we change the QLET T to QLET 'EAGER, and if (WORK-ON X Y) returns a
value before the process computing X finishes, then we may execute a
RETURN-FROM in an illegal position (see [CLtL], p. 41). It would be
unfortunate to consider this an error, when in a sequential execution
there would not be an error, and since depending on the scheduling of
processes the program might or might not reach such a state.
c. TAGBODY and GO: These introduce a similar situation when a GO executed
in one process points to a label in another process. The questions are
whether to kill processes that are made useless because of the GO, and
what to do if the extent of the TAGBODY in the destination process has
already ended. We must certainly ensure that a single point of control
remains in each process, i.e., in
(tagbody
(spawn (go b))
(spawn (go c))
a (foo)
b (bar)
c (baz))
we don't want to end up in three places at once!
∂29-Nov-86 1157 SJG re: closest possible world
[In reply to message rcvd 28-Nov-86 18:11-PT.]
... and this is exactly what Dave and I allow for by saying that
an action can have a set of consequences. But in the text, where
we simply talk about the situation where the bookcase is moved to
some new location, we are assuming that the action is simpler:
the robot simply picks the bookcase up and relocates it.
∂29-Nov-86 1442 WINSLETT@Score.Stanford.EDU re: Reagan Library
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 29 Nov 86 14:42:09 PST
Date: Sat 29 Nov 86 14:40:21-PST
From: Marianne Winslett <WINSLETT@Score.Stanford.EDU>
Subject: re: Reagan Library
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 28 Nov 86 13:57:00-PST
Message-ID: <12258887515.9.WINSLETT@Score.Stanford.EDU>
If you decide that you don't approve of the location, the petition
posted on the MJH stairwell doors is specifically about the location,
and mentions a site near 280 as being a possible alternative.
--Marianne
-------
∂29-Nov-86 2031 YM Thesis
Are you interested in getting a red-bound copy of my thesis?
-Yoni
∂30-Nov-86 1542 M.MCD@LEAR.STANFORD.EDU re: Iran and the Contras
Received: from LEAR.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Nov 86 15:41:58 PST
Date: Sun 30 Nov 86 15:37:25-PST
From: ERIK MCDERMOTT <M.MCD@LEAR.STANFORD.EDU>
Subject: re: Iran and the Contras
To: JMC@SU-AI.ARPA
cc: su-etc@LEAR.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sun 30 Nov 86 13:11:00-PST
Message-ID: <12259160049.11.M.MCD@LEAR.STANFORD.EDU>
Indeed, bribes for hostages seems to make little sense, but the timing
of the arms sales and the subsequent release of hostages suggests that this
was in fact the goal, or at least the main goal. Of course the administration
argues that this is not bribing the kidnappers, but rather influencing those
who can influence them.
Here's a brief chronology:
According to today's Examiner, the arms sales began shortly after the TWA
hijacking in June 1985, after which "the White House determines Iran has
crucial influence toward gaining release of American hostages." Following
an Israeli shipment of weapons to Iran shortly thereafter, Benjamin Weir is
released. In January Reagan decides to break the embargo against Iran. In May
McFarlane vists Tehran and is apparently told that more hostages could be
released if Iran gets more arms. In early July more arms are sent, in late
July Lawrence Martin Jenco is released. In late October, more arms are sent.
On Nov. 2 David Jacobsen is released. On Nov. 13 Reagan in his national address
denies any ransom was paid for hostages...
There must be easier ways to come up with 30 mil.
-------
∂30-Nov-86 2342 LES re: dial costs
[In reply to message rcvd 30-Nov-86 20:39-PT.]
When I looked a couple of months ago, the answer was "No, it is not costing
much," but I'll take another look soon.
∂01-Dec-86 0854 VAL the context seminar
We were planning to speak on contexts this Thursday. Would you like me to
announce it as a joint talk or as two mini-talks? In either case, please give
me 2 or 3 sentences for the abstract.
∂01-Dec-86 0900 JMC
new glasses
∂01-Dec-86 0927 RICHARDSON@Score.Stanford.EDU AI Courses
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 1 Dec 86 09:27:00 PST
Date: Mon 1 Dec 86 09:24:51-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: AI Courses
To: ai.list: ;
Message-ID: <12259354369.12.RICHARDSON@Score.Stanford.EDU>
Plan B: What is your availability during the week of Jan. 5 to discuss AI
courses for next year?
-Anne
-------
∂01-Dec-86 1129 VAL
The man who was active at the last seminar is Peter Ladkin.
∂01-Dec-86 1140 RA lunch
I will go out for lunch today; will be back around 1:30.
∂01-Dec-86 1313 VAL your book
Barbara Bernstein from Ablex asked a few questions she needs answered before
she can send us the contract:
1. What is the title?
2. When will they get the manuscript?
3. What is the situation with permissions from the original publishers and
the fees they may require?
(Should I tell her that it's their problem? Or that we'll take care of the
permissions and they should promise to pay the fees?)
4. Are we satisfied with their usual royalties: 12% for the first 1500 copies
and 15% after that? And how do we want to divide them up?
(I'll be perfectly happy to consider that work as a part of my job for
which I'm being paid by Stanford).
∂01-Dec-86 1501 VAL Nonmonotonic and Commonsense Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
This week we'll meet for the last time before the break to continue the
discussion of John McCarthy's talk on formalizing the notion of context.
I'll try to define contexts as first-order theories of a certain kind
(Richard Waldinger's idea). If there is time left, John will speak about
approximate theories and their relation to contexts.
Date: Thursday, December 4.
Time: 4pm.
Place: Jordan 050.
Vladimir
∂01-Dec-86 2159 GROSOF@Score.Stanford.EDU 2 thoughts about contexts
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 1 Dec 86 21:59:35 PST
Date: Mon 1 Dec 86 21:57:53-PST
From: Benjamin N. Grosof <GROSOF@Score.Stanford.EDU>
Subject: 2 thoughts about contexts
To: jmc@Sail.Stanford.EDU
cc: grosof@Score.Stanford.EDU, val@Sail.Stanford.EDU
Message-ID: <12259491454.18.GROSOF@Score.Stanford.EDU>
December 1, 1986
Hi John,
A couple of thoughts on the subject of contexts.
Benjamin
---------------------
1) It may sometimes useful to view (AI) programs not ordinarily
regarded as having a very helpful high-level declarative semantics as
moving between different contexts whose semantics is cleaner.
For example, many AI programs are viewed from (the rather weak) paradigm of
search. E.g. CMU school, notably lately SOAR (Rosenbloom et al).
Also theorem-proving procedures.
We might try to relate search to contexts in the following way:
View a node (location, etc.) in a search graph of a problem-solving
process as having a declarative semantics provided by a corresponding
context. View the node's descendants as the context's
specializations. View an expansion link from a parent node to a child
node as corresponding to a relation between contexts of immediate
specialization: typically, adopting particular assumptions or making
particular choices. Perhaps view the sub-graph of the overall search
graph defined by all descendants of the given node as providing a
basis for the semantics of the context corresponding to the node. For
example, let the contextual "theory" corresponding to a terminal
(leaf) node be a declarative description of the "solution" or "answer"
or "alternative" that the terminal node represents. Then let the
contextual theory corresponding to an internal node be the disjunction
of the contextual theories corresponding to all of its children, or
alternatively of the contextual theories corresponding to all of its
terminal descendants.
A semantic tree, e.g. of the kind used in analysis of resolution, is
an example of such a search graph. Another example of the use of the
above perspective is in Johan de Kleer's work on problem-solving
methods for which an ATMS (Assumption-based Truth Maintenance System)
is an appropriate aid (see e.g. his AAAI-84 paper). Indeed, the idea
of contexts is used heavily in the notion of ATMS. There contexts are
defined as sets of distinguished Assumed propositions. Generally, if
we are searching a graph for a unique "solution", we might make the
contextual theory of a node be the statement that the "solution" lies
in the sub-grpah of that node's descendants.
Movements between search branches, e.g. backtracking, can be seen as
switches between contexts. The switch from a first context to a
second corresponds to a sometimes non-monotonic step (e.g. switching to
to a sibling rather than a child in a semantic tree) from the first
contextual theory to the second.
--------------------
2) We might want to change logical language and syntax when specializing a
context so as to eliminate arguments of predicates (or functions);
and similarly, when generalizing a context so as to add arguments.
For example, in a succession of increasingly specialized contexts A1,A2,A3:
A1: COLOR(Sky,Blue)
A2: Sky-Color(Blue)
A#: Sky-is-Blue
might all be equivalent in semantic content, but the context can be
exploited to increase syntactic conciseness.
Another example would be
in the more general context:
ON(A,B) & LID(A,C) & BELONGS(A, C) & TRASHCAN(C) & BASE-TO-CAN(B,C) &
BELONGS (B,C)
while in the more specific context:
On(Trashcan-Lid)
Note such reformulations in moving from context to context may involve
positing (or ignoring) objects' existence, as well as giving
definitions of predicates, functions and object constants.
It would be enlightening no doubt to formulate for a few examples in
detail the precise axioms describing such relationships between
contexts and their associated logical languages and theories. The
above sketchy examples are of course only suggestive.
Such reformulations among contexts might help to assuage concerns that
there is something wrong with declarative representations
('representation' in the sense of data structures and algorithms)
because more information, being represented as more wffs, makes for
more computationally unwieldiness. E.g. the motivating critique that
Stan Rosenschein presents to justify his work on compilation and
situated automata. The nice thing about the above kinds of
reformulations is that one doesn't have to move to a very compiled
expressively restricted language right away; one just keeps "a step
ahead" by simplifying syntax incrementally as information accumulates
in moving to a more specialized context.
-------
∂02-Dec-86 0800 JMC
amarel
∂02-Dec-86 0848 TAJNAI@Score.Stanford.EDU SOE Adv.Council
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 2 Dec 86 08:47:58 PST
Date: Tue 2 Dec 86 08:45:36-PST
From: Carolyn Tajnai <TAJNAI@Score.Stanford.EDU>
Subject: SOE Adv.Council
To: fullerton@Sierra.Stanford.EDU
cc: JLH@Sonoma.Stanford.EDU, Linton@Sierra.Stanford.EDU,
Ungar@Sierra.Stanford.EDU, EJM@Shasta.Stanford.EDU, Reid@sonora.dec.com,
Blank@Sierra.Stanford.EDU, m@Sierra.Stanford.EDU,
Lundstrom@Sierra.Stanford.EDU, vrp@Sail.Stanford.EDU,
Oliger@Navajo.Stanford.EDU, Lantz@Score.Stanford.EDU,
Wiederhold@SUMEX-AIM.Stanford.EDU, Nilsson@Score.Stanford.EDU,
Genesereth@SUMEX-AIM.Stanford.EDU, ZM@Sail.Stanford.EDU,
JMC@Sail.Stanford.EDU, Winograd@CSLI.Stanford.EDU,
OR.Stein@Sierra.Stanford.EDU
Message-ID: <12259609367.15.TAJNAI@Score.Stanford.EDU>
Dwain, The following are CSD/CSL/OR faculty who will be involved with
the Forum program on Thursday, Feb. 5.
Blank, Tom
Dantzig, George
Flynn, Mike
Genesereth, Mike
Guibas, Leo
Lantz, Keith
Linton, Mark
Lundstrom, Steve
Manna, Zohar
McCarthy, John
McCluskey, Ed
Nilsson, Nils
Oliger, Joe
Pratt, Vaughan
Reid, Brian
Ungar, David
Wiederhold, Gio
Winograd, Terry
Tom Binford, John Hennessy, and Mark Horowitz are scheduled for
Wednesday, Feb. 4.
Hope this will be helpful information.
Carolyn
-------
∂02-Dec-86 0953 CLT Qlisp meeting reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
Time: Friday Dec 5, 12:15
Place: 352 Margaret Jacks
Topic: Discussion of Qlisp semantics
∂02-Dec-86 1047 @Score.Stanford.EDU:GSMITH@SUMEX-AIM.ARPA AI Planning Retreat
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 2 Dec 86 10:47:06 PST
Received: from SUMEX-AIM.ARPA by SU-SCORE.ARPA with TCP; Tue 2 Dec 86 10:44:02-PST
Date: Tue 2 Dec 86 10:45:42-PST
From: Grace Smith <GSMITH@SUMEX-AIM.ARPA>
Subject: AI Planning Retreat
To: Binford@SU-SCORE.ARPA, Buchanan@SUMEX-AIM.ARPA, Feigenbaum@SUMEX-AIM.ARPA,
Genesereth@SUMEX-AIM.ARPA, McCarthy@SU-SCORE.ARPA, Nilsson@SUMEX-AIM.ARPA,
Rosenbloom@SUMEX-AIM.ARPA, Shortliffe@SUMEX-AIM.ARPA,
Winograd@SU-SCORE.ARPA
cc: Haunga@SUMEX-AIM.ARPA, Puryear@SUMEX-AIM.ARPA, Richardson@SU-SCORE.ARPA,
Gsmith@SUMEX-AIM.ARPA
Message-ID: <12259631232.25.GSMITH@SUMEX-AIM.ARPA>
It's time to begin planning for another one-day, AI mini-retreat,
so would you please let me know as soon as possible which of the
following Saturdays you will not be available.
January 17, 24
February 14, 28
March 7, 14, 21
April 4, 11, 18
Thanks very much,
Grace Smith
-------
∂02-Dec-86 1056 RA
requesting permission
I would like to take the week of new year off to attend the
Linguistic Society meeting in NY and to take care of important
personal matters. This is very important to me and I truly hope
you won't object. I can ask Taleen to substitute for me while I am gone.
Thank you,
∂02-Dec-86 1058 CLT Qlisp meeting reminder
I asked you when we set Friday noon as a regular time
and you said it would be fine.
If you would like another regular time say what it should be.
If you just want to change this Friday,
you shoud negotiate a time with RPG (say Friday afternoon)
and I will send a revised announcement and try to find a place.
∂02-Dec-86 1124 AAAI-OFFICE@SUMEX-AIM.ARPA List of Sponsored Workshops
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 2 Dec 86 11:24:44 PST
Date: Tue 2 Dec 86 11:24:23-PST
From: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: List of Sponsored Workshops
To: Reddy@A.CS.CMU.EDU, Buchanan@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA
cc: aaai-OFFICE@SUMEX-AIM.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Message-ID: <12259638272.39.AAAI-OFFICE@SUMEX-AIM.ARPA>
FYI:
Below is a list of sponsored workshops.
-- Claudia
LIST OF AAAI SPONSORED WORKSHOPS
DATE TITLE SPONSOR AMT
___________________________________________________________________
1984 AI & MEDICINE OSU $5K
1984 MULTI-AGENT PLANNING SRI/SU $5K
1984 NL GENERATION SRI/CSLI $5K
1984 NL UNDERSTANDING & LOGIC PRMGNG SIMON FRASHER $2.2K
1984 NON-MONTONIC PENN $5
1985 SYMBOLIC COUPLING & EXP SYSTEMS BOEING $5K
1985 AI & MEDICINE RUTGERS $7.5K
1985 DAI SU $5K
1985 REASONING ABOUT KNOWLEDGE IBM-SJ $10K
1985 ICAI UNIV LANCASTER $2K
1985 PROBABLITY NASA $5K
1986 AI FOUNDATIONS & METHODOLOGY NM STATE $5K
1986 HIGH LEVEL TOOLS OSU $5K
1986 EXP SYSTEMS IN CAD UNIV SYDNEY $5K
1986 KNOWLEDGE COMPILATION OREGON STATE $5K
1986 COMPLEX LEARNING UNIV LANCASTER $2.5K
1986 KNOWLEDGE ACQUISITION BOEING $5K
1986 DAI BBN $5K
1986 UNCERTAINTY NASA/RCA $5K
1986 AUTONOMOUS AGENTS SRI $10K
1986 AI IN CHEM ENGINEERING COLUMBIA $5K
1986 COMPUTERS & MATHEMATICS IBM/COLUMBIA $5K
1986 AI & EDUCATION UNIV PITTSBURGH $10K
1986 LOGICAL SOLUTIONS TO FRAME PROB UNIV KANSAS $10K
1987 THEORETICAL ISSUES ON NLP NM STATE $5K
1987 KNOWLEDGE ACQUISITION BOEING $5K
1987 SYMBOLIC COUPLING BOEING $7K
1987 MACHINE LEARNING UC-IRVINE $10K
1987 QUALITATIVE REASONING UNIV ILLINOIS $10K
1987 REASONING ABOUT KNOWLEDGE IBM-ALMADEN $10K
-------
∂02-Dec-86 1130 JMC re: Qlisp meeting reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
[In reply to message from CLT rcvd 02-Dec-86 09:53-PT.]
Is there a problem with making Friday's meeting in the afternoon,
say 2pm?
∂02-Dec-86 1153 JJW Seminar time
To: JMC, CLT
∂02-Dec-86 1130 JMC re: Qlisp meeting reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
[In reply to message from CLT rcvd 02-Dec-86 09:53-PT.]
Is there a problem with making Friday's meeting in the afternoon,
say 2pm?
JJW - I have a conflict at that time.
∂02-Dec-86 1153 OKUNO@SUMEX-AIM.ARPA re: Qlisp meeting reminder
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 2 Dec 86 11:53:27 PST
Date: Tue 2 Dec 86 11:53:34-PST
From: Hiroshi "Gitchang" Okuno <Okuno@SUMEX-AIM.ARPA>
Subject: re: Qlisp meeting reminder
To: qlisp@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Tue 2 Dec 86 11:30:00-PST
Message-ID: <12259643585.51.OKUNO@SUMEX-AIM.ARPA>
I have a meeting of AAP (Advanced Architectures Project) at 2:00.
How about in the morning?
- Gitchang -
-------
∂02-Dec-86 1258 CLT Qlisp update
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
Due to various conflicts the meeting time and place has been
revised as follows:
Time: Friday Dec 5, 11:00
Place: 252 Margaret Jacks
Topic: Discussion of Qlisp semantics
∂02-Dec-86 1329 LES Qlisp Meeting
To: CLT, JMC
2pm on Friday is somewhat better for me, though I note there is a talk
on parallel computation in CIS at 2:30pm.
∂02-Dec-86 1331 CLT Qlisp Meeting
To: LES
CC: JMC
11am seems to be the only time everyone else can come.
Is it impossible for you?
∂02-Dec-86 1447 VAL the context seminar
1. I left a draft on your terminal.
2. I called the court about the jury service and they said they don't know
if and when they need me and there's no way to influence that decision. Maybe
we should schedule another seminar anyway (unless we can finish this time),
and, if I can't come, you'll tell me later whether there was anything valuable
in the discussion.
∂02-Dec-86 1455 CHEADLE@Score.Stanford.EDU AI Qual prep course
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 2 Dec 86 14:55:15 PST
Date: Tue 2 Dec 86 14:38:49-PST
From: Victoria Cheadle <CHEADLE@Score.Stanford.EDU>
Subject: AI Qual prep course
To: phd@Score.Stanford.EDU
Office: Margaret Jacks 258, 723-1519
Message-ID: <12259673670.33.CHEADLE@Score.Stanford.EDU>
It has just been brought to my attention that no one has volunteered
to teach CS523 Winter Quarter. This course (Readings in AI) is
primarily intended for those students preparing to take the AI qual in
the Spring. It would be a real shame if the students who are planning
on taking the qual this year don't have the opportunity to benefit
from this course. In the years that it has been offered, the pass
rate for the qual has been significantly higher. I know everyone has
very busy schedules, but I hope a qualified (and soft-hearted) student
will volunteer. They are planning on cancelling the course unless
someone volunteers in the next day or so.
Victoria
-------
∂02-Dec-86 1656 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 2 December 1986
Previous Balance 4.60
Payment(s) 4.60 (check 11/12/86)
-------
Current Charges 4.00 (bicycle lockers)
0.25 (vending machine)
-------
TOTAL AMOUNT DUE 4.25
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 1.0% per month will be
charged on balances remaining unpaid 25 days after bill date above.
An account with a credit balance earns interest of .33% per month,
based on the average daily balance.
∂03-Dec-86 0940 RA Bibel
May from the university of BC called re Dr. Bibel. Her tel. (604) 228 6281.
She said you can reverse the charges when you return her call.
∂03-Dec-86 1523 AIR
May I borrow KCL report (KCL manual)?
Also I would like to spend couple minutes talking about Andrew and EBOS.
Thanks
Arkady
∂03-Dec-86 1615 LAMPING@Sushi.Stanford.EDU Re: Bug
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 3 Dec 86 16:15:01 PST
Date: Wed 3 Dec 86 16:12:34-PST
From: John Lamping <LAMPING@Sushi.Stanford.EDU>
Subject: Re: Bug
To: JMC@Sail.Stanford.EDU
Message-ID: <12259952880.8.LAMPING@Sushi.Stanford.EDU>
The first of your questions that I have a plausible answer for, and no
prize! Mailing labels often show account numbers which contain
"significant" letters from the adressee's name. Assume that the
initial part of the account number for the AMS is formed from the
first letter of the last name, followed by the next consonant of the
last name, followed by the first letter of the first name. (This
might be verifiable by looking at a mailing label). Then assume that
the AMS stores member information alphabetically by account number.
Then an error which destroys or omits the records for 300 consecutive
accounts will have the described behavior. Maybe a disk error on an
ISAM file?
-------
∂03-Dec-86 1647 VAL Reminder: Commonsense and Nonmonotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
We'll continue the discussion of John McCarthy's talk on formalizing the
notion of context. I'll try to define contexts as first-order theories of
a certain kind (Richard Waldinger's idea). If there is time left, John
will speak about approximate theories and their relation to contexts; if
not, we'll have one more context seminar next week.
Date: Thursday, December 4.
Time: 4pm.
Place: Jordan 050.
Vladimir
∂03-Dec-86 1703 GOLDBERG@CSLI.STANFORD.EDU Re: bug appreciation
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 3 Dec 86 17:03:21 PST
Date: Wed 3 Dec 86 16:43:58-PST
From: Jeffrey Goldberg <GOLDBERG@CSLI.STANFORD.EDU>
Subject: Re: bug appreciation
To: JMC@SU-AI.ARPA
cc: su-etc@SU-AI.ARPA, GOLDBERG@CSLI.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 3 Dec 86 15:27:00-PST
Here is my guess.
1. The sorting and processing of the names was done by breaking down
the entire list into different files.
2. A file consisting of (some subset) of Bl* was created.
3. For some unknown (and unimaginable reason) this file was broken
down into smaller files based on FIRST names. (This is the
implausible bit.)
4. Some file or files got trashed in any number of imaginable ways.
I guess the real question is, are 3 and 4 related? Was three an error
that got noticed and interrupted badly just as a file was being
written?
-Jeff Goldberg
-------
∂03-Dec-86 1723 2F09ULL%UKANVAX.BITNET@WISCVM.WISC.EDU Frame workshop.
Received: from WISCVM.WISC.EDU by SAIL.STANFORD.EDU with TCP; 3 Dec 86 17:23:44 PST
Received: from (2F09ULL)UKANVAX.BITNET by WISCVM.WISC.EDU on 12/03/86
at 13:47:55 CST
Date: Wed, 3 Dec 86 12:54 CDT
From: <2F09ULL%UKANVAX.BITNET@WISCVM.WISC.EDU>
Subject: Frame workshop.
To: jmc@su-ai.arpa
X-Original-To: jmc@su-ai.arpa, 2F09ULL
John:
This is to acknowledge your note of Dec. 1.
We have extended the deadline to Dec. 19. We would like a full
length paper (rather than an abstract) by that time, but if this
is not possible, then we will look forward to a full length paper
by Feb. 1 at which time a copy with accompanying documentation will
also be due to Morgan-Kauffman. We will give you all the details
and paperwork when we receive your paper.
Thanks,
Glenn Veach for F. M. Brown
∂03-Dec-86 1742 6058598%PUCC.BITNET@forsythe.stanford.edu SDI reply
Received: from FORSYTHE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 3 Dec 86 17:41:52 PST
Received: by lindy.STANFORD.EDU; Wed, 3 Dec 86 17:33:37 PST
Date: Wed, 3 Dec 86 17:40:23 PST
From: <6058598@pucc.bitnet>
Reply-To: 6058598%PUCC.BITNET@forsythe.stanford.edu
To: jmc@sail.stanford.edu
Subject: SDI reply
Received: by PUCC (Mailer X1.23b) id 6514; Wed, 03 Dec 86 20:38:03 EST
Date: Wed, 03 Dec 86 20:16:53 EST
From: Bill Crandall <6058598@PUCC>
Subject: SDI reply
To: John McCarthy <jmc@sail.stanford.edu>
Thank you for your long and considered reply to my note to
you about software development for the Strategic Defense
Initiative.
I have a number of follow-up questions that I would like
to ask you and hope that I might be able to meet with you when
I return home to the Bay Area for Christmas break. If
you would be able to meet with me at some time between
December 17th and January 2nd, I would be grateful.
If we can't get together, then I hope that I may send you
electronic mail that includes the questions that I'd like to
ask you.
Thanks.
Bill Crandall
∂03-Dec-86 1745 HEWETT@SUMEX-AIM.ARPA Re: bug appreciation
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 3 Dec 86 17:45:46 PST
Date: Wed 3 Dec 86 17:45:51-PST
From: Mike Hewett <HEWETT@SUMEX-AIM.ARPA>
Subject: Re: bug appreciation
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 3 Dec 86 15:27:00-PST
Message-ID: <12259969861.55.HEWETT@SUMEX-AIM.ARPA>
I would guess that the computer crashed during the merge part
of a file-based sort routine and when they brought it back up it
skipped over part of the names.
Mike
-------
∂03-Dec-86 2303 JACOBS@Sushi.Stanford.EDU Re: "America Bashing"
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 3 Dec 86 23:03:10 PST
Date: Wed 3 Dec 86 22:58:13-PST
From: Joseph D. Jacobs <JACOBS@Sushi.Stanford.EDU>
Subject: Re: "America Bashing"
To: JMC@Sail.Stanford.EDU
cc: su-etc@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 3 Dec 86 21:43:00-PST
Message-ID: <12260026725.21.JACOBS@Sushi.Stanford.EDU>
> ... However, for example, I have never heard of
> them opposing people doing paleontology, even though the subset of them
> that are creationists disagree with its results.
I think that it could be more accurately said that creationists
disagree with the current interpretation of the results of
paleontological exploration rather than that they disagree with the
results themselves.
Joseph
-------
∂04-Dec-86 0942 RA going to the bookstore
I am going to the bookstore.
∂04-Dec-86 1100 PALLAS@Sushi.Stanford.EDU Re: "America Bashing"
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 4 Dec 86 10:55:10 PST
Date: Thu 4 Dec 86 10:52:46-PST
From: Joseph I. Pallas <PALLAS@Sushi.Stanford.EDU>
Subject: Re: "America Bashing"
To: su-etc@Sail.Stanford.EDU
cc: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 3 Dec 86 21:43:00-PST
Message-ID: <12260156807.22.PALLAS@Sushi.Stanford.EDU>
JMC claims that attacks on fundamentalists are (to quote)
on an even lower intellectual and moral level than many of the
fundamentalist attacks on "secular humanism".
I can't help being amused by this, because he is admitting
typographically that nobody, including the fundamentalists themselves,
has defined "secular humanism". That hasn't stopped fundamentalists
from collecting under this heading just about anyone who doesn't agree
with their particular views on religion, politics, economics or
morality and denouncing them all as tools of the Devil.
Very intellectual.
joe
Strange bedfellows department: It seems most Communist revolutions
make the purging of intellectuals the first order of business. The
fundamentalists would probably do the same, given the attitudes
they've expressed toward free thought.
-------
∂04-Dec-86 1126 VAL Przymusinski
Przymusinski will be coming at the beginning of February. Ullman says he doesn't
understand why he should pay him.
∂04-Dec-86 1400 JMC
letter to Beckmann
∂04-Dec-86 1435 RA telephone
It turns out that once I pick up your phone in my office there is no way to
transfer it back to you, unless you have another telephone number in which
case I can simply do a transfer.
∂04-Dec-86 1544 RA leaving
It is Thursday and I am leaving in about 15 min.
∂04-Dec-86 2323 ILAN@Score.Stanford.EDU Island dishes
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 4 Dec 86 23:22:54 PST
Date: Thu 4 Dec 86 23:21:00-PST
From: Ilan Vardi <ILAN@Score.Stanford.EDU>
Subject: Island dishes
To: su-etc@Score.Stanford.EDU
cc: jmc@Sail.Stanford.EDU
Message-ID: <12260293019.11.ILAN@Score.Stanford.EDU>
I saw a story on Evening Magazine or 60 minutes a couple of years ago
about a guy on some carribean island who had a sattelite dish and ran
his own radio station using american TV. The only thing I remenber is that
people said it was ruining the local culture and that the most popular
program was all star wrestling from Chicago.
Ilan
-------
∂05-Dec-86 0457 Yuri_Gurevich@um.cc.umich.edu e-address correction
Received: from UMIX.CC.UMICH.EDU by SAIL.STANFORD.EDU with TCP; 5 Dec 86 04:56:59 PST
Received: by umix.cc.umich.edu (5.54/umix-2.0)
id AA14192; Fri, 5 Dec 86 07:04:30 EST
Date: Fri, 5 Dec 86 07:02:26 EST
From: Yuri_Gurevich@um.cc.umich.edu
X-Notice: Our machine name has changed to UM.CC.UMICH.EDU from UMich-MTS.Mailnet
To: janos%techsel.bitnet@WISCVM.WISC.EDU, marek%uky@RELAY.CS.NET,
JMC@SU-AI.ARPA, schmitt%dhdurz2.bitnet@WISCVM.WISC.EDU,
THOMAS%DACTH51.BITNET@WISCVM.WISC.EDU,
trakhte%taurus.bitnet@UCBVAX.BERKELEY.EDU,
iwldald%accumv%mcvax.uucp@SEISMO.CSS.GOV,
vishkin%taurus.bitnet@UCBVAX.BERKELEY.EDU
Message-Id: <1649133@um.cc.umich.edu>
Subject: e-address correction
The new e-address is
Yuri_Gurevich@UM.CC.UMICH.EDU
Regards,
-Yuri
∂05-Dec-86 0900 JMC
phone cate
∂05-Dec-86 0909 cramer@Sun.COM re: news story
Received: from Sun.COM by SAIL.STANFORD.EDU with TCP; 5 Dec 86 09:09:37 PST
Received: from snail.sun.com by Sun.COM (3.2/SMI-3.2)
id AA10578; Fri, 5 Dec 86 09:08:56 PST
Received: from clem.sun.uucp by snail.sun.com (3.2/SMI-3.2)
id AA24585; Fri, 5 Dec 86 09:07:46 PST
Received: by clem.sun.uucp (3.2/SMI-3.2)
id AA00200; Fri, 5 Dec 86 09:09:26 PST
Date: Fri, 5 Dec 86 09:09:26 PST
From: cramer@Sun.COM (Sam Cramer)
Message-Id: <8612051709.AA00200@clem.sun.uucp>
To: JMC@SAIL.STANFORD.EDU
Subject: re: news story
It is in the December 8 issue of The New Republic, page 10.
Sam
∂05-Dec-86 0918 EPPLEY@Score.Stanford.EDU Rutie
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 5 Dec 86 09:18:51 PST
Date: Fri 5 Dec 86 09:16:56-PST
From: LaDonna Eppley <EPPLEY@Score.Stanford.EDU>
Subject: Rutie
To: JMC@Sail.Stanford.EDU
Message-ID: <12260401504.24.EPPLEY@Score.Stanford.EDU>
Rutie has called in that she is running late and probably won't be here
until 10:00.
Thanks
LaDonna
-------
∂05-Dec-86 0936 AAAI-OFFICE@SUMEX-AIM.ARPA [Aravind Joshi <Joshi@cis.upenn.edu>: Workshop proposal]
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 5 Dec 86 09:35:37 PST
Date: Fri 5 Dec 86 09:33:04-PST
From: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: [Aravind Joshi <Joshi@cis.upenn.edu>: Workshop proposal]
To: JMC@SU-AI.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Message-ID: <12260404441.39.AAAI-OFFICE@SUMEX-AIM.ARPA>
John,
Here is that workshop proposal from Joshi.
Claudia
←
---------------
Return-Path: <JOSHI@cis.upenn.edu>
Received: from linc.cis.upenn.edu by SUMEX-AIM.ARPA with TCP; Thu 4 Dec 86 13:02:58-PST
Received: by linc.cis.upenn.edu
id AA16758; Thu, 4 Dec 86 16:00:52 EST
Posted-Date: Thu, 4 Dec 86 16:04 EST
Message-Id: <8612042100.AA16758@linc.cis.upenn.edu>
From: Aravind Joshi <Joshi@cis.upenn.edu>
Subject: Workshop proposal
To: aaai@SUMEX-AIM.Stanford.EDU
Cc: Peters@su.csli.edu, Bill_Rounds%umich-mts.mailnet@mit-multics.arpa,
Patricia Grosse <trisha@cis.upenn.edu>
Date: Thu, 4 Dec 86 16:04 EST
Claudia:
I am sending you our proposal that I talked to you about to complete the
formalities. Please acknowledge. If you need any further information, please
let me know. I will be away between Dec 14 and Jan 1 in the old country. Hope
you had wonderful time in Australia.
Aravind
Mathematical Theories of Language: A workshop to be held in conjuction
with the LSA Institute, July 13 - August 7 1987 at the Stanford University.
A short description of the workshop.
The workshop will meet twice a week for presentation and
discussion of mathematically formulated models of language
systems about which results can be proved that have interesting
interpretations for language. Topics are likely to include
formal languages, automata theory, strong and weak equivalence of
grammatical theories, rigorously defined frameworks for semantic
analysis and their mathematical properties, mathematical
properties of artificial languages as they relate to natural
languages, and mathematical theories of language learning. The
workshop is being organized by Aravind Joshi, Stanley Peters and
William Rounds and is expected to include specially invited
participants as well as participants who will be at the
Linguistic Institute for other activities.
Partial list of possible participants who are likely to be at the LSA
meeting.
Stuart Shieber, SRI
Fernando Pereira, SRI
Kelly Roach, Xerox,PARC
Geoff Pullum, UC Santa Cruz
Gerald Gazdar, U. Sussex
Chris Cully, Stanford
Jon Barwise, Stanford
Ron Kaplan, Xerox, PARC
Barbara Partee, U. Mass
Carl Pollard, Stanford
Tom Wasow, Stanford
Ken Church, AT&T Bell Labs.
Martin Kay, Xerox,PARC
Christopher Mellish, U. Sussex
Francis Pelletier, U.Edmonton
Ray Perrault, SRI
Henry Thompson,U.Edinburgh
[This is a preliminary list. No one has been invited yet.]
Partial list of possible participants who will be specially invited to this
workshop. These participants are not to expected to be at the LSA meeting.
Johan van Benthem, Filosofish Institute, Ryksuniversiteit,Groningen, The Nether
lands.
Ed Barton, AI Lab, MIT, Cambridge
Bill Marsh, Xerox, PARC
Karin Harbush, University des Saarlandes, Sarrbrucken, Germany
Dan Oscherson, MIT, Cambridge / Scott Weinstein, U Penn
.
.
.
We will have about 6 invitees. The estimated expense per person will
be on the average $600. The only costs covered will be the actual
expenses such as travel, lodging , and boarding. No stipends will be
offerd. Actually, the amounts we are proposing will only partially
cover the expenses, because the paprticipants will stay at least
two weeks during the workshop. We think that a number of participants
will agree to come with some contribution towards their expenses. The
exact contribution for each particpant will be determined by the
actual needs and the sources of other funding the participant may have.
We will also invite some advanced graduate students to paprticipate
in the workshop. These students will be encouraged to apply for
the LSA fellwoships.
Budget:
1.Partial expenses of invited participants: 6 X $600 $3600
2.Partial expenses for the non-local organizers
2 X $2300 $4600
3.Other expenses: Copying, etc ($ 200)
Two small receptions ($ 300) $ 500
(The participants who will be at the Institute will not receive
any funds for their expenses. Two small receptions (one in
the middle of the workshop and one at the end) may be small
attraction we can provide for these people besides of course the
honor of being invited !
Total $9700
Amount requested from AAAI $7000
We will try to obtain the balance from other sources.
After the wrokshop, we will submit a report suitable for publication
in the AI magazine.
Submitted by
Aravind K. Joshi, University of Pennsylvania
netaddress: joshi@cis.upenn.edu
Stan Peters, Stanford University
netaddress: peters@su-csli
Bill Rounds, University of Michigan
netaddress: Bill_Rounds%umich-mts.mailnet@mit-multics.arpa
Please send all communication to Aravind Joshi.
-------
∂05-Dec-86 0953 stevens@amadeus.stanford.edu re: news story
Received: from AMADEUS.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 5 Dec 86 09:53:21 PST
Received: by amadeus.stanford.edu; Fri, 5 Dec 86 09:54:32 PST
Date: 5 Dec 1986 0954-PST (Friday)
From: Greg Stevens <stevens@amadeus.stanford.edu>
To: John McCarthy <JMC@sail.stanford.edu>
Cc: stevens@amadeus.stanford.edu
Subject: re: news story
In-Reply-To: John McCarthy <JMC@sail.stanford.edu> /
04 Dec 86 2343 PST.
This is correct (your assumption that I was refering to the Belize
satellite dishes). More specifically, the article is in the Dec. 8
issue. This is the one with the cover story titled "The Idiocy of
Rural Life".
Regards.
∂05-Dec-86 0954 RA Shankar
Shankar was supposed to meet you this afternoon but he won't be able to make
it because he is sick.
∂05-Dec-86 1112 RA mcc invoice
Should there be a consulting charge on the invoice If yes, is it $1,000 for one
day, Nov. 25?
Thanks,
∂05-Dec-86 1150 RA Staff x-mas lunch
I am going out for the staff x-mas lunch.
∂05-Dec-86 1157 CHEADLE@Score.Stanford.EDU Gray Tuesday
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 5 Dec 86 11:56:16 PST
Date: Fri 5 Dec 86 11:53:44-PST
From: Victoria Cheadle <CHEADLE@Score.Stanford.EDU>
Subject: Gray Tuesday
To: JMC@Sail.Stanford.EDU
Office: Margaret Jacks 258, 723-1519
Message-ID: <12260430048.17.CHEADLE@Score.Stanford.EDU>
Here are the online records of your advisees. Please check it over as
soon as possible and send me any corrections/updates by the middle of
next week. As I mentioned earlier, Gray Tuesday will take place
December 16th this year.
