perm filename S83.IN[LET,JMC] blob
sn#719049 filedate 1983-07-02 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00434 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00047 00002
C00048 00003 ∂02-Apr-83 1414 SCOTT@SUMEX-AIM [Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>: Re: [John McCarthy <JMC@SU-AI>:]]
C00050 00004 ∂03-Apr-83 1553 KLC
C00051 00005 ∂03-Apr-83 2002 JMC*
C00052 00006 ∂04-Apr-83 0046 Lowell Wood <LLW at S1-A> Applied Eugenics
C00062 00007 ∂04-Apr-83 1031 ATP.BLEDSOE@UTEXAS-20 [CL.BOYER at UTEXAS-20: [John McCarthy <JMC%SU-AI@USC-ECL>:]]
C00064 00008 ∂04-Apr-83 1716 RINDFLEISCH@SUMEX-AIM Equipment Contract Rebudgeting
C00066 00009 ∂04-Apr-83 1938 LGC Opportunity for data-acquisition
C00070 00010 ∂05-Apr-83 0900 BUCHANAN@SUMEX-AIM systems candidate Gael Buckley
C00071 00011 ∂05-Apr-83 1218 WIEDERHOLD@SRI-AI Title
C00072 00012 ∂05-Apr-83 1346 BUCHANAN@SUMEX-AIM Comtex Introduction
C00073 00013 ∂05-Apr-83 1346 RINDFLEISCH@SUMEX-AIM Stanford Equipment Contract - DSSW Rebudgeting
C00076 00014 ∂05-Apr-83 1350 BUCHANAN@SUMEX-AIM G.Novak's application
C00077 00015 ∂05-Apr-83 1610 PB I haven't forgotten
C00078 00016 ∂05-Apr-83 2109 CL.BOYER at UTEXAS-20 Wilson's Theorem, FYI
C00080 00017 ∂06-Apr-83 1622 CSD.BSCOTT@SU-SCORE
C00081 00018 ∂07-Apr-83 2052 WIEDERHOLD@SRI-AI Addition to ARPA joint proposal
C00083 00019 ∂07-Apr-83 2103 WIEDERHOLD@SUMEX-AIM Re: Addition to ARPA joint proposal
C00085 00020 ∂07-Apr-83 2147 DEK
C00096 00021 ∂07-Apr-83 2142 DEK ARPA
C00098 00022 ∂08-Apr-83 0512 OHLANDER@USC-ISI Re: Stanford Equipment Contract - DSSW Rebudgeting
C00100 00023 ∂08-Apr-83 0825 WIEDERHOLD@SUMEX-AIM Re: Addition to ARPA joint proposal
C00102 00024 ∂08-Apr-83 0847 RINDFLEISCH@SUMEX-AIM DARPA Equipment Rebudgetting
C00106 00025 ∂08-Apr-83 0911 CSD.BSCOTT@SU-SCORE Re: Addition to ARPA joint proposal
C00108 00026 ∂08-Apr-83 1117 BUCHANAN@SUMEX-AIM search comm. mtg today 3:30, chm's conf. room
C00109 00027 ∂08-Apr-83 1128 BUCHANAN@SUMEX-AIM Room Change for 3:30 mtg (sorry) -- Room 301 instead of Chm's Conf.Rm
C00110 00028 ∂08-Apr-83 1342 RPG Tasking
C00111 00029 ∂09-Apr-83 1542 AAAI@SRI-AI MIT Offer
C00114 00030 ∂11-Apr-83 0748 CSD.BSCOTT@SU-SCORE Re: Tasking
C00116 00031 ∂11-Apr-83 1237 CLT on TTY75 (at TV-143) 1237
C00117 00032 ∂11-Apr-83 1633 RINDFLEISCH@SUMEX-AIM Re: DARPA Equipment Rebudgetting
C00119 00033 ∂11-Apr-83 2359 RV AI Qual
C00120 00034 ∂12-Apr-83 1645 NAN@SU-SCORE pmessage
C00121 00035 ∂12-Apr-83 2005 WIEDERHOLD@SUMEX-AIM Re: ARPA Umbrella
C00122 00036 ∂12-Apr-83 2005 WIEDERHOLD@SUMEX-AIM
C00123 00037 ∂12-Apr-83 2016 CSD.RUSSELL@SU-SCORE ai qual
C00124 00038 ∂12-Apr-83 2024 RPG
C00126 00039 ∂12-Apr-83 2113 WIEDERHOLD@SUMEX-AIM
C00128 00040 ∂14-Apr-83 1000 BUCHANAN@SUMEX-AIM next comm. mtg
C00130 00041 ∂14-Apr-83 1002 BUCHANAN@SUMEX-AIM Michael Carey
C00131 00042 ∂14-Apr-83 1414 CSD.BSCOTT@SU-SCORE Re: honor yes, advantage no
C00132 00043 ∂14-Apr-83 1448 NAN@SU-SCORE pmessage
C00133 00044 ∂14-Apr-83 1650 AAAI@SRI-AI.ARPA AAAI Membership
C00134 00045 ∂15-Apr-83 1307 DFH
C00135 00046 ∂15-Apr-83 1415 CSD.ULLMAN@SU-SCORE hrumpf
C00137 00047 ∂15-Apr-83 1455 RINDFLEISCH@SUMEX-AIM Draft Equipment Budget
C00140 00048 ∂15-Apr-83 2053 WIEDERHOLD@SUMEX-AIM Re: Proposal
C00142 00049 ∂16-Apr-83 1111 CLT boris
C00143 00050 ∂16-Apr-83 1335 CSD.BSCOTT@SU-SCORE Boeing
C00144 00051 ∂17-Apr-83 0242 Lowell Wood <LLW at S1-A> Monday Meetings
C00146 00052 ∂17-Apr-83 2238 JACK MINKER <minker.umcp-cs@UDel-TCP> Lerner's 70th Birthday
C00149 00053 ∂17-Apr-83 2249 LGC Data Opportunity Reminder
C00150 00054 ∂18-Apr-83 1145 DFH Govt. publications
C00151 00055 ∂18-Apr-83 1308 ALS
C00153 00056 ∂18-Apr-83 1744 KAHN at USC-ISI Re: dinner on Wednesday?
C00155 00057 ∂19-Apr-83 1016 BUCHANAN@SUMEX-AIM Gael Buckley's papers
C00157 00058 ∂19-Apr-83 1441 ALS
C00158 00059 ∂19-Apr-83 1627 DFH
C00159 00060 ∂19-Apr-83 1916 KAHN at USC-ISI
C00160 00061 ∂19-Apr-83 2015 KAHN@USC-ISI Re: Addition to ARPA joint proposal
C00162 00062 ∂19-Apr-83 2342 RPG
C00164 00063 ∂20-Apr-83 0620 BUCHANAN@SUMEX-AIM correction: Buckley is coming Apr.29 (not May 29)
C00165 00064 ∂20-Apr-83 1238 YOM
C00166 00065 ∂21-Apr-83 1022 WIEDERHOLD@SUMEX-AIM Re: Addition to ARPA joint proposal
C00169 00066 ∂21-Apr-83 1041 WIEDERHOLD@SUMEX-AIM Re: Addition to ARPA joint proposal
C00172 00067 ∂22-Apr-83 1003 RINDFLEISCH@SUMEX-AIM [OHLANDER at USC-ISI: Re: Equipment Contract Budget]
C00178 00068 ∂22-Apr-83 1332 ADMIN.LIBRARY@SCORE Lisp-Anwendungsgebiete, Grundbegriffe, Geschichte
C00180 00069 ∂22-Apr-83 1335 MRC@SCORE Re: TIP phone numbers
C00182 00070 ∂22-Apr-83 1339 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract Budget
C00184 00071 ∂22-Apr-83 1356 RPG Trip(s)
C00186 00072 ∂22-Apr-83 1459 AAAI%SRI-AI.ARPA@SCORE [Louis Robinson <AAAI-OFFICE@SUMEX-AIM.ARPA>: [Michael Genesereth <GENESERETH@SUMEX-AIM.ARPA>: request]]
C00191 00073 ∂22-Apr-83 1525 RINDFLEISCH@SUMEX-AIM Lisp Workstations
C00194 00074 ∂22-Apr-83 1528 RINDFLEISCH@SUMEX-AIM Still haven't received your earlier msg...