A few notes:
We still do not have the necessary software to accomodate the ``new''
teaching requirement (for those who entered 1-84 and after), so that
will not appear. I'll have it available in hardcopy for the meeting.
Remember that 1-86 refers to Autumn Quarter 1986-87, 2-86 Winter
Quarter 1986-87, and so forth.
Kaelbling, Leslie P. Advisor: McCarthy PhD Entry: 1-84
Comp Writ: Pass Comp Prog: Pass Quarter: 3.5
Qual: AI , Cond Pass Qual Qtr.: 3-85 Honors Coop: SRI
Teach (%): 50 Cand Begin: 1-86 Cand End: 1-91
COMP 3-82 MS PASS
COMP 2-84 MS PASS WRITTEN
COMP 3-84 PHD PASS WRITTEN
COMP 4-85 PHD PASS PROGRAM
QUAL 3-85 CONDITIONAL PASS AI --See folder for details.
------
Weening, Joseph S. Advisor: McCarthy PhD Entry: 1-80
Comp Writ: Pass Comp Prog: Pass Quarter: 19
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: McCarthy, 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
EVAL 6/27/1986 Expect solid proposal by Autumn 86-87 or Hold reg.
EVAL 10/01/1986 Hold released.
QUAL 1-81 FAIL MTC
QUAL 1-82 PASS MTC
------
Thanks for your help.
Victoria
-------
∂07-Dec-86 1800 GLB students and EKL
It looks like the students did reasonably well in the homework.
∂07-Dec-86 2213 @SUMEX-AIM.ARPA:MRC@PANDA [Ian Merritt <ihnp4!nrcvax!minnie!ihm@ucbvax.Berkeley.EDU>: Russian PDP-11 ADA compiler for sale?!?]
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 7 Dec 86 22:12:52 PST
Received: from PANDA by SUMEX-AIM.ARPA with Cafard; Sun 7 Dec 86 22:06:44-PST
Date: Sun 7 Dec 86 19:19:27-PST
From: Mark Crispin <MRC%PANDA@SUMEX-AIM.Stanford.EDU>
Subject: [Ian Merritt <ihnp4!nrcvax!minnie!ihm@ucbvax.Berkeley.EDU>: Russian PDP-11 ADA compiler for sale?!?]
To: JMC@SAIL.Stanford.EDU, Wancho@SIMTEL20.ARPA
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12261035477.7.MRC@PANDA>
Can you believe this???
---------------
Return-Path: <@SUMEX-AIM:ihnp4!nrcvax!minnie!ihm@ucbvax.Berkeley.EDU>
Received: from SUMEX-AIM by PANDA with Cafard; Thu 4 Dec 86 10:12:12-PST
Received: from ucbvax.Berkeley.EDU by SUMEX-AIM.ARPA with TCP; Thu 4 Dec 86 09:40:34-PST
Received: by ucbvax.Berkeley.EDU (5.53/1.18)
id AA21806; Thu, 4 Dec 86 08:59:38 PST
Received: by ihnp4.ATT.COM id AA11288; 1 Dec 86 18:06:29 CST (Mon)
Date: 1 Dec 86 15:53:46 PDT
Orig-From: Ian Merritt <minnie!ihm>
From: Ian Merritt <ihnp4!nrcvax!minnie!ihm@ucbvax.Berkeley.EDU>
To: billy@ucbvax.Berkeley.EDU, jumping-jim@ucbvax.Berkeley.EDU,
jsol@ucbvax.Berkeley.EDU, byrne@ucbvax.Berkeley.EDU,
mrc@ucbvax.Berkeley.EDU
Message-Id: <65256006021@minnie.UUCP>
Subject: Russian PDP-11 ADA compiler for sale?!?
Thought you might get a kick out of the enclosed letter if you have't already
seen it. It seems the USSR is marketing an ADA compiler but Never fear! they
didn't forget to attribute the trademark "ADA" to the US.
[Enclosure]
------------------------------------------------------------------------
From decwrl!la780.dec.com!goldsmith@ihnp4.UUCP Sun Nov 30 00:37:04 1986
Received: by ihnp4.ATT.COM id AA00508; 25 Nov 86 19:06:47 CST (Tue)
Received: from rhea.dec.com by decwrl.dec.com (5.54.3/4.7.34)
id AA03620; Tue, 25 Nov 86 10:39:08 PST
Message-Id: <8611251839.AA03620@decwrl.dec.com>
Date: Tuesday, 25 Nov 1986 10:32:48-PST
Orig-From: ihnp4!decwrl!la780.dec.com!goldsmith (Neal E. Goldsmith - LAO 213 417-4240)
From: decwrl!la780.dec.com!goldsmith@ihnp4.UUCP (Neal E. Goldsmith - LAO 213 417-4240)
To: nrcvax!minnie!ihm
Subject: Red ADA...
Sir,
Here is a copy of that ADA letter I was telling you about...
--- Neal
---------------------
Newsgroups: net.lang.ada,net.lang
Path: decwrl!decvax!genrad!panda!talcott!harvard!seismo!mcvax!enea!tut!hmj
Subject: Ada cross-compiler for PDP11 (from USSR)
Posted: 6 Feb 86 17:29:44 GMT
Organization: Tampere University of Tecnology, Finland
Xref: decwrl net.lang.ada:637 net.lang:2056
I suppose this offer we got few weeks ago may be interesting.
The offer was addressed to 'University', Helsinki.
I copied whole letter here as I got it.
Short summary:
They have made Ada compiler producing code for SM4 (actually
PDP11) under OS RW (actually RSX11M). The compiler is cross-compiler
running on Siemens 7.500 series, but by special offer you may get
compiler running on IBM370/CMS or VAX/VMS.
!
Vsesoyuznoe obyedineniye
119034 Moscow,
Starokoniushenny per., 6
Telegr. adres:
Moscow, "Vneshtechnika"
Telex: 411418 "Molot"
Dear Sirs,
We direct you enclosed information on programming system
based on ADA language for Siemens 7500 and SM 4/SM 1420 computers.
Please pass on these material to your specialists and inform
us about the results of their examination.
If the subject would interest you we would be glad to hold
techniocal negotiations whenever it is convinient for you and your
specialists.
Look forward to early responce.
Faithfully yours A.V.Ivanov
Deputy General Director
!
Cross-proramming system on the basis of
ADA* language for SM computer (KSP ADA/SM)
Purpose of the program product and its structure
------------------------------------------------
KSP ADA/SM is meant for the development and maintenance of the
programs of improved reliability for the SM4/SM1420 computers
in the cross-mode. It means, thet hte compiling and program
linkage is made at the so-called host computer Siemens 7.500
(with the 32-bit word), whereas the execution of the program is full-
filed at the so-called target computer SM4/Sm1420 (with 16-bit
word). KSP ADA/SM includes the following program components:
1) control proram-monitor
2) compilete time library support
3) syntax analyzer
4) semantic analyzer
5) code generator
6) run-time programming support
Main functions
--------------
1. Syntactical and semantical control of source modules
(the compilation units) of ADA-programs.
2. Checking of intermodule links of ADA-program with the
same level of control, as in the case of one-module program.
3. Tranfer of compiler ADA-program from the host to the target
computer (SM4 under OS RW) by two means:
- via elekrical communication line,
- via magnetic tape.
------------
* ADA - is the trade mark of US DOD
!
4. Preparation for the execution and run-time support of
ADA-programs by means of dialog with the run-time support
system at the target computer.
5. Monitoring of compiling and receiving of data about the
state of modules of ADA-program at the host computer
(Siemens 7.500 under BS 2000) by means of monitor directives
KSP ADA/SM.
Field of application
--------------------
Development and maintanence of long-living programs of the imp-
roved reliability. For example, the systems for control of
energetics and chemical objects, for the control of NC-machine-
tools etc.
Main features
-------------
1. The full implementation of the compile-time library support,
including multi-access to the module library of ADA-programs.
that is the possibility of simultaneous participation in the
development of ADA-programs to 64 groups of specialists.
2. The possibility of rapid retargeting (at the expense of change
of computer-depended part of code generator and program
support of execution).
3. The possibility of extending the set of intrumens of
compiling period (structure editor, optimiser, etc.) due to
using the standard internal interface analysis-synthesis
(of the "Diana" type).
!
4. The possibility of including in the elektical comminucation
line between host and target computers the dedicated telefon
line, that allows to use the object computer as the remote
working station for the development of ADA-programs. The
following actions are accesible from there:
- monitoring of compiling of ADA-program modules,
- editing of texts of ADA-programs,
- compiling of ADA-programs,
- monitoring of execution and execution of ADA-programs.
Requirements to the basic software
----------------------------------
Host computer: operating system BS 2000 for the computer
Siemens 7.500.
Target computer: operation system OSRW for the computer
SM/SM1420 (or RSX 11.M for PDP 11) woth driver
of electrical communication line.
Requirements to the hardware
----------------------------
Host computer: Siemens 7.500 (model 7.531 and higher).
Target computer: SM4/SM1420 (or PDP 11) with elektrical
communication equipment (sequential interface:
V.24 or RS 232 or "current loop"; 2 modems, which
are compartible with this interface).
Additional possibilities
------------------------
In case of special order it is possible to transfer the host part
of KSP ADA/SM to the computers, which are compartible with IBM 370
under CMS, VAX 11 under VMS.
--
Hannu-Matti Jarvinen
ASENTO - Ada Software ENgineering TOols -project
hmj@tut.UUCP (note .UUCP, tut.ARPA is not the same computer).
------------------------------------------------------------------------
[End enclosure]
Cheerz--
<>IHM<>
-------
∂08-Dec-86 0048 RLG NSF application
i wrote a "proposed plan of research" for NSF. I would value your
critiques when i revise it tonite (monday nite).
specifically, is the tone too vague, too arrogant, or do i run on
too much unnecessarily?
the essay is page 6 of NSF.TXT[1,RLG]
thank you!
∂08-Dec-86 0115 GLB
Thank you. In case you wanted to have an EKL question in the final,
the following is a possibility:
;∀U.MEMBER(X,U)⊃NTH(U,FSTPOSITION(U,X))=X
;∀U N.UNIQUENESS(U)∧N<LENGTH U ⊃ FSTPOSITION(U,NTH(U,N))=N
It is easy, but not trivial.
∂08-Dec-86 0907 NILSSON@Score.Stanford.EDU DARPA
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 8 Dec 86 09:05:36 PST
Date: Mon 8 Dec 86 09:02:29-PST
From: Nils Nilsson <NILSSON@Score.Stanford.EDU>
Subject: DARPA
To: jmc@Sail.Stanford.EDU, les@Sail.Stanford.EDU
cc: nilsson@Score.Stanford.EDU
Message-ID: <12261185306.27.NILSSON@Score.Stanford.EDU>
I phoned Bob Simpson on Friday about the Intelligent Communicating
Agents Proposal. He had not yet read our draft but conceded that in
the interests of getting it in officially before the end of the year,
I should just go ahead and send it in. (It is being sent in under
the provisions of a CBD announcement put out by RADC/DARPA. Although
the CBD announcement doesn't expire until May '87, the proposals
are "bunched" every 3 months, and there is a deadline at the end of
Dec. If we wait longer than that we'll be delayed another 3 months.)
He seemed to imply that he would have the money to fund it out of
87/88 funds mainly. He acknowledged that he was "working on" the
McCarthy proposal also but he didn't connect the two (so I didn't
either). I'm inclined now not to call Saul but to let things take
their cour. You folks should make sure that Bob is going to
accept and fund your proposal and that you be informed about any
funding problems. If a problem arises that can be solved by some
kind of combination two proposals, let me know. -
-------
∂08-Dec-86 0937 AI.ELLIE@MCC.COM 1986 Expenses
Received: from MCC.COM by SAIL.STANFORD.EDU with TCP; 8 Dec 86 09:37:17 PST
Date: Mon 8 Dec 86 11:38:15-CST
From: Ellie Huck <AI.ELLIE@MCC.COM>
Subject: 1986 Expenses
To: jmc@SU-AI.ARPA
cc: ai.scullion@MCC.COM, ai.buchberger@MCC.COM
Message-ID: <12261191818.14.AI.ELLIE@MCC.COM>
Please bill Mary Buchberger for all your 1986 consulting expenses by
December 19, 1986.
-------
∂08-Dec-86 1131 GLB
Another possibility for the exam.
Formulate appropriately in EKL language and prove:
Let U be a list containing only red and green objects.
If V, W are arbitrary lists containing all the red elements,
resp. green elements, of U and only those,
then the length of U is the sum of the lengths of V and W.
This is easy, since we have already in the preloaded files
DISJOINT(A,B)⊃MULTIPLICITY(U,A∪B)=MULTIPLICITY(U,A)+MULTIPLICITY(U,B)
but some attention is required to linguistic details and
a few new lemmas are needed.
∂08-Dec-86 1240 RA going out
I am going to the bookstore.
∂08-Dec-86 1322 CLT Qlisp meeting reminder
To: "@Q.DIS[1,CLT]"@SAIL.STANFORD.EDU
Time: Friday Dec 12, 11:00
Place: 252 Margaret Jacks
Topic: Strategies for parallelizing Lisp programs
Joe will talk about algorithms used for introducing futures
to automatically parallelize pure lisp programs
(work of Halstead students).
In the remaining time we will discuss how such strategies
fit into the Qlisp world -- what happens in the presents
of side-effects, catch/throw, proocess killing.
Discussion of semantics will resume in January
The Lucid contingent is out-of-town this week.
∂08-Dec-86 1455 RA seminar room
To: JMC, VAL
The room you are using now for the Thursday seminar is already spoken for for
next quarter; you were assigned room 161K in Bldg. 160 for next quarter.
Rutie
-----
∂08-Dec-86 1526 RICHARDSON@Score.Stanford.EDU AI Mtg.
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 8 Dec 86 15:26:01 PST
Date: Mon 8 Dec 86 15:23:04-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: AI Mtg.
To: ai.list: ;
Message-ID: <12261254587.22.RICHARDSON@Score.Stanford.EDU>
Would those of you that have not yet responded please send me a message
letting me know your availability in January for the proposed AI meeting?
Thanks,
Anne
-------
∂08-Dec-86 1601 RA tired
I am extremely tired and if you don't mind I'd like to leave now.
Thanks,
∂08-Dec-86 1606 LES Gould number cruncher
I talked to Rob Laudati about the rates on their Gould computer. He says
that they have six people in the group, but that they are being supported
by a grant. The rates for the computer are therefore based on the marginal
costs of running the system. I remarked that their disk rates looked
attractive for archival storage and he responded that that would be fine
-- they have four Eagles that are almost completely empty.
∂08-Dec-86 1751 CLT shopping list
milk
powdered similac (green label) 3 cans
gerbers oatmeal
supper
∂08-Dec-86 2000 JMC
Fredkin
∂09-Dec-86 0800 JMC
cate
∂09-Dec-86 1031 RA credit
I put the credit slip for you airline ticket on your desk.
∂09-Dec-86 1203 GARDNER@CSLI.STANFORD.EDU Book promotion
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 9 Dec 86 12:03:31 PST
Date: Tue 9 Dec 86 11:56:59-PST
From: Anne Gardner <GARDNER@CSLI.STANFORD.EDU>
Subject: Book promotion
To: jmc@SU-AI.ARPA
cc: gardner@CSLI.STANFORD.EDU
Here is a request, made with considerable diffidence. My thesis revision
is finally about to be published by Bradford Books/MIT Press, and the
press would like to know of some "leaders in the field from whom we might
solicit endorsements of [the] book either before or after publication."
Would you feel it appropriate if I suggested your name, and would you be
agreeable to that?
Thanks for considering the question, whatever the decision.
Anne
-------
∂09-Dec-86 1205 RA lunch
I am going out for lunch.
∂09-Dec-86 1519 VAL
May I remind you that you owe me the abstract of your talk. - Vladimir
∂09-Dec-86 1857 DEK Ed Fredkin
Do you know an address where I can reach Fredkin, either electronically
or the old way?
∂09-Dec-86 2145 MODET@Sushi.Stanford.EDU Some questions/comments about the final
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 9 Dec 86 21:45:40 PST
Date: Tue 9 Dec 86 21:43:45-PST
From: Andres Modet <MODET@Sushi.Stanford.EDU>
Subject: Some questions/comments about the final
To: jmc@Sail.Stanford.EDU
cc: modet@Sushi.Stanford.EDU
Message-ID: <12261586034.23.MODET@Sushi.Stanford.EDU>
Here are some questions/comments about the final. I shall try to be at
Thursday's lecture, but I am not sure if it will be possible. Since the tape
won't be available until Saturday, I would appreciate if you could answer
by e-mail.
Problem 1:
Should we care about circularity? It seems that the question is referred
to 'plain' S-expressions (without circular pointers).
The function gmat is not defined for atoms. This means that our implement-
ation is free to return any value when x is an atom, doesn't it? Any way,
there are two possible generalizations of the gmat semantics to atoms, can we
choose any of both?
Problem 2:
x may contain lambda expressions, can the bound ('dummy') variables in x
be renamed when the substitution is performed? This transformation surely
preserves the meaning (as long as no capture within x occurs), but would
make our lifes harder. So, could we assume that x is substituted as it is,
without any internal renaming?
Problem 4:
The backtracking term is a bit misleading, what is really requested is
a non-recursive structure traversal, isn't it?
There is an important unstated assumption: 'All bits are set to nil
previous to the traversal'. Otherwise, if the bits may be either true or
false at the beginning, I do not see any way of solving the problem.
Extra-credit if we use only one bit? (Does not require an answer).
Thank you,
Andres Modet (Modet@Sushi)
-------
∂10-Dec-86 0129 HSU@Sushi.Stanford.EDU gmat
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 10 Dec 86 01:28:58 PST
Date: Wed 10 Dec 86 01:25:29-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: gmat
To: cs306-distribution: ;
cc: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12261626398.13.HSU@Sushi.Stanford.EDU>
Clarification:
gmat[x,y] is true if any CAR link on the path leading to some atom in x
corresponds to a CAR-CDR chain, starting with CAR,
on the path leading to the same atom in y;
and similarly,
any CDR link on the path leading to some atom in x
corresponds to a CAR-CDR chain, starting with CDR,
on the path leading to the same atom in y.
-------
∂10-Dec-86 0441 KE425%RPICICGE.BITNET@WISCVM.WISC.EDU
Received: from WISCVM.WISC.EDU by SAIL.STANFORD.EDU with TCP; 10 Dec 86 04:41:09 PST
Received: from (KE425)RPICICGE.BITNET by WISCVM.WISC.EDU on 12/10/86 at
06:40:19 CST
Date: 10 December 86 06:46-EST
From: KE425%RPICICGE.BITNET@WISCVM.WISC.EDU
To: JMC@SAIL.STANFORD.EDU
Dear Prof. McCarthy:
I am Yu from RPI. I have sent you a letter last week to show
my desire to meet you when I arrive at Stanford on Dec. 11 and 12.
Perhaps I won't get your message before I leave for Stanford. I
think it will be better for me to send you my work in advance.
If convenient, please notify me today whether you are available on
11 and 12. Thank you.
Liang-Yin Yu
∂10-Dec-86 0449 KE425%RPICICGE.BITNET@WISCVM.WISC.EDU
Received: from WISCVM.WISC.EDU by SAIL.STANFORD.EDU with TCP; 10 Dec 86 04:48:43 PST
Received: from (KE425)RPICICGE.BITNET by WISCVM.WISC.EDU on 12/10/86 at
06:47:45 CST
Date: 10 December 86 06:55-EST
From: KE425%RPICICGE.BITNET@WISCVM.WISC.EDU
To: JMC@SAIL.STANFORD.EDU
ENQUIRIES ON SOME FUNDAMENTAL CONCEPTS OF ARTIFICIAL INTELLIGENCE
Liang-Yin Yu
Dec. 9, 1986
Introduction
This treatise is motivated primarily by the critiques given by
Dreyfus and Winograd. The lines below reflect my personal view about
the connections between machine intelligence and human intelligence.
On the other hand I would like to emphasize the true essence of the
word "meaning".
What I consider to be highly valuable for a serious AI researcher
is the way to contemplate those implicit aspects of our thinking
(meta-thinking) and, whether consenting or not, these critiques possess
traits that we should never overlook. I personally was benifited from
these traits when I was pondering the possible structure of human mind.
As Dreyfus has put it, the assumptions underlying our researches have
no experimentally justifiable basis but I am not convinced by what
Dreyfus argued. The "conceptual difficulties" posed by Dreyfus and the
analysis adopted from Wittgenstein is not persuative. I'll explain my
opinions later. On the other hand the connections between Winograd and
Dreyfus are by no means trivial. There are many impressive argument in
Winograd & Flores, though the hermenutics approach lies on an
introspective ground and the theory by Maturana possesses a highly
spectulative air. I am greatly inspired to consider some of the
intricate problems in the AI field, and adoped from some of the
viewpoints of both Dreyfus and Maturana, I try to provide another
viewpoint that may not be so pessimisitic about our AI business. I am
trying to further develope these ideas and what presents here is just
the primitive draft, both in its form and its content.
Structural Process
We have two ways to focus our attention on the man-machine
arguments. Either we would like to reproduce human intelligence on a
machine with basically different structure or we just intend to explore
the essence of intelligence with a mild view? In 1956, at Dartmouth,
following words were put down to indicate the original intention.
Every aspect of learning or any other feature of
intelligence can in principle be so precisely described
that a machine can be made to simulate it.
Certainly we won't assert that the form of intelligence that is so
familiar to us is the only one out of all the possibility neither should
we judge every possibility based on our own image.
Befor we begin the discussions of the critical problems of AI
there is one thing that we should bear in mind. Our brain is composed
of about a hundred billion nurons and in average, every nuron is
connected to 40 to 50 other neurons. This incomprehensible complexity
underlies all our behaviors. With this fundamental fact in our mind, we
can hardly expect those behaviors can be reproduced by the computer
technology we can achieve today.
Psychology and biology have provided evidence that the structure
of our mind is basically one with multi-structure, either functionally
or physically. For instance, brain can perform some "replacing" function
in case some other portion is damaged, and it is well known that there
are many emergent "back-up" structure such as the transition from
underconciousness to conciousness in order to handle some unfamiliar
problems. No matter how many functionally equivalent structures has been
built in, we would expect that computer will have both capacity and
structure of the commensurate complexity as brain.
That the structures of biological system and the present machine
system are different is evident from the fact that the fastest functions
that human can perform is drastically different from that machine can
perform. As Dreyfus explicated in his "biological assumption", the
operation of neuron cannot be assert to digital( we shall discuss the
discrete and continuous aspect of mental process later). Even it does,
the fact can't be used as an evidence for the feasibility of AI, since
there still a huge structure involoved.
Such a drasstic divergence in structures accounts for the
non-algorithmic process of perception. I would like to refer it as
"structural process". Lashley(1942) suggested that image information is
stored in wave form and De Valois provided the hologram as a plausible
model for this image retrieving and conceiving process. Haugeland also
considered it possible. The amazingly fast speed of pattern recognition
of human should be due to this structural process.
Dreyfus has proposed four types of human information processing
which seemed unlikely to be implemented in computer and they backed up
his argument of psychological assumption and epistemological assumption
in AI, that is the rule-based explaination of human behavior and
formalizability of all human knowledge. Let's consider these arguments
to see if they are reliable to be the base for us to forward our step.
Dreyfus's Information Processing
The first one and the most important one Dreyfus discussed is "the
fringes of conciousness". Dreyfus pointed out that it is unlikely that
chess master has counted out all the possible move, like the search
method incorporated in the computer program, before he notices any
"weak point" of his opponent, and he attributed this ability to the
process proceeding in our underconciousness. The result can only be
known as "marginal awareness". The words he adopted from Polanyi
explained the core idea.
This power resides in the are which tends to
function as a background because it extends indeterminately around
the central onject of our attention.
Without discussing some of his out-of date argument like "present
program encoding no long-term strategy at all" let's focus on his
opinion about the process on the fringes of conciousness. As the study
of de Groot showed that the chess master has groups of patterns that
enable him to find the most advantageous move. According to Chase and
Simon(1973) master has about 25,000 to 100,000 patterns while a normal
playerhas only about 1,000 patterns. Apparently the operation undergoing
on the "fringes of conciousness" may be accounted as the structrual
process of pattern recognition also. Master might scan over the board
and through "recognition of resemblance"(according to Dreyfus's
classification) he might match the local pattern around some particular
chessmen(the place he focused his attention).Such a kind of "counting
out" won't exhaust the time master had. What I would like to emphasize
is that such "repertoire" of patterns(or shemata) should depend greatly
on the method knowledge is represented, and what we mean by "the
accmulation of experience" is just the piling-up process of various
schemata.
The second type of information processing, 'ambiguity tolerance",
is extremely important in the developement of linguistics. The
generative grammer can't explain such ambiguity nor can situational
semantics(Winograd 1985). But the term used by Dreyfus should actually
be "context-dependent accuracy" instead of "ambiguity tolerance". About
this point I am agree with both Dreyfus and Winograd that our computer
cannot deal with this semantic problem. It is deep related to the
"origin of meaning" and "the limitation of language" that I shall
discuss later where the problem of learning will briefly reviewed.
The third type Dreyfus mentioned is "essential/inessential
discrimination". Winograd's discussion of "blindness" is closely related
to this. Dreyfus wrote:
Normally essential operations are not around to be
found because they do not exist independently of the
pragmatic context.
Winograd, on the other hand, put down the following words:
Reflective thought is impossible without the kind
of abstraction that produces blindness.
Although basically I consent to both of the authors I must point
out one important insight. Can human always "break the rule"? We can
only think of breaking the rule that comes to our mind but certainly not
those which are beyond our thought. Indeed it is not the machine that
decides but the programmer, bur so is human who cannot the constraint
gene, culture, society and his holistic past history exert on him. In
this case the essential/inessential criterion is also decided
beforehand, but in a much higher level. Again I would refer such a
problem as a problem of knowledge represention, but to achieve a
knowledge base which is commensurate to human and with all the necessary
schemata and link( like the one Quilliqn tried to do), I think we are
trying to play the role of the Creater. Unless we can develope the
learning process as Minsky suggested, perhaps the problems of
"context-dependent accuracy" and "essential/inessential discrimination"
could only remain to be plausible in principle but unachievable in
practice. This may reflect my opinion about the following argument made
by Winograd:
...two very different things: structure-determined
systems...and systems programmed explicitly with a
chosen representation.
But, of course, the "representation" here is not quite the same
thing as Dreysus's "essential/inessential discrimination". This problem
should refer to the limitation of language or more appropriately, the
limitation of representation. We shall discuss it later.
The fourth type of human information processing Dreyfus mentioned
is the one concerning with pattern recognition. I do not intend to go
through it since it is closely related to the structrual perception
process. I only mention a few points. First, there is no problem for the
machine to deal with the first sort of pattern recognition, that is
adaptive "sympton and criterion distinction" of Wittgenstein, through
goal-directed and planning process(if they are sound enough). On the
other hand, while tackling the sceond sort of pattern recignition, the
recognition of resemblance, Dreyfus has made use of the Pudovkin's
experiment, that is what you think of the expression of a man face is
strongly dependent to what you think prior to that picture. I think
Dreyfus has mixed up the associating process with the recognizing
process. To recognize whose face the picture is is not the same as how
the face looks. The later has deep relationship with the mental process
of the observer. The emotion of the observer could only affect what he
think the face is expressing about but has little connection to the
ability to recognize the objective contour of the face. For example,
to recognize something is cloud is not the same as that the shape of the
cloud gave you an impression of a dog. Finally, I think there is no
essential difference between the first sort and the third sort of
pattern recognition, since fuzzy theory can be introduced to handle the
problem of "family resemblance". Before moving forward I would like to
discuss one important concept mention by Dreyfus. Dreyfus said:
...the formalized principle used to explain our behavior
may not be the principle underlying our behavior.
For instance, in the preceeding case, the fuzzy theory may be used
as a formalized theory to deal with the problem of "family resemblances"
but it is not necessarily the principle our perception process follows.
My opinion is that indeed it's quite true for such a "false replacement"
being pervasive in our studies of various subjects but we are not
declaring that we have found the truth of the universe once we set out
to explain something. Functional equivalence is not structural
equivalence. So we are doing our best to explain every phenomenon we can
observe and if the explaination appears to be inadequate time will sifts
it out just as the motion of Mercury turned our attention from Newtonian
mechanics to relativity. I am not trying to overrate physics without the
inkling that it might fall into the pit of "formalizability" problem. I
only want to point out that there perhaps exist more than one
explaination that may encompass our goal space. Functionally equivalent
principle cannot be proved to be wrong anyway.
After discussing the origin of Dreyfus's argument, we may turn to
the core part of his book. This is also the primary concern of Winograd.
Dreyfus put it in the form of "four fumdamental assumptions" and
Winograd took it as "our rationalistic tradition". I tend to put these
concepts in a unified form and try to explicate it with "the limit of
language","structrual process" and "the origin of meaning". Alternate
viewpoints are explored along with the discussions and we may see how
far can present computer techology go with the aid of these alternate
viewpoints.
Basic Assumptions
Dreyfus challenged AI on four basic assumptions, namely
biological, psychological, epistemological and ontological assumptions.
These important assumptions may reveal themself through the following
three questions:
i) Is mental process basically discrete or continuous?
ii) May the objects in the external world be formalized or
represented in our mind?
iii) Ontologically is there such property as objective reality or
context-free objects?
Dreyfus put the first problem in his biological assumption and
discuss the rest in the remaining three assumptions. On the other hand,
Winograd considered the basic difficulties of western tradition, i.e.
the rationalistic tradition, from the works of hermenutics, biology and
linguistics. They all focus on a common concern which is used as the
most radical foundation of scientific investigation. The inadequacy
arises when we are trying to extend this presupposition to the studies
of ourselves, the object that is traditionally studied in philosophy,
psychology and linguistics. Their arguments appear persuasive especially
when considering the multifarious issues arising in our daily usage of
language, but there would be somewhat different "interpretation" of
their arguments. From these viewpoints, we may consider seriously the
limitation of our computer methodology in tackling those problems that
seem "resistant to be formalized", as I shall show, it might direct us
to consider " a totally different way of using computer", both in the
sense of Winograd's "intervention in the background of our heritage" and
in the sense of exploring generalized intellgence.
Fundamentally, all these assumptions or questions can be
considered under the unified frame of formalizability. Formalization may
take various different forms. Basically, it is a process to explain or
formulate whatever attention focuses with finite numbers of context-free
objects whose properties could be uniquely defined without ambiguity.
The concept of context-free objects presuppose that everything which
could be explained in terms of these objects is discrete, and there
exists unique representation that corresponds to each of these objects.
Furthermore, the "rules" that operate upon this representation of
objects must also be formalizable. The assumption and its consequences
are by no means self-evident. However, human beings possess some
features that seem to us natural to make this assumption. First, it is
well known that there are two symbolized process in the conciousness of
human mind, namely image and language. Mental image and internal speech
are almost dominant in human conciousness. Though human are possible to
develope other symbol system, like Hellen Keller, these two symbolic
process are generally considered to be the indispensible composite of
thought. Adopting the tern from Maturana, when we are conveying message
through language, we are interacting within a "consensual domain". We
may perceive such implicit structures, but most of the time we are so
overwhelmed by the message conveying through the words that we take it
for granted that we are engaging in a formalizable process.
We can see that this assumption is inherent in our way of using
language. Language itself is a collection of discrete unit. We can
perceive continuous phenomenon, but my opinion is that they can only be
manipulated under structural process. If the subject tend to express
such a continuous phenomenon in the form of language, ambiguity arises.
Most of the difficulties AI researchers confronted are the seeming
inability to handle the continuous transition between meanings, i.e.
attempting using discrete concepts to express the meaning of our
continuous meaning. Some refer this problem as common sense
representation. Though in principle we can cover up this continuous
spectrum of meaning with discrete grids of careful chosen concepts, we
are nonetheless setting out for a possible eternal journey. On the other
hand, the capability of human to deal with the problems of situated
meaning of language is being "already" in the situation. While avoiding
to play the role of the Creater, we might be curious enough to explore
the precondition of "being already in the situation". One may argue
that perhaps we are undergoing a complex "pattern" recognition process
in order to decide where we are, what we are doing, how we get here and
what are we intending to do. The argument, "infinite regress" of Dreyfus
and "structure coupling" of Winograd seem particularly persuasive here.
So we must handle the experience accumulation process artificially in
order to facsimile human behavior. This is certainly not a short cut.
Even this can be done, we are still facing the problem of continuous
transformatin, as Maturane put it.
Learning is not a process of accumualtion of
representations of the environment; it is a continuous
process of transformation of behavior through continuous
change in the capacity of the nervous syatem to synthesize
it.
So we are forced to ask ourselves again that can we represent the
continuous process? Much can be done and much can be said. Whatever it
comes out, we should expect it to be gigantic work.
Winograd summarized his opinions about natural language processing
as "understanding as pattern recognition". As a matter of fact, this
problem is pervasive in all AI research fields. The true state of the
problem is that we cannot endow the computer with "meaning". We
communicate with computer through a group of symbols. We represent
"knowledge" in the computer in a form of combinations of symbols.
Finally, we interpretate the results as "meaningful". The computer can't
"interpretate" language and it is "incable of making commitments", as
Winograd put in his book. However, it may not doomed to be so though it
certainly seems so presently. This problem is a most difficult one not
only in AI but in philosophy and psychology as well. It is the origin of
of meaning that we are really expected to solve. This is also the key to
solve the problem of "being in the world".
Origin of Meaning
Maturana said in his article, "neurophysiology of cognition", that
Machines differ from living systems not in the principles
used for their function, but in their reference....
the relevance of their functioning is determined by how
they satisfy the desire of their maker. Living systems...
necessarily have a self-refered domain of interactions.
The artificial cognition systems can be of two kinds:
(1)those that will describe their interaction in our
terms, that is, recognize what we recognize.
(2)those that will make descriptions in their own terms
but which we have to map afterward into ours.
In either case, I think, there is no need for imitating
what occurs in our brain. In the first case, however, it
would be essential to give the machine a domain of
interactions like ours, not our description of it.
We have noticed that Maturana proposed "to give the machine a
domain of interactions like ours, not our description of it". The base
of our being able to interact with external world is the various sensors
we possess. These parts of our physical body enable us not only to
explore what the external world looks like but to indentify ourselves as
a whole. In Dreyfus's words, this is the ability in "organizing and
unifying our experience of objects". One may lose his ability to hear or
or to see or both, but he may never lose his tact. Psychologist Piaget,
Luria and Bruner all suggested that the neonate responds directly to the
stimulus rather than to an internal repreaentation of the stimulus. They
also suggested that there is a distinct stage of presymbolic behavior.
It is undoubtful that sensing ability is important for our being in the
world, but what we see here is that they also possess a definite
relationship to the oringin of meaning. Dreyfus argued that "it is the
sort of intelligence which we share with animals, such as pattern
recognition that has resistedmachine simulation". So we encounter
difficulties in the " front ends" of our perspective intellgent machine.
Contrarily to Dreyfus who took it as an evidence for the doomed fate of
AI, I think it is a problem of structure. I do not intend to say that
the structural process such as vision is amiable to our computer, but
if computer is to have any "intelligence" that will not be ridiculed as
"pattern recognition", I insist that it should be given sensors
compatible to their human counter parts so as to find out "who it is".
Serle has perposed a condition which he called "the Chinese room".
He argued that if we cage a person in a room and feed all the necessary
information of Chinese in the form of rules to him. Then even he can
respond well in whatever case Chinese is needed, the man can not be
refered as "knowing" Chinese, though he "knows" Chinese in the Turing's
sense. This dilemma is exactly the case of our computer, but the point
is the man in the room is deprived of all the possibility to access
the external world. If he cannot make the analogy for himself he indeed
can only be refered as possessing the ability of manipulating a group of
symbols without any semantic meaning. The experience with another
language may enable the man to make analogy between what he has already
bore and the one he now encounter. This the exactly starting point of
the meaning. In case that he cannot infer anything from what he already
had, there is no possibility for him to get anything meaningful from the
rules.
It has been argued by Haugeland that "it is far from apparent what
a learning system needs to acquire". So if we would like to solve the
problems of AI through learning system, we must figure out what is the
mind structure and the goal encoded into the neonate by gene. In this
case we have to deal the part where structural process is undergoing in
the first place, that is, the part where continuous portion of our mind
process may proceed. In this regard, I expect the method proposed by
connectionist and model such as hologram might lead us one step ahead.
It is hard to tell whether the one dealing with knowledge representation
in the full range of a human adult and then approximating everything by
discrete steps easier or the one tackling initial structure with least
knowledge then followed by a complete learning process is easier. As we
may see that even in the later case, we still have to acquire a concrete
model of human knowledge representation first. This problem is
discussed in "the limit of language" later.
Though I incline to see some progress, the fundamental
difficulties are indeed the greatest challenge human has ever
encountered and, most likely, it may come out with little results.
Dreyfus has mentioned the opinion of Heidderg that western philosophy
has reached its climax in AI, namely to study mental process in a
formalized way. I would say that at least we can realized the limitation
(if it exists) of our rationalistic tradition and the potential of
ourseves. It is not necessarily to retreat now.
In the next section I would like to explore the drawback of our
methodology in "knowledge representation" nowadays.
The limit of Language and its Implication
No matter which way we chose to approach the true essence of
intelligence, there is always a fundamental problem that is most
critical among others. We might call it "knowledge representation" as
most people called it. It appeared to be in the form of phenomenology as
Dreyfus put it, that is we have to organize almost infinite numbers of
"aspect" in order to explain our world as viewed by mind. I would refer
all the problems as "the problem of representation".
When we speak of "representation" in AI, most people would think
of "knowledge representation". The later is the very concept that can
hardly escape the stern critiques of Dreyfus and Winograd. What I mean
by representation is in a broder sense, which may be defined as the form
of information in our mind that enable us to engage in whatever we would
like to know or think. We must agree that we must possess some relevant
material in our mind before we may set out to think. This relevant
material is just what I mean by "representation".
I have suggest the crucial role our sensors play, as Dreyfus and
Maturana did. It may be the most important factor responsible for the
origin of meaning though not the only one. Even we get over this
difficult point we won't have much progress in dealing with human
intelligence with the method of knowledge representation today. As I
have mentioned that we communicate with the computer through language,
either artificial one or, hopefully, the same language we use in our
daily life. What we have created in this case is the "naming" process,
that the process to give names to everything. I believe that language
is a way we approximate what we really have in our mind. The reason that
we almost have an infinite way to express our thought is we can never
make it exactly what "occured" to our mind( not the "thought" in our
mind). We can't create the expression of fire or the feeling of being
burned by just giving something the name "fire". We have to "see" the
phenomenon of burning and "touch" the flame to get a complete idea of
"fire". When we are talking about the "feeling" we have, we are refering
the something in our mind. I do not quite agree with Dreyfus that it is
agnostic in nature. I think this "continuous" aspect of our inner
representation has a deep relationship with the structural process.
In order to explore the essence of meaning and to save our works in the
representation we might have to consider some more appropriate form in
storing our initial "knowledge" of the world.
I believe that the phenomenonlogical aspect of our knowledge
representation is due to the effort trying to express our true knowledge
about the world exclusively in terms of language. We have to consider
the limitation of our language first.
Conclusion
I am atempting to discuss some of the intricate problems involved
in AI, not the tecniques used to implement various problems but the one
that underlies what we really think intelligence is.
(A detailed conclusion will be worked later)
∂10-Dec-86 0800 JMC
approximate theories
∂10-Dec-86 0914 RA procedural question
I need to ask Binford for his updated CV can I tell him what it's for?
Thanks,
∂10-Dec-86 0954 RA Joan Treichel
Treichel, (202) 332 8624, is writing a book about human memory for Doubleday
and would like to have your opinion re the comparison between computer memory
and human memory. She would appreciate your calling her back, collect if you
wish.
∂10-Dec-86 1234 PHY
To: "@THEORY.DIS[1,PHY]"@SAIL.STANFORD.EDU
I think it would be good for the `theory group' of professors to have
a lunch meeting once per quarter, so that we can coordinate plans
and perhaps tell Nils that `the theory group would like to do thus and so'.
Therefore I'm taking the ball in my hands (or by the horns or something)
and calling a meeting for next Tuesday, December 16, at the faculty
club at noon. We will all have to be on campus anyway for Gray Tuesday,
and there's no ordinary faculty lunch scheduled for that day.
I'll try to get a private dining room. Hope to see you there.
- Don Knuth
∂10-Dec-86 1242 VAL Commonsense and Non-Monotonic Reasoning Seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
CONCEPTS DEFINED VIA APPROXIMATE THEORIES
John McCarthy
Thursday, December 11, 4pm
Jordan 050
Some important concepts for AI including "it can", "it believes" and
counterfactuals may be precisely definable in theories that approximate
reality in a generalized sense. Useful approximate theories of action are
typically non-deterministic even when they approximate deterministic
systems. The concepts are useful to the extent that the approximate
theory answers questions about the real world, but they often become
imprecise when attempts are made to define them directly in real world
terms. The lecture will discuss the sense of approximation, give
some examples, and make connections with the previous discussion of
contexts. Some of the material is discussed in my paper "Ascribing
Mental Qualities to Machines".
∂10-Dec-86 1532 RA CS 306 final
SU TV people wanted to know whether it
will it be possible to give TV students an extra day to turn in their final
since they got it one day later than the rest of the class.
∂10-Dec-86 1603 DEK Pierre Baldi
To: "@THEORY.DIS[1,PHY]"@SAIL.STANFORD.EDU
I just had a call from Pierre Baldi, who applied a couple weeks ago
to our department for a position. He has just gotten his PhD in
math from Caltech. The purpose of his call was to mention that
he will be in the Bay Area tomorrow and Friday. So I invited him
to come to AFLB tomorrow; we can make his acquaintance after that meeting.
(His advisor at Caltech was Richard Wilson, who is the great
expert in combinatorial designs. Pierre says his own interests are
very close to CS; he is not "pure combinatorialist". The same
description would have applied to me in 1963, because my Caltech
PhD was also in the area of combinatorial designs!)
∂10-Dec-86 1640 PHYSICSLIB@Sierra.Stanford.EDU Re: Preserving information
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 10 Dec 86 16:39:55 PST
Date: Wed 10 Dec 86 16:35:34-PST
From: Henry E. Lowood <PHYSICSLIB@Sierra.Stanford.EDU>
Subject: Re: Preserving information
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 10 Dec 86 12:02:00-PST
Message-ID: <12261792075.34.PHYSICSLIB@Sierra.Stanford.EDU>
TO: John McCarthy
FR: Henry Lowood
Yes, I can get a copy of the full report from the NRC to you. If the
department--or somebody in it--would be willing to help us solve the
problem of devising an appropriate and practical format, that would of
course be something we would use. I hope I haven't given you another
impression; our problem is that received wisdom in the archival
community is moving in a different direction, it seems. Personally,
that doesn't concern me much; if we can do better, we should.
Henry
-------
∂11-Dec-86 0900 JMC
Fredkin
∂11-Dec-86 1100 JMC
pony bill
∂11-Dec-86 1209 RA contex[f86,jmc]
The material from your notebook is now in that file. It is not TEXed but if you
want it TEXed, let me know.
∂11-Dec-86 1439 RA
Question about CS306 final
Ratan, TV student in CS306 has a question about the final (408) 942 7836.
∂11-Dec-86 1540 RINDFLEISCH@SUMEX-AIM.ARPA Computer Museum AI Exhibit
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 11 Dec 86 15:38:30 PST
Date: Thu 11 Dec 86 15:35:03-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Computer Museum AI Exhibit
To: Les@SU-AI.ARPA, Tucker@SUMEX-AIM.ARPA, Feigenbaum@SUMEX-AIM.ARPA,
Buchanan@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA, Nilsson@SU-SCORE.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA
Message-ID: <12262043201.37.RINDFLEISCH@SUMEX-AIM.ARPA>
Friends, Oliver Strimpel from the Computer Museum in Boston called me about
help with an AI Exhibit they are preparing to open next June. He wants to have
a hands-on display of PARRY for the exhibit. It's disturbing to me that he
thinks of PARRY as the "best" available example of a working natural language
system and I've already pointed him to people like Barbara Grosz, Bonny Webber,
Terry Winograd, John Seeley Brown, and Gary Hendricks for more modern examples.
Anyway, to follow up the PARRY bit there are some technical and policy
decisions I'd like comments on:
1) Can/do we want to help Strimpel with PARRY? -- Colby punted to us. I'd say
"no" except that I think EAF (and maybe others) has had contact/interest in the
museum exhibit. If we decide to help, a policy issue would be how much control
we would want over the AI context in which PARRY might be presented before we
would help. Technically, I think PARRY has run most recently at SAIL and
SUMEX. The SUMEX (actually KI-TENEX) version was relegated to the archives
some time past and may or may not be resurrectable for use. What is the PARRY
situation at SAIL?
2) Is there a way Strimpel could have museum patrons run PARRY on one of our
machines. While this may be a reasonable community service, it seems very
difficult to me from an administrative point of view. We would have to arrange
network (TELENET or ARPANET) access, it would per force distract from academic
and research work, and such a thing seems tangential to Stanford's interests
w/r to the recent provostial policies about outside use of university
computers. Could we we get a version of PARRY that he could run on some
machine in the Boston area?
Tom R.
-------
∂11-Dec-86 1540 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU Star Wars
Received: from UCBVAX.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 11 Dec 86 15:38:55 PST
Received: by ucbvax.Berkeley.EDU (5.53/1.19)
id AA06984; Thu, 11 Dec 86 15:36:47 PST
Received: by ucscc (5.51/4.7)
id AA17383; Sun, 7 Dec 86 15:53:42 PST
Received: by ucscd (5.51/4.7)
id AA20828; Sun, 7 Dec 86 15:53:33 PST
Date: Sun, 7 Dec 86 15:53:33 PST
From: ucscc!ucscd.beeson@ucbvax.Berkeley.EDU (20012000)
Message-Id: <8612072353.AA20828@ucscd>
To: jmc@su-ai.edu, jmc@sail.stanford.edu@ucbvax.Berkeley.EDU
Subject: Star Wars
If you have time and interest, please find all the weak places in the
arguments in the attached essay on Star Wars. Your bias is different than
mine and hence you will find errors I have overlooked.
\input vanilla.sty
\title
The Truth About Star Wars
\endtitle
\author
M. Beeson
\endauthor
\def\bul{$\bullet$\ }
There are two aspects of the truth about Star Wars: technological and
political. The truth is really very simple; I shall set it out in
six statements, which will then be explained.
\medskip
\bul {\sl Star Wars can never work as a defensive system.}
\medskip
\bul {\sl Star Wars might work as an offensive system.}
\medskip
\bul {\sl Nuclear war is suicide and everybody knows it these days.}
\medskip
\bul {\sl Therefore the threat of nuclear war is not a credible threat.}
\medskip
\bul {\sl The threat of nuclear war has been the cornerstone of American
foreign policy since 1945.}
\medskip
\bul {\sl The effect of Star Wars is to make the threat of
nuclear war almost credible again.}
\medskip
The implications of these facts for the future of civilization and current
politics are then taken up, in three sections as follows:
\medskip
\bul {\sl Star Wars and Arms Control don't mix.}
\medskip
\bul {\sl The Choice: Star Wars or arms control.}
\medskip
\bul {\sl What are our leaders thinking?}
\medskip
\subheading{Star Wars can never work as a defensive system}
\medskip
This is a purely technological point, and it is necessary to pay close
attention to some details of weapons systems.
The purpose of Star Wars is to shoot down attacking missles; the plan
is to do that with high-powered laser beams. If you are going to shoot
down missiles, you have to do it during the ``boost phase'', while their
rockets are still firing. This lasts only about two minutes and is over
before the missiles leave the atmosphere. The reason you have to hit
them in the boost phase is this: after the rockets have fired, ten
warheads (the SALT II limit) are released from each missile, so you have
ten times as many targets. Moreover, if Star Wars were in place, perhaps
hundreds of ``decoy'' warheads, which look like warheads but are not,
would also be released. In addition, if SALT II is abrogated, perhaps
there would be fifty real warheads. Moreover, it would be relatively
inexpensive to shorten the boost phase to as little as forty seconds.
All these points are agreed upon by all sides in the debate over Star Wars.
Since the earth is round, and since you have to hit the missiles just after
they leave the ground, and since they will be launched from points well
inside the USSR, your lasers have to be located in space in order to
hit the missiles.
Therefore the original plan called for a number, perhaps a couple of
hundred, of orbiting ``battle stations'' from which the lasers would
be fired.
{\it This plan has now been given up}, because such battle stations
would be vulnerable to ``space mines'', which are small explosive
satellites which would sneak up unobserved and orbit near the
battle stations until exploded by a signal from the ground.
Instead, the current Star Wars plan calls for ``pop-up'' weapons.
That is, rockets would launch the laser battle stations into space
just before they were needed.
Now the crux of the matter is this: how are you going to have time to
notice the onset of an attack, get the pop-up missiles ready to fire,
decide through some human process (or, God forbid, quickly through a
computer) to fire them, fire them, get them above the atmosphere,
all during the time that the attacking missles are still in the
atmosphere, during the first forty seconds of the attack? This is
clearly and obviously impossible. You don't have to be an expert
to see that.
None of the arguments presented here depend upon it, but this seems
to be the place to point out that current plans for the Star Wars
system require a nuclear explosion to power the lasers. Vast amounts
of power are required and no other way has been suggested that might
be practical to generate the power. Whether laser beams powered by
nuclear explosions are practical is one of the things being researched.
\medskip
\subheading{Star Wars might work as an offensive system}
\medskip
A little background information about weapons systems is necessary for
this point as well. Until now, most ballistic missiles have not been
accurate enough to reliably destroy ``enemy'' missiles in their silos.
This is true of the several hundred land-based ICBM's that both sides
now have deployed. The next generation of missiles, however, will
incorporate new technology which permits almost pinpoint accuracy.
(The technology involves the computerized comparison of a camera image
of the scene below with stored satellite photographs, permitting the
computation of last-minute course adjustments.) The Pershing II
missiles recently deployed in Europe are the first missiles to be
deployed with this increased accuracy. Submarine-based missiles, too,
will achieve the same accuracy. At present they are less accurate
than land-based missiles due to imperfect knowledge of the exact
location of the submarine, but in the next generation of missiles,
they will connect to a worldwide
satellite navigation network as soon as they break the ocean surface,
and accurate position information will be supplied to on-board computers.
Each missile has ten warheads, each of which can be independently and
accurately guided to a different target.
This ``MIRV-ing'' (Multiple independently-targetable re-entry vehicles)
of the missiles means that you get ten chances to knock out each of your
opponent's missiles, assuming both sides have the same number. This
gives the first-striker an advantage. At present, however, it is still
much too dangerous to undertake a first strike, since if even a few
missiles are not knocked out, you would suffer tremendous damage from
the retaliation. Even the submarine-based missles would be enough to
deliver an unacceptable retaliatory blow, and at present there is no way
to keep track of their locations, let alone destroy them.
This is where Star Wars comes in. Suppose you were planning a first strike.
You could launch your pop-up laser battle stations simultaneously with
your attacking missiles. They would then be in position and ready to
shoot down the retaliatory missiles. There would be many fewer of these,
perhaps only the submarine-based missiles and a small percentage of the
land-based missles, so your task would be much easier. As the attacker,
you would have had time enough to prepare everything. It just might work!
\medskip
\subheading{Nuclear war is suicide and everybody knows it}
\medskip
There has been so much media attention to this point that it hardly
bears repeating. Even the hawks in the Reagan administration no longer
talk about surviving nuclear war. What with Chernobyl and the nuclear
winter, who could possibly doubt the point that nuclear war is suicide?
Every child of seven knows it. Every politician knows it. Every
general knows it. Every citizen knows it. Every Russian knows it, too.
\medskip
\subheading{Therefore the threat of nuclear war is not a credible threat}
\medskip
Who would believe that we would commit suicide? That threat is credible
only if we can make them believe we might be crazy enough to commit
suicide. This is Richard Nixon's ``madman theory'' of international
politics. In today's world even this is not enough; nobody would be crazy
enough to deliberately initiate a nuclear war under any circumstances.
\medskip
\subheading{The threat of nuclear war has been the cornerstone of American
foreign policy since 1945}
\medskip
Indeed, that is the political meaning of possessing nuclear weapons at all.
The use of nuclear weapons has been threatened perhaps some tens of times
since 1945. Nixon threatened to use them in Vietnam. Kennedy threatened
to use them during the Cuba confrontation in 1963.
The NATO policy on the defense of Europe rests on the threat of using
nuclear weapons in response to an attack by conventional weapons. It is
for this reason that the United States has refused to make a declaration
of ``no first use''. It is exactly the fact that this threat is not a
credible one that gave rise to the decision to install the Pershing II
missles in Europe: suppose that a conventional attack on Europe were
made, would the US risk Soviet nuclear retaliation by being the first
to use nuclear weapons? It is in fact no more believable that
NATO would risk the destruction of Europe by firing nuclear weapons from
European soil in that situation. The truth is that there is no
credible defense of Europe at present. Naturally this makes the
generals and politicians uncomfortable.
The Soviet Union and the United States have a continuing interest in
maintaining their influence or control over various countries in
South America, Asia, and Africa, where there are continually
``regional conflicts'' between groups supported by the US and the USSR.
The current situations in Nicaragua, Afghanistan, and Angola are
entirely typical. What restrains these conflicts from developing into
full-scale conventional warfare? The fear of escalation to nuclear war;
or to put it another way, the mutual threat of nuclear war.
As the perception grows that nuclear war is not a credible threat,
the superpowers will feel less and less restraint.
\medskip
\subheading{The effect of Star Wars is
to make the threat of nuclear war almost credible again}
\medskip
Once one understands the technological and political points made above,
this conclusion about the effect of Star Wars is obvious. With a
Star Wars system in place, it just might be believable that a desperate
US President might order a first strike. I am not suggesting that
a US President would do so; only that the presence of a Star Wars system
might make an opponent {\it believe} that he might do so. In other words,
it might return the credibility of the threat of nuclear war to whatever
value it had, say, in 1970.
The fundamental paradox underlying the arms race is that the threat of
nuclear war has never been a very credible threat. In the effort to
make the threat more real, the nations involved have piled up ten thousand
warheads, when in the event a few dozen would probably suffice to wreak
sufficient destruction, and anyway only a few hundred missiles are
available to {\it deliver} the warheads. The security of the world
at this moment rests fundamentally on our {\it trusting each other} not to
push the button.
Everyone saw what happened to the radioactive cloud that left Chernobyl.
Everyone has heard about the nuclear winter. Everyone knows that even
after a completely successful first strike, the ``victor'' will also
suffer. Nevertheless, the uncertainties involved and the nature of the
threat mean that if the symmetry between the two superpowers can be
destroyed, the threat is ``more credible''.
Finally, the possibility exists that the Star Wars system could be
used {\it directly} as a weapon against targets on the ground. The same
powerful laser beams that can incinerate a missile can also be directed
at the ground, where they can start fires in a few seconds. They will
be able, if the system can be built at all, to refocus and fire again in
a few seconds, and so in ten minutes each battle station could start
hundreds of fires, creating Dresden-style fire storms in major cities.
This possibility has so far not hit the newspapers, but has been discussed
in the scientific press. The Holocaust does not have to be nuclear,
except for the weapons that power the lasers. These explode in space,
and so do not return fallout to the families of the attackers. This
terrible possibility adds a new dimension to the ``credible threat'' posed
by Star Wars.
\medskip
\subheading{Star Wars and arms control don't mix}
\medskip
Suppose Star Wars could be successfully built.
What is the best strategic response?
One response is simply to build and deploy more missiles, to increase
the number of missiles that one expects the Star Wars system to miss.
Another response is to build your own Star Wars system, thus restoring
the symmetry of the arms race. Let us suppose for the moment that
for one reason or another the Soviets do not try to build their own Star
Wars system; will the method of building more missiles be effective?
You can try to analyze the situation quantitatively, and in the
Appendix I have tried to do so, since I am trained as a scientist.
However, there are so many unknown quantities in the equations that
one can't really do better than common sense: the side that has
Star Wars would quite like it if there were a limit on the number of missiles
that the other side could accumulate, so that there were no danger of
their getting too many to shoot down. (This is true regardless of
whether one thinks of Star Wars offensively or defensively.) Thus
at first glance it might seem that one could mix arms control and
Star Wars.
On the other hand, consider the matter from the point of view of the
side without Star Wars, who necessarily will consider Star Wars as an
offensive system threatening their retaliatory force. They will feel
safer with a large missile force. Would it make sense for them to
negotiate a treaty limiting or reducing the size of missile forces but
permitting the deployment of Star Wars? Such a limit would make
it easier for Star Wars to shoot down the few missiles that might
survive a first strike, and moreover would give the designers of Star
Wars some certainty as to how many missiles Star Wars must be prepared
to cope with. As Gorbachev put it after the Iceland summit:
``Only a madman would agree to that.''
Notice that we have left the question unanswered whether the response
of building more missiles would be effective: nobody knows, just as
nobody knows if Star Wars can be made to work at all. But one can
say with certainty at least this much: if Star Wars works, it can only
be offensively, and if it is deployed, the response will be to build
more missiles, as well as to shorten the boost phase, add decoys, and
increase the number of warheads per missile.
\medskip
\subheading{The Choice: Star Wars or arms control}
\medskip
Since they don't mix, we have to choose one or the other. What are
the consequences of the choice?
If we choose Star Wars, we can look forward to ``more of the same'':
a world politics based on two superpowers competing for domination of the
rest of the world, in a game where the scope of direct conflict is
limited by the threat of nuclear war. We can look forward to an
indefinite future of further escalation in the arms race. Star Wars
itself will violate the ABM treaty, and SALT II with its limits on the
number of missiles and more importantly on the number of warheads per
missile, will also go, since an obvious response to threat of space-based
laser weapons is more warheads per missile. Thus all progress so far
achieved in arms control will be lost, and the arms race will be restrained
only by economics. The tremendous economic waste of the arms race will
be intensified on both sides. With conflicts in Africa, Asia,
South America, and the Middle East proceeding
as they have been, arms budgets in these areas will continue to consume
resources that could be devoted to public health, agriculture, and industry.
What will not go on as before is the arms race itself. The new applications
of computers to missile guidance means that the arms race is becoming
inherently more unstable, in the sense that there is more and more incentive
to be the first-striker. Even without Star Wars, this is true, since the
next generation of missiles will be able to destroy missiles still in their
silos. It becomes a case of ``Use 'em or lose 'em!''. Even if the temptation
to strike first is resisted, the temptation to establish a policy of
``launch on warning'' will be more difficult to resist. This means a policy
whereby the missiles would be launched when satellite photos and radar both
confirmed an enemy attack, without waiting for the attack to materialize.
Since the warning time is about twenty minutes, the decision to fire would
have to be made by an officer on duty, or by a computer. The dangers
of accidental war will be multiplied greatly if this comes to pass.
What is the alternative? Does arms control offer us any real hope?
What has to be understood here is another technological point: The
capacity to destroy civilization, once achieved, cannot be lost. Even
if all nuclear weapons and missiles are destroyed, more can be built
if desired. The only thing that can be negotiated is this: How long
does it take, starting now, to destroy civilization? At present
it takes about half an hour. If all warheads and missiles were
destroyed, that time would be lengthened to several weeks. If in
addition all factories for making missiles and warheads were destroyed,
it might be lengthened to several months. If in the meantime economic
and cultural ties are strengthened, there is hope that they might be
strong enough to reverse any trend to war during those months.
What would then replace the threat of nuclear war as a restraint on
conventional conflict? Only the threat of disturbing the carefully
achieved network of treaties and co-operation; in other words, the
fear of returning to the dangerous and chaotic situation of the 1980's.
\subheading{What are our leaders thinking?}
\medskip
In the foregoing we have made several points which we believe must be
accepted by every person with the patience to read them, regardless of
their politics or their opinion about the Russians. Namely: Star Wars
doesn't make defensive sense, but might make offensive sense. Its main
effect will be to make the threat of nuclear war more credible by raising
the possibility of a first strike in which the retaliatory strike is
totally destroyed.
These facts, simple as they are, are not widely known today, and lie
hidden behind the mass of contradictory and confusing statements about
Star Wars that are reported in the press. Star Wars has been sold to
the American public as a ``defensive'' weapon, which appeals to the
desire of the people to stop threatening the Russians with doom, without
feeling vulnerable in the process. However, polls show that 80\% of the
public would prefer a treaty banning weapons from space. The impetus
for Star Wars does not come from the public.
Who is really behind Star Wars and what are their motivations and intentions?
One group in favor of Star Wars, of course, is
the defense contractors and researchers who will reap the
benefits; and the initial proponents were directors of the national
research laboratories at Livermore and Los Alamos, who stand to benefit
the most. Nevertheless, such an unpopular program could not survive
without political support at the highest levels, from the President,
the Joint Chiefs of Staff, and the advisors of the President. These
people have maintained their support in spite of the fact that it meant
rejecting generous offers for the reduction and elimination of the weapons
Star Wars is advertised as ``defending against''.
Are these people aware of the points made above, about the offensive
nature of Star Wars and its effect on the credibility of the threat
which underlies American foreign policy? It is difficult to believe
that they are not, but it is perhaps not impossible. After all, a welter
of contradictory statements have been issued since the
1983 speech announced the Star Wars program: originally the system was
supposed to protect cities and make the whole concept of deterrence
obsolete, but by June 1985, the President's science
advisor, George Keyworth, was advertising Star Wars as a ``deterrent
against first strike''. Early on it was advertised as a tool to
``force the Russians to the bargaining table'', but Weinberger
says (November 6, 1985) it's not a bargaining chip, and neither research
nor deployment can be given up under any circumstances, even if it means
abrogating the ABM treaty (which it certainly will), and in Iceland,
Reagan's refusal to give up Star Wars prevented a major reduction in
strategic arms from being negotiated.
There are two possibilities: the administration, the Joint Chiefs, and
the civilian advisors are as confused as these statements indicate, or
they are deliberately trying to confuse the public. In this paper
I shall not argue for either alternative, since I do not think either
can be convincingly proved. Moreover, a third possibility is that
some of them are confused, and some of them are deliberately trying to
confuse the others {\it and} the public.
If our leaders are confused, we should replace them. Let us therefore
assume that they are not confused. In that case, what can be their
intentions?
Until recently, it was possible to imagine that President Reagan's
intention was to negotiate an historic arms-control agreement, using
Star Wars as a bargaining chip. After the failed summit at Iceland,
it is clear that a deliberate choice is being made to pursue Star Wars,
along with the accelerated arms race and abandonment of treaties that
it implies, rather than negotiate the elimination of nuclear weapons.
The only motive for such a choice can be the pursuit of power in
South America, Africa, Asia, and the Middle East, backed by the
renewed threat of nuclear force. Regardless of the view one takes
of the causes of and blame for the
worldwide conflict between East and West, one must recognize
that the choice is being made to continue and renew the conflict.
We are witnessing an historic event: the deliberate choice on the part
of American leadership to reject a world in which the threat of nuclear
war is reduced or eliminated, in favor of a world in which a renewed
and more credible threat of nuclear war is available as an instrument of
policy. Let us all understand the choice our leadership is making for
us!
\bye
In order to answer that question, we first have to consider the simpler
question when Star Wars is not present. Given missiles each carrying
ten warheads (or more, if SALT II goes out the window), is it safer
to have a large or a small force of missiles on each side? That is,
under what conditions is it most dangerous to undertake a first strike?
The answer here is that each missile on the first-striker's side can
knock out on the average $10p$ missiles on the other side, where $p$
is the probability of knocking out a single missle that you aim at.
This is independent of the size of the arsenals. Nobody knows what $p$
is; as long as the MX is not deployed, everyone believes $10p$ is
substantially less than one, so there isn't much temptation to a first
strike. This situation will change for the worse with the deployment
of the MX, with advances in satellite navigation technology for
submarine--based missiles, and if SALT II's demise replaces 10 by 50.
How does Star Wars change the equation? Let us suppose that a single laser
battle station can destroy $N$ enemy missiles on the average when used
in conjunction with a first strike.
≠
∂11-Dec-86 1600 RA leaving
It's Thursday and I am leaving early. Do you need the Time letter today?
∂11-Dec-86 1721 HSU@Sushi.Stanford.EDU Re: problem 2
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 11 Dec 86 17:21:00 PST
Date: Thu 11 Dec 86 17:04:37-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: Re: problem 2
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Thu 11 Dec 86 15:08:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12262059506.24.HSU@Sushi.Stanford.EDU>
I have forwarded the message to the students. If there is anything else
needs to be done, please let me know.
Yung-jen
-------
∂11-Dec-86 1904 HSU@Sushi.Stanford.EDU re: problem 2
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 11 Dec 86 19:03:59 PST
Date: Thu 11 Dec 86 19:01:47-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: re: problem 2
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Thu 11 Dec 86 18:58:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12262080837.24.HSU@Sushi.Stanford.EDU>
Sure. Do you want to have a hardcopy as well?
Yung-jen
-------
∂11-Dec-86 1936 LES re: Computer Museum AI Exhibit
To: Rindfleisch@SUMEX-AIM.STANFORD.EDU
CC: Tucker@SUMEX-AIM.STANFORD.EDU, Feigenbaum@SUMEX-AIM.STANFORD.EDU,
Buchanan@SUMEX-AIM.STANFORD.EDU, JMC@SAIL.STANFORD.EDU,
Nilsson@SCORE.STANFORD.EDU
[In reply to message sent Thu 11 Dec 86 15:35:03-PST.]
A few weeks ago I received a request from an acquaintance of Bill Faught
(the last PARRY hacker) who wanted to get the source files and make it
run on a PC. I told him that he would need a giant PC and that it would
require too much work for me to locate all the sources.
I subsequently talked to Bill Faught, who is now at Intellicorp. He said
that he wouldn't mind putting the sources into an orderly form if someone
wanted to do something with it. He also said that he would come around to
visit some time. I haven't heard from him since.
I agree that we should not offer machine cycles to the Computer Museum. I
suggested to Strimpel when he was here that he could get free DEC-10s or
20s from a number of sources, but he said that he couldn't afford to
maintain them. (Of course, if he collected enough of them, he could throw
them away when they break . . .)
In summary, it appears that we could make the PARRY sources available, but
not much more. If someone hacked it to run on a maintainable workstation,
it might be usable by the Computer Museum, but whoever did that would
likely be aiming at commercial exploitation and might not want to give it
to the Museum. Hmm, but we might make that a condition of giving them the
sources . . .
Les
∂12-Dec-86 0236 DEGENHARDT@Sushi.Stanford.EDU CS 306: Questions about the final
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 12 Dec 86 02:35:54 PST
Date: Fri 12 Dec 86 02:33:49-PST
From: Jon Degenhardt <DEGENHARDT@Sushi.Stanford.EDU>
Subject: CS 306: Questions about the final
To: jmc@Sail.Stanford.EDU
cc: degenhardt@Sushi.Stanford.EDU
Message-ID: <12262163127.23.DEGENHARDT@Sushi.Stanford.EDU>
Admittedly, these questions should have been asked during class. However,
talkback from TV land wasn't working, and my main question, which is about
GMAT, wasn't formulated until after class anyway.
After listening to your description of GMAT and re-reading it's description,
I wasn't convinced I knew exactly what it was supposed to do. Here is an
alternate description. Please let me know if this sounds right.
GMAT [x,y] is true if for every path in x leading to an atom there
is a "generalized path" in y leading to the same atom. A path P is
a "generalized path" of a path Q if P contains the same sequence of
cars and cdrs as Q, though possibly having additional cars and cdrs
mixed in. A path of the form <...-car-...-cdr-...-car-...> is a
generalized path of <car-cdr-car>. Thus, <cdr-car-cdr-cdr-cdr-car>
is a generalized path of <car-cdr-car>.
This assumes no merged or circular structures.
Problem 2 (Checksub)
There are 2 questions here. The first is what to do when asked to
substitute for a bound variable. Since such substitutions are
meaningless, they can either be disallowed (ie. no legal solution)
or ignored (since we now have to match equivalent lambda expressions).
For example:
Checksub [(car x), y, (lambda (y) y), (lambda (y) y))]
The second is really just a comment. By allowing the result form
to contain renamed bound variables, captures can occur with variables
not found in x. For example:
Checksub [x, y, (lambda (x) (fun x y w)), (lambda (w) (fun w x w))]
is not valid because w was captured.
Thanks for the help,
Jon
-------
∂12-Dec-86 0800 JMC
Givan recommendation
∂12-Dec-86 0901 PERLAKI@Score.Stanford.EDU Rutie
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 12 Dec 86 09:01:42 PST
Date: Fri 12 Dec 86 09:00:10-PST
From: Agnes M. Perlaki <PERLAKI@Score.Stanford.EDU>
Subject: Rutie
To: mccarthy@Score.Stanford.EDU
Message-ID: <12262233460.20.PERLAKI@Score.Stanford.EDU>
Rutie called to say that she will be late this morning.
-Agi
-------
∂12-Dec-86 1209 RLG reminder
today is the last work day before the recommendations come due.
thanx...
∂12-Dec-86 1253 RA Barbara Vrabec
Vrabec, Tom Engel Productions (312) 664 8141 called re: conference call to
be set for Wed., Thurs., or Friday for you, Tom Engel and Ellie Gray.
∂12-Dec-86 1314 RA going out
I am going to the bookstore.
∂12-Dec-86 1439 NILSSON@Score.Stanford.EDU
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 12 Dec 86 14:39:11 PST
Date: Fri 12 Dec 86 14:37:30-PST
From: Nils Nilsson <NILSSON@Score.Stanford.EDU>
To: JMC@Sail.Stanford.EDU
cc: NILSSON@Score.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 10 Dec 86 14:39:00-PST
Message-ID: <12262294868.20.NILSSON@Score.Stanford.EDU>
Ok, I'll talk to terry about nygard. -Nils
-------
∂12-Dec-86 1531 RA leaving
I don't feel well and am going home; have a nice weekend.
∂12-Dec-86 1605 VAL Nonmonotonic reasoning seminar
To: "@CS.DIS[1,VAL]"@SAIL.STANFORD.EDU
No more meetings until January 8. Happy 1987!
Vladimir
∂13-Dec-86 1113 CLT igor
To: LES
CC: JMC
currently plans to arrive around jan 5
do we have a place for him to sit?
∂13-Dec-86 1114 CLT qlisp mailing list
q.dis[1,clt] is the mailing list
I have been using in case you want
to call a meeting.
∂13-Dec-86 1355 LES re: igor
To: CLT
CC: JMC
[In reply to message from CLT rcvd 13-Dec-86 11:13-PT.]
I suggest having him occupy the former Greep desk. Another alternative
is the third desk in 302, with Rabinov and Shankar.
∂13-Dec-86 1401 CLT igor
To: LES
CC: JMC
It seems to me that the former greep desk is better since
(a) that room is bigger
(b) it contains the lisp machine
(c) it doesn't contain the noisy rt
but maybe igor should be given the opportunity
to choose when he arrives.
∂13-Dec-86 1421 bulwinkl@rocky.stanford.edu re: don't want old magazines??
Received: from ROCKY by SAIL.STANFORD.EDU with TCP; 13 Dec 86 14:21:32 PST
Received: by rocky.stanford.edu; Sat, 13 Dec 86 14:16:13 PST
Date: 13 Dec 1986 1416-PST (Saturday)
From: Jeff Soesbe <bulwinkl@rocky.stanford.edu>
To: John McCarthy <JMC@sail.stanford.edu>
Subject: re: don't want old magazines??
In-Reply-To: John McCarthy <JMC@SAIL.STANFORD.EDU> /
12 Dec 86 2110 PST.
yeah!! those would be great, especially for animal pictures...if
you've got extras, I'll take em off your hands...
jef
∂13-Dec-86 1807 BRONSTEIN@Score.Stanford.EDU I'm always amazed...
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 13 Dec 86 18:07:03 PST
Date: Sat 13 Dec 86 18:05:32-PST
From: Alexandre Bronstein <BRONSTEIN@Score.Stanford.EDU>
Subject: I'm always amazed...
To: jmc@Sail.Stanford.EDU
Message-ID: <12262594886.9.BRONSTEIN@Score.Stanford.EDU>
...at how much weight you carry!
5 lines of yours saying that you agree with some of L.B.'s reasonning,
and we already have 3 or 4 messages agreeing partly with him. (For
the last week, he'd been defending his points alone.)
I guess the other half of the conclusion is that most people are real
chicken when it comes to fighting a mob..
Alex
-------
∂14-Dec-86 1450 bulwinkl@rocky.stanford.edu re: don't want old magazines??
Received: from ROCKY by SAIL.STANFORD.EDU with TCP; 14 Dec 86 14:50:16 PST
Received: by rocky.stanford.edu; Sun, 14 Dec 86 14:45:10 PST
Date: 14 Dec 1986 1445-PST (Sunday)
From: Jeff Soesbe <bulwinkl@rocky.stanford.edu>
To: John McCarthy <JMC@sail.stanford.edu>
Subject: re: don't want old magazines??
In-Reply-To: John McCarthy <JMC@SAIL.STANFORD.EDU> /
13 Dec 86 1453 PST.
ok...great!!!
thank you very much,
jef
∂14-Dec-86 1652 PHYSICSLIB@Sierra.Stanford.EDU Re: same author
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 14 Dec 86 16:52:15 PST
Date: Sun 14 Dec 86 16:52:40-PST
From: Henry E. Lowood <PHYSICSLIB@Sierra.Stanford.EDU>
Subject: Re: same author
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 13 Dec 86 16:41:00-PST
Message-ID: <12262843763.9.PHYSICSLIB@Sierra.Stanford.EDU>
TO: John McCarthy
Yes, I assume the listing of one author as two was a mistake made by a
cataloger, either here or at Library of Congress. We can ask to have it
corrected and send the information to LC, as well.
Henry
P. S. I sent you a copy of the NRC report.
-------
∂15-Dec-86 0804 RICHARDSON@Score.Stanford.EDU AI Courses
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Dec 86 08:00:46 PST
Date: Mon 15 Dec 86 07:56:14-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: AI Courses
To: ai.list: ;
Message-ID: <12263008254.16.RICHARDSON@Score.Stanford.EDU>
Please, folks! I would appreciate an idea of your availability the first
two weeks of Jan. for a 2-hour meeting to discuss AI courses. (Not to be
confused with the AI retreat.)
-Anne
-------
∂15-Dec-86 0900 JMC
Planlunch Waldinger
∂15-Dec-86 0936 RA telephone answering machine
I finally got an answer from ITS. They recommend getting a Panasonic
and Macy's has it now on sale for $135.
Shall I go and get for you?
Rutie
-----
∂15-Dec-86 1040 @Score.Stanford.EDU:GSMITH@SUMEX-AIM.ARPA AI Planning Retreat
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Dec 86 10:40:08 PST
Received: from SUMEX-AIM.ARPA by SU-SCORE.ARPA with TCP; Mon 15 Dec 86 10:38:32-PST
Date: Mon 15 Dec 86 10:36:49-PST
From: Grace Smith <GSMITH@SUMEX-AIM.ARPA>
Subject: AI Planning Retreat
To: McCarthy@SU-SCORE.ARPA
cc: Gsmith@SUMEX-AIM.ARPA
Message-ID: <12263037487.24.GSMITH@SUMEX-AIM.ARPA>
I haven't yet received an answer from you regarding availability of dates
for the AI planning retreat. The days now available with everyone else are
Feb. 28
Mar. 7, 14
Apr. 4, 11, 18
I would appreciate a reply from you.
Thanks,
Grace Smith
-------
∂15-Dec-86 1103 REGES@Score.Stanford.EDU couple of things
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Dec 86 11:03:52 PST
Date: Mon 15 Dec 86 11:02:07-PST
From: Stuart Reges <REGES@Score.Stanford.EDU>
Subject: couple of things
To: jmc@Sail.Stanford.EDU
Office: CS-TAC 22, 723-9798
Message-ID: <12263042092.35.REGES@Score.Stanford.EDU>
I would like to enter your EMAIL address into the departmental database. I
remember that you had a strong objection to doing so because the program
wouldn't allow you to say more than just a single EMAIL address. We now have
a facility for general text comments, so I can add a comment about JMC versus
JMC-LISTS. Would you agree to having your EMAIL field there with a text
comment?