C00195 00075 ∂22-Apr-83 1609 PB library data base
C00198 00076 ∂22-Apr-83 1630 Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
C00200 00077 ∂22-Apr-83 2015 CSD.ROWE@SCORE Re: Omni
C00201 00078 ∂23-Apr-83 0730 Woody Bledsoe <ATP.Bledsoe@UTEXAS-20>
C00202 00079 ∂23-Apr-83 1402 KAHN@USC-ISI Re: Addition to ARPA joint proposal
C00204 00080 ∂23-Apr-83 2351 Lowell Wood <LLW at S1-A> Trips
C00208 00081 ∂24-Apr-83 0004 Lowell Wood <LLW at S1-A> IPTO S-1 Support
C00212 00082 ∂24-Apr-83 1603 CSD.MCGRATH@SCORE [BUNDY@RUTGERS: New Lectureships]
C00216 00083 ∂24-Apr-83 1656 CLT
C00218 00084 ∂25-Apr-83 0328 host USC-ECL
C00222 00085
C00224 00086 ∂25-Apr-83 1100 JK
C00228 00087 ∂25-Apr-83 1159 GOTELLI@SCORE Re: new account
C00229 00088 ∂25-Apr-83 1259 ZAVEN@SCORE DARPA, RADC
C00230 00089 ∂25-Apr-83 1432 RPG Buying all the 3600's the first year
C00232 00090 ∂25-Apr-83 1609 CSD.FOYSTER@SCORE Denelcor HEP
C00234 00091 ∂25-Apr-83 1653 DFH Chicago travel expenses
C00235 00092 ∂26-Apr-83 1433 DFH phone call from Prof. Schwartz (sp?)
C00236 00093 ∂27-Apr-83 0408 KDF%MIT-OZ@MIT-MC metaphors
C00237 00094 ∂27-Apr-83 0616 Communications Satellite <COMSAT @ MIT-MC> Msg of Wednesday, 27 April 1983 06:56 EDT
C00239 00095 ∂27-Apr-83 0839 GENESERETH@SUMEX-AIM Invitation
C00242 00096 ∂27-Apr-83 1006 FEIGENBAUM@SUMEX-AIM BBC's HORIZON program on AI
C00246 00097 ∂27-Apr-83 1244 GAVAN%MIT-OZ@MIT-MC metaphors
C00248 00098 ∂27-Apr-83 1305 GAVAN%MIT-OZ@MIT-ML metaphors
C00250 00099 ∂27-Apr-83 1316 DAM%MIT-OZ@MIT-ML metaphors
C00253 00100 ∂27-Apr-83 1328 DAM%MIT-OZ@MIT-MC metaphors
C00256 00101 ∂27-Apr-83 1328 RICKL%MIT-OZ@MIT-ML Re: metaphors
C00262 00102 ∂27-Apr-83 2016 JMC*
C00263 00103 ∂27-Apr-83 2055 Christopher C. Stacy <CSTACY @ MIT-MC>
C00264 00104 ∂28-Apr-83 0743 CSD.BSCOTT@SCORE
C00265 00105 ∂28-Apr-83 1042 BUCHANAN@SUMEX-AIM COMTEX introduction
C00267 00106 ∂29-Apr-83 1401 CSD.ULLMAN@SCORE Adcom
C00269 00107 ∂29-Apr-83 1600 CSD.ULLMAN@SCORE
C00270 00108 ∂29-Apr-83 1600 BUCHANAN@SUMEX-AIM comtex intro
C00272 00109 ∂29-Apr-83 1636 CSD.ULLMAN@SCORE
C00273 00110 ∂29-Apr-83 1719 MAILER failed mail returned
C00275 00111 ∂30-Apr-83 1128 CSD.BSCOTT@SCORE Diana Hall
C00278 00112 ∂30-Apr-83 1258 JJW Slander?
C00284 00113 ∂01-May-83 2317 FEIGENBAUM@SUMEX-AIM IBM
C00286 00114 ∂02-May-83 0918 NILSSON%SRI-AI.ARPA@SCORE Lou Robinson Resignation
C00291 00115 ∂02-May-83 1022 BOSACK@SCORE [Ed Franceschini <FRANCESCHINI.CMCL1@NYU.ARPA>: RE: We will have a user in your area....]
C00294 00116 ∂02-May-83 1037 BOSACK@SCORE Your NYC User
C00295 00117 ∂02-May-83 1032 DFH Statistical Abstract of US
C00297 00118 I see that SCORE has become very prompt in forwarding mail.
C00299 00119 ∂02-May-83 1949 SIEGMAN@SU-SIERRA Re: C.I.T. and lunch
C00301 00120 ∂03-May-83 0134 MRC@SCORE Re: Abstract for talk "How Flavors Differ from Smalltalk"
C00303 00121 ∂03-May-83 0757 SIEGMAN@SU-SIERRA
C00312 00122 ∂03-May-83 1555 BSCOTT@SCORE
C00313 00123 ∂03-May-83 1610 RINDFLEISCH@SUMEX-AIM Equipment Contract Rebudgeting
C00319 00124 ∂03-May-83 1627 SIEGMAN@SU-SIERRA
C00320 00125 ∂03-May-83 1654 RINDFLEISCH@SUMEX-AIM
C00321 00126 ∂03-May-83 1702 MAILER failed mail returned
C00323 00127 ∂04-May-83 0921 AAAI-OFFICE@SUMEX-AIM Re: Lou Robinson Resignation
C00329 00128 ∂04-May-83 1000 JMC*
C00330 00129 ∂04-May-83 1600 JMC*
C00331 00130 ∂04-May-83 2121 RPG*
C00332 00131 ∂04-May-83 2125 ULLMAN@SCORE CF
C00333 00132 ∂04-May-83 2126 PERILLO%SRI-NIC.ARPA@SCORE Re: user in NYC
C00336 00133 ∂04-May-83 2127 jlh@Shasta McSuns
C00338 00134 ∂04-May-83 2134 PERILLO@SRI-NIC Re: user in NYC
C00340 00135 ∂04-May-83 2145 @USC-ECLC:LLW@S1-A Re-Transmitting. . .
C00344 00136 ∂04-May-83 2326 RMS%MIT-OZ%MIT-MC.ARPA@SCORE My abstract
C00345 00137 Does the following mean that TCP incoming mail is working, and I can
C00348 00138 ∂05-May-83 0026 ME mail
C00349 00139 ∂05-May-83 0245 @USC-ECLC,@MIT-MC:RJF@MIT-MC
C00351 00140 ∂05-May-83 1307 DFH additional phone
C00353 00141 ∂05-May-83 1524 minker.umcp-cs@UDel-Relay Lerner's 7oth Birthday and IJCAI
C00369 00142 ∂06-May-83 1519 RESTIVO@SCORE latest Prolog
C00372 00143 ∂06-May-83 1524 DFH air reservations
C00373 00144 ∂06-May-83 1833 ME Prancing Pony Bill
C00375 00145 ∂07-May-83 1244 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN essences
C00381 00146 ∂07-May-83 1247 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN
C00383 00147 ∂07-May-83 1253 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN metaphor
C00391 00148 ∂07-May-83 1651 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN
C00393 00149 ∂07-May-83 1716 @MIT-MC:BATALI%MIT-OZ%MIT-MC@SU-DSN Unnatural kinds
C00399 00150 ∂07-May-83 1733 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN metaphor
C00414 00151 ∂07-May-83 2001 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN failed mail return
C00416 00152 ∂07-May-83 2211 @MIT-MC:KDF%MIT-OZ%MIT-MC@SU-DSN visit
C00417 00153 ∂08-May-83 0247 RMS BBOARD message
C00418 00154 ∂08-May-83 1528 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN Re: natural kinds
C00425 00155 ∂08-May-83 1728 Mailer@SCORE Message of 8-May-83 10:49:32
C00426 00156
C00440 00157 ∂08-May-83 1816 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN natural kinds
C00442 00158 ∂08-May-83 2341 @MIT-MC:MINSKY%MIT-OZ%MIT-MC@SU-DSN visit
C00443 00159 ∂09-May-83 0850 OHLANDER@USC-ISI Stanford Proposal.