Also, can you give me some contact information for Ray Strong and Robert Smith,
the industrial lecturers for winter and spring?
-------
∂15-Dec-86 1141 RA Taleen
Would you like Taleen to substitute for me when I am gone the last week
of December?
∂15-Dec-86 1212 REGES@Score.Stanford.EDU CS326
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Dec 86 12:12:40 PST
Date: Mon 15 Dec 86 11:37:06-PST
From: Stuart Reges <REGES@Score.Stanford.EDU>
Subject: CS326
To: jmc@Sail.Stanford.EDU
Office: CS-TAC 22, 723-9798
Message-ID: <12263048461.22.REGES@Score.Stanford.EDU>
The enrollment estimate for your Winter course is 22. It is listed as a
non-lab course, which means it requires less than 6 hours/week of programming.
The TA allocation is, therefore, 25%, (a 25% TA works 10 hours/week).
Please contact me if the lab/non-lab classification is incorrect or if you
feel your course deserves more TA support.
-------
∂15-Dec-86 1527 CRISPIN@SUMEX-AIM.ARPA Re: regulating packet switching
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Dec 86 15:26:50 PST
Date: Mon 15 Dec 86 15:23:29-PST
From: Mark Crispin <Crispin@SUMEX-AIM.ARPA>
Subject: Re: regulating packet switching
To: JMC@SU-AI.ARPA
cc: su-etc@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sun 14 Dec 86 18:30:00-PST
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12263089673.31.CRISPIN@SUMEX-AIM.ARPA>
I must disagree with JMC. I have set up a type of computer-computer
telephone call network (called PANDAnet, but heavily influenced in
design by the old Dialnet project) that presently includes 9 nodes,
of which 5 are presently online. This sort of network is quite
wonderful (especially in terms of economics) but the lack of a central
management is not always A Good Thing.
With a central management, such as ARPANET, a coherant policy is more
or less established across the entire network. Without a central
management, policies are often at the whim of the management of the
individual nodes. Dialnet class networks are highly dependent upon
management that is committed to cooperation. As soon as charging
becomes a consideration things get far more complicated than it would
with a single central management.
-------
∂16-Dec-86 0837 AAAI-OFFICE@SUMEX-AIM.ARPA Re: Space Station Automation Workshop
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 16 Dec 86 08:37:28 PST
Date: Tue 16 Dec 86 08:34:04-PST
From: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: Re: Space Station Automation Workshop
To: JMC@SU-AI.ARPA
cc: AAAI-OFFICE@SUMEX-AIM.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 15 Dec 86 18:02:00-PST
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Message-ID: <12263277283.37.AAAI-OFFICE@SUMEX-AIM.ARPA>
I'll do it whn I come back from my vacation in the new year.
OK?
Claudia
-------
∂16-Dec-86 0932 RA Maria Finitzo
Finitzo is with Tom Engel Production and she called re the tape they are
making Jan. 7 which you'll be in. She is sending you a list of questions
and would like to set a tel. conference call for Wed. or Thurs. this week.
Please let me know which day and time will be convenient for you (they are
two hours ahead of us). Her tel. is (312) 664 8141. Please call her or
let me know and I will call her.
Thanks,
∂16-Dec-86 0952 RA Dina Bolla
Betty from Dina Bolla called re your trip to Los Alamos in Feb.
She can get you the following:
South West Air 2/11, SF to Albuquerque 1:35pm
2/14 Alb. to SF 3:10pm. There is no direct flight to Los Almos. This
is a special fare ticket and must be issued by Jan. 17. Please let her
or me know what you would like to do. Would you like to rent a car? She
said that the weather might be rough and thought there may be a shuttle
service from Albuquerque to Los Alamos.
∂16-Dec-86 0957 RA tel. ansering machine
I got you your machine. I need a check from you for $144.45. I will use the
receipt to reimburse you. Do you want to charge it to your NSF account or
your unrestricted?
∂16-Dec-86 1053 STEINBERGER@SRI-KL.ARPA re: Decommissioning reactors
Received: from SRI-KL.ARPA by SAIL.STANFORD.EDU with TCP; 16 Dec 86 10:52:56 PST
Date: Tue 16 Dec 86 10:50:32-PST
From: Richard Steinberger <STEINBERGER@SRI-KL.ARPA>
Subject: re: Decommissioning reactors
To: JMC@SAIL.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Tue 16 Dec 86 09:41:00-PST
Message-ID: <12263302126.35.STEINBERGER@SRI-KL.ARPA>
You're right! I did intend to say "touted." I can remember reading literature
from Oak Ridge Labs where the phrase ("too cheap to meter") was used. I don't
have the original document, alas.
Of course the Worldwatch has its biases, but as you might suspect, I tended
to agree with the tenor of this announcement. Whether any nuclear
plants will in fact run (safely) for 30+ years remains to be seen. There
have certainly been far more mishaps and higher operating costs than anyone
would have or did predict at the start of the commercial nuclear power era.
I'm sure you're aware that no new reactors have been built or planned in this
country for some time. Of course you might argue that this is due to
over-regulation or excessive environmental concerns, but I think that even
the business press is beginning to disagree with that. In short, nuclear
power's role as a major power source for the next say 50 years is far from
guaranteed.
As to solar and wind power: You are quite correct here, in a sense. Neither
of these technologies has made overall total energy contributions anywhere
near the level (measured in any convenient energy units) of nuclear power.
But neither has had (nor has) the 30+ years of government subsidies
that created commercial nuclear power. This is not to agree that solar,
wind, water and conservation are worthless technologies. Rather they are
best suited (in general) to highly decentralized application (e.g. individual
homes having photocells, thermosyphon hot water systems or small-scale
wind or water power systems). Are all/any of these "economic" at present?
Some are, many are not; at least the way we currently do our accounting.
It is very likely that the cost/savings ratio for very many conservation
steps that most homeowners can do is considerably less than adding a comp-
arable amount of nuclear power to supply the energy.
Thanks for your comment.
Regards,
Ric Steinberger
-------
∂16-Dec-86 1126 RPG Gray Tuesday Reminder
∂15-Dec-86 0905 CHEADLE@Score.Stanford.EDU Gray Tuesday Reminder
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 15 Dec 86 09:05:17 PST
Date: Mon 15 Dec 86 09:01:46-PST
From: Victoria Cheadle <CHEADLE@Score.Stanford.EDU>
Subject: Gray Tuesday Reminder
To: faculty@Score.Stanford.EDU
Office: Margaret Jacks 258, 723-1519
Message-ID: <12263020183.18.CHEADLE@Score.Stanford.EDU>
Remember that Gray Tuesday is tomorrow from 2:30-5:00 pm, in Jacks
146. If you aren't planning to attend the meeting, please let me
know how your advisees are doing. If you feel they aren't making
reasonable progress, also state what you'd like accomplished, either by
Black Friday or before (as it will be almost 6 months before the Black
Friday meeting).
Victoria
-------
∂16-Dec-86 1131 RPG Weening
John, I also noticed that a note would suffice when I sent you the
message. Here is what I sent Cheadle:
Victoria, I've got the bad flu that is going around, and I think
I should not attend the Gray Tuesday meeting. I have been working with
Joe fairly closely, and I think he's doing better now, and certainly his
progress is satisfactory.
-rpg-
∂16-Dec-86 1227 RA Dr. Flaherty
Please call Dr. Flaherty from RPI (I don't know what RPI stands for and
forgot to ask) (518) 266 6348, re: his visit.
∂16-Dec-86 1314 KOHEN@Sushi.Stanford.EDU re: Decommissioning reactors
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 16 Dec 86 13:14:09 PST
Date: Tue 16 Dec 86 13:01:06-PST
From: Abraham Kohen <KOHEN@Sushi.Stanford.EDU>
Subject: re: Decommissioning reactors
To: JMC@Sail.Stanford.EDU
cc: STEINBERGER@SRI-KL.ARPA, su-etc@Sail.Stanford.EDU, KOHEN@Sushi.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Tue 16 Dec 86 09:41:00-PST
Message-ID: <12263325897.26.KOHEN@Sushi.Stanford.EDU>
I too recall the "too cheap to meter" slogan as being in the
ORNL/ERDA/DOE brochures and also in the demo booth at the 1964(?)
World's Fair.
The 30-year lifetime of nuclear power plants was based on
designing the major components to last for 30 years. It was
assumed that embrittlement would ensue after 30 years and the
plants would be decommissioned or the major parts rebuilt.
Unfortunately, the design estimates were under-consertvative
and embrittlement occurred at much earlier times than the nominal
30 years. So a number of reactors have had to shutdown prematurely
and have had to "retrofit" major components.
No commercial plant in the US has operated for 30 years. A number
of plants have been shutdown indefinitely after a much shorter
lifetime.
Like JMC, I am a proponent of nuclear power, but the economics
are a lot worse than many in the nuclear industry would admit to.
Abe
-------
∂16-Dec-86 1630 RPG State of Affairs
To: LES@SAIL.STANFORD.EDU
CC: JJW@SAIL.STANFORD.EDU, JMC@SAIL.STANFORD.EDU,
CLT@SAIL.STANFORD.EDU, edsel!ejg@NAVAJO.STANFORD.EDU,
edsel!arg@NAVAJO.STANFORD.EDU, edsel!hbs@NAVAJO.STANFORD.EDU
We got the kernel running on the Alliant today. In a day or two
it will be stablized. From a kernel we load in Lisp code to make a
full Common Lisp.
I don't think there is any more hacking we can do beyond stablizing the
kernel until the link is operational.
-rpg-
∂16-Dec-86 1700 RA be late tomorrow
To: JMC@SAIL.STANFORD.EDU
CC: eppley@SCORE.STANFORD.EDU
I will come in late tomorrow; probably around 10:00.
Rutie
----
∂16-Dec-86 1822 LES re: dial costs
[In reply to message rcvd 30-Nov-86 20:39-PT.]
In response to your earlier request, I just reviewed the DIAL log for the
last six months and find no systematic abuses. All calls were to area
codes 415, 408, or 800 with two exceptions: a 13-Sep call by GVB to
Washington DC and a 25-Jun call by WOL to Northern New Jersey.
GVB was an account that we later discovered had been taken over by a
cracker who was trying to get into various DoD computers. The logged call
was apparently one of those attempts. He did most of his (apparently
unsuccessful) trials via Arpanet. As soon as Marty noticed him (some
Saturday in September as I recall) the account was flushed.
I don't know why Pierre Wolper called New Jersey, but I don't plan to
inquire unless you think it might be important.
∂17-Dec-86 0800 JMC
Empire Distributing
∂17-Dec-86 0820 STEINBERGER@SRI-KL.ARPA re: Nuclear Power
Received: from SRI-KL.ARPA by SAIL.STANFORD.EDU with TCP; 17 Dec 86 08:20:50 PST
Date: Wed 17 Dec 86 08:18:36-PST
From: Richard Steinberger <STEINBERGER@SRI-KL.ARPA>
Subject: re: Nuclear Power
To: JMC@SAIL.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 17 Dec 86 01:03:00-PST
Message-ID: <12263536613.24.STEINBERGER@SRI-KL.ARPA>
John,
Clearly you and I hold differing opinions on the commercial power issue
and neither one of us is likely to be swayed much by the other's perspect-
ive. It is perhaps best if we drop this discussion for a while or
at least until one of us discovers some "new fact" that just can't wait.
In any case, I have enjoyed hearing from you and the other "pro-
nukes." I have been reminded that I need to sharpen my "debating" skills
and bone up a bit on the nuclear power issue.
Merry Christmas,
Richard Steinberger
-------
∂17-Dec-86 0905 coraki!pratt@Sun.COM
Received: from SUN.COM by SAIL.STANFORD.EDU with TCP; 17 Dec 86 09:05:20 PST
Received: from sun.Sun.COM by Sun.COM (3.2/SMI-3.2)
id AA03451; Wed, 17 Dec 86 09:04:31 PST
Received: by sun.Sun.COM (3.2/SMI-3.2)
id AA00516; Wed, 17 Dec 86 09:04:07 PST
Received: by coraki.uucp (3.2/SMI-1.2)
id AA11968; Wed, 17 Dec 86 07:12:05 PST
Date: Wed, 17 Dec 86 07:12:05 PST
From: coraki!pratt@Sun.COM (Vaughan Pratt)
Message-Id: <8612171512.AA11968@coraki.uucp>
To: John McCarthy <sail.stanford.edu!JMC@Sun.COM>
In-Reply-To: message of 16 Dec 86 2319 PST.
<8612170719.AA02133@Sun.COM>
Same seasons as here: long vacation in summer, Dec-Jan for primary and
secondary, Dec-Feb for tertiary. Used to be trimester universally but
I think there has been some shift to semesters since I left.
Considerable variation between states in the pri/sec split and their
durations in years (as of 1969 anyway).
-v
∂17-Dec-86 0948 TREITEL@Sushi.Stanford.EDU re: NATO bucks
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Dec 86 09:48:24 PST
Date: Wed 17 Dec 86 09:45:50-PST
From: Richard Treitel <TREITEL@Sushi.Stanford.EDU>
Subject: re: NATO bucks
To: JMC@Sail.Stanford.EDU
cc: su-etc@Sushi.Stanford.EDU, treitel@Sushi.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 17 Dec 86 00:45:00-PST
Message-ID: <12263552494.26.TREITEL@Sushi.Stanford.EDU>
But John, that is exactly the calculation I was trying to do. My initial
figure of $300bn was the entire military budget, including all known overheads.
Susbequent calculations were of proportions, not absolute amounts, and so the
overhead would have been pro-rated.
And yes, I do find it anomalous that US troops are in Western Europe, just as
I find it anomalous that Russian troops are in Eastern Europe (or, more
accurately, Central Europe). The first anomaly is caused by the second.
Anybody who thinks that Europe could be defended by men stationed in America
should contemplate the number of C-5s required to move a third of a million
troops to West Germany in time to prevent it being overrun. By the time
Lockheed can build that many, nobody will remember what NATO is.
- Richard
-------
∂17-Dec-86 1156 YEFF@Sushi.Stanford.EDU NatGeo
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Dec 86 11:56:04 PST
Date: Wed 17 Dec 86 11:49:34-PST
From: jeff <YEFF@Sushi.Stanford.EDU>
Subject: NatGeo
To: jmc@Sail.Stanford.EDU
Message-ID: <12263575017.25.YEFF@Sushi.Stanford.EDU>
just picked them up....thanx mucho!!!
jef
-------
∂17-Dec-86 1157 RICHARDSON@Score.Stanford.EDU Annual Faculty Report
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Dec 86 11:57:29 PST
Date: Wed 17 Dec 86 11:55:54-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: Annual Faculty Report
To: jmc@Sail.Stanford.EDU
Message-ID: <12263576171.9.RICHARDSON@Score.Stanford.EDU>
The Dean's office has sent me a second request for Annual Faculty Reports.
May I ask that you send that to me and I will forward it?
Thanks,
Anne
-------
∂17-Dec-86 1237 @Score.Stanford.EDU:ROSENBLOOM@SUMEX-AIM.STANFORD.EDU Graduate Cognitive Science Program
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Dec 86 12:36:46 PST
Received: from SUMEX-AIM.STANFORD.EDU by SU-SCORE.ARPA with TCP; Wed 17 Dec 86 12:30:29-PST
Date: Wed 17 Dec 86 12:28:33-PST
From: Paul Rosenbloom <ROSENBLOOM@SUMEX-AIM.STANFORD.EDU>
Subject: Graduate Cognitive Science Program
To: Binford@SCORE.STANFORD.EDU, Buchanan@SUMEX-AIM.STANFORD.EDU,
Feigenbaum@SUMEX-AIM.STANFORD.EDU, Genesereth@SUMEX-AIM.STANFORD.EDU,
McCarthy@SCORE.STANFORD.EDU, Nilsson@SUMEX-AIM.STANFORD.EDU,
Rosenbloom@SUMEX-AIM.STANFORD.EDU, Shortliffe@SUMEX-AIM.STANFORD.EDU,
Winograd@SCORE.STANFORD.EDU
Message-ID: <12263582114.42.ROSENBLOOM@SUMEX-AIM.STANFORD.EDU>
I am in the process of developing a revised brochure for the
interdisciplinary graduate program in cognitive science and would like
to list relevant AI faculty members who are willing to have their name
associated with it. This does not involve any obligations except for
the willingness to be contacted occasionally by interested students.
If you are willing to do so, please let me know. For your reference,
I have included the one paragraph description that currently exists
for the program. -- Paul
Computer Science, Linguistics, Philosophy and Psychology are
participating in an interdisciplinary program of Cognitive Science.
The program is intended to provide students with interdisciplinary
education as well as a deeper concentration in one of the constituent
disciplines. Doctoral students in one of the four associated
departments are eligible to participate in the Cognitive Science
Program. Students who complete the requirements within theiry
department and the Cognitive Science requirements will have the Ph.D.
granted from the parent department with a special designation in
Cognitive Science. To receive the field designation, students must
complete 30 units of approved courses, 18 of which must be taken in
two disciplines outside the student's major.
-------
∂17-Dec-86 1358 RA Chicago conference operator
Please call the local operator and ask them for the Chicago Conference
operator no. 511 when you are available for the conference call.
∂17-Dec-86 1409 RA trip to Portland
Franklin at Dina Bolla would like to know whether you an Carolyn are still
on for the trip to Portland 12/31, 5:30pm and back 1/4/87, 5:00. Please
let me know and I'll call him back.
Thanks,
∂17-Dec-86 1418 berke@CS.UCLA.EDU approximate theories and naming
Received: from [128.97.2.2] by SAIL.STANFORD.EDU with TCP; 17 Dec 86 14:18:17 PST
Received: by zeus.CS.UCLA.EDU (Sendmail 5.51/5.14)
id AA28153; Wed, 17 Dec 86 14:18:43 PST
Date: Wed, 17 Dec 86 14:18:43 PST
From: berke@CS.UCLA.EDU (Peter Berke)
Message-Id: <8612172218.AA28153@zeus.CS.UCLA.EDU>
To: jmc@sail.stanford.edu
Subject: approximate theories and naming
Cc: VIDAL@usc-ecl.arpa, berke@CS.UCLA.EDU, vidal@locus.ucla.edu
Dear Dr. McCarthy,
It was very good to meet you. After our conversation and your seminar on
approximate theories I wrote down the following thoughts:
I was unsure from your discussion of theories approximating each
other, whether the theories had to be approximate. Can we
assume that two agents can have (and, in the sense that they have
to define their own variables, must have) their own, well-defined
theories? They just don't match when compared by some objective
external measurement (supposing one available).
Quantitative differences are differences in the values of shared
variables.
Qualitiative differences are differences in what the variables
measure, differences in the definition of the variables in a
system.
One theory quantitatively approximates another when it has a
"shared variable," that is, when each has a variable that is
defined in the same way, but the measurements may differ. If
this is the only kind of difference between the theories, then
the theories are scale-approximations of each other.
One theory qualitiatively approximates another when its variables
only partly correspond to the variables in the other. That is,
the variable definitions overlap in some way. In this sense,
quantitative approximation is a form of qualitative
approximation, in that quantitative approximation occurs when
variable definitions overlap exactly between systems.
At the other extreme are two theories whose variable definitions
don't overlap at all. If experience plays a role in the way we
as independent agents formulate our theories, as my work
indicates it must, then two agents with two non-overlapping
theories are two agents with no common experiences: e.g., alien
beings who could not communicate, and possibly not even perceive
each other. Since agents have to grow their symbols in
reaction to experience, agents with completely identical theories
must have had the same experiences, which since, no two things
can have the same experiences, these don't exist. Since beings
must be in some sense finite, I think it is possible for them to
not have anything in common.
This sense of overlap between the definitions of variables or
symbols, however, may end up being Frege's sense, though it may
be useful initially to create a formal system which may
extensionally equate this sense of overlap with syntactic
identity of the definitions. Especially if the agents will be
working together.
In a sense, DNA gives us a certain kind of shared symbology to
start with. How this shared symbology gets encoded into DNA would
be nice to explain by a hierarchy of experience-encoding
mechanisms which ultimately distribute critical shared knowledge
to every cell in an organism, and thus to its offspring.
Some of these thoughts may overlap with those in the papers you
have given to me, but I have not had the time to read them. I
intend to carefully read them over Christmas.
In the paper I gave you on Naming and Knowledge, I try to outline
the complexity of structures that must be incorporated into an
independent being that reasons. Since this complexity exceeds
that of previous formal work, I explicitly avoided any discussion
of communication between agents.
Communication between agents with formal systems "in their heads"
has to be some kind of approximate pattern matching. The problem
with this is that the approximate pattern matching has to arise
from the same mechanisms that give rise to the internal
structures. Thus, my view of communication with another being
should be also generated by what I have come to call structural
reaction to environmental stimuli. I'm not saying I have
accomplished this, merely that I am heading in that direction.
This is outlined in the four-page document that I gave you on my
research interests.
I would greatly appreciate your reaction to the papers I gave
you, as I respect your opinion and your work. The naming paper
is too long to submit to IJCAI, so I'm not sure what to do with
it. It needs to be rewritten, but I feel at this draft that its
errors are in the writing and not in the material content of the
paper. Since it makes no mention of functions or variables,
Church naturally has certain problems with it. It represents a
shift from Frege back to Mill on the small but critical point of
common names. This developed because I needed to assume names
for functions and variables, so I tried to do as much as possible
with naming and found that I didn't need to mention functions and
variables if I allowed ambiguous reference, i.e., Mill's common
names.
If you have any recommendation as to where I might submit it for
publication, I would appreciate it. I don't know anyone whose
interests overlap logic, language, and adaptation; the interplay
between logical and physical structures. Except of course, my
advisor, Dr. Jacques J. Vidal at UCLA, who, I hope you don't
mind, I am sending a copy of this message.
Let me thank you again for our conversation and the invitation to
your seminar.
Sincerely,
Peter Berke
∂17-Dec-86 1458 LIBRARY@Score.Stanford.EDU book on hold
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Dec 86 14:56:59 PST
Date: Wed 17 Dec 86 14:55:18-PST
From: C.S./Math Library <LIBRARY@Score.Stanford.EDU>
Subject: book on hold
To: jmc@Sail.Stanford.EDU
Message-ID: <12263608831.29.LIBRARY@Score.Stanford.EDU>
STANFORD UNIVERSITY
MATH & COMPUTER SCIENCE LIBRARY
LIBRARY@SCORE
723-4672
12/17/86
The following item is being held for you at the Math & Computer Science
Library:
CALL# QA303.G74
AUTHOR Grabiner, Judith V.
TITLE The origins of Cauchy's rigorous calculus.
It should be picked up by 12/24/86.
-------
∂17-Dec-86 2210 ILAN@Score.Stanford.EDU re: I don't know about you guys
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 17 Dec 86 22:10:39 PST
Date: Wed 17 Dec 86 22:08:57-PST
From: Ilan Vardi <ILAN@Score.Stanford.EDU>
Subject: re: I don't know about you guys
To: JMC@Sail.Stanford.EDU
cc: helen@Psych.Stanford.EDU, su-etc@Score.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 17 Dec 86 21:43:00-PST
Message-ID: <12263687772.22.ILAN@Score.Stanford.EDU>
Though I agree with John McCarthy, I heard Phil Donahue use the word
``irregardless'', this makes a strong case for Helen Cunningham.
-------
∂18-Dec-86 0734 CHANDRA%OSU-20@ohio-state.ARPA Support Requested for Workshop
Received: from OHIO-STATE.ARPA by SAIL.STANFORD.EDU with TCP; 18 Dec 86 07:34:12 PST
Return-Path: <CHANDRA%OSU-20@ohio-state.ARPA>
Received: from OSU-20 (osu-20.ARPA) by ohio-state.ARPA (4.12/6.1.OSU-CIS)
id AA00294; Thu, 18 Dec 86 10:30:58 est
Message-Id: <8612181530.AA00294@ohio-state.ARPA>
Date: Thu 18 Dec 86 10:32:27-EST
From: B. Chandrasekaran <Chandra%OSU-20@ohio-state.ARPA>
Subject: Support Requested for Workshop
To: JMC@SU-AI.arpa
Cc: hendler@brillig.umd.edu, Chandra%OSU-20@ohio-state.ARPA
John,
As Chair of the Workshop described in the enclosed description,
and on behalf of the Program Committee, I'd like to request AAAI
support for this Workshop. We expect that a portion of the grant
will go to support graduate student attendance, as has been the case
with the two previous workshops for which I had obtained AAAI support
(AIM workshop, 1984; and High Level Tools workshop, 1986). We also
expect that the grant will help us to keep registration expenses low,
and will also help us to support the travel of speakers who may
not have travel support available. (A good number of prospective
attendees are from the West Coast.)
Thanks for your attention to this matter.
Sincerely,
Chandra
------------------
Workshop on
Theoretical Issues in Conceptual Information Processing
June 4-5, 1987
University of Maryland, College Park, MD
Over the last several years, a series of workshops under the title,
"Theoretical Issues in Conceptual Information Processing" have been
held at Yale, Georgia Institute of Technology, and most recently in
Philadelphia in conjunction with the AAAI Annual meeting. Issues common to
representation and organization of knowledge and memory for natural language
understanding, planning, problem solving, and explanation have been among
the topics that have been covered in these workshops.
The goal of the investigations under this label has been understanding
intelligence and cognition
computationally, rather than merely the construction of performance
programs or formalization per se. The approaches
are united by a concern with representation, organization and
processing of conceptual knowledge. Another common element has been
an emphasis on empirical investigation of these phenomena by construction
of computer programs that perform these tasks in a computationally
tractable manner.
We plan to hold the next workshop in the series at the University of
Maryland during June 4-5, 1987. We expect this workshop to adopt a format
of both presented papers as well as a number of panel discussions.
Another feature of the workshop is expected to be a significant
participation by graduate students.
Workshop Chair: B. Chandrasekaran, The Ohio State University
Program Commitee: Jim Hendler, University of Maryland (Chair)
Richard Alterman, Brandeis U.
Jaime Carbonell, CMU
Michael Dyer, UCLA
-------
∂18-Dec-86 0926 RA Re: answering machine
[Reply to message recvd: 16 Dec 86 18:01 Pacific Time]
I called ITS and they said that the person who can explain how to install the
machine Sol in Instalation. Would you like to talk to him directly?
His tel. is 5-0524.
Rutie
------
∂18-Dec-86 0948 VAL
How about discussing the course some time this afternoon?
∂18-Dec-86 0957 RA Hotel in Chicago
Franklin from Dina Bolla called to let you know that the Merriot Hotel in
Chicago is confirmed for Jan. 6 for one night. He wanted to know whether
you are going for a special meeting because this way you might get a
special rate. Otherwise the rate is $140/night.
∂18-Dec-86 1023 VAL re: reply to message
[In reply to message rcvd 18-Dec-86 10:02-PT.]
2pm is fine.
∂18-Dec-86 1041 edsel!sunvalleymall!jlz@navajo.stanford.edu Lisp Journal announcement
Received: from NAVAJO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Dec 86 10:41:10 PST
Received: by navajo.stanford.edu; Thu, 18 Dec 86 10:40:24 PST
Received: from sunvalleymall.edsel.uucp by edsel.uucp (2.2/SMI-2.0)
id AA05043; Thu, 18 Dec 86 10:26:55 pst
Received: by sunvalleymall.edsel.uucp (1.1/SMI-3.0DEV3)
id AA07683; Thu, 18 Dec 86 10:34:42 PST
Date: Thu, 18 Dec 86 10:34:42 PST
From: edsel!sunvalleymall!jlz@navajo.stanford.edu (Jan Zubkoff)
Message-Id: <8612181834.AA07683@sunvalleymall.edsel.uucp>
To: navajo!JMC%SAIL.STANFORD.EDU@navajo.stanford.edu
Cc: sunvalleymall!jlz@navajo.stanford.edu
In-Reply-To: John McCarthy's message of 17 Dec 86 1614 PST
Subject: Lisp Journal announcement
Thanks for your correction and comments. I will be speaking to
the publisher today.
---jan---
∂18-Dec-86 1448 RA IBM RT loan
Mary from Pat Devaney office called re IBM RT loan you have through the
Ramus project. She says that it is important that you get in touch with
her today or tomorrow.
Her tel. is 3-3504.
∂18-Dec-86 1439 NSH Munindar Paul Singh
I am not aware of such a person. Since I was
working in the institute (tower), my contacts
were limited. Let me know if you would like
me to ask some one more knowledgeable.
Shankar
∂18-Dec-86 1553 RA leaving
It is Thursday and I am leaving early.
∂18-Dec-86 1621 RICHARDSON@Score.Stanford.EDU Annual Faculty Report
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Dec 86 16:21:50 PST
Date: Thu 18 Dec 86 16:19:45-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: Annual Faculty Report
To: jmc@Sail.Stanford.EDU
Message-ID: <12263886347.28.RICHARDSON@Score.Stanford.EDU>
John,
Our records here indicate that you did not turn in a faculty report last
year so I double-checked with SOE. Their records concur.
Below is an on-line version of the form.
-Anne
-----
23-Oct-86 11:45:01-PDT,3955;000000000001
Return-Path: <AUSTIN-KITZMILLER@Sierra.Stanford.EDU>
Received: from Sierra.Stanford.EDU by SU-SCORE.ARPA with TCP; Thu 23 Oct 86 11:43:43-PDT
Date: Thu 23 Oct 86 10:12:17-PDT
From: Shari I. Austin-Kit <AUSTIN-KITZMILLER@Sierra.Stanford.EDU>
Subject: Faculty Report
To: richardson@Score.Stanford.EDU
cc: austin-kitzmiller@Sierra.Stanford.EDU
Message-ID: <12249128465.14.AUSTIN-KITZMILLER@Sierra.Stanford.EDU>
STANFORD UNIVERSITY-SCHOOL OF ENGINEERING
ANNUAL FACULTY REPORT FOR ACADEMIC YEAR 1985-86
Dear Colleague,
It is time again for a Faculty Report. This office finds it very
useful to have the information outlined below, and I appreciate your
taking time to fill out the form carefully. I realize that this represents
only a summary of your contributions to the School and misses completely
your goodwill and spirit which are equally important to our mission.
Please give this completed form to your departmental secretary by December
1, 1986. Thanks for your help with this chore and for your contributions
to the School and the University.
Cordially,
Jim Gibbons
Dean
(Please note: Information requested pertains to the period 9/1/85 to 8/31/86
only.)
Name _____________________________________________________________________
Last First Middle
Academic Rank ____________________________________________________________
Department _______________________________________________________________
Teaching - Please indicate by quarter, course title, number of units and
enrollment. Also include course or curriculum development, computer
education software tutorials, specially prepared television presentations
or other relevant work.
Academic Advising
1. Number of freshman advisees. __________
2. Number of other undergraduate advisees.__________
3. Number of graduate advisees. __________
Supervision of Ph.D Candidates
1. Number of students for which you
are principal dissertation advisor. __________
2. Number of students for which you are
on reading committee. __________
Publications (Please indicate nature of work, such as books, monographs,
journals, technical reports, etc., giving title, date, pages and
publisher or issuing agency. Include only items actually published and
for archival journals include papers accepted for publication. Do not
include papers submitted for publication.)
Books and contributions to books.
Archival Journal Articles
Refereed Symposia Publications.
Technical Reports.
Presentations at Meetings and Symposia.
Research Projects
Project title and Names of Principal Approx. annual
Funding Source and co-Principal dollar value of
Investigators.(if project for which
any). you are responsible.
University Services Other Than Teaching and Research. (Include administrative
duties and other committee work.)
Professional Activities Outside the University. (Include offices in
professional organizations, services to government agencies or industry,
editorship of journals, invited presentations, and outside administrative
or public service.)
Honors and Awards
Other. (Describe below any relevant activities or make any comments that
do not fit under previous categories.)
-------
-------
-------
∂18-Dec-86 1700 JMC
go
∂18-Dec-86 1716 BUCHANAN@SUMEX-AIM.STANFORD.EDU Re: Computer Museum AI Exhibit
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Dec 86 17:16:26 PST
Date: Thu 18 Dec 86 17:12:51-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.STANFORD.EDU>
Subject: Re: Computer Museum AI Exhibit
To: Rindfleisch@SUMEX-AIM.STANFORD.EDU
cc: Les@SAIL.STANFORD.EDU, Tucker@SUMEX-AIM.STANFORD.EDU,
Feigenbaum@SUMEX-AIM.STANFORD.EDU, JMC@SAIL.STANFORD.EDU,
Nilsson@SCORE.STANFORD.EDU
In-Reply-To: <12262043201.37.RINDFLEISCH@SUMEX-AIM.ARPA>
Message-ID: <12263896015.49.BUCHANAN@SUMEX-AIM.STANFORD.EDU>
Tom,
PARRY would not be a bad exhibit as showing state of the art in
natural language processing circa 1968. But to present it otherwise
would be false. I think Strimple would not want to put misleading or
wrong labels on his exhibits, and would yield to persuasion from us.
Is there any machine still running that would run PARRY, or would
someone need to massage its Lisp? Strimpel would like not to pay anyone
to do such things, which seems unreasonable. That is, he wants it done
by "free" labor at a university. Much better if it is run on a machine in
Boston.
Alternative: use the PC version of ELIZA that is floating around for
the exhibit. I used to have a copy, maybe still do.
bgb
-------
∂18-Dec-86 1751 LES re: Computer Museum AI Exhibit
To: BUCHANAN@SUMEX-AIM.STANFORD.EDU, Rindfleisch@SUMEX-AIM.STANFORD.EDU
CC: Tucker@SUMEX-AIM.STANFORD.EDU, Feigenbaum@SUMEX-AIM.STANFORD.EDU,
JMC@SAIL.STANFORD.EDU, Nilsson@SCORE.STANFORD.EDU
[In reply to message from BUCHANAN@SUMEX-AIM.STANFORD.EDU sent Thu 18 Dec 86 17:12:51-PST.]
Parry is still running on SAIL under the pseudonym "Harry," but there are
now just two copies of the WAITS operating system in the universe and it
is unreasonable to expect either of them to become a Parry server for the
museum.
Lets face it, Eliza is _not_ an alternative to Parry -- about the only
things that they have in common are that they are both interactive natural
language programs that are mildly amusing. Eliza usually starts repeating
itself within five exchanges, while Parry can go on for hundreds of rounds
without being repetitious. Incidentally, while Parry began life in the
late '60s, it continued to evolve through about 1974.
I still believe that Parry could be made to run on a workstation with a
modest effort. Pehaps we (or even I) should approach OTL to see if they
can find anyone out there who is interested in licensing it and capable of
doing something with it -- there is one known candidate already. We could
stipulate that educational institutions, including museums, would get it
at bargain rates.
Les
∂18-Dec-86 1821 RINDFLEISCH@SUMEX-AIM.STANFORD.EDU re: Computer Museum AI Exhibit
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Dec 86 18:21:24 PST
Date: Thu 18 Dec 86 18:18:10-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.STANFORD.EDU>
Subject: re: Computer Museum AI Exhibit
To: LES@SAIL.STANFORD.EDU, BUCHANAN@SUMEX-AIM.STANFORD.EDU
cc: Tucker@SUMEX-AIM.STANFORD.EDU, Feigenbaum@SUMEX-AIM.STANFORD.EDU,
JMC@SAIL.STANFORD.EDU, Nilsson@SCORE.STANFORD.EDU,
Rindfleisch@SUMEX-AIM.STANFORD.EDU
In-Reply-To: Message from "Les Earnest <LES@SAIL.STANFORD.EDU>" of Thu 18 Dec 86 17:51:00-PST
Message-ID: <12263907903.83.RINDFLEISCH@SUMEX-AIM.STANFORD.EDU>
Bob Tucker has found there is a copy of PARRY that runs on our 2020. These
machines have no residual market value and Strimpel may be able to get DEC to
loan one from their grave yard.
Tom R.
PS to JMC: Colby is the one who directed Strimpel to talk to us about getting
access to PARRY.
-------
∂18-Dec-86 2122 G.BEESON@CSLI.STANFORD.EDU commonsense reasoning
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Dec 86 21:22:29 PST
Date: Thu 18 Dec 86 21:21:55-PST
From: Michael Beeson <G.BEESON@CSLI.STANFORD.EDU>
Subject: commonsense reasoning
To: jmc@CSLI.STANFORD.EDU
Has anyone ever made a serious attempt to catalog or inventory the
styles of commonsense reasoning and the categories of knowledge to be
represented in your hypothetical "commonsense data base"? I have in
mind using ordinary English as the representation language, for an
extensive survey of the domain eventually to be formalized in the
hypothetical formalism which you say we should develop for formalizing
common sense. How can we develop a suitable formalism if we don't know
exactly what it is supposed to do?
-------
∂18-Dec-86 2129 G.BEESON@CSLI.STANFORD.EDU commonsense reasoning & knowledge
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Dec 86 21:29:12 PST
Date: Thu 18 Dec 86 21:28:39-PST
From: Michael Beeson <G.BEESON@CSLI.STANFORD.EDU>
Subject: commonsense reasoning & knowledge
To: jmc@SU-AI.ARPA
Has anyone ever tried to make a reasonably detailed survey of the kinds
of knowledge and reasoning you understand by "common sense"? You say we
should develop a formalism in which common sense can be represented, but
I say we should first make a fairly extensive first attempt at it in
English, writing down in a fair amount of detail what categories of
knowledge we have in mind, e.g. knowledge about physical objects and their
characteristics, about simple kinds of actions and their effects and
their pre-conditions, about goals and plans of others, what others know,
cultural knowledge, basic intellectual knowledge, etc., etc.
Has anyone tried to do this?
-------
∂18-Dec-86 2134 G.BEESON@CSLI.STANFORD.EDU commonsense and natural language
Received: from CSLI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Dec 86 21:34:05 PST
Date: Thu 18 Dec 86 21:33:35-PST
From: Michael Beeson <G.BEESON@CSLI.STANFORD.EDU>
Subject: commonsense and natural language
To: jmc@SU-AI.ARPA
I've been reading Just and Carpenter, "The Psychology of Reading and
Language Comprehension", which I can recommend to you as full of interesting
examples and psychological experiments. For example, on p. 117 we find
the following question which is used to test whether fourth-grade children
have learned the vocabulary word "commended":
"When father heard that Lisa had ripped up the letter from Steve,
father COMMENDED her for it".
I leave to you the amusing exercise of formalizing the knowledge and reasoning
involved in answering the question "What do you think father thought of
Steve?" You will see that it is quite involved and a knowledge of the
meaning of "commended" is only a small part of it, but presumably the
fourth-graders have no trouble with it!
-------
∂18-Dec-86 2135 WINSLETT@Score.Stanford.EDU more reagan library
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 18 Dec 86 21:35:19 PST
Date: Thu 18 Dec 86 21:33:17-PST
From: Marianne Winslett <WINSLETT@Score.Stanford.EDU>
Subject: more reagan library
To: jmc@Sail.Stanford.EDU
Message-ID: <12263943425.30.WINSLETT@Score.Stanford.EDU>
I found out last night that my husband, a closet gung-ho Reagan
supporter, also vehemently opposes the Reagan Library because of
the site they have chosen. So there are at least two such people.
Also, I had thought they were putting it next to Behavioral Sciences
and now read in the CR that it will be smack in the middle of things,
up on the hillside. That's much worse than the plan that I thought
they had selected.
--Marianne
-------
∂18-Dec-86 2206 CHANDRA%OSU-20@ohio-state.ARPA Workshop grant request again
Received: from OHIO-STATE.ARPA by SAIL.STANFORD.EDU with TCP; 18 Dec 86 22:05:28 PST
Return-Path: <CHANDRA%OSU-20@ohio-state.ARPA>
Received: from OSU-20 (osu-20.ARPA) by ohio-state.ARPA (4.12/6.1.OSU-CIS)
id AA10553; Fri, 19 Dec 86 01:02:14 est
Message-Id: <8612190602.AA10553@ohio-state.ARPA>
Date: Fri 19 Dec 86 01:03:48-EST
From: B. Chandrasekaran <Chandra%OSU-20@ohio-state.ARPA>
Subject: Workshop grant request again
To: JMC@SU-AI.arpa
Cc: hendler@brillig.umd.edu
I am told by Jim Hendler that University of Maryland Institute
for Automation will match the AAAI contribution dollar per dollar
for the conduct of the workshop upto $5,000.
-------
∂19-Dec-86 0658 hendler@brillig.umd.edu ~m
Received: from BRILLIG.UMD.EDU by SAIL.STANFORD.EDU with TCP; 19 Dec 86 06:58:01 PST
Received: by brillig.umd.edu (5.9/4.7)
id AA18464; Fri, 19 Dec 86 09:56:04 EST
Date: Fri, 19 Dec 86 09:56:04 EST
From: Jim Hendler <hendler@brillig.umd.edu>
Message-Id: <8612191456.AA18464@brillig.umd.edu>
To: JMC@su-ai.arpa
Subject: ~m
From: B. Chandrasekaran <Chandra%OSU-20@ohio-state.arpa>
Subject: Workshop grant request again
To: JMC@su-ai.arpa
I am told by Jim Hendler that University of Maryland Institute
for Automation will match the AAAI contribution dollar per dollar
for the conduct of the workshop upto $5,000.
Actually, that's the "University of Maryland Institute for Advanced Computer
Studies,' but the monetary information is correct.
∂19-Dec-86 0941 ANDY@Sushi.Stanford.EDU Re: animal rights
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 19 Dec 86 09:40:51 PST
Date: Fri 19 Dec 86 09:38:16-PST
From: Andy Freeman <ANDY@Sushi.Stanford.EDU>
Subject: Re: animal rights
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 19 Dec 86 00:02:00-PST
Message-ID: <12264075405.15.ANDY@Sushi.Stanford.EDU>
Harinder Singh claims that animal-free cosmetic companies exist.
I'm not sure their existence proves much, since freedom of
information will let them use the results of testing by
other companies if the FDA doesn't make such information
available.
I'm willing to believe that they'd claim to be free of that
too, but I wouldn't believe it unless I sued and they didn't
use animal tests as a defense.
In the absence of changes to the liability laws (that would
let a consumer products company hide behind "don't stick this
in your eyes" labels), I don't think the comparison is fair.
With such changes, people could vote with their dollars.
(Animal testing is expensive, so I'm biasing the test. Of
course, I'm more interested in the legal changes ....)
-andy
-------
[In reply to message rcvd 18-Dec-86 16:09-PT.]
Is philpa.[e80,jmc] the list of papers for inclusion in the book, or the
reading list for the course, or both?
∂19-Dec-86 1127 RA Dr. Gibbon
Mary Cloutier from Gibbon's office would like you to come by Gibbon's office
for 10 min. at 3:15. She said it was important but could not tell me
what it was about. Her tel. 3-3938. His office is Terman 214.
∂19-Dec-86 1358 RA meeting with Gibbon
Are you going to make it to your meeting with Gibbon? If not, please let
me know.
Thanks,
Rutie
------
∂19-Dec-86 1419 RA leave early
I need to leave at 4:20 today for a doctor appointment. Is there anything
you would like me to do before I leave?
∂19-Dec-86 1422 HSU@Sushi.Stanford.EDU finals and grade sheet
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 19 Dec 86 14:20:34 PST
Date: Fri 19 Dec 86 14:18:00-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: finals and grade sheet
To: jmc@Sail.Stanford.EDU
cc: hsu@Sushi.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12264126327.29.HSU@Sushi.Stanford.EDU>
Prof. McCarthy,
We have finished grading problems 1, 3 and 5. Abhay is grading problems
2 and 4 now. He'll probably finish them by tomorrow. Will you be in your
office any time during the weekend, so that I can bring the finals to you
for reviewing problem 6 (the essay question)? I'll also bring the handouts
and the accumulated grades of the students. By the way, do you have the
grade sheet with you? Rutie said she seem to have given it to you, but
she is not sure. We'll have to return the grades by Monday noon.
Yung-jen
-------
∂19-Dec-86 1503 HSU@Sushi.Stanford.EDU re: finals and grade sheet
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 19 Dec 86 15:03:39 PST
Date: Fri 19 Dec 86 14:58:16-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: re: finals and grade sheet
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 19 Dec 86 14:52:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12264133659.29.HSU@Sushi.Stanford.EDU>
Thanks for your prompt reply, I'll let you know when I got the finals back
from Abhay.
Yung-jen
-------
∂20-Dec-86 1705 OKUNO@SUMEX-AIM.STANFORD.EDU Re: junet
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 20 Dec 86 17:05:10 PST
Date: Sat 20 Dec 86 17:04:59-PST
From: Hiroshi "Gitchang" Okuno <Okuno@SUMEX-AIM.STANFORD.EDU>
Subject: Re: junet
To: JMC@SAIL.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 20 Dec 86 16:59:00-PST
Message-ID: <12264418871.22.OKUNO@SUMEX-AIM.STANFORD.EDU>
Please use the following address:
ito%aoba.tohoku.junet%utokyo-relay.csnet@relay.cs.net
If you want to send him an express mail, please use the following instead:
ito#aoba.tohoku.junet%nttca.uucp@shasta.stanford.edu.
The former is official and the latter is via NTT's private link.
If you have a question, please ask me at any time.
Good luck,
- Gitchang -
-------
∂21-Dec-86 1919 HSU@Sushi.Stanford.EDU final
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Dec 86 19:19:14 PST
Date: Sun 21 Dec 86 18:53:56-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: final
To: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12264700848.20.HSU@Sushi.Stanford.EDU>
Prof. McCarthy:
I haven't got the finals bact from Abhay Mehta. I've been trying to
contact him since yesterday, and I finally got hold of him just now;
however, he said he still needed about 3 hours before he could finish
the grading. Will you be in your office tomorrow? I will put the
finals in your office as soon as I got them. The grades are due
tomorrow noon, but it should not take too long to grade problem 6.
(I have done some programs to calculate the total scores once we got
the final scores and decided the appropriate weights for the homeworks,
midterm and final. So, when we have everything, I can prepare a sorted
total scores for you to assign the final grades.)
Yung-jen
-------
∂21-Dec-86 1927 HSU@Sushi.Stanford.EDU handouts
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Dec 86 19:27:31 PST
Date: Sun 21 Dec 86 19:24:55-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: handouts
To: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12264706487.20.HSU@Sushi.Stanford.EDU>
Prof. McCarthy:
I have the hardcopy of the handouts ready, and I'll give them to you
tomorrow. But I wasn't able to transfer those files to the [206,jmc]
directory due to access protections. Those files are in my directory
<hsu.cs306> on SUSHI. The file names are
F86H*.* (for the handouts)
*.hlp (for the help messages about the problem sets)
Yung-jen
-------
∂21-Dec-86 2319 HSU@Sushi.Stanford.EDU re: final
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Dec 86 23:18:55 PST
Date: Sun 21 Dec 86 23:16:17-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: re: final
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sun 21 Dec 86 21:04:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12264748606.18.HSU@Sushi.Stanford.EDU>
I'm still waiting for Abhay's message, and I've asked him to send you
a note when he's done. I will ask Claire Stager about the possibility
of sending in the grade sheet late tomorrow afternoon.
Yung-jen
-------
∂21-Dec-86 2337 AMEHTA@Sushi.Stanford.EDU Final
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 21 Dec 86 23:37:03 PST
Date: Sun 21 Dec 86 23:34:20-PST
From: Abhay Mehta <AMEHTA@Sushi.Stanford.EDU>
Subject: Final
To: jmc@Sail.Stanford.EDU
cc: hsu@Sushi.Stanford.EDU
Message-ID: <12264751894.12.AMEHTA@Sushi.Stanford.EDU>
I have placed 23 graded finals in Hsu's drawer. These finals are completely
graded except for problem 6. There is one additional final that I have
put in Gianluigis folder (only problems 2 & 4 graded)
Abhay
-------
∂22-Dec-86 0934 BJORK@Score.Stanford.EDU Broken terminal
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 09:34:11 PST
Date: Mon 22 Dec 86 09:28:56-PST
From: Steven Bjork <BJORK@Score.Stanford.EDU>
Subject: Broken terminal
To: jmc@Sail.Stanford.EDU
Message-ID: <12264860138.30.BJORK@Score.Stanford.EDU>
Hi, I'd like to get Carolyn's broken terminal (sorry for the wait).
I couldn't find the phone number for your house, though. Can you
tell me what would be a good time to drop over? Thanks.
--Steve
-------
∂22-Dec-86 1000 JMC
Kwan Y. Wong
∂22-Dec-86 1100 PALLAS@Sushi.Stanford.EDU Re: morals and ethics
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 11:00:19 PST
Date: Mon 22 Dec 86 10:40:43-PST
From: Joseph I. Pallas <PALLAS@Sushi.Stanford.EDU>
Subject: Re: morals and ethics
To: JMC@Sail.Stanford.EDU
cc: su-etc@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 20 Dec 86 17:04:00-PST
Message-ID: <12264873203.15.PALLAS@Sushi.Stanford.EDU>
I would be delighted to hear about JMC's experience. I can anticipate
that part of his response will refer to "guidelines" restricting
advertising among members of a profession (which I will agree are not
motivated by ethics). I'm intensely curious about what other examples
of illegal conspiracy in restraint of trade are often given the label
"ethics".
joe
P.S. Perhaps it's not relevant, but I find it odd that "restraint of
trade" is a sin when practiced by associations, but a virtue when
practiced by corporations.
-------
∂22-Dec-86 1124 pratt@navajo.stanford.edu mailing list
Received: from NAVAJO.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 11:24:28 PST
Received: by navajo.stanford.edu; Mon, 22 Dec 86 11:23:35 PST
Date: Mon, 22 Dec 86 11:23:35 PST
From: Vaughan Pratt <pratt@navajo.stanford.edu>
Subject: mailing list
To: thfac@navajo.stanford.edu
Mail sent to thfac@navajo will reach the following people. Any other
names I should put on? Anyone want off?
-v
rwf@sail.stanford.edu
guibas@src.dec.com
dek@sail.stanford.edu
zm@sail.stanford.edu
mayr@score.stanford.edu
jmc@sail.stanford.edu
papa@score.stanford.edu
pratt@navajo.stanford.edu
ullman@navajo.stanford.edu
∂22-Dec-86 1141 HSU@Sushi.Stanford.EDU grade sheet
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 11:41:10 PST
Date: Mon 22 Dec 86 11:38:29-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: grade sheet
To: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12264883721.14.HSU@Sushi.Stanford.EDU>
I contacted Claire Stager at CS TAC, she says it's okay that we send
the grade sheet to her this afternoon (I can deliver it in person.)
I have the finals in my office, and I'll bring them to you later.
See you then.
Yung-jen
-------
∂22-Dec-86 1405 RA finals
I put the finals on your desk.
∂22-Dec-86 1405 HSU@Sushi.Stanford.EDU
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 14:05:44 PST
Date: Mon 22 Dec 86 14:03:01-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 22 Dec 86 14:03:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12264910033.14.HSU@Sushi.Stanford.EDU>
I've given the finals to Rutie, did you get them? I still have 3 more
finals with me, I'll bring them to you later.
Yung-jen
-------
∂22-Dec-86 1501 JMC
Professor Samuel Holtzman
∂22-Dec-86 1606 shashank@su-whitney.arpa Re: ethics and restraint of trade
Received: from WHITNEY.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 16:03:14 PST
Received: by su-whitney.arpa with Sendmail; Mon, 22 Dec 86 15:59:41 pst
Date: Mon, 22 Dec 86 15:59:41 pst
From: Shashank Shekhar <shashank@su-whitney.arpa>
Subject: Re: ethics and restraint of trade
To: JMC@SAIL.STANFORD.EDU, su-etc@SAIL.STANFORD.EDU
The 5 day payment restraint is in response to the 5 days they give you
to pay after you have bought the stocks. They basically make money on
after all the deal because on average clients sell the stocks at higher
prices than they bought the stock at. So they keep larger sum for five days
than they spend at the time of purchase.
shashank
∂22-Dec-86 1628 PHYSICSLIB@Sierra.Stanford.EDU Re: For Henry Lowood
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 16:28:25 PST
Date: Mon 22 Dec 86 16:28:09-PST
From: Henry E. Lowood <PHYSICSLIB@Sierra.Stanford.EDU>
Subject: Re: For Henry Lowood
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sat 20 Dec 86 02:37:00-PST
Message-ID: <12264936452.11.PHYSICSLIB@Sierra.Stanford.EDU>
TO: John McCarthy
Thanks for the comments. I would appreciate knowing what Dr. Wong at
IBM San Jose has to say about the report. I noticed a predominance of
chemists and conservators in the makeup of the committees, which would
tend to support your verdict that magnetic media were perhaps not given
so much attention.
A program set up with ITS would be worth looking into, particularly if they
can spare the two FTE (we would have to add them--a doubtful prospect). I
see no overwhelming obstacles to such a program, though I should talk to
Roxanne Nilan a bit more about it.
Incidentally, I was puzzled that optical recording came off so poorly in the
report; I guess they were judging by current capabilities, though here again
there wasn't that much meat on the bones of the conclusion.
Henry
P.S. Re: my cryptic comment about chemists--this suggests a traditional
emphasis on paper. On second thought, they probably know a fair amount
about mag tape, too.
-------
∂22-Dec-86 1707 GLB
We have now an example of applying Functional Interpretation to an EKL proof:
we did it on an example by Kreisel. EKL behaves well and we may be able
to do it in an entirely automatical way.
∂22-Dec-86 1731 SCHMIDT@SUMEX-AIM.STANFORD.EDU Re: ethics and restraint of trade
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 17:30:54 PST
Date: Mon 22 Dec 86 17:30:26-PST
From: Christopher Schmidt <SCHMIDT@SUMEX-AIM.STANFORD.EDU>
Subject: Re: ethics and restraint of trade
To: shashank@WHITNEY.STANFORD.EDU
cc: JMC@SAIL.STANFORD.EDU, su-etc@SAIL.STANFORD.EDU
In-Reply-To: Message from "Shashank Shekhar <shashank@su-whitney.arpa>" of Mon 22 Dec 86 16:08:54-PST
Message-ID: <12264947790.34.SCHMIDT@SUMEX-AIM.STANFORD.EDU>
The 5 day payment restraint is in response to the 5 days they give you
to pay after you have bought the stocks. They basically make money on
after all the deal because on average clients sell the stocks at higher
prices than they bought the stock at. So they keep larger sum for five days
Where are the average clients' yachts?
--Christopher
P.S. More on 'ethics': Once a lawyer hired my dad (a physician) as
an expert witness in a lawsuit and reneged on the agreed-upon fee.
When my dad reported the dishonest behavior to the ABA, he was told
(by a representative) that it would be unethical for the ABA to do
anything which might negatively affect the business of a fellow lawyer.
-------
∂22-Dec-86 1815 @SUMEX-AIM.STANFORD.EDU:MRC@PANDA fundamentalist textbook case
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 18:15:47 PST
Received: from PANDA by SUMEX-AIM.STANFORD.EDU with Cafard; Mon 22 Dec 86 18:13:17-PST
Date: Mon 22 Dec 86 18:08:25-PST
From: Mark Crispin <MRC%PANDA@SUMEX-AIM.Stanford.EDU>
Subject: fundamentalist textbook case
To: SU-etc@SUMEX-AIM.Stanford.EDU, JMC@SAIL.Stanford.EDU
Postal-Address: 1802 Hackett Ave.; Mountain View, CA 94043-4431
Phone: +1 (415) 968-1052
Message-ID: <12264954705.7.MRC@PANDA>
February 1987's issue of Playboy has a list of the textbook items that
the fundamentalists in Greeneville, TN objected to:
1) discussion of the Renaissance, because "a central idea of the Renaissance
was a belief in the dignity and worth of human beings"
2) discussion of Leonardo da Vinci, because his paintings glorify man instead
of God [sic]
3) a science fiction story titled "A Visit to Mars", because (according to
[Vicki] Frost [the lead plaintiff]) it deals with thought transference, or
telepathy, a supernatural ability that is properly God's alone
4) a passage in "Anne Frank: The Diary of a Young Girl", because it implies
that all religions are equal. (Anne tells a friend, "Oh, I don't mean you
have to be orthodox. . . .I just mean some religion. It doesn't matter
what. Just to believe in something.")
5) a mention of Roman Catholicism. To make such a mention acceptable, the
children "would have to be exposed to the error of it."
6) a text suggesting that children use "the powerful and magical eye inside
[their] head[s]" -- their imagination. The children of Christians, says
Frost, "cannot violate their religious beliefs by participating in an
occult practice. . . .I cannot cope with my child closing his eyes and
going through a supernatural experience. Our children's imaginations have
to be bounded."
7) Shakespeare's "Macbeth", because it mentions magic and witchcraft
8) "The Wizard of Oz", for concluding that people have a power within
themselves to change the way they are
9) The fairy tale "Cinderella", because it mentions magic
10) Stories about dinosaurs, because their existance indicates that the earth
is older than the Bible tells us.
These textbooks, by the way, are used by fifth and sixth graders at
Jerry Falwell's Lynchburg Academy without complaint.
The judge in the case held that their objections were "sincerely held
religous convictions...[that might cause the children if exposed to] adopt
the views of a feminist, a humanist, a pacifist, an anti-Christian, a
vegatarian or an advocate of 'one-world government.'" He gave the plaintiff's
children the right to cut the classes using this book rather than be exposed
to such godless material.
In other words, in theory, by labelling one's "sincerely held convictions"
as "religious", we should expect that feminists may demand that their children
be allowed to cut classes at which male accomplishments are discusses (there
goes all history classes), that pacifists may demand that their children not
study anything to do with war, or that vegatarian children may cut all science
classes that metion meat as part of the food chain.
After all, "secular humanism" has been labelled by the fundamentalists as
a religion. Maybe it's time to ban bibles from bookstores; after all, it's one
of the most pornographic books ever published. Better yet, let's run the bible
stores, Christian Science Reading Rooms, and their ilk out of town. The American
Atheist Bookstore in Denver has been the subject of repeated harassment by local
Christian groups; recently their insurance was cancelled (the explanation given
was "we don't insure abortion clinics any more either").
-------
∂22-Dec-86 1840 HSU@Sushi.Stanford.EDU grades
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 18:40:45 PST
Date: Mon 22 Dec 86 18:38:06-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: grades
To: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12264960108.14.HSU@Sushi.Stanford.EDU>
After a closer look at my records, I found that Joe Hsy should be given
an A instead of the A+ as I suggested. Do you want me to change the record?
Yung-jen
-------
∂22-Dec-86 1957 DEK affirming Landau
I'm not sure what you wish me to affirm... Was your note intended
for Leo (the search committee chairman)? Or did you mean that
I should affirm the principle that anybody named Landau is
probably good? Or did you mean that I should reread the January
Notices (which I tossed some time ago)? Or did you mean that
I should try to ingratiate myself with this person? (If she's
related to Edward, she might be as rich as he was!)
∂22-Dec-86 2143 CLT visit
∂20-Dec-86 1714 JMC visit
To: ito%aoba.tohoku.junet%utokyo-relay.csnet@RELAY.CS.NET
CC: CLT@SAIL.STANFORD.EDU
You will be welcome. I will be away on the 6th and 7th, but other,
especially Carolyn Talcott, will be here, and I look forward to talking
with you when I get back. She is managing the Qlisp project. Our seminar
is on Friday at 11 in case you can stay for it.
Who is visiting and when
∂22-Dec-86 2044 DEK excuse me,...
I see that you meant the Notices from January 87 (not 86). Wow,
how time flies, January is upon us again.
The article there does look very good.
∂22-Dec-86 2349 HSU@Sushi.Stanford.EDU re: grades
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 23:49:40 PST
Date: Mon 22 Dec 86 23:46:58-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: re: grades
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 22 Dec 86 18:42:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12265016338.10.HSU@Sushi.Stanford.EDU>
Claire had left when I got there today, so I'll send the grade sheet in
first thing tomorrow. (They need to do some processing before the noon
deadline at the registrar's.)
The main reason why I withdraw my suggestion was that the most difficult
problem in the EKL assignment was counted as an extra credit problem, but
Joe didn't do that one (which probably took at least 10hrs of work for
people who did solve it). Therefore I think he should get an A instead.
If you'd allow me the liberty to change his grade, then I'll do it. Or
will you be in your office early tomorrow morning?
I'm enclosing a complete record of the scores below.
Yung-jen
----------------------------------------------------------------------
Name Homeworks Extra Midterm Final
Basu, Julie 99 0 58 88
Cohen, Edith 94 70 70 87
Cole, Mary 61 0 44 84
Degenhardt, Jon 76 0 34 89
Fogel, Efi 99 70 67 91
Gaudinat, Bruno 95 0 22 94
Gunning, Chris 77 0 96 85
Henke, Andrea 87 20 36 76
Hsy, Joe 100 0 91 97
Jog, Vidya 99 0 54 77
Lawson, Tom 36 0 12 42
Lazanas, Anthony 100 40 71 97
Luo, Jay 87 70 74 80
Maher, Timothy 85 0 61 88
Malik, Neel 99 0 54 82
Modet, Andres 100 70 98 98
Nelson, Phil 61 0 77 87
Ramchandani, Ratan 84 0 84 97
Rogers, Chris 70 0 46 89
Roland, David 58 0 14 62
Van Gael, Jan 98 50 51 88
Waitz, Tony 86 0 32 93
Wold, Per 69 0 30 99
Zell, Andreas 100 70 62 88
Remote TV students
Kunin, Rick 98 0 53
Shanmugam, Alagappan 32
Stout, John 40
Tung, Charles 48
Dropped?
Bourg, John 0 17
Thomas, Danny 0 46
Powell, Chris 0 31
-------
∂22-Dec-86 2354 HSU@Sushi.Stanford.EDU Tom Lawson
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 22 Dec 86 23:54:16 PST
Date: Mon 22 Dec 86 23:51:35-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: Tom Lawson
To: jmc@Sail.Stanford.EDU
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12265017176.10.HSU@Sushi.Stanford.EDU>
I just got a message from Tom Lawson saying that he did turn in his
EKL assignment on time (it might have been lost somewhere on the TV routes.)
He would like to know if we can take that into consideration w.r.t. his
grade (he does not know he got an C yet). What's your opinion?
Yung-jen
-------
∂23-Dec-86 0011 HSU@Sushi.Stanford.EDU re: Tom Lawson
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 23 Dec 86 00:11:51 PST
Date: Tue 23 Dec 86 00:09:11-PST
From: Yung-Jen Hsu <Hsu@Sushi.Stanford.EDU>
Subject: re: Tom Lawson
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Mon 22 Dec 86 23:59:00-PST
Phone: (415) 856-8499
Office: MJH450 (415) 723-3088
Message-ID: <12265020381.10.HSU@Sushi.Stanford.EDU>
Thanks for your prompt response. I'll give you a copy of the grade sheet
tomorrow.
Yung-jen
-------
∂23-Dec-86 0900 JMC
Dentist at noon.
∂23-Dec-86 0902 SB Terminal update
The terminal works fine here at CSD, so I suspect a problem
with the mux or (unlikely) a SAIL problem. I would like to
try moving the port that Carolyn's terminal is on on the mux
to a different line. When would be a convenient time for me
to come out?
--Steve
∂23-Dec-86 1046 RA Dr. Flaherty
You got another call from Dr. Flaherty at RPI (518) 266 6348. He wants to
come and visit you in January.
∂23-Dec-86 1126 BJORK@Score.Stanford.EDU A thought on the terminal problem.
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 23 Dec 86 11:26:15 PST
Date: Tue 23 Dec 86 11:24:06-PST
From: Steven Bjork <BJORK@Score.Stanford.EDU>
Subject: A thought on the terminal problem.
To: jmc@Sail.Stanford.EDU
Message-ID: <12265143245.23.BJORK@Score.Stanford.EDU>
It may be that the physical connector on the logic board in the
terminal has a problem. The next time I'm out perhaps I should
take a look at it.
Seasons Greetings, and best wishes for the new year.
--Steve
-------
∂23-Dec-86 1644 RA leaving
I all of a sudden have a bad pain in my ear and am going to see a doctor.
See you tomorrow
∂24-Dec-86 0929 CLT dinner
ok, did you make any plans for tomorrow?
∂24-Dec-86 1459 RA vacation
To: JMC@SAIL.STANFORD.EDU, ZM@SAIL.STANFORD.EDU, VAL@SAIL.STANFORD.EDU,
CLT@SAIL.STANFORD.EDU, LES@SAIL.STANFORD.EDU,
JJW@SAIL.STANFORD.EDU, AIR@SAIL.STANFORD.EDU,
shankar@SCORE.STANFORD.EDU
I will be on vacation until Jan. 5. Taleen will substitute for me during
that time.
Rutie
-----
∂24-Dec-86 1545 PHYSICSLIB@Sierra.Stanford.EDU Keys for Math Library?
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 24 Dec 86 15:45:33 PST
Date: Wed 24 Dec 86 15:45:06-PST
From: Henry E. Lowood <PHYSICSLIB@Sierra.Stanford.EDU>
Subject: Keys for Math Library?
To: cn.sci@Forsythe.Stanford.EDU, cn.phm@Forsythe.Stanford.EDU
cc: library@Score.Stanford.EDU, physicslib@Sierra.Stanford.EDU,
jmc@Sail.Stanford.EDU, dek@Sail.Stanford.EDU, knuth@Score.Stanford.EDU
Message-ID: <12265452903.12.PHYSICSLIB@Sierra.Stanford.EDU>
TO: Eleanor Goodchild, Science Libraries
Paul Mosher, Research Services
FR: Henry Lowood
Today (Christmas Eve), Don Knuth called me at home with a surprising
bit of news. His key to the Math-CS Library no longer works. What surprised
me was:
1) That Prof. Knuth, Prof. McCarthy and other faculty members have keys. This
was apparently arranged with Harry Llull. I am not aware that anyone else in
the library system was informed. Certainly, I was not.
2) That Harry seems to have had the library rekeyed before he left. I have
a vague recollection of this (I think it happened in the summer), but I had
no idea that he was effectively (and perhaps intentionally?) locking out
the same faculty by doing this.
I have no recollection of being told of any of this business. Unfortunately,
the victim turned out to be Prof. Knuth, which I consider very unfortunate. As
I understand the arrangement Harry made with the faculty (per Knuth), the keys
were a quid pro quo for committee work rendered on behalf of the library.
It's too late to do anything about this holiday weekend, but I wonder if we
could handle this situation on the up-and-up by stating an open and clear
library policy on keys. Would the library administration object to providing
keys to tenured faculty from the departments directly served by the library?
(That is, could we formalize what was already the policy de facto under
Harry?) Should we wait until the new (permanent) librarian is on board to
confront this issue? Can I go ahead and issue new keys to profs. Knuth,
McCarthy, and ?? in advance of resolving these issues permanently? (I have
no idea who has keys at the moment.)
I am reluctant to handle this situation as Harry did, i.e., to hand out keys
without informing anyone. That is asking for this embarassing situation today
to be repeated again in the future.
Henry
cc: Lim, Llull, Knuth, McCarthy
-------
∂24-Dec-86 1828 LES
.<< OK, this works on Maple >>
.device xgp;
.font 1 "buck75";
.center
.select 1
FOO
∂25-Dec-86 2129 ME psych and bboards
∂25-Dec-86 1239 JMC psych
All bboard messages to psych seem not to arrive. Mightn't it be a good
idea to remove that computer from the su-etc list?
ME - No. Just ignore the rejection messages. I assume they'll come
back up sometime. If you really think they should be removed, ask JR.
∂25-Dec-86 2217 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU star wars
Received: from UCBVAX.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 25 Dec 86 22:17:28 PST
Received: by ucbvax.Berkeley.EDU (5.53/1.19)
id AA26035; Thu, 25 Dec 86 22:17:47 PST
Received: by ucscc (5.51/4.7)
id AA09432; Thu, 25 Dec 86 21:33:07 PST
Received: by ucscd (5.51/4.7)
id AA24639; Thu, 25 Dec 86 21:33:01 PST
Date: Thu, 25 Dec 86 21:33:01 PST
From: ucscc!ucscd.beeson@ucbvax.Berkeley.EDU (20012000)
Message-Id: <8612260533.AA24639@ucscd>
To: ucbvax!sail.stanford.edu!jmc
Subject: star wars
If you have time and interest, please criticize the attached essay.
It is intended to be logically sound in the sense that the conclusions
should be independent of the reader's general political position.
Perhaps an unattainable aim?
\input vanilla.sty
\title
The Truth About Star Wars
\endtitle
\author
M. Beeson
\endauthor
\def\bul{$\bullet$\ }
There are two aspects of the truth about Star Wars: technological and
political. The truth is really very simple; I shall set it out in
six statements, which will then be explained.
\medskip
\bul {\sl Star Wars can never work as a defensive system.}
\medskip
\bul {\sl Star Wars might work as an offensive system.}
\medskip
\bul {\sl Nuclear war is suicide and everybody knows it these days.}
\medskip
\bul {\sl Therefore the threat of nuclear war is not a credible threat.}
\medskip
\bul {\sl The threat of nuclear war has been the cornerstone of American
foreign policy since 1945.}
\medskip
\bul {\sl The effect of Star Wars is to make the threat of
nuclear war almost credible again.}
\medskip
The implications of these facts for the future of civilization and current
politics are then taken up, in three sections as follows:
\medskip
\bul {\sl Star Wars and Arms Control don't mix.}
\medskip
\bul {\sl The Choice: Star Wars or arms control.}
\medskip
\bul {\sl What are our leaders thinking?}
\medskip
\subheading{Star Wars can never work as a defensive system}
\medskip
This is a purely technological point, and it is necessary to pay close
attention to some details of weapons systems.
The purpose of Star Wars is to shoot down attacking missles; the plan
is to do that with high-powered laser beams. If you are going to shoot
down missiles, you have to do it during the ``boost phase'', while their
rockets are still firing. This lasts only about two minutes and is over
before the missiles leave the atmosphere. The reason you have to hit
them in the boost phase is this: after the rockets have fired, ten
warheads (the SALT II limit) are released from each missile, so you have
ten times as many targets. Moreover, if Star Wars were in place, perhaps
hundreds of ``decoy'' warheads, which look like warheads but are not,
would also be released. In addition, if SALT II is abrogated, perhaps
there would be fifty real warheads. Moreover, it would be relatively
inexpensive to shorten the boost phase to as little as forty seconds.
All these points are agreed upon by all sides in the debate over Star Wars.
Since the earth is round, and since you have to hit the missiles just after
they leave the ground, and since they will be launched from points well
inside the USSR, your lasers have to be located in space in order to
hit the missiles.
Therefore the original plan called for a number, perhaps a couple of
hundred, of orbiting ``battle stations'' from which the lasers would
be fired.
{\it This plan has now been given up}, because such battle stations
would be vulnerable to ``space mines'', which are small explosive
satellites which would sneak up unobserved and orbit near the
battle stations until exploded by a signal from the ground.
Instead, the current Star Wars plan calls for ``pop-up'' weapons.
That is, rockets would launch the laser battle stations into space
just before they were needed.
Now the crux of the matter is this: how are you going to have time to
notice the onset of an attack, get the pop-up missiles ready to fire,
decide through some human process (or, God forbid, quickly through a
computer) to fire them, fire them, get them above the atmosphere,
all during the time that the attacking missles are still in the
atmosphere, during the first forty seconds of the attack? This is
clearly and obviously impossible. You don't have to be an expert
to see that.
None of the arguments presented here depend upon it, but this seems
to be the place to point out that current plans for the Star Wars
system require a nuclear explosion to power the lasers. Vast amounts
of power are required and no other way has been suggested that might
be practical to generate the power. Whether laser beams powered by
nuclear explosions are practical is one of the things being researched.
\medskip
\subheading{Star Wars might work as an offensive system}
\medskip
A little background information about weapons systems is necessary for
this point as well. Until now, most ballistic missiles have not been
accurate enough to reliably destroy ``enemy'' missiles in their silos.
This is true of the several hundred land-based ICBM's that both sides
now have deployed. The next generation of missiles, however, will
incorporate new technology which permits almost pinpoint accuracy.
(The technology involves the computerized comparison of a camera image
of the scene below with stored satellite photographs, permitting the
computation of last-minute course adjustments.) The Pershing II
missiles recently deployed in Europe are the first missiles to be
deployed with this increased accuracy. Submarine-based missiles, too,
will achieve the same accuracy. At present they are less accurate
than land-based missiles due to imperfect knowledge of the exact
location of the submarine, but in the next generation of missiles,
they will connect to a worldwide
satellite navigation network as soon as they break the ocean surface,
and accurate position information will be supplied to on-board computers.
Each missile has ten warheads, each of which can be independently and
accurately guided to a different target.
This ``MIRV-ing'' (Multiple independently-targetable re-entry vehicles)
of the missiles means that you get ten chances to knock out each of your
opponent's missiles, assuming both sides have the same number. This
gives the first-striker an advantage. At present, however, it is still
much too dangerous to undertake a first strike, since if even a few
missiles are not knocked out, you would suffer tremendous damage from
the retaliation. Even the submarine-based missles would be enough to
deliver an unacceptable retaliatory blow, and at present there is no way
to keep track of their locations, let alone destroy them.
This is where Star Wars comes in. Suppose you were planning a first strike.
You could launch your pop-up laser battle stations simultaneously with
your attacking missiles. They would then be in position and ready to
shoot down the retaliatory missiles. There would be many fewer of these,
perhaps only the submarine-based missiles and a small percentage of the
land-based missles, so your task would be much easier. As the attacker,
you would have had time enough to prepare everything. It just might work!
\medskip
\subheading{Nuclear war is suicide and everybody knows it}
\medskip
There has been so much media attention to this point that it hardly
bears repeating. Even the hawks in the Reagan administration no longer
talk about surviving nuclear war. What with Chernobyl and the nuclear
winter, who could possibly doubt the point that nuclear war is suicide?
Every child of seven knows it. Every politician knows it. Every
general knows it. Every citizen knows it. Every Russian knows it, too.
\medskip
\subheading{Therefore the threat of nuclear war is not a credible threat}
\medskip
Who would believe that we would commit suicide? That threat is credible
only if we can make them believe we might be crazy enough to commit
suicide. This is Richard Nixon's ``madman theory'' of international
politics. In today's world even this is not enough; nobody would be crazy
enough to deliberately initiate a nuclear war under any circumstances.
\medskip
\subheading{The threat of nuclear war has been the cornerstone of American
foreign policy since 1945}
\medskip
Indeed, that is the political meaning of possessing nuclear weapons at all.
The use of nuclear weapons has been threatened perhaps some tens of times
since 1945. Nixon threatened to use them in Vietnam. Kennedy threatened
to use them during the Cuba confrontation in 1963.
The NATO policy on the defense of Europe rests on the threat of using
nuclear weapons in response to an attack by conventional weapons. It is
for this reason that the United States has refused to make a declaration
of ``no first use''. It is exactly the fact that this threat is not a
credible one that gave rise to the decision to install the Pershing II
missles in Europe: suppose that a conventional attack on Europe were
made, would the US risk Soviet nuclear retaliation by being the first
to use nuclear weapons? It is in fact no more believable that
NATO would risk the destruction of Europe by firing nuclear weapons from
European soil in that situation. The truth is that there is no
credible defense of Europe at present. Naturally this makes the
generals and politicians uncomfortable.
The Soviet Union and the United States have a continuing interest in
maintaining their influence or control over various countries in
South America, Asia, and Africa, where there are continually
``regional conflicts'' between groups supported by the US and the USSR.
The current situations in Nicaragua, Afghanistan, and Angola are
entirely typical. What restrains these conflicts from developing into
full-scale conventional warfare? The fear of escalation to nuclear war;
or to put it another way, the mutual threat of nuclear war.
As the perception grows that nuclear war is not a credible threat,
the superpowers will feel less and less restraint.
\medskip
\subheading{The effect of Star Wars is
to make the threat of nuclear war almost credible again}
\medskip
Once one understands the technological and political points made above,
this conclusion about the effect of Star Wars is obvious. With a
Star Wars system in place, it just might be believable that a desperate
US President might order a first strike. I am not suggesting that
a US President would do so; only that the presence of a Star Wars system
might make an opponent {\it believe} that he might do so. In other words,
it might return the credibility of the threat of nuclear war to whatever
value it had, say, in 1970.