C00466 00160 ∂09-May-83 1017 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN visit and seminar
C00467 00161 ∂09-May-83 0921 YEARWOOD@SCORE Student Support for Summer 83
C00469 00162 ∂09-May-83 1136 ME
C00472 00163 ∂09-May-83 1404 DFH flight times
C00474 00164 ∂09-May-83 1532 minker.umcp-cs@UDel-Relay Special Session on Non-Monotonic Logic
C00479 00165 ∂09-May-83 1800 JMC*
C00480 00166 ∂09-May-83 1840 JMC*
C00481 00167 ∂09-May-83 1926 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN visit
C00484 00168 ∂09-May-83 2040 GENESERETH@SUMEX-AIM Re: Keyworth
C00485 00169 ∂09-May-83 2102 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN [JMC: visit ]
C00487 00170 ∂09-May-83 2120 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN account
C00489 00171 ∂09-May-83 2254 @MIT-MC:MINSKY%MIT-OZ%MIT-MC@SU-DSN natural kinds
C00497 00172 ∂09-May-83 2355 @MIT-MC:MINSKY%MIT-OZ%MIT-MC@SU-DSN
C00503 00173 ∂10-May-83 0035 ISAACSON@USC-ISI Control-Z: Inquir at OZ
C00505 00174 ∂10-May-83 0047 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN [JMC: <ctrl>z ]
C00507 00175 ∂10-May-83 0108 RESTIVO@SCORE PROLOG Digest V1 #2
C00527 00176 ∂10-May-83 0111 @LBL-CSAM.ARPA:uw-beaver!ubc-vision!reiter@LBL-CSAM Lerner
C00530 00177 ∂10-May-83 0326 @USC-ECL,@MIT-ML:GAVAN%MIT-OZ%MIT-ML@SU-DSN natural kinds
C00538 00178 ∂10-May-83 1041 DCL Umbrella
C00540 00179 ∂10-May-83 1100 JMC*
C00541 00180 ∂10-May-83 1104 RPG Varia concerning Ohlander
C00542 00181 ∂10-May-83 1154 DFH Ohlander
C00543 00182 ∂10-May-83 1309 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN paradigms
C00544 00183 ∂10-May-83 1442 TOB
C00547 00184 ∂10-May-83 1554 @MIT-MULTICS.ARPA:JCMa@MIT-OZ my mail address
C00549 00185 ∂10-May-83 1657 ZM DARPA proposal
C00551 00186 ∂10-May-83 1738 RITA@SCORE Package for you
C00553 00187 ∂10-May-83 1814 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN paradigms
C00556 00188 ∂10-May-83 1848 FWH
C00560 00189 ∂10-May-83 1848 FWH SOW
C00561 00190 ∂10-May-83 1849 FWH SOW
C00563 00191 ∂10-May-83 2250 NOVAK@SUMEX-AIM Petition
C00564 00192 ∂11-May-83 0224 POURNE@MIT-MC test
C00565 00193 ∂11-May-83 1105 Mailer@SCORE Message of 8-May-83 10:49:32
C00567 00194 ∂11-May-83 1109 @MIT-MC:LEVITT%MIT-OZ%MIT-MC@SU-DSN paradigms
C00569 00195 ∂11-May-83 1333 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN natural kinds
C00574 00196 ∂11-May-83 1613 @MIT-MC:MINSKY%MIT-OZ%MIT-MC@SU-DSN natural kinds
C00582 00197 ∂11-May-83 2015 Hewitt%MIT-OZ@MIT-MC your seminar
C00583 00198 ∂11-May-83 2023 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN paradigms
C00586 00199 ∂12-May-83 0220 GOLUB@SCORE Re: DARPA scope document
C00587 00200 ∂12-May-83 0515 Solomon.Datanet@MIT-MULTICS.ARPA memo for weekend seminar
C00614 00201 ∂12-May-83 0728 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN counterfactuals & etc.
C00616 00202 ∂12-May-83 1554 RV qual details
C00617 00203 ∂12-May-83 1829 NILSSON%SRI-AI.ARPA@SCORE Online abstracts and articles
C00620 00204 ∂12-May-83 1903 NILSSON%SRI-AI.ARPA@SCORE Online abstracts and articles
C00623 00205 ∂12-May-83 1903 NILSSON%SRI-AI.ARPA@SCORE Agenda Item
C00629 00206 ∂13-May-83 0816 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN [CSTACY: forwarded]
C00630 00207 ∂13-May-83 1001 Solomon.Datanet@MIT-MULTICS.ARPA Saturday and Sunday
C00632 00208 ∂13-May-83 1037 DBL ai qual
C00633 00209 ∂13-May-83 1108 DFH AI Qual
C00634 00210 ∂13-May-83 1349 DFH AI Qual
C00635 00211 ∂13-May-83 1447 DFH Ben Grosof
C00636 00212 ∂13-May-83 1518 minker.umcp-cs@UDel-Relay Special Session on Non-Monotonic Logic
C00640 00213 ∂13-May-83 1614 @MIT-MC:LEVITT%MIT-OZ%MIT-MC@SU-DSN oops -- corrected back to data types
C00645 00214 ∂13-May-83 1701 DFH Computer Facilities Meeting
C00646 00215 ∂13-May-83 1707 MAILER failed mail returned
C00647 00216 ∂14-May-83 0201 ME mail
C00648 00217 ∂14-May-83 0238 @MIT-MC:HDT%MIT-OZ%MIT-MC@SU-DSN via mail rms
C00650 00218 ∂14-May-83 2122 DEK DARPA scope document
C00651 00219 ∂14-May-83 2320 Kanef%HP-HULK.HP-Labs%RAND-RELAY.ARPA@SCORE Circumscription paper
C00654 00220 ∂15-May-83 0747 FEIGENBAUM@SUMEX-AIM
C00656 00221 ∂16-May-83 0900 JMC*
C00657 00222 ∂16-May-83 0927 BSCOTT@SU-SCORE.ARPA ARPA Umbrella
C00658 00223 ∂16-May-83 0947 DFH
C00659 00224 ∂16-May-83 1048 DFH Meeting
C00660 00225 ∂16-May-83 1237 ME leased line
C00661 00226 ∂16-May-83 2009 JMC*
C00662 00227 ∂16-May-83 2205 RPG Dialups???
C00663 00228 ∂17-May-83 1833 ME leased line
C00664 00229 ∂17-May-83 2139 KARP@SUMEX-AIM Facilities committee meeting Thursday
C00665 00230 ∂17-May-83 2249 RPG
C00666 00231 ∂18-May-83 0742 OHLANDER@USC-ISI Stanford SOW
C00678 00232 ∂18-May-83 0900 JMC*
C00679 00233 ∂18-May-83 1000 JMC*
C00680 00234 ∂18-May-83 1000 JMC*
C00681 00235 ∂18-May-83 1132 BSCOTT@SU-SCORE.ARPA Re: SOW
C00682 00236 ∂18-May-83 1259 GENESERETH@SUMEX-AIM Re: Keyworth is out
C00684 00237 ∂18-May-83 1441 NILSSON%SRI-AI.ARPA@SU-SCORE.ARPA Letter
C00688 00238 ∂18-May-83 2242 WIEDERHOLD@SUMEX-AIM Re: DARPA scope document
C00689 00239 ∂19-May-83 0006 LLW@S1-A Second Choice
C00691 00240 ∂19-May-83 0112 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #3
C00714 00241 ∂19-May-83 1002 NOVAK@SUMEX-AIM Why MCC chose Austin
C00717 00242 ∂19-May-83 1022 DMC AI qual
C00718 00243 ∂19-May-83 1928 @MIT-MC:LEVITT%MIT-OZ%MIT-MC@SU-DSN no learning systems yet??
C00721 00244 ∂19-May-83 2316 WIEDERHOLD@SUMEX-AIM Re: Stanford Proposal.
C00723 00245 ∂20-May-83 0030 ARK@SU-SCORE.ARPA SCORE Modem Service
C00727 00246 ∂20-May-83 0455 OHLANDER@USC-ISI Re: Stanford Proposal.
C00729 00247 ∂20-May-83 0814 DIFFERDING@SUMEX-AIM Reducing number of Altos
C00730 00248 ∂20-May-83 0828 TVR
C00733 00249 ∂20-May-83 0934 NOVAK@SU-SCORE.ARPA Altos
C00734 00250 ∂20-May-83 0943 WINSLETT@SU-SCORE.ARPA alto reduction
C00735 00251 ∂20-May-83 0948 trickey@Diablo Altos
C00736 00252 ∂20-May-83 1012 RPG travel strategy
C00737 00253 ∂20-May-83 1142 NOVAK@SUMEX-AIM [NAME AAAI-OFFICE <AAAI-OFFICE@SUMEX-AIM.ARPA>: Potential Exhibitors for AAAI-83]
C00740 00254 ∂20-May-83 1147 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN representing sequences by sets
C00742 00255 ∂20-May-83 1217 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN role of logic in AI
C00744 00256 ∂20-May-83 1332 YOUM@SU-SIERRA Removal of ALTO
C00745 00257 ∂20-May-83 1341 SHARON@SU-SCORE.ARPA tickets
C00746 00258 ∂20-May-83 1447 BSCOTT@SU-SCORE.ARPA Layoff
C00750 00259 ∂20-May-83 1448 BSCOTT@SU-SCORE.ARPA layoff letter
C00751 00260 ∂20-May-83 1600 SHARON@SU-SCORE.ARPA visitor
C00753 00261 ∂20-May-83 2206 shore@Shasta Altos
C00755 00262 ∂21-May-83 0107 WIEDERHOLD@SUMEX-AIM [OHLANDER at USC-ISI: Re: Stanford Proposal.]
C00758 00263 ∂21-May-83 1223 DIETTERICH@SUMEX-AIM Re: reducing number of Altos
C00763 00264 ∂23-May-83 0009 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #4
C00770 00265 ∂23-May-83 0815 BSCOTT@SU-SCORE.ARPA Re: phone conversation
C00771 00266 ∂23-May-83 0846 REG
C00775 00267 ∂23-May-83 1035 DIETTERICH@SUMEX-AIM
C00779 00268 ∂23-May-83 1038 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN role of logic in AI
C00789 00269 ∂23-May-83 1712 CAB Altos
C00791 00270 ∂23-May-83 1724 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract
C00793 00271 ∂23-May-83 1729 @MIT-MC:KDF%MIT-OZ%MIT-MC@SU-DSN role of logic in AI
C00802 00272 ∂23-May-83 2329 RMS%MIT-MC.ARPA@SU-SCORE.ARPA
C00803 00273 ∂24-May-83 1049 KARP@SUMEX-AIM Facilities committee
C00804 00274 ∂24-May-83 1057 BSCOTT@SU-SCORE.ARPA ARPA Contract MDA903-80-C-0102
C00806 00275 ∂24-May-83 1658 PACK@SU-SCORE.ARPA Philosophy-of-science
C00810 00276 ∂25-May-83 1452 MDP@SU-SCORE.ARPA PROLOG Digest V1 #5
C00823 00277 ∂27-May-83 1459 RPG Common Lisp
C00825 00278 ∂29-May-83 0026 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #6
C00835 00279 ∂29-May-83 1853 ROODE@SRI-NIC [ROODE at SRI-NIC (David Roode): x.25 interfaces]
C00840 00280 ∂31-May-83 0640 OHLANDER@USC-ISI Re: Common Lisp
C00841 00281 ∂31-May-83 1302 @SU-SCORE.ARPA:LFW@MIT-ML
C00842 00282 ∂31-May-83 1514 TGD Minski's Learning Meaning
C00843 00283 ∂31-May-83 1657 NOVAK@SU-SCORE.ARPA Alto use
C00844 00284 ∂31-May-83 2011 @MIT-MC:BLOCK%MIT-OZ%MIT-MC@SU-DSN Re: role of logic in AI
C00845 00285 ∂31-May-83 2324 RPG Common Lisp
C00848 00286 ∂03-Jun-83 0024 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #7
C00862 00287 ∂04-Jun-83 2200 ME Prancing Pony Bill
C00864 00288 ∂06-Jun-83 0041 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #8
C00880 00289 ∂06-Jun-83 2224 ME uncooperative user
C00884 00290 ∂07-Jun-83 0029 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #9
C00897 00291 ∂07-Jun-83 1453 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA ARPA proposal
C00899 00292 ∂08-Jun-83 1353 BSCOTT@SU-SCORE.ARPA Re: ARPA proposal
C00900 00293 ∂08-Jun-83 1446 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA Re: ARPA proposal
C00902 00294 ∂09-Jun-83 0947 WIEDERHOLD@SUMEX-AIM revision of proposal
C00903 00295 ∂10-Jun-83 0000 JMC*
C00904 00296 ∂10-Jun-83 1020 ULLMAN@SU-SCORE.ARPA facilities committee
C00906 00297 ∂10-Jun-83 1026 DUMAS%SUMEX-AIM.ARPA@SU-SCORE.ARPA prolog
C00908 00298 ∂10-Jun-83 1031 TOB
C00910 00299 ∂10-Jun-83 2147 FEIGENBAUM@SUMEX-AIM Guibas appointment
C00912 00300 ∂12-Jun-83 0222 @SU-SCORE.ARPA:unmvax!hector@Berkeley Cprolog for VAX's
C00915 00301 ∂12-Jun-83 2146 DEK Leo Guibas
C00917 00302 ∂13-Jun-83 1638 PKANERVA@SUMEX-AIM My dissertation
C00920 00303 ∂13-Jun-83 2010 @SU-SCORE.ARPA:HFISCHER@USC-ECLB Anybody Trying to stuff Prolog into an 8088?