The fundamental paradox underlying the arms race is that the threat of
nuclear war has never been a very credible threat. In the effort to
make the threat more real, the nations involved have piled up ten thousand
warheads, when in the event a few dozen would probably suffice to wreak
sufficient destruction, and anyway only a few hundred missiles are
available to {\it deliver} the warheads. The security of the world
at this moment rests fundamentally on our {\it trusting each other} not to
push the button.
Everyone saw what happened to the radioactive cloud that left Chernobyl.
Everyone has heard about the nuclear winter. Everyone knows that even
after a completely successful first strike, the ``victor'' will also
suffer. Nevertheless, the uncertainties involved and the nature of the
threat mean that if the symmetry between the two superpowers can be
destroyed, the threat is ``more credible''.
Finally, the possibility exists that the Star Wars system could be
used {\it directly} as a weapon against targets on the ground. The same
powerful laser beams that can incinerate a missile can also be directed
at the ground, where they can start fires in a few seconds. They will
be able, if the system can be built at all, to refocus and fire again in
a few seconds, and so in ten minutes each battle station could start
hundreds of fires, creating Dresden-style fire storms in major cities.
This possibility has so far not hit the newspapers, but has been discussed
in the scientific press. The Holocaust does not have to be nuclear,
except for the weapons that power the lasers. These explode in space,
and so do not return fallout to the families of the attackers. This
terrible possibility adds a new dimension to the ``credible threat'' posed
by Star Wars.
\medskip
\subheading{Star Wars and arms control don't mix}
\medskip
Suppose Star Wars could be successfully built.
What is the best strategic response?
One response is simply to build and deploy more missiles, to increase
the number of missiles that one expects the Star Wars system to miss.
Another response is to build your own Star Wars system, thus restoring
the symmetry of the arms race. Let us suppose for the moment that
for one reason or another the Soviets do not try to build their own Star
Wars system; will the method of building more missiles be effective?
You can try to analyze the situation quantitatively, and in the
Appendix I have tried to do so, since I am trained as a scientist.
However, there are so many unknown quantities in the equations that
one can't really do better than common sense: the side that has
Star Wars would quite like it if there were a limit on the number of missiles
that the other side could accumulate, so that there were no danger of
their getting too many to shoot down. (This is true regardless of
whether one thinks of Star Wars offensively or defensively.) Thus
at first glance it might seem that one could mix arms control and
Star Wars.
On the other hand, consider the matter from the point of view of the
side without Star Wars, who necessarily will consider Star Wars as an
offensive system threatening their retaliatory force. They will feel
safer with a large missile force. Would it make sense for them to
negotiate a treaty limiting or reducing the size of missile forces but
permitting the deployment of Star Wars? Such a limit would make
it easier for Star Wars to shoot down the few missiles that might
survive a first strike, and moreover would give the designers of Star
Wars some certainty as to how many missiles Star Wars must be prepared
to cope with. As Gorbachev put it after the Iceland summit:
``Only a madman would agree to that.''
Notice that we have left the question unanswered whether the response
of building more missiles would be effective: nobody knows, just as
nobody knows if Star Wars can be made to work at all. But one can
say with certainty at least this much: if Star Wars works, it can only
be offensively, and if it is deployed, the response will be to build
more missiles, as well as to shorten the boost phase, add decoys, and
increase the number of warheads per missile.
\medskip
\subheading{The Choice: Star Wars or arms control}
\medskip
Since they don't mix, we have to choose one or the other. What are
the consequences of the choice?
If we choose Star Wars, we can look forward to ``more of the same'':
a world politics based on two superpowers competing for domination of the
rest of the world, in a game where the scope of direct conflict is
limited by the threat of nuclear war. We can look forward to an
indefinite future of further escalation in the arms race. Star Wars
itself will violate the ABM treaty, and SALT II with its limits on the
number of missiles and more importantly on the number of warheads per
missile, will also go, since an obvious response to threat of space-based
laser weapons is more warheads per missile. Thus all progress so far
achieved in arms control will be lost, and the arms race will be restrained
only by economics. The tremendous economic waste of the arms race will
be intensified on both sides. With conflicts in Africa, Asia,
South America, and the Middle East proceeding
as they have been, arms budgets in these areas will continue to consume
resources that could be devoted to public health, agriculture, and industry.
What will not go on as before is the arms race itself. The new applications
of computers to missile guidance means that the arms race is becoming
inherently more unstable, in the sense that there is more and more incentive
to be the first-striker. Even without Star Wars, this is true, since the
next generation of missiles will be able to destroy missiles still in their
silos. It becomes a case of ``Use 'em or lose 'em!''. Even if the temptation
to strike first is resisted, the temptation to establish a policy of
``launch on warning'' will be more difficult to resist. This means a policy
whereby the missiles would be launched when satellite photos and radar both
confirmed an enemy attack, without waiting for the attack to materialize.
Since the warning time is about twenty minutes, the decision to fire would
have to be made by an officer on duty, or by a computer. The dangers
of accidental war will be multiplied greatly if this comes to pass.
What is the alternative? Does arms control offer us any real hope?
What has to be understood here is another technological point: The
capacity to destroy civilization, once achieved, cannot be lost. Even
if all nuclear weapons and missiles are destroyed, more can be built
if desired. The only thing that can be negotiated is this: How long
does it take, starting now, to destroy civilization? At present
it takes about half an hour. If all warheads and missiles were
destroyed, that time would be lengthened to several weeks. If in
addition all factories for making missiles and warheads were destroyed,
it might be lengthened to several months. If in the meantime economic
and cultural ties are strengthened, there is hope that they might be
strong enough to reverse any trend to war during those months.
What would then replace the threat of nuclear war as a restraint on
conventional conflict? Only the threat of disturbing the carefully
achieved network of treaties and co-operation; in other words, the
fear of returning to the dangerous and chaotic situation of the 1980's.
\subheading{What are our leaders thinking?}
\medskip
In the foregoing we have made several points which we believe must be
accepted by every person with the patience to read them, regardless of
their politics or their opinion about the Russians. Namely: Star Wars
doesn't make defensive sense, but might make offensive sense. Its main
effect will be to make the threat of nuclear war more credible by raising
the possibility of a first strike in which the retaliatory strike is
totally destroyed.
These facts, simple as they are, are not widely known today, and lie
hidden behind the mass of contradictory and confusing statements about
Star Wars that are reported in the press. Star Wars has been sold to
the American public as a ``defensive'' weapon, which appeals to the
desire of the people to stop threatening the Russians with doom, without
feeling vulnerable in the process. However, polls show that 80\% of the
public would prefer a treaty banning weapons from space. The impetus
for Star Wars does not come from the public.
Who is really behind Star Wars and what are their motivations and intentions?
One group in favor of Star Wars, of course, is
the defense contractors and researchers who will reap the
benefits; and the initial proponents were directors of the national
research laboratories at Livermore and Los Alamos, who stand to benefit
the most. Nevertheless, such an unpopular program could not survive
without political support at the highest levels, from the President,
the Joint Chiefs of Staff, and the advisors of the President. These
people have maintained their support in spite of the fact that it meant
rejecting generous offers for the reduction and elimination of the weapons
Star Wars is advertised as ``defending against''.
Are these people aware of the points made above, about the offensive
nature of Star Wars and its effect on the credibility of the threat
which underlies American foreign policy? It is difficult to believe
that they are not, but it is perhaps not impossible. After all, a welter
of contradictory statements have been issued since the
1983 speech announced the Star Wars program: originally the system was
supposed to protect cities and make the whole concept of deterrence
obsolete, but by June 1985, the President's science
advisor, George Keyworth, was advertising Star Wars as a ``deterrent
against first strike''. Early on it was advertised as a tool to
``force the Russians to the bargaining table'', but Weinberger
says (November 6, 1985) it's not a bargaining chip, and neither research
nor deployment can be given up under any circumstances, even if it means
abrogating the ABM treaty (which it certainly will), and in Iceland,
Reagan's refusal to give up Star Wars prevented a major reduction in
strategic arms from being negotiated.
There are two possibilities: the administration, the Joint Chiefs, and
the civilian advisors are as confused as these statements indicate, or
they are deliberately trying to confuse the public. In this paper
I shall not argue for either alternative, since I do not think either
can be convincingly proved. Moreover, a third possibility is that
some of them are confused, and some of them are deliberately trying to
confuse the others {\it and} the public.
If our leaders are confused, we should replace them. Let us therefore
assume that they are not confused. In that case, what can be their
intentions?
Until recently, it was possible to imagine that President Reagan's
intention was to negotiate an historic arms-control agreement, using
Star Wars as a bargaining chip. After the failed summit at Iceland,
it is clear that a deliberate choice is being made to pursue Star Wars,
along with the accelerated arms race and abandonment of treaties that
it implies, rather than negotiate the elimination of nuclear weapons.
The only motive for such a choice can be the pursuit of power in
South America, Africa, Asia, and the Middle East, backed by the
renewed threat of nuclear force. Regardless of the view one takes
of the causes of and blame for the
worldwide conflict between East and West, one must recognize
that the choice is being made to continue and renew the conflict.
We are witnessing an historic event: the deliberate choice on the part
of American leadership to reject a world in which the threat of nuclear
war is reduced or eliminated, in favor of a world in which a renewed
and more credible threat of nuclear war is available as an instrument of
policy. Let us all understand the choice our leadership is making for
us!
\bye
In order to answer that question, we first have to consider the simpler
question when Star Wars is not present. Given missiles each carrying
ten warheads (or more, if SALT II goes out the window), is it safer
to have a large or a small force of missiles on each side? That is,
under what conditions is it most dangerous to undertake a first strike?
The answer here is that each missile on the first-striker's side can
knock out on the average $10p$ missiles on the other side, where $p$
is the probability of knocking out a single missle that you aim at.
This is independent of the size of the arsenals. Nobody knows what $p$
is; as long as the MX is not deployed, everyone believes $10p$ is
substantially less than one, so there isn't much temptation to a first
strike. This situation will change for the worse with the deployment
of the MX, with advances in satellite navigation technology for
submarine--based missiles, and if SALT II's demise replaces 10 by 50.
How does Star Wars change the equation? Let us suppose that a single laser
battle station can destroy $N$ enemy missiles on the average when used
in conjunction with a first strike.
≠
∂25-Dec-86 2218 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU star wars
Received: from UCBVAX.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 25 Dec 86 22:17:57 PST
Received: by ucbvax.Berkeley.EDU (5.53/1.19)
id AA26045; Thu, 25 Dec 86 22:18:33 PST
Received: by ucscc (5.51/4.7)
id AA09519; Thu, 25 Dec 86 21:49:07 PST
Received: by ucscd (5.51/4.7)
id AA24684; Thu, 25 Dec 86 21:49:03 PST
Date: Thu, 25 Dec 86 21:49:03 PST
From: ucscc!ucscd.beeson@ucbvax.Berkeley.EDU (20012000)
Message-Id: <8612260549.AA24684@ucscd>
To: sail.stanford.edu!jmc@ucbvax.Berkeley.EDU
Subject: star wars
You say that my arguments are "fairly standard"; and also that you hadn't
finished my essay; or did you mean you hadn't finished commenting on it,
perhaps. I had two points: (1) Star Wars is offensive not defensive,
and (2) its purpose is to enable continued use of the threat of nuclear
war as a political instrument. Did you specifically disagree with either
of these two propositions? and if so, with which one? Or did you disagree
with the general tone of the article? If you specifically disagree with
either of these conclusions I am interested in pinpointing the propositions
on which disagreement depends, as an exercise in the formalization of
commonsense reasoning. If on the other hand you agree with the two propositions,
I am interested in editing the essay until the tone no longer rubs you wrong.
I think that point (1) is fairly standard in knowledgeable circles, but
point (2) I thought of for myself, although it didn't require great
creativity, so I suppose others have also. Point (1) however is not
at all widely known even among university faculty, let alone among the
general public, and has been often contradicted by various officials,
so in the interest of an informed population it ought to be repeated more
widely.
∂25-Dec-86 2217 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU disregard previous message
Received: from UCBVAX.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 25 Dec 86 22:17:44 PST
Received: by ucbvax.Berkeley.EDU (5.53/1.19)
id AA26040; Thu, 25 Dec 86 22:18:17 PST
Received: by ucscc (5.51/4.7)
id AA09462; Thu, 25 Dec 86 21:38:58 PST
Received: by ucscd (5.51/4.7)
id AA24646; Thu, 25 Dec 86 21:38:56 PST
Date: Thu, 25 Dec 86 21:38:56 PST
From: ucscc!ucscd.beeson@ucbvax.Berkeley.EDU (20012000)
Message-Id: <8612260538.AA24646@ucscd>
To: sail.stanford.edu!jmc@ucbvax.Berkeley.EDU
Subject: disregard previous message
I just sent you another copy of the star wars essay, believing
erroneously that you had not received it, because a copy I did send
with an old address (su-ai) was returned to me, and I had forgotten
that I had already sent another copy to the correct address.
∂26-Dec-86 0637 @REAGAN.AI.MIT.EDU:Hewitt@XX.LCS.MIT.EDU response to McDermott
Received: from REAGAN.AI.MIT.EDU by SAIL.STANFORD.EDU with TCP; 26 Dec 86 06:36:52 PST
Received: from DUE-PROCESS.AI.MIT.EDU by REAGAN.AI.MIT.EDU via CHAOS with CHAOS-MAIL id 16594; Fri 26-Dec-86 09:37:21 EST
Date: Fri, 26 Dec 86 09:37 EST
From: Carl Hewitt <Hewitt@XX.LCS.MIT.EDU>
Subject: response to McDermott
To: JMC@SU-AI.ARPA
cc: Hewitt@XX.LCS.MIT.EDU
Message-ID: <861226093712.2.HEWITT@DUE-PROCESS.AI.MIT.EDU>
John,
Are you writing a response to McDermott's "Critique of Pure Reason"?
Enclosed please find a copy of my response. Commensts appreciated.
Thanks,
Carl
∂26-Dec-86 1149 VAL reply to message
[In reply to message rcvd 26-Dec-86 10:47-PT.]
Here is our reply:
When such a prominent advocate of the use of logic in AI
as Drew McDermott becomes discouraged and publishes a recantation,
it behooves us logic users to pay careful attention. Even if we
are not also discouraged, we may find sharper formulations of the
problems we face than are provided by the long term opponents of
rationalism like the Dreyfuses, Searle and Winograd and the opponents
of logic within AI like Minsky and Schank.
McDermott's main summary of the logicist position is correct.
We also agree that the progress within the logic approach to AI has
been rather slow, but we do not share McDermott's views on the causes
of it.
It is a mistake, in our opinion, to explain the difficulties by
saying that the inferences AI researchers are interested in ``are not
deductions''. The question is not which inferences are deductions, but
which inferences can be usefully carried out by deduction. If we learn
that the sides of a certain triangle are equal, and conclude that its
angles are equal too, is this a deduction? For someone who has never
studied deductive geometry, this is an inference based on visual images
and intuition, but we know that it can be also interpreted as a deduction,
and attempts to incorporate the knowledge of geometric properties of the
world into AI systems would have been severely handicapped if we did not
have deductive geometry. When we examine an ALGOL program and see that it
checks whether its input is a prime number, is this a deduction? This was
not a deduction before a formal semantics for ALGOL was invented, but now
it is, and this fact is crucial for designing systems that can operate
with programs in an intelligent way.
In this spirit, for example, we approach the planning problem.
True, corporate managers do not think of planning as a deductive process.
But, within a model of the situation, one indeed may be able to prove
that a plan will work. If the reasoning within the model needs to be
elaborate then this may constitute most of the thought. McDermott views
proving plans as incompatible with replanning during execution; but there
is no contradiction here: the robot experiments at SRI based on the
STRIPS formalism included proving plans, monitoring their execution, and
replanning and reproving if necessary.
Every time when another aspect of the world becomes a part of
deductive mathematics, be it the numeric aspect, or geometric, or
topological, or mechanical, or algorithmic, or probabilistic, - this
is always the result of many years of foundational studies, often
requiring major inventions, both conceptual and technical, that could
be made only after long periods of painfully slow progress. The aspects
of the world that AI research is concerned with are not exceptions.
Developing adequate formalisms will be a work of years and will probably
require a touch of genius.
We prefer the term ``common sense physics'' to ``naive physics'',
because common sense physics does not have to be naive. Actually both
terms miss the point which is the epistemological situation. Common sense
physical reasoning must make do with an epistemological situation
characterized by
a. No clear boundary between the phenomenon being observed
and the rest of the physical world. To the extent that such a boundary
can be established it is a result rather than a premiss of the
common sense reasoning. Non-monotonic reasoning is used here by
humans and will be essential for machines.
b. Lack of quantitative knowledge of the physical characteristics
of the situation in which reasoning and perhaps action is required.
c. Lack of precise physical theory relevant to the situation.
d. Inability to do the numerical computations that would be
relevant even if the information were available.
Physicists' physics sometimes provides an island in the
sea of common sense physics of a phenomenon. If we contemplate
dropping a barometer off a building and have a watch, we can use
the fact to tell how tall the building is. However, the relevant
physics formula $s=1/over 2 gt↑2$ must be selected by common sense
physical reasoning which also controls the time observed from the
watch and the computation of $s$.
We think some of the proposals in Hayes's Naive Physics
Manifesto are on the wrong track,
because basing the entire formalism on the space time tubes unnecessarily
encumbers most common sense physics reasoning. But it is not obvious
how to set up a general logical framework for common sense physics.
The recognition of the role of non-monotonic reasoning in AI was
an important step. Unfortunately, logicians had not studied this area
until that time. Formalized non-monotonic reasoning is less than ten
years old, and there is very active exploration both of its computational
aspects and of its possibilities for extending formalisms and reasoning.
Not so long ago, McDermott's criticism of non-monotonic reasoning because
of its ``intractability'' would be valid, but today it is a mistake to
dismiss the cases which reduce to computable algorithms by saying in
parentheses that they ``are of no interest to us here''.
McDermott says that almost all computerized inference is
non-monotonic; it would be more accurate to say that almost all
computerized inference contains non-monotonic parts.
Moreover,
it is sometimes easier to write a program which does the
reasoning correctly than to write a specification for such a program.
This situation has historic precedents. In the times of Euler, it was
easier to sum infinite series than to explain what these computation meant.
In the early days of computing, it was easier to write a compiler than to
explain in what sense it is correct. These observations show once again
that the study of non-monotonic reasoning is still in a very early stage,
but they do not suggest that it should be abandoned.
Another conclusion is that it is important to study the relation
between non-monotonic formalisms and logic programming. The results of
this work have so far been encouraging. The traditional semantics of
logic programs without negation is based on models minimal in exactly
the same sense as in the theory of circumscription. Recent work on the
semantics of stratified programs with negation is closely related to
prioritized circumscription. As Reiter pointed out at the 1984
Non-monotonic Reasoning Workshop, the blocks world axioms based on
circumscription translated readily into a Prolog program.
Concerning the applicability of default logics to the temporal
domain we can say that ``Fred's death'' is an important technical
problem, moderately difficult, but we are puzzled by McDermott's use
of this example as an argument against the logicist position in general.
Our group alone has produced two new formalisms that solve this problem.
One solution is based on chronological minimization and uses pointwise
circumscription. The other is based on including causality in the
language as a new primitive concept and does not require modifying or
generalizing the definition of circumscription. This last solution is
as simple as can be asked. Moreover, the result of circumscription
in this case can be determined simply by predicate completion.
If McDermott works at it, he can undoubtedly come up with some
new problems that will give difficulty. While this one was solved
in a few months in a quite general way, he may be lucky and find
a really difficult problem that will take years to solve. We logicists
will keep at it, however, and the only real way to beat us is to think
of a different way of doing things that is a lot better and isn't just
a subset of
logic in disguise. As soon as the people who attempt to devise a new
formalism that is capable of expressing general common sense knowledge
and also expresses knowledge in a communicable form succeed, we'll bet
they'll find themselves reintroducing most of logic. As long as logic
is the only game in town, and it really is, repeated attempts will be
made to use it.
Incidentally, in our opinion there is a major problem in the
application of logic to AI that McDermott has missed but may be at
the basis of his and other people's doubts about logic. We call it
the {\it context problem}. A fundamental idea of logic is that the language
and usually the domain are fixed before reasoning begins.
Human language and reasoning isn't like that. We never have an
absolutely basic formalism. Our formalisms are always subject to
extension. It seems likely that this phenomenon can and must be
accomodated in logic before we can make a truly general purpose database of
common sense knowledge. Our current approach involves introducing
contexts as objects and including them as extra parameters in formulas.
- Vladimir Lifschitz and John McCarthy
∂26-Dec-86 1213 VAL CS326
Here is an attempt to arrange the contents in a systematic way. How do you like
the idea of following a plan of this sort instead of reading papers in a
chronological order? If we do it this way then we'll cover sometimes different
parts of the same paper in different parts of the course. For instance, the
philosophical problems paper may be read at the very beginning, except that
the automaton part would be postponed until later.
1. Expressing knowledge in logical languages
Epistemological adequacy. Reification. Situation calculus.
STRIPS and logic programming. Logical tools.
2. Non-monotonic reasoning
Minimizing abnormality. Computing circumscription. Priorities.
Pointwise circumscription. Axiomatic description of circumscription
policies.
3. Reasoning about action
Frame problem. Qualification problem. Overweak disjunctions.
Temporal minimization. Causality.
4. Reasoning about mental qualities
Ascribing mental qualities to machines. Automaton metaphysics.
Modalities and modal logic. Wise men, Mr. S and Mr. P.
∂28-Dec-86 1522 netlibd@anl-mcs.ARPA index
Received: from ANL-MCS.ARPA by SAIL.STANFORD.EDU with TCP; 28 Dec 86 15:21:08 PST
Received: by anl-mcs.ARPA (4.12/4.9)
id AA00092; Sun, 28 Dec 86 17:20:36 cst
Date: Sun, 28 Dec 86 17:20:36 cst
From: netlibd@anl-mcs.ARPA (Netlib)
Message-Id: <8612282320.AA00092@anl-mcs.ARPA>
Subject: index
Apparently-To: JMC@SAIL.STANFORD.EDU
===== general NETLIB index =====
Welcome to netlib, a system for distribution of mathematical software
by electronic mail. This index is the reply you'll get to:
mail netlib@anl-mcs.arpa
send index.
To examine the full index for any library send a request of the form:
send index from eispack.
To search for all software with certain keywords:
find cubic spline.
To search for somebody in Gene Golub's address list:
Who is Joan Doe?
displays entries containing "Joan" and "Doe". (no spelling correction!)
Here are some addition forms a request may take...
send dgeco from linpack
(Retrieves routine DGECO and all routines it calls from the LINPACK library.)
send only dgeco from linpack
(Retrieves just DGECO and not subsidiary routines.)
send dgeco but not dgefa from linpack
(Retrieves DGECO and subsidiaries, but excludes DGEFA and subsidiaries.)
send list of dgeco from linpack
(Retrieves just the file names rather than the contents;
this can be helpful when one already has an entire library and just
wants to know what pieces are needed in a particular application.)
whois france
(Retrieves all addresses of people in the database living in France.)
You may include several requests in a single piece of mail, but put
each on a separate line.
Send the requests to "netlib@anl-mcs.arpa" even though replies appear to be
coming from "netlibd@anl-mcs.arpa". You'll be talking to a program,
so don't expect it to understand much English. If your Unix system
doesn't talk directly to research, you might try forwarding through
ihnp4 (AT&T Bell Labs in Chicago) or mcvax (Math Centrum in Amsterdam).
Someone will be paying for long distance phone calls, so be reasonable!
Those with access, possibly through some gateway, to the arpanet can
use netlib@anl-mcs.arpa (at Argonne National Lab) instead of research!netlib
(at Bell Labs in Murray Hill, New Jersey).
The default precision is double; to get single, prefix the library name
with "s". However, if the library only comes in one precision, that's what
you will be sent. To save space we remove sequence numbers and maintain a
central set of machine dependent constants. Otherwise the codes, which are
almost all in Fortran, are as received from the authors. Bugs found in
core libraries like eispack will receive prompt attention; in general, we
will forward comments (and annual lists of recipients) to the code authors.
The "Caveat receptor" you see means: this code may not be worth more than
you are paying for it!
-------quick summary of contents---------
alliant - set of programs collected from Alliant users
apollo - set of programs collected from Apollo users
benchmark - various benchmark programs and a summary of timings
bihar - Bjorstad's biharmonic solver
bmp - Brent's multiple precision package
conformal - Schwarz-Christoffel codes by Trefethen; Bjorstad+Grosse
core - machine constants, blas
domino - communication and scheduling of multiple tasks; Univ. Maryland
eispack - matrix eigenvalues and vectors
elefunt - Cody and Waite's tests for elementary functions
errata - corrections to numerical books
fishpack - separable elliptic PDEs; Swarztrauber and Sweet
fitpack - Cline's splines under tension
fftpack - Swarztrauber's Fourier transforms
fmm - software from the book by Forsythe, Malcolm, and Moler
fn - Fullerton's special functions
go - "golden oldies" gaussq, zeroin, lowess, ...
harwell - MA28 sparse linear system
hompack - nonlinear equations by homotopy method
itpack - iterative linear system solution by Young and Kincaid
lanczos - Cullum and Willoughby's Lanczos programs
laso - Scott's Lanczos program for eigenvalues of sparse matrices
linpack - gaussian elimination, QR, SVD by Dongarra, Bunch, Moler, Stewart
lp - linear programming
machines - short descriptions of various computers
microscope - Alfeld and Harris' system for discontinuity checking
minpack - nonlinear equations and least squares by More, Garbow, Hillstrom
misc - everything else
ode - ordinary differential equations
odepack - ordinary differential equations from Hindmarsh
paranoia - Kahan's floating point test
pchip - hermite cubics Fritsch+Carlson
pltmg - Bank's multigrid code; too large for ordinary mail
port - the public subset of PORT library
pppack - subroutines from de Boor's Practical Guide to Splines
quadpack - univariate quadrature by Piessens, de Donker, Kahaner
siam - typesetting macros for SIAM journal format
slatec - machine constants and error handling package from the Slatec library
specfun - transportable special functions
toeplitz - linear systems in Toeplitz or circulant form by Garbow
toms - Collected Algorithms of the ACM
y12m - sparse linear system (Aarhus)
--------a bit more detail--------
MISC Contains various pieces of software collected over time and:
the source code for the netlib processor itself;
the paper describing netlib and its implementation;
the abstracts list maintained by Richard Bartels.
GO Golden Oldies: routines that have been widely used,
but aren't available through the standard libraries.
Nominations welcome!
CORE Machine constants (i1mach,r1mach,d1mach), blas (level 1 and 2)
EISPACK A collection of Fortran subroutines that compute the eigenvalues
and eigenvectors of nine classes of matrices. The package can
determine the eigensystems of complex general, complex Hermitian,
real general, real symmetric, real symmetric band, real symmetric
tridiagonal, special real tridiagonal, generalized real, and
generalized real symmetric matrices. In addition, there are two
routines which use the singular value decomposition to solve
certain least squares problems.
Developed by the NATS Project at Argonne National Laboratory.
(d.p. refer to eispack, s.p. refer to seispack)
LINPACK A collection of Fortran subroutines that analyze and solve linear
equations and linear least squares problems. The package solves
linear systems whose matrices are general, banded, symmetric
indefinite, symmetric positive definite, triangular, and tridiagonal
square. In addition, the package computes the QR and singular value
decompositions of rectangular matrices and applies them to least
squares problems.
Developed by Jack Dongarra, Jim Bunch, Cleve Moler and Pete Stewart.
(all precisions contained here)
TOMS Collected algorithms of the ACM. When requesting a specific
item, please refer to the Algorithm number.
FMM Routines from the book Computer Methods for Mathematical
Computations, by Forsythe, Malcolm, and Moler.
Developed by George Forsythe, Mike Malcolm, and Cleve Moler.
(d.p. refer to fmm, s.p. refer to sfmm)
FNLIB Wayne Fullerton's special function library.
(all precisions contained here)
FFTPACK A package of Fortran subprograms for the Fast Fourier
Transform of periodic and other symmetric sequences
This package consists of programs which perform Fast Fourier
Transforms for both complex and real periodic sequences and
certain other symmetric sequences.
Developed by Paul Swarztrauber, at NCAR.
FISHPACK A package of Fortran subprograms providing finite difference
approximations for elliptic boundary value problems.
Developed by Paul Swarztrauber and Roland Sweet.
QUADPACK A package for numerical computation of definite univariate integrals.
Developed by Piessens, Robert(Appl. Math. and Progr. Div.- K.U.Leuven)
de Donker, Elise(Appl. Math. and Progr. Div.- K.U.Leuven
Kahaner, David(National Bureau of Standards) (slatec version)
TOEPLITZ A package of Fortran subprograms for the solution of systems
of linear equations with coefficient matrices of Toeplitz or
circulant form, and for orthogonal factorization of column-
circulant matrices.
Developed by Burt Garbow at Argonne National Laboratory,
as a culmination of Soviet-American collaborative effort.
(d.p. refer to toeplitz, s.p. refer to stoeplitz)
PPPACK Subroutines from: Carl de Boor, A Practical Guide to Splines,
Springer Verlag. This is an old version, from around the time the book
was published. We will install a newer version as soon as we can.
ITPACK Iterative Linear System Solver based on a number of methods:
Jacobi method, SOR, SSOR with conjugate gradient acceleration
or with Chebyshev (semi-iteration - SI) acceleration.
Developed by Young and Kincaid and the group at U of Texas.
BIHAR Biharmonic solver in rectangular geometry and polar coordinates.
These routines were obtained from Petter Bjorstad,
Veritas Research, Oslo Norway in July 1984.
LANCZOS procedures computing a few eigenvalues/eigenvectors of a large (sparse)
symmetric matrix. Jane Cullum and Ralph Willoughby, IBM Yorktown.
LASO A competing Lanczos package. David Scott.
CONFORMAL contains routines to solve the "parameter problem" associated
with the Schwarz-Christoffel mapping. Includes:
SCPACK (polygons with straight sides) from Nick Trefethen.
CAP (circular arc polygons) from Petter Bjorstad and Eric Grosse.
FITPACK A package for splines under tension. (an early version)
For a current copy and for other routines, contact:
Alan Kaylor Cline, 8603 Altus Cove, Austin, Texas 78759, USA
HARWELL Currently contains on the sparse matrix routine MA28 from the
Harwell library. from Iain Duff, AERE Harwell
BENCHMARK contains benchmark programs and the table of Linpack timings.
MACHINES contains information on high performance computers that
are or soon to be made available
MINPACK A package of Fortran programs for the solution of systems of
nonlinear equations and nonlinear least squares problems.
Five algorithmic paths each include a core subroutine and an
easy-to-use driver. The algorithms proceed either from an analytic
specification of the Jacobian matrix or directly from the problem
functions. The paths include facilities for systems of equations
with a banded Jacobian matrix, for least squares problems with a
large amount of data, and for checking the consistency of the
Jacobian matrix with the functions.
Developed by Jorge More', Burt Garbow, and Ken Hillstrom at
Argonne National Laboratory.
(d.p. refer to minpack, s.p. refer to sminpack)
PORT The public subset of the PORT library. Includes the latest version
of Gay's NL2SOL nonlinear least squares. The rest of the PORT3
library is available by license from AT&T.
Y12M calculation of the solution of systems of linear systems of
linear algebra equations whose matrices are large and sparse.
authors: Zahari Zlatev, Jerzy Wasniewski and Kjeld Schaumburg
PCHIP is a fortran package for piecewise cubic hermite inter-
polation of data. It features software to produce a monotone and
"visually pleasing" interpolant to monotone data.
Fred N. Fritsch, Lawrence Livermore National Laboratory
LP Linear Programming - At present, this consists of one subdirectory,
data: a set of test problems in MPS format, maintained by David Gay.
For more information, try a request of the form
send index for lp/data
ODE various initial and boundary value ordinary differential equation
solvers: colsys, dverk, rkf45, ode
ODEPACK The ODE package from Hindmarch and others.
This is the single precision verison to get dp refer to dodepack.
Alan Hindmarch, Lawrence Livermore National Laboratory
ELEFUNT is a collection of transportable Fortran programs for testing
the elementary function programs provided with Fortran compilers. The
programs are described in detail in the book "Software Manual for the
Elementary Functions" by W. J. Cody and W. Waite, Prentice Hall, 1980.
SPECFUN is an incomplete, but growing, collection of transportable
Fortran programs for special functions, and of accompanying test
programs similar in concept to those in ELEFUNT.
W.J. Cody, Argonne National Laboratory
PARANOIA is a rather large program, devised by Prof. Kahan of Berkeley,
to explore the floating point system on your computer.
SLATEC library DoE policy apparently prohibits us from distributing this.
Contact the National Energy Software Center or your congressman.
HOMPACK is a suite of FORTRAN 77 subroutines for solving nonlinear systems
of equations by homotopy methods. There are subroutines for fixed
point, zero finding, and general homotopy curve tracking problems,
utilizing both dense and sparse Jacobian matrices, and implementing
three different algorithms: ODE-based, normal flow, and augmented
Jacobian.
DOMINO is a set of C-language routines with a short assembly language
interface that allows multiple tasks to communicate and schedules
local tasks for execution. These tasks may be on a single processor
or spread among multiple processors connected by a message-passing
network. (O'Leary, Stewart, Van de Geijn, University of Maryland)
Apollo A set of programs collected from Apollo users.
Alliant A set of programs collected from Alliant users.
∂28-Dec-86 1533 netlibd@anl-mcs.ARPA index from fnlib
Received: from ANL-MCS.ARPA by SAIL.STANFORD.EDU with TCP; 28 Dec 86 15:32:50 PST
Received: by anl-mcs.ARPA (4.12/4.9)
id AA00246; Sun, 28 Dec 86 17:32:15 cst
Date: Sun, 28 Dec 86 17:32:15 cst
From: netlibd@anl-mcs.ARPA (Netlib)
Message-Id: <8612282332.AA00246@anl-mcs.ARPA>
Subject: index from fnlib
Apparently-To: JMC@SAIL.STANFORD.EDU
c To get d1mach, mail netlib
c send d1mach from core
Caveat receptor. (Jack) dongarra@anl-mcs, (Eric Grosse) research!ehg
Compliments of netlib Sun Dec 28 17:32:07 CST 1986
===== FNLIB index =====
Special function library developed by Wayne Fullerton.