C00921 00304 ∂14-Jun-83 0017 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #10
C00934 00305 ∂14-Jun-83 1030 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA Informal Session Dedicated to Lerner
C00937 00306 ∂14-Jun-83 1300 @SU-SCORE.ARPA,@MIT-MC:Hewitt%MIT-OZ@MIT-MC please change my name on this list
C00938 00307 ∂14-Jun-83 1315 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA Proposal
C00940 00308 ∂14-Jun-83 1410 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA ExCom Draft Agenda
C00944 00309 ∂14-Jun-83 1749 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA Draft revised umbrella proposal
C00985 00310 ∂14-Jun-83 1815 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA Draft revised umbrella proposal
C01026 00311 ∂15-Jun-83 0828 DFH AI Qual
C01027 00312 ∂15-Jun-83 1338 BSCOTT@SU-SCORE.ARPA ARPA Umbrella
C01029 00313 ∂15-Jun-83 1514 GROSOF@SU-SCORE.ARPA out of town
C01030 00314 ∂16-Jun-83 1035 DFH
C01034 00315 ∂16-Jun-83 1100 JMC*
C01035 00316 ∂16-Jun-83 1349 ULLMAN@SU-SCORE.ARPA
C01036 00317 ∂16-Jun-83 1415 DIETTERICH@SUMEX-AIM RE: Learning Meaning
C01039 00318 ∂16-Jun-83 1426 CAB Altos
C01042 00319 ∂16-Jun-83 1706 CLT talk
C01043 00320 ∂16-Jun-83 1657 BSCOTT@SU-SCORE.ARPA Umbrella proposal
C01044 00321 ∂16-Jun-83 1632 JJW EKL proofs
C01046 00322 ∂16-Jun-83 1615 JK
C01047 00323 ∂16-Jun-83 1548 JK
C01049 00324 ∂16-Jun-83 2017 JK trip
C01050 00325 ∂17-Jun-83 0545 OHLANDER@USC-ISI Re: Umbrella Proposal
C01052 00326 ∂17-Jun-83 0809 BSCOTT@SU-SCORE.ARPA Re: Umbrella Proposal
C01053 00327 ∂17-Jun-83 0940 WIEDERHOLD@SUMEX-AIM [OHLANDER at USC-ISI: Re: Umbrella proposal]
C01057 00328 ∂17-Jun-83 1113 DFH This morning's phone messages
C01059 00329 ∂17-Jun-83 1324 YM PhD reading committee
C01060 00330 ∂17-Jun-83 1337 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA Award Suggestion
C01063 00331 ∂18-Jun-83 0108 minker.umcp-cs@UDel-Relay Panel on Non-Monotonic Logic
C01067 00332 ∂18-Jun-83 0603 BUNDY@RUTGERS.ARPA
C01068 00333 ∂18-Jun-83 1152 GENESERETH@SUMEX-AIM
C01069 00334 ∂19-Jun-83 1417 FEIGENBAUM@SUMEX-AIM profile of you
C01071 00335 ∂19-Jun-83 1437 FEIGENBAUM@SUMEX-AIM Xerox and AI
C01073 00336 ∂19-Jun-83 1456 CL.BOYER@UTEXAS-20 correctness of program for computing pi or e
C01074 00337 ∂19-Jun-83 1557 FEIGENBAUM@SUMEX-AIM
C01076 00338 ∂20-Jun-83 0014 FAHLMAN@CMU-CS-C lisp on perq
C01081 00339 ∂20-Jun-83 0800 JMC*
C01082 00340 ∂20-Jun-83 1052 RJB βxMAIL modification
C01083 00341 ∂20-Jun-83 1107 Mailer failed mail returned
C01085 00342 ∂20-Jun-83 1400 JMC*
C01086 00343 ∂20-Jun-83 1533 DFH Facilities Committee Meeting
C01087 00344 ∂20-Jun-83 1554 RPG Interesting Orals
C01088 00345 ∂20-Jun-83 2109 HURD@SU-SCORE.ARPA Cuthbert C. Hurd is back on SCORE
C01092 00346 ∂20-Jun-83 2242 FEIGENBAUM@SUMEX-AIM
C01093 00347 ∂20-Jun-83 2247 FEIGENBAUM@SUMEX-AIM Re: Facilities Committee Meeting
C01095 00348 ∂21-Jun-83 0102 BUNDY@RUTGERS.ARPA Re: circumscription
C01096 00349 ∂21-Jun-83 1024 JRP lunch
C01097 00350 ∂21-Jun-83 1135 JRP lunch
C01099 00351 ∂21-Jun-83 1231 WIEDERHOLD@SUMEX-AIM Chris Tucci's programming project for the comp.
C01101 00352 ∂21-Jun-83 1331 ARK Re: Chris Tucci's programming project for the comp.
C01103 00353 ∂21-Jun-83 1526 BSCOTT@SU-SCORE.ARPA lay-off letter
C01105 00354 ∂21-Jun-83 1529 @MIT-MC:perlis.umcp-cs@UDel-Relay logic
C01110 00355 ∂21-Jun-83 1646 PW Chris Tucci's final project
C01112 00356 ∂21-Jun-83 1752 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN First Order Logic and Epistemological Adequacy
C01120 00357 ∂21-Jun-83 1753 REM@MIT-MC Belated voice notes (January, TCP not yet up at SU),
C01129 00358 ∂21-Jun-83 1813 REM@MIT-MC Belated voice notes (January, TCP not yet working then), crunch
C01132 00359 ∂21-Jun-83 1854 REM@MIT-MC Belated voice notes (Feb., TCP not yet working then), causality debugging
C01137 00360 ∂21-Jun-83 2350 RPG
C01138 00361 ∂22-Jun-83 0140 ARK Chris Tucci's CS293 Project
C01140 00362 ∂22-Jun-83 0700 JMC*
C01141 00363 ∂22-Jun-83 0805 JRP 1979 paper
C01142 00364 ∂22-Jun-83 0912 DFH my schedule today
C01143 00365 ∂22-Jun-83 0933 ATP.BLEDSOE@UTEXAS-20 Re: strange mail
C01144 00366 ∂22-Jun-83 0935 DFH Facilities Committee Meeting
C01145 00367 ∂22-Jun-83 1107 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract
C01148 00368 ∂22-Jun-83 1331 ullman@Diablo Re: Equipment Contract
C01150 00369 ∂22-Jun-83 1204 RJB
C01152 00370 ∂22-Jun-83 1451 DFH Arpa pre-award audit
C01153 00371 ∂22-Jun-83 1500 JMC*
C01154 00372 ∂22-Jun-83 1521 BSCOTT@SU-SCORE.ARPA Auditors
C01155 00373 ∂22-Jun-83 1532 DFH summer support for Luigi
C01156 00374 ∂22-Jun-83 1547 BSCOTT@SU-SCORE.ARPA Auditors
C01157 00375 ∂22-Jun-83 2029 JMC*
C01158 00376 ∂22-Jun-83 2041 perlis.umcp-cs@UDel-Relay Re: First Order Logic and Epistemological Adequacy
C01168 00377 ∂22-Jun-83 2205 LLW@S1-A Festschrift, Etc.
C01171 00378 ∂22-Jun-83 2218 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract
C01173 00379 ∂22-Jun-83 2221 LLW@S1-A Treasures Of The East
C01175 00380 ∂23-Jun-83 0131 TUCCI@SU-SCORE.ARPA CS 293 etc
C01178 00381 ∂23-Jun-83 0844 MWALKER@SU-SCORE.ARPA
C01190 00382 ∂23-Jun-83 1059 DFH lunch today
C01191 00383 ∂23-Jun-83 1159 RV Let's talk
C01192 00384 ∂23-Jun-83 1306 RSC Afternoon meeting
C01193 00385 ∂23-Jun-83 1422 ULLMAN@SU-SCORE.ARPA meeting
C01194 00386 ∂23-Jun-83 2240 FEIGENBAUM@SUMEX-AIM Re: making life complicated for ourselves
C01195 00387 ∂23-Jun-83 2322 ULLMAN@SU-SCORE.ARPA
C01197 00388 ∂24-Jun-83 0024 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #11
C01217 00389 ∂24-Jun-83 0954 RSC Prolog
C01218 00390 ∂24-Jun-83 1002 WIEDERHOLD@SUMEX-AIM Re: making life complicated for ourselves
C01219 00391 ∂24-Jun-83 1021 RPG Programming Ability Evaluation
C01223 00392 ∂24-Jun-83 1349 ARK Comp Programming Project
C01224 00393 ∂24-Jun-83 1545 ATP.BLEDSOE@UTEXAS-20 ATP Volume - New information
C01230 00394 ∂25-Jun-83 0019 FEIGENBAUM@SUMEX-AIM copy of msg to Gabriel
C01232 00395 ∂26-Jun-83 0220 LLW@S1-A Wednesday
C01233 00396 ∂26-Jun-83 1600 JMC*
C01234 00397 ∂26-Jun-83 1734 RWW your part of the proposal
C01235 00398 ∂26-Jun-83 2354 HST ICAI in Karlsruhe
C01236 00399 ∂27-Jun-83 0003 HST karlsruhe
C01237 00400 ∂27-Jun-83 1015 BSCOTT@SU-SCORE.ARPA NSF
C01238 00401 ∂27-Jun-83 1223 PAPA@SU-SCORE.ARPA Going Away
C01239 00402 ∂27-Jun-83 1658 RINDFLEISCH@SUMEX-AIM Architecture Proj Hardware
C01241 00403 ∂27-Jun-83 2111 CLT susie
C01242 00404 ∂27-Jun-83 2234 Bonnie%UPenn.UPenn@UDel-Relay AAAI executive board meeing
C01244 00405 ∂28-Jun-83 0015 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #12
C01251 00406 ∂28-Jun-83 1038 NILSSON@SRI-AI.ARPA Re: AAAI executive board meeing
C01252 00407 ∂28-Jun-83 1042 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA meeting
C01253 00408 ∂28-Jun-83 1322 DFH phone messages
C01254 00409 ∂28-Jun-83 1517 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA AAAI Executive Director
C01257 00410 ∂28-Jun-83 1550 @SU-SCORE.ARPA:BOBROW.PA@PARC-MAXC.ARPA Re: AAAI Executive Director
C01259 00411 ∂28-Jun-83 1639 DFH phone message
C01260 00412 ∂28-Jun-83 2242 JBR@S1-A
C01261 00413 ∂29-Jun-83 1618 DFH phone message
C01262 00414 ∂29-Jun-83 2016 CLT
C01263 00415 ∂30-Jun-83 0700 JMC*
C01264 00416 ∂30-Jun-83 0750 CLT matadero shell
C01265 00417 ∂30-Jun-83 1100 DFH ARPA budget
C01267 00418 ∂30-Jun-83 1400 JMC*
C01268 00419 ∂30-Jun-83 1604 @SU-SCORE.ARPA,@MIT-MC:MINSKY%MIT-OZ@MIT-MC AAAI Executive Director
C01270 00420 ∂30-Jun-83 1618 AAAI-OFFICE@SUMEX-AIM COnference accomodations
C01275 00421 ∂30-Jun-83 2000 JMC*
C01276 00422 ∂30-Jun-83 2324 HST AFTER IJCAI
C01278 00423 ∂01-Jul-83 0129 ARK Facilities Committee
C01279 00424 ∂01-Jul-83 1204 ullman%SU-HNV.ARPA@SU-SCORE.ARPA conferences
C01280 00425 ∂01-Jul-83 1207 ullman@Diablo parallel LISP
C01281 00426 ∂01-Jul-83 1210 ullman@Diablo Weyrauch
C01284 00427 ∂01-Jul-83 1347 jlh@Shasta Re: Equipment Contract
C01286 00428 ∂01-Jul-83 1415 ullman@Diablo Re: Parallelizing LISP
C01287 00429 ∂01-Jul-83 1427 ullman@Diablo Re: Equipment Contract
C01290 00430 ∂01-Jul-83 1430 ullman%SU-HNV.ARPA@SU-SCORE.ARPA
C01293 00431 ∂01-Jul-83 1507 lantz%SU-HNV.ARPA@SU-SCORE.ARPA
C01294 00432 ∂01-Jul-83 1528 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Werner finked out. Wants to wait a few weeks.
C01295 00433 ∂02-Jul-83 1133 BSCOTT@SU-SCORE.ARPA Final Report
C01296 00434 ∂02-Jul-83 1956 JMC
C01297 ENDMK
C⊗;
∂02-Apr-83 1414 SCOTT@SUMEX-AIM [Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>: Re: [John McCarthy <JMC@SU-AI>:]]
Received: from SUMEX-AIM by SU-AI with PUP; 02-Apr-83 14:05 PST
Date: Sat 2 Apr 83 13:50:13-PST
From: Carli Scott <SCOTT@SUMEX-AIM.ARPA>
Subject: [Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>: Re: [John McCarthy <JMC@SU-AI>:]]
To: jmc@SU-AI.ARPA
Mail-from: SU-NET-TEMP host SCORE rcvd at 30-Mar-83 1627-PST
Date: Wed 30 Mar 83 15:52:33-PST
From: Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>
Subject: Re: [John McCarthy <JMC@SU-AI>:]
To: SCOTT@SUMEX-AIM.ARPA
cc: CSD.BSCOTT@SU-SCORE.ARPA
In-Reply-To: Your message of Wed 30 Mar 83 07:41:14-PST
Very good, John. Thanks for your comments. -- Betty
-------
-------
∂03-Apr-83 1553 KLC
∂03-Apr-83 1515 JMC elephant
"He meant what he said, and he said what he meant, for an elephant is
faithful one hundred percent".
or
"He said what he meant, and he meant what he said, for an elephant is
faithful until he is dead".
-------------
I think the former is due to Dr. Suess, in ``Horton Hears a Who,''
or perhaps another Horton book. Is the latter a well-known saying somewhere?
I hope this is not some obscure kind of threat.
-Ken
∂03-Apr-83 2002 JMC*
eat
∂04-Apr-83 0046 Lowell Wood <LLW at S1-A> Applied Eugenics
Received: from S1-A by SU-AI with NCP/FTP; 4 Apr 83 00:46:02 PST; for: jmc
Date: 04 Apr 1983 0045-PST
From: Lowell Wood <LLW at S1-A>
Subject: Applied Eugenics
To: jmc at SU-AI
CC: LLW at S1-A
[John: The world surely needs more unabashed reactionaries and
class enemies like yourself! The final sentence of your last
paragraph is right on the button, vis-a-vis Hertz Fellowships.
Have you got Joe Weening married off yet? Lowell]
∂04-Apr-83 1031 ATP.BLEDSOE@UTEXAS-20 [CL.BOYER at UTEXAS-20: [John McCarthy <JMC%SU-AI@USC-ECL>:]]
Received: from SU-SCORE by SU-AI with PUP; 04-Apr-83 10:31 PST
Received: from UTEXAS-20 by SU-SCORE.ARPA with TCP; Mon 4 Apr 83 10:12:24-PST
Date: 4 Apr 1983 1205-CST
From: Woody Bledsoe <ATP.Bledsoe@UTEXAS-20>
Subject: [CL.BOYER at UTEXAS-20: [John McCarthy <JMC%SU-AI@USC-ECL>:]]
To: jmc@SU-AI
cc: ATP.Bledsoe@UTEXAS-20
John,
We echo Bob Boyer's remark below: talk on whatever you like at
our Centenial.
I hope you are able to come. Please let me know. Woody
---------------
Return-path: <CL.BOYER@UTEXAS-20>
Mail-From: CL.BOYER created at 1-Apr-83 12:42:56
Date: Friday, 1 April 1983 12:42-CST
From: CL.BOYER at UTEXAS-20
To: Woody Bledsoe <ATP.Bledsoe at UTEXAS-20>
Subject: [John McCarthy <JMC%SU-AI@USC-ECL>:]
I'd rather hear McCarthy talk on most anything than hear
anyone else talk on anything.
-------
∂04-Apr-83 1716 RINDFLEISCH@SUMEX-AIM Equipment Contract Rebudgeting
Received: from SUMEX-AIM by SU-AI with PUP; 04-Apr-83 17:15 PST
Date: Mon 4 Apr 83 17:15:10-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Equipment Contract Rebudgeting
To: Feigenbaum@SUMEX-AIM.ARPA, Ullman@SU-HNV.ARPA, JMC@SU-AI.ARPA,
CSL.JLH@SU-SCORE.ARPA, RPG@SU-AI.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA, CSD.BScott@SU-SCORE.ARPA
We (SPO) have heard from DSSW about the equipment contract with some
good news and some bad news. The good news is that they will not do
a preaward audit. The bad news is that we have to redo the budget in
terms of current (auditable) equipment quotes and university staff
benefit and indirect cost rates. DSSW claims they can let a contract
within about 45 days of receiving the new budget. This contract will
be for 3 years starting about June 1983.
I want to ask Ron Ohlander about the possibility of moving some of the
future year funding into the first year since we are already a year late
in getting started and badly need some of the equipment. Are there any
other issues we should include to minimize the separate msgs to
Ohlander?
Tom R.
-------
∂04-Apr-83 1938 LGC Opportunity for data-acquisition
In a recent bboard msg, you expressed the following opinion:
"... The probability that a central authority is needed to avoid rival
networks is low enough to take the chance."
I find this judgment very interesting as a potentially valuable real-life
datum for commonsense decision-theory. I am convinced that many good
epistemological and decision-theoretic insights can be gained from the
retro/introspective analysis of particular real-life judgements, carried out not
long after the judgments were made. So, I hope you'll be willing to try to
make such an analysis in this case. I'm well aware that introspective access
to the workings of our reasoning faculties is severely limited, but I still
think it's worth trying to glean what information we can from this source. Our
chances are best here when we focus on specific real-life judgments.
I assume that your opinion as quoted above reflects serious (though perhaps
not fully conscious) underlying qualitative judgements concerning:
1) the commonsense likelihood that harmfully competing rival networks would
develop at Stanford, provided that there was no centrally imposed universal
network. [Queries: How was degree of harm estimated here, and in what
seat-of-the-pants units? How detailed (or undetailed) an analysis of
harm-involving scenarios was involved? In what seat-of-the-pants units was
likelihood estimated? What kinds of factual information, if provided, might
change this judgment, and how?]