Port framework routines, such as i1mach and fdump, are in the
"core" library. version as of 21 Feb 84
ACOS arc cosine
ACOSH arc hyperbolic cosine
AI Airy function
AID evaluate derivative of Airy Ai function
AIDE evaluate derivative of exp scaled Airy Ai function
AIE exp scaled Airy function
AINT truncate real to an integral value
ALBETA log beta function
ALGAMS log gamma function with sign, a subroutine
ALI logarithmic integral
ALNGAM log abs gamma function
ALNREL relative error logarithm ALOG(1+X)
ALOG natural logarithm
ALOG10 common logarithm
ASIN arc sine
ASINH arc hyperbolic sine
ATAN arc tangent
ATAN2 quadrant correct arc tangent y/x
ATANH arc hyperbolic tangent
BESI0 modified Bessel function of first kind, order zero
BESI0E exp scaled modified Bessel fn of first kind, order zero
BESI1 modified Bessel function of first kind, order one
BESI1E exp scaled modified Bessel fn of first kind, order one
BESJ0 Bessel function of first kind, order zero
BESJ1 Bessel function of first kind, order one
BESK0 modified Bessel function of third kind, order zero
BESK0E exp scaled modified Bessel fn of third kind, order zero
BESK1 modified Bessel function of third kind, order one
BESK1E exp scaled modified Bessel fn of third kind, order one
BESKES sequence of N exp scaled modified Bessel fns of third kind
BESKS sequence of N modified Bessel fns of third kind
BESY0 Bessel function of second kind, order zero
BESY1 Bessel function of second kind, order one
BETA beta function
BETAI incomplete beta function
BI Airy function of second kind, the Bairy function
BID evaluate derivative of Airy Bi function
BIDE evaluate derivative of exp scaled Airy Bi function
BIE exp scaled Airy function of second kind
BINOM binomial fac(n)/(fac(m)*fac(n-m))
C8LGMC complex log gamma correction term for all x
C9LGMC complex log gamma correction term
C9LN2R complex relative error logarithm from second order
CABS absolute value of complex number
CACOS complex arc cosine
CACOSH complex arc hyperbolic cosine
CARG argument (angle) of complex number
CASIN complex arc sine
CASINH complex arc hyperbolic sine
CATAN complex arc tangent
CATAN2 complex quadrant correct arc tangent y/x
CATANH complex arc hyperbolic tangent
CBETA complex beta function
CBRT cube root
CCBRT complex cube root
CCOS complex cosine
CCOSH complex hyperbolic cosine
CCOT complex cotangent
CEXP complex exponential function
CEXPRL complex relative error exp from first order (CEXP(X)-1)/X
CGAMMA complex gamma function
CGAMR complex reciprocal gamma function
CHI hyperbolic cosine integral
CHU confluent hypergeometric function, logarithmic solution
CI cosine integral
CIN cosine integral inverse
CINH hyperbolic cosine integral inverse
CLBETA complex log beta function
CLNGAM complex log gamma function
CLNREL complex relative error logarithm CLOG(1+Z)
CLOG complex natural logarithm
CLOG10 complex common logarithm
COMP1 compare single and dble prec functions
COMP2 compare single and dble prec functions
COMP3 compare single and dble prec functions
COS cosine
COSDG cosine x in degrees
COSH hyperbolic cosine
COT cotangent
CPSI complex digamma function
CSEVL evaluate Chebyshev series
CSIN complex sine
CSINH complex hyperbolic sine
CSQRT complex square root
CTAN complex tangent
CTANH complex hyperbolic tangent
D9ATN1 dble prec arc tangent from first order -X)/X**3
D9GAML dble prec gamma under and overflow limits
D9LGMC dble prec log gamma correction term
D9LN2R dble prec relative error logarithm from second order
D9PAK pack dble prec number
D9UPAK subroutine to unpack dble prec number
DACOS dble prec arc cosine
DACOSH dble prec arc hyperbolic cosine
DAI dble prec Airy function
DAID evaluate derivative of Airy Ai function
DAIDE evaluate derivative of exp scaled Airy Ai function
DAIE dble prec exp scaled Airy function
DASIN dble prec arc sine
DASINH dble prec arc hyperbolic sine
DATAN dble prec arc tangent
DATAN2 dble prec quadrant correct arc tangent y/x
DATANH dble prec arc hyperbolic tangent
DAWS Dawson-s function
DBESI0 dble prec modified Bessel function of first kind, order zero
DBESI1 dble prec modified Bessel function of first kind, order one
DBESJ0 dble prec Bessel function of first kind, order zero
DBESJ1 dble prec Bessel function of first kind, order one
DBESK0 dble prec modified Bessel function of third kind, order zero
DBESK1 dble prec modified Bessel function of third kind, order one
DBESKS dble prec sequence of N modified Bessel fns of third kind
DBESY0 dble prec Bessel function of second kind, order zero
DBESY1 dble prec Bessel function of second kind, order one
DBETA dble prec beta function
DBETAI dble prec incomplete beta function
DBI dble prec Airy function of second kind, the Bairy function
DBID evaluate derivative of Airy Bi function
DBIDE evaluate derivative of exp scaled Airy Bi function
DBIE dble prec exp scaled Airy function of second kind
DBINOM dble prec binomial fac(n)/(fac(m)*fac(n-m))
DBSI0E dble prec exp scaled modified Bessel fn of first kind, order zero
DBSI1E dble prec exp scaled modified Bessel fn of first kind, order one
DBSK0E dble prec exp scaled modified Bessel fn of third kind, order zero
DBSK1E dble prec exp scaled modified Bessel fn of third kind, order one
DBSKES dble prec sequence of N exp scaled modified Bessel fns of third kind
DCBRT dble prec cube root
DCHI hyperbolic cosine integral
DCHU dble prec confluent hypergeometric function, logarithmic solution
DCI dble prec cosine integral
DCIN dble prec cosine integral inverse
DCINH dble prec hyperbolic cosine integral inverse
DCOS dble prec cosine
DCOSDG dble prec cosine x in degrees
DCOSH dble prec hyperbolic cosine
DCOT dble prec cotangent
DCSEVL evaluate dble prec Chebyshev series
DDAWS dble prec Dawson-s function
DE1 dble prec exponential integral E-sub-1
DEI dble prec exponential integral Ei
DERF dble prec error function
DERFC dble prec complementary error function
DEXP dble prec exponential function
DEXPRL dble prec relative error exp from first order (DEXP(X)-1)/X
DFAC dble prec factorial
DGAMI dble prec incomplete gamma function
DGAMIC dble prec complementary incomplete gamma function
DGAMIT dble prec Tricomi-s incomplete gamma function
DGAMMA dble prec gamma function
DGAMR dble prec reciprocal gamma function
DINT truncate dble prec to an integral value
DLBETA dble prec log beta function
DLGAMS dble prec log gamma function with sign, a subroutine
DLI dble prec logarithmic integral
DLNGAM dble prec log abs gamma function
DLNREL dble prec relative error logarithm DLOG(1+X)
DLOG dble prec natural logarithm
DLOG10 dble prec common logarithm
DPOCH dble prec Pochhammer-s generalized symbol
DPOCH1 dble prec (poch(a,x)-1)/x
DPSI dble prec psi or digamma function
DSHI dble prec hyperbolic sine integral
DSI dble prec sine integral
DSIN dble prec sine
DSINDG dble prec sine x in degrees
DSINH dble prec hyperbolic sine
DSPENC dble prec Spence-s function, related to dilogarithm
DSQRT dble prec square root
DTAN dble prec tangent
DTANH dble prec hyperbolic tangent
E1 exponential integral E-sub-1
EI exponential integral Ei
ERF error function
ERFC complementary error function
EXP exponential function
EXPREL relative error exp from first order (EXP(X)-1)/X
FAC factorial
GAMI incomplete gamma function
GAMIC complementary incomplete gamma function
GAMIT Tricomi-s incomplete gamma function
GAMMA gamma function
GAMR reciprocal gamma function
INITDS initialize dble prec orthogonal polynomial series
INITS Initialize orthogonal polynomial series
POCH Pochhammer-s generalized symbol
POCH1 (poch(a,x)-1)/x
PSI psi or digamma function
R9ATN1 arc tangent from first order -X)/X**3
R9GAML gamma under and overflow limits
R9LGMC log gamma correction term
R9LN2R relative error logarithm from second order
R9PAK pack floating point number
R9UPAK subroutine to unpack floating point number
RAND portable uniform random numbers
RANDGS portable normal random numbers
RANDOM portable uniform random numbers
RANF portable uniform random numbers
SHI hyperbolic sine integral
SI sine integral
SIN sine
SINDG sine x in degrees
SINH hyperbolic sine
SPENC Spence-s function, related to dilogarithm
SQRT square root
TAN tangent
TANH hyperbolic tangent
∂28-Dec-86 1540 netlibd@anl-mcs.ARPA sin from fnlib
Received: from ANL-MCS.ARPA by SAIL.STANFORD.EDU with TCP; 28 Dec 86 15:39:37 PST
Received: by anl-mcs.ARPA (4.12/4.9)
id AA00344; Sun, 28 Dec 86 17:38:56 cst
Date: Sun, 28 Dec 86 17:38:56 cst
From: netlibd@anl-mcs.ARPA (Netlib)
Message-Id: <8612282338.AA00344@anl-mcs.ARPA>
Subject: sin from fnlib
Apparently-To: JMC@SAIL.STANFORD.EDU
c To get d1mach, mail netlib
c send d1mach from core
Caveat receptor. (Jack) dongarra@anl-mcs, (Eric Grosse) research!ehg
Compliments of netlib Sun Dec 28 17:38:46 CST 1986
function sin (x)
c august 1980 edition. w. fullerton, los alamos scientific lab.
c
c this routine is based on the algorithm of cody and waite
c in argonne tm-321, software manual working note number 1.
c
dimension sincs(10)
external aint, csevl, inits, r1mach, sqrt
c
c series for sin on the interval 0.00000e+00 to 2.46740e+00
c with weighted error 2.06e-19
c log weighted error 18.69
c significant figures required 18.10
c decimal places required 19.19
c
data sin cs( 1) / -0.3749911549 5587317584 0e0/
data sin cs( 2) / -0.1816031552 3725020186 4e0/
data sin cs( 3) / 0.0058047092 7459863355 9e0/
data sin cs( 4) / -0.0000869543 1177934075 7e0/
data sin cs( 5) / 0.0000007543 7014808885 1e0/
data sin cs( 6) / -0.0000000042 6712966505 6e0/
data sin cs( 7) / 0.0000000000 1698042294 5e0/
data sin cs( 8) / -0.0000000000 0005012057 9e0/
data sin cs( 9) / 0.0000000000 0000011410 1e0/
data sin cs( 10) / -0.0000000000 0000000020 6e0/
c
c pihi + pilo = pi. pihi is exactly representable on all machines
c with at least 8 bits of precision. whether it is exactly
c represented depends on the compiler. this routine is more
c accurate if it is exactly represented.
data pihi / 3.140625e0 /
data pilo / 9.676535897 9323846e-4 /
data pirec / 0.3183098861 8379067e0 /
data pi2rec / 0.63661977236 7581343e0 /
data ntsn, xsml, xwarn, xmax / 0, 3*0.0 /
c
if (ntsn.ne.0) go to 10
ntsn = inits (sincs, 10, 0.1*r1mach(3))
c
xsml = sqrt (6.0*r1mach(3))
xmax = 1.0/r1mach(4)
xwarn = sqrt (xmax)
c
10 y = abs(x)
if (y.gt.xmax) call seteru (
1 42hsin no precision because abs(x) is big, 42, 2, 2)
if (y.gt.xwarn) call seteru (
1 54hsin answer lt half precision because abs(x) is big,
2 54, 1, 1)
c
sin = x
if (y.lt.xsml) return
c
xn = aint (y*pirec+0.5)
n2 = amod (xn, 2.0) + 0.5
sgn = x
if (n2.ne.0) sgn = -sgn
f = (y-xn*pihi) - xn*pilo
c
sin = f + f*csevl (2.0*(f*pi2rec)**2-1.0, sincs, ntsn)
if (sgn.lt.0.0) sin = -sin
if (abs(sin).gt.1.0) sin = sign (1.0, sin)
c
return
end
function sqrt (x)
c june 1977 edition. w. fullerton, c3, los alamos scientific lab.
dimension sqrt2(3)
external alog, r1mach, r9pak
data sqrt2(1) / 0.7071067811 8654752e0 /
data sqrt2(2) / 1.0 e0 /
data sqrt2(3) / 1.4142135623 7309505 e0 /
c
data niter / 0 /
c
if (niter.eq.0) niter = 1.443*alog(-0.104*alog(0.1*r1mach(3)))+ 1.
c
if (x.le.0.) go to 20
c
call r9upak (x, y, n)
ixpnt = n/2
irem = n - 2*ixpnt + 2
c
c the approximation below has accuracy of 4.16 digits.
sqrt = .261599e0 + y*(1.114292e0 + y*(-.516888e0 + y*.141067e0))
c
do 10 iter=1,niter
sqrt = sqrt + 0.5*(y - sqrt**2) / sqrt
10 continue
c
sqrt = r9pak (sqrt2(irem)*sqrt, ixpnt)
return
c
20 if (x.lt.0.) call seteru (21hsqrt x is negative, 21, 1, 1)
sqrt = 0.0
return
c
end
function aint (x)
c december 1983 edition. w. fullerton, c3, los alamos scientific lab.
c
c aint (x) extracts the integer part of x and returns the result as a
c floating point number. this routine is ordinarily a built-in
c function. it is supplied here merely for completeness.
c
external alog, i1mach, r1mach
data npart, scale, xbig, xmax / 0, 3*0.0 /
c
if (npart.ne.0) go to 10
ibase = i1mach(10)
xmax = 1.0/r1mach(4)
xbig = amin1 (float(i1mach(9)), xmax)
scale = ibase**int (alog(xbig)/alog(float(ibase))-0.5)
npart = alog (xmax)/alog (scale) + 1.0
c
10 if (x.lt.(-xbig) .or. x.gt.xbig) go to 20
c
aint = int(x)
return
c
20 xscl = abs(x)
if (xscl.gt.xmax) go to 50
c
do 30 i=1,npart
xscl = xscl/scale
30 continue
c
aint = 0.0
do 40 i=1,npart
xscl = xscl*scale
ipart = xscl
part = ipart
xscl = xscl - part
aint = aint*scale + part
40 continue
c
if (x.lt.0.0) aint = -aint
return
c
50 call seteru (67haint abs(x) may be too big to be represented as
1 an exact integer, 67, 1, 1)
aint = x
return
c
end
function alog (x)
c june 1977 edition. w. fullerton, c3, los alamos scientific lab.
dimension alncs(6), center(4), alncen(5)
external csevl, inits, r1mach
c
c series for aln on the interval 0. to 3.46021d-03
c with weighted error 1.50e-16
c log weighted error 15.82
c significant figures required 15.65
c decimal places required 16.21
c
data aln cs( 1) / 1.3347199877 973882e0 /
data aln cs( 2) / .0006937562 83284112e0 /
data aln cs( 3) / .0000004293 40390204e0 /
data aln cs( 4) / .0000000002 89338477e0 /
data aln cs( 5) / .0000000000 00205125e0 /
data aln cs( 6) / .0000000000 00000150e0 /
c
data center(1) / 1.0 /
data center(2) / 1.25 /
data center(3) / 1.50 /
data center(4) / 1.75 /
c
data alncen( 1) / 0.0e0 /
data alncen( 2) / +.2231435513 14209755 e+0 /
data alncen( 3) / +.4054651081 08164381 e+0 /
data alncen( 4) / +.5596157879 35422686 e+0 /
data alncen( 5) / +.6931471805 59945309 e+0 /
c
c aln2 = alog(2.0) - 0.625
data aln2 / 0.0681471805 59945309e0 /
data nterms / 0 /
c
if (nterms.eq.0) nterms = inits (alncs, 6, 28.9*r1mach(3))
c
if (x.le.0.) call seteru (
1 29halog x is zero or negative, 29, 1, 2)
c
call r9upak (x, y, n)
c
xn = n - 1
y = 2.0*y
ntrval = 4.0*y - 2.5
if (ntrval.eq.5) t = ((y-1.0)-1.0) / (y+2.0)
if (ntrval.lt.5) t = (y-center(ntrval))/(y+center(ntrval))
t2 = t*t
c
alog = 0.625*xn + (aln2*xn + alncen(ntrval) + 2.0*t +
1 t*t2*csevl(578.0*t2-1.0, alncs, nterms) )
c
return
end
function csevl (x, cs, n)
c april 1977 version. w. fullerton, c3, los alamos scientific lab.
c
c evaluate the n-term chebyshev series cs at x. adapted from
c r. broucke, algorithm 446, c.a.c.m., 16, 254 (1973). also see fox
c and parker, chebyshev polys in numerical analysis, oxford press, p.56.
c
c input arguments --
c x value at which the series is to be evaluated.
c cs array of n terms of a chebyshev series. in eval-
c uating cs, only half the first coef is summed.
c n number of terms in array cs.
c
dimension cs(1)
c
if (n.lt.1) call seteru (28hcsevl number of terms le 0, 28, 2,2)
if (n.gt.1000) call seteru (31hcsevl number of terms gt 1000,
1 31, 3, 2)
if (x.lt.(-1.1) .or. x.gt.1.1) call seteru (
1 25hcsevl x outside (-1,+1), 25, 1, 1)
c
b1 = 0.
b0 = 0.
twox = 2.*x
do 10 i=1,n
b2 = b1
b1 = b0
ni = n + 1 - i
b0 = twox*b1 - b2 + cs(ni)
10 continue
c
csevl = 0.5 * (b0-b2)
c
return
end
function inits (os, nos, eta)
c april 1977 version. w. fullerton, c3, los alamos scientific lab.
c
c initialize the orthogonal series so that inits is the number of terms
c needed to insure the error is no larger than eta. ordinarily, eta
c will be chosen to be one-tenth machine precision.
c
c input arguments --
c os array of nos coefficients in an orthogonal series.
c nos number of coefficients in os.
c eta requested accuracy of series.
c
dimension os(nos)
c
if (nos.lt.1) call seteru (
1 35hinits number of coefficients lt 1, 35, 2, 2)
c
err = 0.
do 10 ii=1,nos
i = nos + 1 - ii
err = err + abs(os(i))
if (err.gt.eta) go to 20
10 continue
c
20 if (i.eq.nos) call seteru (28hinits eta may be too small, 28,
1 1, 2)
inits = i
c
return
end
function r9pak (y, n)
c december 1979 edition. w. fullerton, c3, los alamos scientific lab.
c
c pack a base 2 exponent into floating point number x. this routine
c is almost the inverse of r9upak. it is not exactly the inverse,
c because abs(x) need not be between 0.5 and 1.0. if both r9pak and
c 2.0**n were known to be in range, we could compute
c r9pak = x * 2.0**n .
c
external i1mach, r1mach
data nmin, nmax / 2*0 /
data aln210 / 3.321928094 887362 e0 /
c
if (nmin.ne.0) go to 10
aln2b = 1.0
if (i1mach(10).ne.2) aln2b = r1mach(5)*aln210
nmin = aln2b*float(i1mach(12))
nmax = aln2b*float(i1mach(13))
c
10 call r9upak (y, r9pak, ny)
c
nsum = n + ny
if (nsum.lt.nmin) go to 40
if (nsum.gt.nmax) call seteru (
1 31hr9pak packed number overflows, 31, 2, 2)
c
if (nsum.eq.0) return
if (nsum.gt.0) go to 30
c
20 r9pak = 0.5*r9pak
nsum = nsum + 1
if (nsum.ne.0) go to 20
return
c
30 r9pak = 2.0*r9pak
nsum = nsum - 1
if (nsum.ne.0) go to 30
return
c
40 call seteru (32hr9pak packed number underflows, 32, 1, 0)
r9pak = 0.0
return
c
end
subroutine r9upak (x, y, n)
c august 1980 portable edition. w. fullerton, los alamos scientific lab
c
c unpack floating point number x so that x = y * 2.0**n, where
c 0.5 .le. abs(y) .lt. 1.0 .
c
absx = abs(x)
n = 0
y = 0.0
if (x.eq.0.0) return
c
10 if (absx.ge.0.5) go to 20
n = n - 1
absx = absx*2.0
go to 10
c
20 if (absx.lt.1.0) go to 30
n = n + 1
absx = absx*0.5
go to 20
c
30 y = sign (absx, x)
return
c
end
subroutine seteru (messg, nmessg, nerr, iopt)
common /cseter/ iunflo
integer messg(1)
data iunflo / 0 /
c
if (iopt.ne.0) call seterr (messg, nmessg, nerr, iopt)
if (iopt.ne.0) return
c
if (iunflo.le.0) return
call seterr (messg, nmessg, nerr, 1)
c
return
end
subroutine seterr (messg, nmessg, nerr, iopt)
c
c this version modified by w. fullerton to dump if iopt = 1 and
c not recovering.
c seterr sets lerror = nerr, optionally prints the message and dumps
c according to the following rules...
c
c if iopt = 1 and recovering - just remember the error.
c if iopt = 1 and not recovering - print, dump and stop.
c if iopt = 2 - print, dump and stop.
c
c input
c
c messg - the error message.
c nmessg - the length of the message, in characters.
c nerr - the error number. must have nerr non-zero.
c iopt - the option. must have iopt=1 or 2.
c
c error states -
c
c 1 - message length not positive.
c 2 - cannot have nerr=0.
c 3 - an unrecovered error followed by another error.
c 4 - bad value for iopt.
c
c only the first 72 characters of the message are printed.
c
c the error handler calls a subroutine named fdump to produce a
c symbolic dump. to complete the package, a dummy version of fdump
c is supplied, but it should be replaced by a locally written version
c which at least gives a trace-back.
c
integer messg(1)
external i1mach, i8save
c
c the unit for error messages.
c
iwunit=i1mach(4)
c
if (nmessg.ge.1) go to 10
c
c a message of non-positive length is fatal.
c
write(iwunit,9000)
9000 format(52h1error 1 in seterr - message length not positive.)
go to 60
c
c nw is the number of words the message occupies.
c
10 nw=(min0(nmessg,72)-1)/i1mach(6)+1
c
if (nerr.ne.0) go to 20
c
c cannot turn the error state off using seterr.
c
write(iwunit,9001)
9001 format(42h1error 2 in seterr - cannot have nerr=0//
1 34h the current error message follows///)
call e9rint(messg,nw,nerr,.true.)
itemp=i8save(1,1,.true.)
go to 50
c
c set lerror and test for a previous unrecovered error.
c
20 if (i8save(1,nerr,.true.).eq.0) go to 30
c
write(iwunit,9002)
9002 format(23h1error 3 in seterr -,
1 48h an unrecovered error followed by another error.//
2 48h the previous and current error messages follow.///)
call eprint
call e9rint(messg,nw,nerr,.true.)
go to 50
c
c save this message in case it is not recovered from properly.
c
30 call e9rint(messg,nw,nerr,.true.)
c
if (iopt.eq.1 .or. iopt.eq.2) go to 40
c
c must have iopt = 1 or 2.
c
write(iwunit,9003)
9003 format(42h1error 4 in seterr - bad value for iopt//
1 34h the current error message follows///)
go to 50
c
c test for recovery.
c
40 if (iopt.eq.2) go to 50
c
if (i8save(2,0,.false.).eq.1) return
c
c call eprint
c stop
c
50 call eprint
60 call fdump
stop
c
end
subroutine fdump
return
end
subroutine e9rint(messg,nw,nerr,save)
c
c this routine stores the current error message or prints the old one,
c if any, depending on whether or not save = .true. .
c
integer messg(nw)
logical save
external i1mach, i8save
c
c messgp stores at least the first 72 characters of the previous
c message. its length is machine dependent and must be at least
c
c 1 + 71/(the number of characters stored per integer word).
c
integer messgp(36),fmt(14),ccplus
c
c start with no previous message.
c
data messgp(1)/1h1/, nwp/0/, nerrp/0/
c
c set up the format for printing the error message.
c the format is simply (a1,14x,72axx) where xx=i1mach(6) is the
c number of characters stored per integer word.
c
data ccplus / 1h+ /
c
data fmt( 1) / 1h( /
data fmt( 2) / 1ha /
data fmt( 3) / 1h1 /
data fmt( 4) / 1h, /
data fmt( 5) / 1h1 /
data fmt( 6) / 1h4 /
data fmt( 7) / 1hx /
data fmt( 8) / 1h, /
data fmt( 9) / 1h7 /
data fmt(10) / 1h2 /
data fmt(11) / 1ha /
data fmt(12) / 1hx /
data fmt(13) / 1hx /
data fmt(14) / 1h) /
c
if (.not.save) go to 20
c
c save the message.
c
nwp=nw
nerrp=nerr
do 10 i=1,nw
10 messgp(i)=messg(i)
c
go to 30
c
20 if (i8save(1,0,.false.).eq.0) go to 30
c
c print the message.
c
iwunit=i1mach(4)
write(iwunit,9000) nerrp
9000 format(7h error ,i4,4h in )
c
call s88fmt(2,i1mach(6),fmt(12))
write(iwunit,fmt) ccplus,(messgp(i),i=1,nwp)
c
30 return
c
end
subroutine eprint
c
c this subroutine prints the last error message, if any.
c
integer messg(1)
c
call e9rint(messg,1,1,.false.)
return
c
end
integer function i8save(isw,ivalue,set)
c
c if (isw = 1) i8save returns the current error number and
c sets it to ivalue if set = .true. .
c
c if (isw = 2) i8save returns the current recovery switch and
c sets it to ivalue if set = .true. .
c
logical set
c
integer iparam(2)
c iparam(1) is the error number and iparam(2) is the recovery switch.
c
c start execution error free and with recovery turned off.
c
data iparam(1) /0/, iparam(2) /2/
c
i8save=iparam(isw)
if (set) iparam(isw)=ivalue
c
return
c
end
subroutine s88fmt( n, w, ifmt )
c
c s88fmt replaces ifmt(1), ... , ifmt(n) with
c the characters corresponding to the n least significant
c digits of w.
c
integer n,w,ifmt(n)
c
integer nt,wt,digits(10)
c
data digits( 1) / 1h0 /
data digits( 2) / 1h1 /
data digits( 3) / 1h2 /
data digits( 4) / 1h3 /
data digits( 5) / 1h4 /
data digits( 6) / 1h5 /
data digits( 7) / 1h6 /
data digits( 8) / 1h7 /
data digits( 9) / 1h8 /
data digits(10) / 1h9 /
c
nt = n
wt = w
c
10 if (nt .le. 0) return
idigit = mod( wt, 10 )
ifmt(nt) = digits(idigit+1)
wt = wt/10
nt = nt - 1
go to 10
c
end
∂28-Dec-86 1603 ANDY@Sushi.Stanford.EDU Forbs education cover story
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 28 Dec 86 16:03:21 PST
Date: Sun 28 Dec 86 16:00:22-PST
From: Andy Freeman <ANDY@Sushi.Stanford.EDU>
Subject: Forbs education cover story
To: jmc@Sail.Stanford.EDU
Message-ID: <12266504259.18.ANDY@Sushi.Stanford.EDU>
Did you see it (they misspelled the cover and put the missing e on
the table of contents page)? Do you have any public coments?
thanks,
-andy
-------
∂28-Dec-86 1612 PHYSICSLIB@Sierra.Stanford.EDU re: Keys for Math Library?
Received: from SIERRA.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 28 Dec 86 16:12:45 PST
Date: Sun 28 Dec 86 16:12:28-PST
From: Henry E. Lowood <PHYSICSLIB@Sierra.Stanford.EDU>
Subject: re: Keys for Math Library?
To: JMC@Sail.Stanford.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Wed 24 Dec 86 16:02:00-PST
Message-ID: <12266506460.14.PHYSICSLIB@Sierra.Stanford.EDU>
TO: John McCarthy
Thanks for the information. I will see what I can do.
Henry
-------
∂28-Dec-86 1615 CLT shopping list
small aligator baggies
huggies
tasters choice
spagetti
toilet paper
eggs
boutique kleenex
rice
yoghurt - raspberry, boysenberry, strawberry
similac green label
meat and veg for supper
∂28-Dec-86 1755 FEIGENBAUM@SUMEX-AIM.STANFORD.EDU Re: argument against logic in AI
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 28 Dec 86 17:55:28 PST
Date: Sun 28 Dec 86 17:55:13-PST
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.STANFORD.EDU>
Subject: Re: argument against logic in AI
To: JMC@SAIL.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Fri 26 Dec 86 16:35:00-PST
Message-ID: <12266525167.14.FEIGENBAUM@SUMEX-AIM.STANFORD.EDU>
John, I dont have any paper of this kind. (In fact, I avoid writing papers
of this kind. If I have something scientifically contributory to say, I write;
I dont write telling people what they shouldn't do).
Minsky has strong opinions on this question. Perhaps he has written something.
Ed
-------
∂28-Dec-86 2307 ME DMWAITS?
∂28-Dec-86 0031 JMC
To: LES, ME
Who has WAITS terminals at home?
ME - You mean DMWAITS terminals (DMs with the WAITS keyboard)?
I can check my list tomorrow.
∂29-Dec-86 0844 ito%aoba.tohoku.junet%utokyo-relay.csnet@RELAY.CS.NET visit
Received: from RELAY.CS.NET by SAIL.STANFORD.EDU with TCP; 29 Dec 86 08:44:09 PST
Received: from utokyo-relay by csnet-relay.csnet id aa01548; 29 Dec 86 11:07 EST
Received: by u-tokyo.junet (4.12/4.9J-1[JUNET-CSNET])
id AA29078; Wed, 24 Dec 86 02:18:49+0900
Received: by nttlab.ntt.junet (4.12/5.0N) with TCP; Wed, 24 Dec 86 00:32:38 jst
Received: by aoba.tohoku.junet (1.1/6.1J) ; Tue, 23 Dec 86 22:32:35 JST
Message-Id: <8612231332.AA01200@aoba.tohoku.junet>
Date: Tue, 23 Dec 86 22:32:35 JST
From: Takayasu ITO <ito%aoba.tohoku.junet%utokyo-relay.csnet@RELAY.CS.NET>
To: JMC%SAIL.STANFORD.EDU%csnet-relay.csnet%u-tokyo.junet@RELAY.CS.NET
Subject: visit
Dear Professor McCarthy,
Thank you for your e-mail.When I arrive at Stanford I will try to get touch
with Dr C Talcott. Also I will try to attend your seminar on Friday at 11.
with best regards,Takayasu Ito
∂29-Dec-86 0922 TALEEN@Score.Stanford.EDU nils would like to speak w/you
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 29 Dec 86 09:22:45 PST
Date: Mon 29 Dec 86 09:20:28-PST
From: Taleen Marashian <TALEEN@Score.Stanford.EDU>
Subject: nils would like to speak w/you
To: jmc@Sail.Stanford.EDU
Message-ID: <12266693603.25.TALEEN@Score.Stanford.EDU>
Prof. McCarthy,
Nils would like to see you when you come in this morning.
-Taleen
-------
∂29-Dec-86 1005 CLT
perhaps you also got a copy
∂28-Dec-86 2221 GLB
JMC, CLT and Shankar are meeting to discuss a course on Automatic Theorem
Proving and they would like you to partecipate, monday 29 or tuesday 30.
GLB
____
------------------------------
I am free only at 5 --- how abou meeting with JMC et al monday at 5
and meeting with you tuesday at 5?
JK
∂29-Dec-86 1458 Bobrow.pa@Xerox.COM A personal favor
Received: from XEROX.COM by SAIL.STANFORD.EDU with TCP; 29 Dec 86 14:56:43 PST
Received: from Cabernet.ms by ArpaGateway.ms ; 29 DEC 86 11:51:56 PST
Date: 29 Dec 86 11:51 PST
Sender: Bobrow.pa@Xerox.COM
From: Danny Bobrow <Bobrow.pa@Xerox.COM>
Subject: A personal favor
To: JMC@SAIL.STANFORD.EDU,McCarthy@SUMEX-AIM.STANFORD.EDU
cc: Bobrow.pa@Xerox.COM,oshea.pa@Xerox.COM
Message-ID: <861229-115156-2121@Xerox>
John,
I am preparing a lecture on programming languageto be given via video
to the new president and other high level management of Xerox. I want
to include short segments of interviews from wise old men in the field.
You obviously fall into that category. Would you be willing to devote
about an hour on Wednesday afternoon January 7 to doing this (actually
any time after 11AM). The TV crew would be happy to come to you (at
Stanford or home). If the idea is OK, but the date is not, let me know,
and Tim Oshea who is producing this extavaganza will be in touch to
arrange a better date.
danny
∂29-Dec-86 1625 CLT
this afternoon
I want to keep the meeting upstairs,
so when its time for Timothy's bottle
I can slip out and take care of it and come back.
So, please don't make any suggestions about
everyone going downstairs.
∂29-Dec-86 1855 LES Gosper account
Do you think that CSD should continue to offer a free SAIL account to Gosper?
If not, do you want to cover the cost out of sponsored research or unrestricted
funds? Incidentally, his last login was in July.
∂29-Dec-86 2220 Bobrow.pa@Xerox.COM re: A personal favor
Received: from XEROX.COM by SAIL.STANFORD.EDU with TCP; 29 Dec 86 22:19:27 PST
Received: from Cabernet.ms by ArpaGateway.ms ; 29 DEC 86 18:32:42 PST
Date: 29 Dec 86 18:32 PST
Sender: Bobrow.pa@Xerox.COM
From: Danny Bobrow <Bobrow.pa@Xerox.COM>
Subject: re: A personal favor
In-reply-to: John McCarthy <JMC@SAIL.STANFORD.EDU>'s message of 29 Dec
86 16:46 PST
To: JMC@SAIL.STANFORD.EDU
cc: Bobrow.pa@Xerox.COM,oshea.pa@Xerox.COM
Message-ID: <861229-183242-2656@Xerox>
This was intended as a one shot with no intention of selling the result.
What is your schedule next week?
danny
∂30-Dec-86 0840 PERLAKI@Score.Stanford.EDU message
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 08:37:35 PST
Date: Tue 30 Dec 86 08:35:16-PST
From: Agnes M. Perlaki <PERLAKI@Score.Stanford.EDU>
Subject: message
To: JMC@Sail.Stanford.EDU
cc: RA@Sail.Stanford.EDU
Message-ID: <12266947519.21.PERLAKI@Score.Stanford.EDU>
Debbie Stewart called from the Development Office (Encina Hall) re: arranging
for professorship dinner. Please call asap at 5-4344.
-Agi
-------
∂30-Dec-86 0956 RICHARDSON@Score.Stanford.EDU AI Faculty Mtg.
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 09:56:41 PST
Date: Tue 30 Dec 86 09:54:14-PST
From: Anne Richardson <RICHARDSON@Score.Stanford.EDU>
Subject: AI Faculty Mtg.
To: ai.list: ;
Message-ID: <12266961894.10.RICHARDSON@Score.Stanford.EDU>
Third time's a charm (I hope)! Please let me know your availability on
Monday, Jan. 12, Wednesday, Jan. 14, and Friday, Jan. 16 (in the afternoons)
for a meeting of the AI faculty.
Thanks,
Anne
-------
∂30-Dec-86 1033 CLT shopping list
prime rib - 4rib roast ordered from draegers
(ask for cooking time)
potatoes for baking
6 sweet 2 idaho
veg- asparagus?
half pint whipping cream
gerbers oatmeal
If you want to get some pat\'e and crackers (or similar stuff)
from NM for starter that would be fine.
∂30-Dec-86 1034 CERF@A.ISI.EDU Digital Library Meeting Agenda
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 10:34:35 PST
Date: 30 Dec 1986 13:31-EST
Sender: CERF@A.ISI.EDU
Subject: Digital Library Meeting Agenda
From: CERF@A.ISI.EDU
To: Bibliotects:
Cc: cerf@A.ISI.EDU
Message-ID: <[A.ISI.EDU]30-Dec-86 13:31:54.CERF>
Happy New Year!
Looking forward to seeing you next week. Please confirm attendance
by sending message to CERF@A.ISI.EDU or leaving message at
(703) 620-8990. If you cannot attend, please similarly advise.
Many thanks,
Vint
3rd Digital Library Working Group Meeting
January 5-6, 1987
Washington, DC
Agenda
January 5
1000 Introductions/Objectives/Agenda - Cerf
1015 Current Status Report - Cerf
Communications of the ACM
American Association of Artificial Intelligence
Dow Jones Information Systems
Plans: Compuserve, IEEE, NTIS?
-------------------------------
1030 System Architecture - Cerf
Fundamental notion: printable documents
Discuss 3-level recursion concept. Are all functions
replicated? Including registration, etc? Optional?
Copyright Royalty Accounting?
Is this the right structure? Did we leave out an
essential function? Discuss reformatting, registration,
repositories, import/export, accounting. Knowbotics?
1200 lunch
1300 Technical Grab Bag No. 1
Document representation issue -
Postscript. JMC observes it isn't ASCII so hard
to use as a convenient basis for searching and
sorting of textual material. Should we consider
multiple representations? Which?
Library vs File System
Relationship of multiple repositories and distributed
file systems? Naming, finding, cataloging, indexing,
access controls, linking of private/public libraries.
Write once optical memories
cataloging/indexing in such systems. Hypertext/XANADU
concepts?
Mostly disconnected model.
Private/personal document manipulation system is mostly
disconnected from the larger environment. When/how does
the work-station link into this environment?
References to other documents within a document?
Automatic retrieval?
Knobots.
Can they really be programs? How should they be
represented? Interpreted? How much context do they
carry with them?
Information authentication.
Can we use certificiate concepts such as encrypted
checksums produced by knowtary public? When invoked?
At registration? Before registration (by originator?)
Public keying technology seems a likely source of
technical assistance.
Document Reading tools
JMC observes that we need better computer-based reading
equipment. Dynabooks. Where should we take this notion?
Is there current development or research work now going
on that we should look at? Can we list desirable
features?
1500 Break
1530 Organizational Participation
What organizations should initially participate?
Is this a function of the type of information we
first include in the library? What role would
university libraries play? Technical information
publishers? Government organizations such as the
Library of Congress? National Library of Medicine?
Would we want to sponsor work station access at
existing library facilities? Where else?
Should there be a series of trials? Should they involve
communications carriers, information services such as
Compuserve, DJ/NRS, other?
How can we harness the current enthusiasm for desktop
publishing? Provide early channels for capturing such
documents in the digital library?
1730 Break for dinner (Clydes?)
January 6
1000 Technical Grab Bag No. 2
Communications requirements.
Is ISDN 64 kb/s adequate? For access? What about for
inter-repository and other system component exchange?
Redundancy.
What policies should we follow to assure preservation
of information entered into the digital library system?
Coping with technology changes.
How should we deal with the problem of archiving digital
information only to have the technology of storage or
retrieval change out from under us? Can we really afford
to copy the entire system periodically to migrate to new
technology base? What about backward compatibility with
earlier display, access, retrieval methods, equipment?
How has this issue appeared and been addressed in present
day library systems?
Research topics.
Can we make a list of known, hard-core research issues
that need to be addressed? Printing and display
technology? Cataloging and Indexing? Document
representation? Searching, sorting, extracting
algorithms? Distributed database techniques?
Natural language methods of access? Scaling issues?
1200 Lunch
1300 Copyright and other legal questions
Is digital representation special? "Digital back rights?"
How can we find out? Who to ask? Note current IEEE and
ACM practice with respect to copying and payment of
royalties (Copyright Clearing Center facility)
Information liability, accuracy, timeliness? (especially
medical information!). Quality control???
1400 Legislative Opportunities
Are there identifiable legislative actions which would
be advantageous to pursue? E.g. a tax on information
products/services to pay for creation of digital library?
Technical registration/format requirements for items
submitted to NTIS?
1500 Break
1530 Project Components
How can we describe the work breakdown structure of
the project? There are certain components that need
development (access engines, knowbotics, registration
and cataloging system,...). There are research areas
that need study (document representation, display and
printing methods, copyright and other legislative areas),
etc. Can we make a list of tasks that should be included
in the project plan?
Project phasing. At the November session, we broke the
project into four phases:
printable information, simple access
Other digital information (e.g. parts catalogs)
Natural language access
Access by inferential description
Is that the right breakdown?
1630 Adjourn
∂30-Dec-86 1110 TALEEN@Score.Stanford.EDU msg. from Vint Cerf
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 11:09:59 PST
Date: Tue 30 Dec 86 11:07:44-PST
From: Taleen Marashian <TALEEN@Score.Stanford.EDU>
Subject: msg. from Vint Cerf
To: jmc@Sail.Stanford.EDU
Message-ID: <12266975273.23.TALEEN@Score.Stanford.EDU>
Prof. McCarthy,
Please call Vint Cerf at (703) 620-8990. He would like to know if you'd
wish to attend a meeting in Washington on Monday.
-Taleen
-------
∂30-Dec-86 1115 TALEEN@Score.Stanford.EDU express letter
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 11:15:45 PST
Date: Tue 30 Dec 86 11:13:28-PST
From: Taleen Marashian <TALEEN@Score.Stanford.EDU>
Subject: express letter
To: jmc@Sail.Stanford.EDU
Message-ID: <12266976319.23.TALEEN@Score.Stanford.EDU>
There is a letter regarding your SRA interview in Chicago. It explains
the arrangements and includes interview questions.
This arrived through express mail this morning. I didn't know when you will
be in today, so I thought it best to send you a message (in case you read
your mail at home or elsewhere). The letter is in your "in" box.