2) the degree of harm that would be involved in having a centrally imposed
universal network. [Queries: Similar to those for 1) How broad-brush was the
analysis here? etc.]
3) The degree of harm-discounted-by-uncertainty in case 1) as compared with
the degree of harm-discounted-by-uncertainty in case 2). [Queries: Within what
conceptual framework was the qualitative comparison of discounted harms carried
out? What quantities or other representations were actually computed and
compared? What kinds of factual information, if provided, might change this
judgment, and how?]
Does this framework of analysis seem appropriate? If not, what framework might
be better?
∂05-Apr-83 0900 BUCHANAN@SUMEX-AIM systems candidate Gael Buckley
Received: from SUMEX-AIM by SU-AI with PUP; 05-Apr-83 08:59 PST
Date: Tue 5 Apr 83 08:55:30-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: systems candidate Gael Buckley
To: search-committee: ;
She is unable to come for a visit until Friday, Apr. 29. Her resume is
on file so please take a look at it before she comes. We will be making
apptmts for her later.
bgb
-------
∂05-Apr-83 1218 WIEDERHOLD@SRI-AI Title
Received: from SUMEX-AIM by SU-AI with PUP; 05-Apr-83 12:15 PST
Received: from SRI-AI.ARPA by SUMEX-AIM.ARPA with TCP; Tue 5 Apr 83 12:05:33-PST
Date: Tue 5 Apr 83 08:23:27-PST
From: Gio <Wiederhold@SRI-AI.ARPA>
Subject: Title
To: rpg%sail@SUMEX-AIM.ARPA, jmc%sail@SUMEX-AIM.ARPA
cc: csd.betty@SU-SCORE.ARPA
What is the title of the proposal to ARPA? Since we dont have a number
I need a name to reference it in my task description.
Is
BASIC AND APPLIED RESEARCH IN COMPUTER SCIENCE too braod?
Gio
-------
∂05-Apr-83 1346 BUCHANAN@SUMEX-AIM Comtex Introduction
Received: from SUMEX-AIM by SU-AI with PUP; 05-Apr-83 13:46 PST
Date: Tue 5 Apr 83 13:43:44-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: Comtex Introduction
To: jmc@SU-AI.ARPA
John,
I have a rough draft of the introduction to the SAIL memo reprints
by Comtex. I'd very much appreciate comments on it -- it's only 4 pages,
so I hope it's not too much of a burden.
Les and I still need to pick out a few photos to accompany it. After
I get your comments, and suggestions from Les & Ed, then I'll revise it.
thanks,
bgb
-------
∂05-Apr-83 1346 RINDFLEISCH@SUMEX-AIM Stanford Equipment Contract - DSSW Rebudgeting
Received: from SUMEX-AIM by SU-AI with PUP; 05-Apr-83 13:42 PST
Date: Tue 5 Apr 83 13:41:37-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Stanford Equipment Contract - DSSW Rebudgeting
To: Ohlander%USC-ISI@SUMEX-AIM.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA, Feigenbaum@SUMEX-AIM.ARPA, Ullman@SU-HNV.ARPA,
JMC@SU-AI.ARPA, CSL.JLH@SU-SCORE.ARPA, RPG@SU-AI.ARPA
Ron, we are being asked by DSSW to refigure the budget on the equipment
modernization contract -- a fairly reasonable request in light of the
delay. In doing this though, we have a question for you. Since we are
a year late in getting started and badly need the equipment, is there
any chance of accelerating some of the funding over the 3-year term
(which will presumably start about 6/83 now)? That is, could we
rebudget some of the equipment originally planned for year 2 into year
1? As you know, the proposal as originally submitted a year ago was
to run for 3 years (4/82 - 3/85) at about $1.1M per year.
We certainly don't want to change things around at the cost of further
delay but we don't know how much flexibility you have in the way you
planned to fund this equipment. We'd appreciate your comments...
Tom R.
-------
∂05-Apr-83 1350 BUCHANAN@SUMEX-AIM G.Novak's application
Received: from SUMEX-AIM by SU-AI with PUP; 05-Apr-83 13:50 PST
Date: Tue 5 Apr 83 13:48:33-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: G.Novak's application
To: search-committee: ;
We have a formal application from Gordon Novak for the systems faculty
slot. Since he is here, we can set up interviews at your leisure.
Probably it would be a good idea to set up a special seminar too.
Elyse will coordinate interviews -- pls respond to her. I'll set up
a Friday seminar.
bgb
-------
∂05-Apr-83 1610 PB I haven't forgotten
I'm still interested in dealing with the library data base thing, but
I have to deal with an unforeseen crisis this week. I'll bug you when
I can breathe again. (Like next week).
--peter
∂05-Apr-83 2109 CL.BOYER at UTEXAS-20 Wilson's Theorem, FYI
Received: from USC-ECL by SU-AI with NCP/FTP; 5 Apr 83 21:09:47 PST
Received: from UTEXAS-20 by USC-ECL; Tue 5 Apr 83 21:08:06-PST
Date: Tuesday, 5 April 1983 23:04-CST
From: CL.BOYER at UTEXAS-20
To: JMC%SU-AI at USC-ECL
Subject: Wilson's Theorem, FYI
A student of ours, David Russinoff, just extended our proof
of Fermat's Theorem to a proof of Wilson's Theorem using our
theorem-prover. It took him a week from the time he first
used started using our program (which was also when he first
started using LISP). On the other hand, he already had a
PhD in number theory and had studied our work for quite a
while. He had to formulate about 45 intermediate lemmas.
∂06-Apr-83 1622 CSD.BSCOTT@SU-SCORE
Received: from SU-SCORE by SU-AI with PUP; 06-Apr-83 16:22 PST
Date: Wed 6 Apr 83 16:25:27-PST
From: Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA, TOB@SU-AI.ARPA, ZM@SU-AI.ARPA, DCL@SU-AI.ARPA,
Wiederhold@SUMEX-AIM.ARPA
cc: CSD.BScott@SU-SCORE.ARPA
The ARPA proposal is ready to be duplicated for mailing. Do any of you want
to see it before this is done?
Betty
-------
∂07-Apr-83 2052 WIEDERHOLD@SRI-AI Addition to ARPA joint proposal
Received: from SUMEX-AIM by SU-AI with PUP; 07-Apr-83 20:52 PST
Received: from SRI-AI.ARPA by SUMEX-AIM.ARPA with TCP; Thu 7 Apr 83 20:48:49-PST
Date: Thu 7 Apr 83 20:42:40-PST
From: Gio <Wiederhold@SRI-AI.ARPA>
Subject: Addition to ARPA joint proposal
To: csd.betty@SU-SCORE.ARPA, csd.atkinson@SU-SCORE.ARPA
cc: jmc%sail@SUMEX-AIM.ARPA, rpg%sail@SUMEX-AIM.ARPA, dek%sail@SUMEX-AIM.ARPA,
wiederhold@SUMEX-AIM.ARPA
Kahn has requested for Don Knuth to come in with a task under the proposal.
I'll get the material and integrste it into the introduction and background.
(RPG: Dick can you send me pointers to the latest versions? )
The budget addition is as follows
Don 25% all three years
100% vis fac. member
2 students
ravel 2 dom. trips /year
normal sec/computing use/other
Could you add that to the aggregate budget?
Thanks. Gio
-------
∂07-Apr-83 2103 WIEDERHOLD@SUMEX-AIM Re: Addition to ARPA joint proposal
Received: from SUMEX-AIM by SU-AI with PUP; 07-Apr-83 21:02 PST
Date: Thu 7 Apr 83 21:02:54-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: Re: Addition to ARPA joint proposal
To: csd.betty@SU-SCORE.ARPA, csd.atkinson@SU-SCORE.ARPA, jmc%sail@SUMEX-AIM.ARPA,
rpg%sail@SUMEX-AIM.ARPA, dek%sail@SUMEX-AIM.ARPA
In-Reply-To: Your message of Thu 7 Apr 83 20:53:01-PST
just found out that umbrella proposal is complete.
I am in quandary.
there maybe enough flexibility in the proposal to keep Don's request
(about 80.000/year) in it, especially since I understand that Zohar
manna's task is being trimmed.