-Taleen
-------
∂30-Dec-86 1313 STEINBERGER@SRI-KL.ARPA Re: Evil Empire
Received: from SRI-KL.ARPA by SAIL.STANFORD.EDU with TCP; 30 Dec 86 13:13:12 PST
Date: Tue 30 Dec 86 13:12:10-PST
From: Richard Steinberger <STEINBERGER@SRI-KL.ARPA>
Subject: Re: Evil Empire
To: JMC@SAIL.STANFORD.EDU
cc: steinberger@SRI-KL.ARPA, su-etc@SCORE.STANFORD.EDU
In-Reply-To: Message from "John McCarthy <JMC@SAIL.STANFORD.EDU>" of Sun 28 Dec 86 13:35:00-PST
Message-ID: <12266997926.10.STEINBERGER@SRI-KL.ARPA>
Let me begin to respond to JMC's comments about the Soviet Union by stating
that I have not read the pamphlet he suggests. To this I would like to
suggest that arguing for a "Yes-they-are -- No-they're-not" an "Evil Empire"
would be essentially a pointless debate. Rather we should be exploring ways
that the two most powerful nations on earth can begin to resolve differences
without threatening to annhilate all life on the planet. Having said this, I
would like to respond to JMC.
There should be no question that the Soviet government has committed and
continues to commit variety of nefarious acts since its inception some 70
years ago. Their domestic policies have varied from severe forms of repres-
sion and imprisonment to 'milder' forms of containment and control. Their
foreign policies have attempted to place many smaller countries inside a
Soviet sphere of influence. (In this effort, the United States has also
made similar attempts with similar results. Both the U.S. and the U.S.S.R.
need to learn that the vast majority of the world's nations do not want to be
either country's pawn.) There is no question that virtually everyone living
in the U.S. and in many other countries would prefer to remain where they are
if given the choice to move to the U.S.S.R. Furthermore most people would
strongly resist any Soviet-launched efforts to alter their native forms of
governance.
The average Russian citizen (only some 50% of the Soviet population is
Russian) *does* support 'his' government. (To a lesser extent many non-Soviet
citizens also support 'their' government.) This fact is largely ignored by
most amateur 'Soviet' watchers. While one could claim that they are
brainwashed, I believe there is little evidence of substance to support this.
Clearly the Soviet government is the only government capable of offering any
protection from the 30,000+ U.S. missiles that are pointed at 'the homeland.'
The claim that the U.S.S.R. is some form of 'Evil Empire' is designed to
conjure up various theological images of eternal battles of the 'forces of
good' (presumably the U.S.A.) and the 'forces of the devil' (presumably the
U.S.S.R.). By and large this claim has had its intended effect. Many
Americans are now convinced that their worst fears of the Soviet Union are
true. The conclusion we as Americans are lead to is that only by building
more powerful and expensive weapons can the Soviet menace be stopped.
Similar anti-American fears are exploited in Soviet citizens.
The dirty little secret in this tragic state of affairs is that each
country's government is dependent on the other's committing or seeming to
commit periodic acts of aggression. When no such act is visible one is
invented. *Both* governments play this game and maintain to a sizable degree
their legitimacy by 'exploiting the nastiness' of the other. Each side
justifies perpetuating the arms race by pointing at the other side.
Whether the U.S.S.R. is more 'evil' than the U.S. is not the point.
We are both pointing enormous numbers of nuclear and non-nuclear weapons
at each other. In foreign policy we both interpret regional conflicts
(e.g. Central America) as us-versus-them. If there is ever to be anything
approaching peace in the world both governments and both sets of populations
are going to need to undergo some serious value adjustments. First and
foremost should be there *realization* that there is in fact only one earth:
one home for the entire plant and animal population. This and other necessary
changes can only begin to come about when both governments are overwhelmed by
a profound genuine desire on the part of the citizenry for peace. Clearly
this is not now happening. Peace, alas, is perceived as a dangerous
alternative.
------------------------------------------------------------------------------
We have a rare opportunity to take a step forward in the search for peace.
We can stop testing nuclear weapons as the Soviets have done for well over a
year. Sadly, current policy is that it is not on our 'national interest' to
do so. We are continuing to search for some new kind of weapons system
(a.k.a. star wars) that will securely and safely contain or destroy the Soviet
population and government. This delusion can only lead to much
disappointment, wasted monies and careers, and add yet another dangerous twist
to the arms race.
A true detente, if we ever attain it, *is* going to be painful. We and
the Soviets will have to give up beliefs we have cherished for years. Many
jobs (including mine) may well be lost and entire industries will perish.
*Both* citizenries and governments will change; and I would bet for the
better. Psychologically many people will have the opportunity for growth
experiences. It's going to be risky, perhaps very risky. When this risk
appears more hopeful than the risk of continuing of current policies, an
appropriate shift in values towards peace may occur.
So we can continue to point fingers and yell, 'Evil Empire.' Or we
can examine ourselves and our government and ask, 'Do we want peace? Really?
What positive, but cautious, steps can we take?' The sooner we choose the
second alternative, the safer and saner our world will become.
-Ric Steinberger
Steinberger@sri-kl.arpa
-------
∂30-Dec-86 1355 KOHEN@Sushi.Stanford.EDU Re: Evil Empire
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 13:55:31 PST
Date: Tue 30 Dec 86 13:53:02-PST
From: Abraham Kohen <KOHEN@Sushi.Stanford.EDU>
Subject: Re: Evil Empire
To: STEINBERGER@SRI-KL.ARPA
cc: JMC@Sail.Stanford.EDU, su-etc@Score.Stanford.EDU, KOHEN@Sushi.Stanford.EDU
In-Reply-To: <12266997926.10.STEINBERGER@SRI-KL.ARPA>
Message-ID: <12267005366.11.KOHEN@Sushi.Stanford.EDU>
Pure hogwash.
Since when is mass murder referred to as "severe form of repression?"
What are the US counterparts to the Soviet actions in Hungary,
Czechslovakia, and Afghanistan?
Abe
-------
∂30-Dec-86 1438 TALEEN@Score.Stanford.EDU book held for you at Green library
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 14:38:37 PST
Date: Tue 30 Dec 86 14:36:15-PST
From: Taleen Marashian <TALEEN@Score.Stanford.EDU>
Subject: book held for you at Green library
To: jmc@Sail.Stanford.EDU
Message-ID: <12267013234.34.TALEEN@Score.Stanford.EDU>
Davies' HUMAN SACRIFICE is being held at Green's Loan Desk. It must be
picked up by tomorrow. I had problems when I went over there to try to
check it out. I cannot check out books with my i.d. card (I can only
use the library). Even if I had the right type of card, it expired a few
days ago (and it takes a few days to have a new one issued). I then attempted
fill out a proxy card, but this didn't work either. I need your signature and
and a letter from you stating that I am working for you. I have the card
with me, but I don't know if you find this worth it since Rutie will be back
shortly. Perhaps it IS worth it if you want this book. It's your choice.
Please let me know what you think.
Thanks,
Taleen
-------
∂30-Dec-86 1441 ANDY@Sushi.Stanford.EDU Re: Evil Empire
Received: from SUSHI.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 14:41:20 PST
Date: Tue 30 Dec 86 14:37:32-PST
From: Andy Freeman <ANDY@Sushi.Stanford.EDU>
Subject: Re: Evil Empire
To: STEINBERGER@SRI-KL.ARPA
cc: JMC@Sail.Stanford.EDU, su-etc@Score.Stanford.EDU
In-Reply-To: <12266997926.10.STEINBERGER@SRI-KL.ARPA>
Message-ID: <12267013467.16.ANDY@Sushi.Stanford.EDU>
Richard Steinberger <STEINBERGER@SRI-KL.ARPA> writes:
Clearly the Soviet government is the only government capable of
offering any protection from the 30,000+ U.S. missiles that are
pointed at 'the homeland.'
Many countries are not threatened by U.S. missles yet few have the
Soviet government. (You can exchange the labels and the statement
is still true.)
The thrust of Ric Steinberger's message is that moving away from
M.A.D. and us-vs-them requires structural changes in both the U.S. and
the U.S.S.R. I agree. I disagree with his (implied) claim that it is
irrelevant whether or not the U.S.S.R. (or the U.S.) is an evil
empire; it is only relevant if you wish to change the status quo.
If one of the two is an evil empire (or if both are equally evil and
one is an empire and the other isn't), then the each country is going
to have to change differently for Ric's utopia to occur.
Ric asked whether or not the U.S. (and Soviet) people really want
peace. (Ric and I agree that the present 40 years without active
hostilities between the U.S. and the U.S.S.R. are not "peace". I
suspect that we both prefer them to 40 years of war though.) JMC says
that the answer to Ric's real question, ``What positive, but cautious,
steps can we take?'', depends on the nature of the `stepper' and the
other side.
Ric, what steps toward peace can the U.S. take that are independent of
whether or not the U.S.S.R. is an evil empire? (If you'd like to
avoid the definition of "evil empire", then what steps are independent
of what differences between the U.S. and the U.S.S.R?) If the
U.S.S.R. (or the U.S.) is an evil empire, what kind of "peace" is
possible?
-andy
ps - Ric advocates passing through intermediate "riskier" stages on
the path to peace. If the status quo is immorally risky, then how
much additional risk is moral and how long is it moral to undertake
that additional risk?
-------
∂30-Dec-86 1500 CERF@A.ISI.EDU Cancellation of Meeting
Received: from A.ISI.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 15:00:14 PST
Date: 30 Dec 1986 17:59-EST
Sender: CERF@A.ISI.EDU
Subject: Cancellation of Meeting
From: CERF@A.ISI.EDU
To: Bibliotects:
Message-ID: <[A.ISI.EDU]30-Dec-86 17:59:00.CERF>
Folks,
as all of you should be aware, Bob and I decided that in the absence
of a draft document for you to comment upon, we would be doing you
a disservice to consume your time here.
The game plan now is to produce a draft document, distribute it to
you for comment and then coordinate a meeting to review your comments.
This should be a more efficient use of your time which we consider
valuable and unique.
Happy New Year from NRI to all of you!
Vint Cerf
∂30-Dec-86 1743 ucscc!ucscd.beeson@ucbvax.Berkeley.EDU star wars continued
Received: from UCBVAX.Berkeley.EDU by SAIL.STANFORD.EDU with TCP; 30 Dec 86 17:42:47 PST
Received: by ucbvax.Berkeley.EDU (5.53/1.19)
id AA02910; Tue, 30 Dec 86 17:41:25 PST
Received: by ucscc (5.51/4.7)
id AA09576; Mon, 29 Dec 86 08:55:41 PST
Received: by ucscd (5.51/4.7)
id AA08378; Mon, 29 Dec 86 08:55:35 PST
Date: Mon, 29 Dec 86 08:55:35 PST
From: ucscc!ucscd.beeson@ucbvax.Berkeley.EDU (20012000)
Message-Id: <8612291655.AA08378@ucscd>
To: jmc@su-ai.edu, jmc@sail.stanford.edu@ucbvax.Berkeley.EDU
Subject: star wars continued
Thank you for giving my Star Wars essay as much attention as you have.
Here are my reactions to the points you made.
In this file your points are not indented and my replies are indented.
The main point I am unclear about is the extent to which you really do
attribute bad motives to the Reagan Administration and previous
U.S. administrations.
I purposely refrained from making a judgment as to whether their
motives are good or bad. I maintain that their motive in pushing
Star Wars is to have a (more) credible nuclear threat to back up
political power. Whether this is a good thing to do or not is a
much more complicated matter; one's conclusion will probably
depend on an assessment of the intentions and motivations of the
present and future leaders of the USSR, and on an assessment of
their reactions to the policies, statements, and actions of the US,
all of which is notoriously complicated as is proved by the lack
of agreement even among rational people.
My own opinion, which is not directly relevant to the essay, is
that a foreign policy based on threats will probably not lead to
the reduction in fear and building of mutual trust and interdependence
which I regard as desirable and probably essential to the future
development of civilization.
1. I don't believe that all your first points are agreed by all sides.
There are many schemes for destroying missiles being developed. Yours
is only one. It is intended to use more than one. Another is
smart rocks. There may even be a few actually secret schemes. Not all
conceivable countermeasures can be simultaneously adopted.
Well, I read everything that passes my hands on SDI, and I went to
a debate at Stanford (you did, too) at which a representative of
SDI was present, and there it was agreed that orbiting weapons were
no longer under consideration; and if the weapons have to be launched
from the ground, then the system isn't defensive. So while there
may be many schemes, they have to be ground-based or space-based,
and my argument applies to the ground-based schemes, and SDI has
given up on space-based schemes (they have, haven't they? unless
they changed their minds again since that debate), so there you are.
Incidentally, you don't mention Soviet efforts at SDI. Did you read the
statement of the 30 emigre Soviet scientists about this?
No, do you have a copy? This is the first I heard of such a statement.
2. Whether SDI makes first strike easier depends on whether offensive
missiles are decommissioned while SDI is being commissioned. Its
advocates propose to negotiate this.
But the dependence goes the wrong way. If offensive missiles are
decommissioned while SDI is being commissioned, then there are fewer
targets to hit; of course proportionately fewer attacking missles also,
but the number of missles that SDI has to deal with, namely the
ones that are not knocked out in the first strike, goes down, so SDI
has an easier time of it and the first strike is more successful.
Consequently if SDI is deployed (by either side) the rational response
is to INCREASE the number of missles deployed, not decrease them.
Hence SDI would make reduction or elimination of missles more difficult.
3. I'm surprised that you have such confidence in what the media say.
The question of whether nuclear war is suicide depends on several
questions. (1) Can you knock out the other guy's offensive weapons?
(2) What do you regard as suicide? Unfortunately, once wars get
started, and ancient wars were often more genocidal than modern ones,
the sides often fight until one can't fight any more. Often the
winner has committed suicide in the sense of suffering damage
worse than losing without fighting, but it happens anyway.
(3) I rather believe that the plan to stop a Soviet tank attack
in Europe with nuclear weapons would be carried out. It would then
be up to the Soviets to escalate further.
The question here is whether the threat of nuclear war is still
credible or not. You say it is, particularly with regard to the
defense of Europe, and particularly if a genuine first-strike
capability could be attained, or if the opponent could at least
be made to believe you had a first-strike capability. The latter
point is exactly what I am suggesting is the purpose of Star Wars;
at present you CAN'T knock out the other guy's offensive weapons
reliably enough to prevent a crippling retaliation. The European
situation is very complicated and I have read hundreds of pages
of analysis by people of all different opinions, and probably you
have too; most recently the current Bulletin of Atomic Scientists
contains an interesting reply to Bundy et. al's proposal that the
US adopt a no-first-strike policy in Europe.
The meaning of "credible" is as follows: Your threat is credible
if you can get other nations to give in on some issue out of fear
of your nuclear attack. Now you maintain here that our threat
is credible, and below you maintain that perhaps it never was credible;
so in one or the other instance, you must be using a different
meaning of "credible" than the one I am using. Or else, you might
be saying that although the US COULD have intimidated other nations
by the threat of nuclear attack, it has never actually DONE SO.
More below.
The Administration has been bullied into not talking about survivability,
but many don't agree. It's a shame we have no civil defense program.
The Swiss don't agree that civil defense is impossible and have
implemented quite effective civil defense. Nuclear winter is
questionable, though possible. Also the Soviets talk out of both
sides of their mouths on nuclear survivability. They have an
extensive nuclear civil defense program that no-one is allowed
to discuss with foreigners, and its doctrine is survivability.
Even with the best civil defense, millions will die in a full
scale nuclear war, but civil defense might result in saving
80 percent rather than 40 percent of the population.
The issue of civil defense is not directly relevant to the points
made in my paper. Since you raise the matter, I'll tell you
my opinion anyway: the existence
of civil defense adds to the climate of fear, and I consider
the fear of nuclear war one of the main psychological pollutants
in modern society. I have perhaps for that reason been too ready
to accept the arguments of those who say an effective civil defense
is impossible; but they do seem sound to me. Morever, IF effective
civil defense is possible, it would require much practice in advance,
and tremendous co-operative preparation. For that much effort, you
might as well train the population in mass non-violent resistance
techniques so that no invader could ever manage the country; then
you could throw away all weapons. No doubt you will find this
statement naive and perhaps it is, but I wonder which is more difficult
to make effective: civil defense or non-violence training.
Chernobyl tends to prove the opposite of your point.
Almost the worst possible nuclear reactor accident killed 31
people directly. There will also be as many as a few thousand added to
the 9 million Ukrainians who will die of cancer in the next 30 years.
Some tens of square miles are contaminated and may or may not be
practically decontaminatable. Two of the remaining three reactors in the
plant are running again, and the third will probably be running within a
year. The Soviets have no plan to abandon this kind of reactor, although
perhaps their new ones will entirely be pressurized water reactors. Civil
defense in the U.S. would make an enormous difference; it could even make
the worst nuclear winter scenarios survivable. Unfortunately, even
if nuclear winter proves to be a fantasy, the direct effects of a
major attack would probably kill more people proportionally than
did the Khmer Rouge in Cambodia. I mention Cambodia to point out
that society in some sense survives the death of even a quarter
of the population, and also that after all that death, they are
still fighting each other.
Good points.
I am very doubtful that the Soviets will ever find an attack
on the U.S. a plausible proposition. For one thing,
they have gotten too many unpleasant surprises in their surrogate
wars with Israel for the political leaders to have confidence
in assurances that Western military technology has been dominated.
Nevertheless, it is better to be safe than sorry, and I
regard being sure that the Soviets won't regard military success
as likely as on the side of safety.
A good argument for a DEFENSIVE system. But, as I argue,
Star Wars is OFFENSIVE. An offensive system makes the Soviets
feel more threatened. Making your opponent feel threatened
is not a smart thing to do. The smart thing to do is
to follow the "THREE-PART RECIPE FOR PEACE":
(1) Make sure he knows he can't successfully attack you, and then
(2) Make him feel assured that you won't attack him, and then
(3) Try to develop mutual trust and economic interdependence.
I don't regard an arms race
itself as dangerous unless politics changes for the worse.
Here we have a significant area of disagreement. The arms race is,
in my opinion, very dangerous, and also it has far-reaching negative
effects even if it doesn't lead to war.
Let me first address its danger, i.e. the chance that it will lead
to nuclear war. You seem to think there is
little reason to believe the arms race will lead to war. The only
argument I have heard you advance is that so far there hasn't been
a war. That argument is fallacious: cf. the man who jumped off
a tall building and remarked as he passed the tenth floor, "So far,
so good." The question has been approached mathematically by
Richardson and his followers, who wrote down systems of coupled
differential equations for arms races, and identified conditions of
stability and instability. (Do you know this work? if not I'll give
you references.) It seems that the Arab-Israeli race is unstable,
the Iran-Iraq race was unstable, and the US-USSR race is stable.
However, the coefficients in the equations depend on the technology
involved; and what worries me is that changes in technology could
shift the US-USSR race into the unstable regime of Richardson's
equations. Those equations accurately describe dozens of arms
races that took place since 1870, and once we cross the
line in phase space to the unstable regime we are done for, just
like the Iraqis and Iranians.
Instability in the sense of Richardson's equations is
not technically the same thing as instability in another sense that
worries me: a complex system is called stable if it tends to return
to its equilibrium when slightly disturbed, unstable otherwise.
The worldwide military system (a single system, nota bene, due to
the dependencies of one army's decisions on the actions of another) is
becoming more tightly coupled every year, thanks e.g. to satellite
observation links. The much-talked-about false alarms at NORAD
never resulted in war. Why not? because to launch an attack,
the generals insist on confirming data from satellite infrared
detectors and from radar showing attacking missles. They never
got both at once. But if one side has a first-strike capability
and the time for evaluating the available data goes down to
ten minutes...I don't see how you can say that isn't dangerous!
Bertrand Russell once said, "Mankind has never abstained from any
folly of which he was capable". That's the fundamental reason
why the arms race is dangerous.
It is important not to exaggerate the so-called arms race by
remembering that we are now spending about six percent of our
GNP on arms, including military pensions, whereas we spent
ten percent around 1960. The Soviet Union spends a substantially
larger percentage.
Now we come to the effects of the arms race (other than
the possible effect of war). First the economic effects.
Your numbers for the US are substantially correct although
I think you have the high seven years too early.
Abraham S. Becker, senior economist for RAND, writes in
RAND report R-2752-AF prepared for the Air Force in 1981,
that "the American defense/GNP ratio rose to a high of 9 or 9.5 %
in 1967-68 and then declined..." to 6% in 1981. That high percentage
during the war was paid for by deficit spending, which later
produced inflation. The inflation was thus an effect of
the arms race (or actually of war itself).
Well, even if you think that inflation was not a big problem,
six percent of the GNP would go a long way if directed toward
desirable social aims or towards segments of the
economy with a higher multiplier than defense industry, which has
a low multiplier and hence is (myth to the contrary) a poor way
in which to stimulate the economy.
Just to be a little more specific about the human cost of the
arms race: Ruth Leger Sivard presents the following possible uses
of the $50 billion 1981 dollars that would be saved by a 50%
reduction in the US and USSR nuclear arsenals (per year; figures
in billions of $):
military R&D transferred to civilian needs 10
vocational training of disabled persons 2
safe water for all within the decade 7
(unsafe water is a factor in 80% of
infectious disease)
direct food aid for the hungry 4
assistance to small farms 6
primary schools and teacher training 5
global effort on nuclear waste disposal 1
five-fold increase in tree planting 2
worldwide clean-air program 5
vaccine protection for all children 1
(WHO estimates 10 children per minute
die of diseases they could be immunized
against for $3 per child)
family planning, maternal health services 2
job training, apprenticeship programs 5
And even if you think we can afford an arms race, the effect of
the arms race in the last decade has been to stimulate an arms race
of unprecedented magnitude in the Third World, where nations are
spending much greater proportions of their GNP on weapons. There
the price is not education and jobs, it is malnutrition and disease.
Instead of drilling wells so everyone can have clean water, they
are stockpiling weapons. (Not atypical is Ethiopia's long-term
barter agreement made in 1981 to get Russian weapons and pay for
them with future agricultural exports.)
Many of these Third-World arms races may turn out to
be Richardson-unstable and lead to war, too, even if the US-USSR
one remains stable. (They are coupled arms races; this is
mathematically complicated. Richardson analyzed N-country arms races
and found them inherently more unstable than 2-country races; he
compared it to dogs. You can have three dogs, no two of which will
fight, but when you put all three together, they fight. He
predicts sudden changes of alliance, such as Nixon brought about
with China. But two coupled arms races is not the same as a four-
country race, and the situation hasn't been mathematically analyzed.)
High as the price of the arms race is in terms of direct danger
and economic drain, the psychological price is more important yet.
I long for a world in which my children need not feel threatened.
The original SDI speech touched on the same feeling. So long as the
arms race goes on, the threat of doomsday lurks in the back of
everybody's mind, creating fear and suspicion and preventing the
establishment of a truly creative and thriving global civilization.
I think you are mistaken about both Nixon and Kennedy
having threatened to use nuclear weapons. Kennedy threatened
to blockade Cuba and put our armed forces on alert. He made
no nuclear threat. Nixon made no nuclear threat in Vietnam,
although he did conventionally bomb North Vietnam. The discussion
of use of nuclear weapons in Vietnam was during the Eisenhower
Administration when some military people suggested helping the
French. Eisenhower vetoed the idea. Truman vetoed the idea
of using nuclear weapons in Korea.
This needs more research. I have a few sources which I haven't
had time to examine. It may be possible to document explicit
threats. However, even if no explicit threats were made, clearly
the mere existence of a vast nuclear arsenal constitutes an implicit
threat, and I think it would be naive to suppose that politicians
and military men had never tried to exploit that threat to achieve
some aim. Indeed if that is not so, what good are the weapons anyway?
A few dozen would be adequate deterrent. (This according to Kissinger,
who declared that the idea of receiving a few dozen Soviet bombs was
enough to deter Kennedy from using nuclear weapons in the Cuban crisis).
Western Europe has more population and more industry
than the Soviet Union. It could maintain a credible defense
if its politicians couldn't rely on the U.S. I tend to support
the people who propose U.S. withdrawal from NATO even if it
meant German nuclear weapons.
I don't have an opinion on this issue.
As I said in my previous message I don't agree with
your symmetrical treatment of the U.S. and the Soviet Union.
The Soviet Union really is a very peculiar society ---
oppressive, feudal, far more nationalistic than any other,
and with no really legitimate way method of internal political
struggle. Such oligarchies in the past eventually either
turn into monarchies or break down in civil war. The half life
may be a hundred years, however. Note that democracies don't
go to war with each other whereas communist countries do go
to war with each other.
The nature of Soviet society is only marginally relevant to
the desirability of concluding arms control or reduction
agreements with the USSR. It is only necessary to make sure,
supposing that we don't have any basis of mutual trust, that
the treaties are verifiable and that there is no incentive to
break the treaty and plenty of incentive to honor it.
(Since you bring it up, however, I don't see that the Soviet
society is feudal or more nationalistic than our own, although
it is certainly oppressive.)
I don't see how you can imagine that the radioactive
cloud from Chernobyl will deter anybody from anything. The
effects perceptible to Europeans were either media effects
or the result of actions of their governments. I'm not saying
the governments were entirely wrong to restrict use of
certain foods --- merely that there was nothing the public
perceived directly. Compare this with the actual devastation
of World War II, and you'll see why its long term psychological
effect will be trivial.
Good point, well taken.
The Soviets have been developing the Star Wars technologies for
some years and are supposed to be ahead of us. That's
one of the things the emigre statement said. If they decide to
deploy it, they won't announce the fact; they'll do it as secretly
as possible. They deploy missiles without announcement also.
A large part, maybe all, of the arms control negotiations have
been carried out with the Soviets accepting U.S. estimates of
their missile deployment and offering none of their own.
All the more reason for negotiating a treaty to make sure
these weapons never are deployed by either side.
I am dubious of further arms control agreements, because even if
we detect violations, our arms control enthusiasts are likely
to continue finding innocent explanations, as for the Krasnoyarsk
phased array radar and the new missile types.
The violations, if they were violations, are insignificant
compared to the real limits imposed by the treaty; limits on
the number of missles and the number of warheads per missile.
This sort of provocative toe-over-the-line game is played by
both sides, isn't it? Our own SDI tests so far have already
violated the ABM treaty.
Whether or not Krasnoyarsk violates the ABM treaty
is to be decided by a
commission set up in Geneva under the terms of the treaty
just for the purpose of deciding about alleged violations.
Let's wait and see. The other issue is whether the SS-25 is a
new missle (forbidden by SALT-II) or a modification of an old.
Looks like the Soviets are in the wrong here, from what I read.
Anyway it seems a little hypocritical for the US to complain about
violations of a treaty which it has never ratified.
I don't agree that the Soviet offers are generous. I doubt that even
they would put it that way.
How about "attractive"? They seem attractive to me.
It seems to me that destroying all nuclear weapons without creating
mutual trust makes the situation far more unstable than it is today.
If it takes approximately a year to re-establish them, then it is
possible to hope to win a race by several months, which might suffice
to destroy the other side's ability to ever produce nuclear weapons.
This is a very difficult problem which will have to be solved
someday; why not now? The difficulty is that it's hard to establish
mutual trust while we are all threatened by nuclear missiles; yet
without mutual trust we can't actually eliminate the missles. The
answer is step-by-step reductions in weapons, and at the same time
many co-operative programs, e.g. in space exploration, massive
student exchange programs, professional exchange programs in every
professional area, and co-operative aid and development projects
in the third world (why should that be impossible?) instead of
surrogate wars.
Such trust is in principle possible. We don't fear the other democracies'
attacking us with nuclear weapons, and they don't fear our attacking them,
even though there are points of rivalry and conflict. However, we
must simply wait for the Soviet Union to evolve some democracy,
however long it takes --- maybe several hundred years.
You don't adduce any evidence for why we must wait until they adopt
our governmental form before we trust them. That point certainly
requires proof!
Since the Administration and armed forces thinkers don't agree with your
analysis of the situation, your speculations as to their motives are
beside the point. It seems, however, that you are more ready to ascribe
bad motives to them than to the Politbureau that divided Poland with
Hitler in 1939, invaded Finland in 1940, Hungary in 1956, Czechoslovakia
in 1968 and Afghanistan in 1979.
Not to mention Stalin's purges and the Gulag; nor the extermination
of the Greek communists by the British after WWII, the many invasions
of various Central American countries by the US earlier in this
century, etc. I'm not much impressed with any government's conduct.
I ascribe only one motive to the Administration in my paper: the
motive of wanting to build Star Wars in order to use the threat of
nuclear war in the successful exercise of political power in various
parts of the world. I tried to refrain from making a judgment
about whether that is good or bad and concentrate on establishing
it as a fact.
Of course, I consider it to be at best misguided, since it does
not work in the direction of increasing mutual trust and directly
opposes my three-point recipe.
≠
∂30-Dec-86 1759 VAL
Please look at p. 4 of cs326[1,val] and give me your comments.
∂31-Dec-86 0834 FEIGENBAUM@SUMEX-AIM.STANFORD.EDU Play about Turing's life
Received: from SUMEX-AIM.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Dec 86 08:34:14 PST
Date: Wed 31 Dec 86 08:34:05-PST
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.STANFORD.EDU>
Subject: Play about Turing's life
To: nilsson@SCORE.STANFORD.EDU, reges@SCORE.STANFORD.EDU,
dek@SAIL.STANFORD.EDU, jmc@SAIL.STANFORD.EDU, ullman@SCORE.STANFORD.EDU
cc: feigenbaum@SUMEX-AIM.STANFORD.EDU
Message-ID: <12267209447.12.FEIGENBAUM@SUMEX-AIM.STANFORD.EDU>
Friends,
when I was in London In November, I attended a performance of an extraordinary
play called "Breaking the Code" (in the West End, to a packed theater). It is
about Turing and is based on the recent biography by Hodges. It is faithful to
Turing's life and the biography. It treated the computer science directly and
without compromise (for example, Turing gives forth with a monologue about
the Decidability Problem; and he discusses his intuitions about the possibilities for machine intelligence). It is also uncompromising and sympathetic with
regard to Turing's homosexuality, his unconventionality, and his
unwillingness to bend to society's conventions.
Given the importance of Turing to Computer Science, and the quality of this
play, and emotional experience of this play, I suggest that CSD enter into
a "joint venture" with the Drama Department at Stanford to do this play at
one of next year's Drama Department performances. Our role would be to
help fund their production (from departmental unrestricted funds and by
some fund-raising, perhaps among ourselves). I also suggest that we enlist
the aid and cooperation of the gay students' organizations, perhaps making
these organizations part of the "joint venture".
What do you people think? Please feel free to pass this message along to
selected individuals who you think might be interested/helpful.
Best new year's wishes to all of you,
Ed Feigenbaum
-------
∂31-Dec-86 0900 CLT
jacket to cleaners
∂31-Dec-86 0900 JMC
book
∂31-Dec-86 0900 JMC
phone about dentist appointment
∂31-Dec-86 1446 TALEEN@Score.Stanford.EDU 3:00
Received: from SCORE.STANFORD.EDU by SAIL.STANFORD.EDU with TCP; 31 Dec 86 14:46:17 PST
Date: Wed 31 Dec 86 14:43:59-PST
From: Taleen Marashian <TALEEN@Score.Stanford.EDU>
Subject: 3:00
To: jmc@Sail.Stanford.EDU
Message-ID: <12267276786.27.TALEEN@Score.Stanford.EDU>
Because the department will "officially" close at 3:00 p.m. this afternoon,
I will be leaving then. Have a Happy (and safe) New Year's Eve (and Day).
May 1987 be a wonderful year for you and your "expanded" family (that
card with Timothy on the terminal is really cute!).
Best Wishes,
Taleen
-------
∂31-Dec-86 1529 Bobrow.pa@Xerox.COM re: A personal favor
Received: from XEROX.COM by SAIL.STANFORD.EDU with TCP; 31 Dec 86 15:27:38 PST
Received: from Cabernet.ms by ArpaGateway.ms ; 31 DEC 86 10:28:02 PST
Date: 31 Dec 86 10:27 PST
Sender: Bobrow.pa@Xerox.COM
From: Danny Bobrow <Bobrow.pa@Xerox.COM>
Subject: re: A personal favor
In-reply-to: John McCarthy <JMC@SAIL.STANFORD.EDU>'s message of 29 Dec
86 23:16 PST
To: JMC@SAIL.STANFORD.EDU
cc: Bobrow.pa@Xerox.COM,oshea.pa@Xerox.COM
Message-ID: <861231-102802-1333@Xerox>
Thanks John. Tim Oshea will be in touch about arrangements
danny
∂31-Dec-86 1530 STEINBERGER@SRI-KL.ARPA Evil empire, etc.
Received: from SRI-KL.ARPA by SAIL.STANFORD.EDU with TCP; 31 Dec 86 15:30:48 PST
Date: Wed 31 Dec 86 15:31:03-PST
From: Richard Steinberger <STEINBERGER@SRI-KL.ARPA>
Subject: Evil empire, etc.
To: jmc@SAIL.STANFORD.EDU
cc: steinberger@SRI-KL.ARPA, su-etc@SCORE.STANFORD.EDU
Message-ID: <12267285354.9.STEINBERGER@SRI-KL.ARPA>
JMC States:
"Finally, what agreements are possible with the Soviet Union? Cultural,
sure. Economic, probably, if genuinely mutual. Scientific, likewise.
Arms control, maybe. My hesitation about arms control agreements stems in a
large measure from the fact that our arms control community is intent on
giving the Soviet Union a license to violate them, by finding every
possible benign intepretation of the evidence for violation, and saying
that if violations did occur, they aren't important. In the Soviet
system, this gives their hawks every possible internal political
advantage. I'd like to swap some of their territorial secrecy for some of our
technology, and this would make other agreements more feasible. I'd
like to swap something to get them to give up their draft."
Rather than disagree with several points he tries to make in earlier portions
of the text I would like to concentrate on the above passage for it is here
that I see the closest points of agreement with my own position. I think John
is wise and forthright in suggesting that even though he feels that "evil
empire" is not an inappropriate epithet to apply to the U.S.S.R., we should
still attempt to proceed with cultural, scientific, and economic agreements.
Arms control agreements are to be approached with the greatest reluctance
if at all; allegedly the U.S. arms control negotiators are intent on giving
the Soviets the "upper hand." I would like to suggest that the situation is
not hopeless. In particular, a comprehensive nuclear test ban treaty would
be a logical step to take. The articles I have read by geophysicists
(e.g. Sykes and Evernden, Scientific American, October, 1982) and other
professionials familiar with monitoring nuclear tests have convinced me that
it would be impossible for the Soviets to 'sucessfully' cheat. Their
government has expressed and demonstrated a willingness to pursue just such
an initiative. Our own government has held back in an attempt to develop
SDI weaponry.
JMC and others are correct in feeling that we are facing a precarious balance
of power vis-a-vis the Soviet government. I agree that the utmost caution
and care ought to be used in reaching any arms control agreements. Rather
than argue on what form this caution ought to take diplomatically and
proceedurally, I would like to suggest that it is important that some
concensus be reached in America on the desirability of sincere (by both
governments) arms reduction talks. (It ought to be obvious but in case it's
not, I'll state that I am in no form calling for unilateral disarmament by
the U.S.). Whether the U.S.S.R. represents an "evil" empire or some other
kind ought not prevent arms controls talks from occuring.
It seems that a vital question is the following: "Can the balance of power
[I know some would claim that the U.S. is in a weaker position at present,
but I shall not now pursue this argument] that exists between the two
major world powers be preserved as the volume and lethality of weapons each
side possesses is reduced over time?" I believe that if we and the Soviets
want this to happen that it can in fact occur; and both nations would benefit
in a variety of ways. If one does not believe that this is possible then
I conclude that they believe it is less risky to continue to pursue the
arms race with accelerating speed and cost. Is this an incorrect interpret-
ation?
If arms reduction talks can become more 'productive' and cultural, economic,
scientific agreements continue in a calmer atmosphere, then it is more likely
that both countries will become better places to live. A big 'if' some would
argue, but I think that the risks are worth taking. When weighed against
an unknown number of future years of continued arms competitions, threats
and counterthreats, cautious detente seems the wisest choice.
------------------------------------------------------------------------------
Some people have suggested that the Soviets are so evil, malicious, nefarious,
dastardly, cruel, and heartless (take your choice or take them all), we can-
not and ought not deal with them under any circumstances. Often supporting
evidence is given and I have been challenged to provide comparable acts
committed by the U.S. government (such as the treatment of most of the native
Americans). I don't see that this can lead to any particularily productive
exchange of opinions (that's what were doing isn't it?). I would like to
suggest that the more fruitful topics of discussion, if people are interested
in continuing, concern the nature of the Soviet state (JMC's original
submission), and the possibility and/or desirability of reaching any form of
agreements with them.
Ric Steinberger
Steinberger@sri-kl.arpa
-------
∂31-Dec-86 1529 oshea.pa@Xerox.COM Video lectures for new Xerox President
Received: from XEROX.COM by SAIL.STANFORD.EDU with TCP; 31 Dec 86 15:27:43 PST
Received: from Semillon.ms by ArpaGateway.ms ; 31 DEC 86 11:06:01 PST
Date: Wed, 31 Dec 86 11:05:33 PST
From: oshea.pa@Xerox.COM
Subject: Video lectures for new Xerox President
To: John McCarthy <JMC@SAIL.STANFORD.EDU>
cc: Bobrow.pa@Xerox.COM
Message-ID: <861231-110601-1340@Xerox>
Dear Professor McCarthy,
As Danny Bobrow mentioned we would be very grateful if you would
consent to a short videotaped interview to be included in the one hour
programme "Programming Languages: A Selected Intellectual and Social
History". Other interviewees include Professors Nygaard and Perlis.
I am afraid that I do not have a budget for payment of presenters or
interviewees. I would be happy to put a fixed time limit on our
quotation from any interview with you of say three or five minutes.
Currently there is no plan to use these videos for any other purpose
than the education of the President and CEOs of Xerox Corporation. If
the first programmes are successful within Xerox then further video
lectures may be prepared for resale by the Corporate Training Division.
I'm happy to attempt to secure a flat fee for you in the event that
Xerox resell a video lecture in which you are quoted. What fee would you
consider appropriate for a five minute interview?
If you are willing to proceed on this basis I can arrange for a video
production team and interviewer to visit you at your convenience on
Friday 9th of January.
Obviously the history and evolution of Lisp will from an important part
of this lecture and my judgement is that Danny is in position to educate
the most senior staff in the Corporation with regard to programming
languages.
If you were willing to donate some time (the interviews to date have
taken about an hour to record) we would be very much in your debt.
Sincerely,
Tim O'Shea