Please advise.
I7ll be in tomorrow morning.
Gio
-------
∂07-Apr-83 2147 DEK
To: JMC%SU-AI@USC-ECL, GIO%SU-AI@USC-ECL,
kahn%USC-ISI@USC-ECL
% I've put together a description of research that may be of interest
% for ARPA funding. This would supplement my existing NSF grant, and
% it would take the place of the ONR funding that will be disappearing
% in September because my contact at ONR says that the Navy is not
% interested in supporting this sort of research at the moment.
% (He has given me an ``honorable discharge,'' however!)
% Basically, I am hoping that ARPA will pay for 25% of my AA research,
% the other 75% to be covered by NSF.
% The rest of this message is a proposal in TeX form that I hope will be
% fairly easy to incorporate into our larger proposal.
\input basic
\ctrline{\bf Research in the Analysis of Algorithms}
\vskip 20pt
Professor Knuth and his associates propose to do research of three
kinds: (1) To create efficient new computer methods for important
practical problems. (2) To create better mathematics in order to
determine how efficient such methods are. (3) To extend programming
methodology so that such algorithms can be implemented more quickly
and reliably than with current techniques. These three areas
of research mutually support each other; advances in each one
lead to advances in the other two.
More specifically, under (1) we are currently working on (a) new
algorithms for digital raster graphics, on (b) new algorithms
for the combinatorial matching problem, and on (c) new algorithms
for the ``all nearest neighbor'' problem.
(1a) Curved lines can be satisfactorily approximated by cubic spline
functions, where the coordinates of the points $(x,y)$ on the curve
are given by the formulas
$$\eqalign{ x ⊗= x↓0 + x↓1 t + x↓2 t↑2 + x↓3 t↑3 \cr
y ⊗= y↓0 + y↓1 t + y↓2 t↑2 + y↓3 t↑3 \cr}$$
as $t$ varies from 0 to 1. However, existing methods of plotting
such curves require a substantial amount of computer time. We believe
that it is possible to plot any cubic spline with a very efficient
new method that involves only addition, subtraction, and testing the
sign of a register, with all computations on integers. If our
preliminary guesses prove to be true, the new method will be ideal
for implementation in hardware, because it computes the curve
in ``real time'' in much the same way as well-known methods for
plotting circles, and because it requires only three times as
much computational machinery as is necessary for circles.
(1b) The combinatorial matching problem consists of determining
whether it is possible to arrange $2n$ elements into $n$ pairs,
subject to the restriction that all pairs must be chosen from
a given set of ``feasible'' pairs. This fundamental problem
has been shown to have numerous applications, and to be a stepping
stone to the solution of other problems for which a brute-force
search of the possibilities would take thousands of years to
complete. It is problems like these where good ideas save many
orders of magnitude of computer time. We believe that it is possible
to find a faster matching algorithm than those already known,
by developing sophisticated new data structures for representing
the manipulations inside the computer.
(1c) The ``all nearest neighbors'' problem consists of taking
a set of $n$ points in $d$-dimensional space and, for each point,
determining which other point is closest to it. We believe that
it will be possible to solve this problem with a new method
whose running time is a constant times $n \log n$; the constant
depends on the dimension $d$. The best methods currently known
for this problem require running time proportional to
$n (\log n)↑{d-1}$. Furthermore this new method should lead to improved
algorithms for other geometric optimization problems, such as
the minimum spanning tree that connects a given set of points.
Under area (2) we are currently looking at problems of random
mappings, which are fundamental to numerous applications, including
cryptanalysis. A mapping on a set can be thought of as a collection
of individuals each of whom is pointing a finger at another
individual. In the applications it is important to study questions
of the following type: Start with an individual $x↓0$ chosen
at random; let $x↓0$ point to $x↓1$, who points to $x↓2$, and so
on. Eventually there will be a cycle, i.e., $x↓{m-1}$ will point
to $x↓m$, where $x↓m$ has already been named ($x↓m=x↓k$ for some
$k$ less than $m$). The problem is to figure out how large $m$ is
likely to be. For example, if $m$ is not too large, it is possible
to ``crack'' certain codes; furthermore, this problem relates to
the efficiency of a variety of other algorithms. We are looking
at special cases of the problem that have never been solved;
in particular, we want to know the answer when half of the individuals
are pointed to by two others, while the other half are never pointed to
at all. We believe that we are on the verge of a breakthrough in
understanding this sort of problem, and that this understanding will
have a payoff in the design of new computer methods.
Finally, under area (3) we believe that our current research is likely to
have a significant effect on the way all large-scale computer software
will be written in the future. Professor Knuth has made successful
preliminary studies with a new system called WEB, which promises to
improve greatly the present state of the art in the documentation of
computer programs. We believe that this not only makes programs easier to
read and to maintain, but that it makes them more reliable and of higher
quality in the first place; yet it takes no longer, and the new system is
actually more pleasant to use than existing methods. These are obviously
strong claims, but the preliminary experiences have indeed been enormously
encouraging. Therefore we plan to greatly increase the usage of WEB in
our own software tools, and to use it exclusively in the new algorithms
that we develop, so that this project will serve as a testbed for
the new methodology. In this way we believe that the example WEB programs
we write will provide a model for a new style of programming that may
well improve the quality of all major computer programs.
The WEB system is a combination of a traditional computer programming
language with a document-formatting language. When the two are blended
together in the right way, we obtain a system that is much more powerful
than when either system is used by itself, and it leads to the
advantages claimed above. The concept is new and cannot be explained
easily in a few paragraphs, but we believe that it is exciting enough
that it merits support. During the next year or two we should be able to
prepare reports that will convince a substantial number of people outside
Stanford to being using this method.
\vfill\end
∂07-Apr-83 2142 DEK ARPA
I talked to Bob Kahn on Tuesday about the problems I have been having
to continue the AA project at its present level, since the ONR is
cancelling out. He was very encouraging and suggested that I write a
proposal, tentatively to be included with the others for which you
are principal investigator. However, if you disagree, it could be
packaged some other way. He prefers amalgamation.
I tried to reach you in New York, but I guess the message didn't
get through. Bob said there was a PI's meeting in Monterey early
next week, so I have had to hurry to get something written. (I'm
going to be out of town this whole weekend, starting Friday.)
I talked to Gio, and he said he and Dick Gabriel would be happy
to graft me in to the existing proposal.
So I'm going to send a draft of what I propose to do to him, to you,
and to Bob Kahn. That draft should be the next thing in your mail
file, if all goes well....
∂08-Apr-83 0512 OHLANDER@USC-ISI Re: Stanford Equipment Contract - DSSW Rebudgeting
Received: from SUMEX-AIM by SU-AI with PUP; 08-Apr-83 05:12 PST
Received: from USC-ISI by SUMEX-AIM.ARPA with TCP; Fri 8 Apr 83 05:10:25-PST
Date: 8 Apr 1983 0506-PST
Sender: OHLANDER at USC-ISI
Subject: Re: Stanford Equipment Contract - DSSW Rebudgeting
From: OHLANDER at USC-ISI
To: Rindfleisch at SUMEX-AIM
Cc: Feigenbaum at SUMEX-AIM, Ullman%SU-HNV at SUMEX-AIM
Cc: JMC%SU-AI at SUMEX-AIM, CSL.JLH at SU-SCORE
Cc: RPG%SU-AI at SUMEX-AIM
Message-ID: <[USC-ISI] 8-Apr-83 05:06:23.OHLANDER>
In-Reply-To: Your message of Tue 5 Apr 83 13:41:37-PST
Tom,
As soon as the contract is signed you should be getting a
large chunk of 82 and 83 money. The 84 money will follow shortly after
October. This money can be spent as rapidly as you want as long as
the equipment is that specified in the contract. The purpose of breaking
it out by year was due to the fact that we could only provide specific
amounts over a 3-year time frame. Now that it has taken so long to get
the contract in place, the money has piled up. By all means, spend the
money as soon as you can.
Ron
∂08-Apr-83 0825 WIEDERHOLD@SUMEX-AIM Re: Addition to ARPA joint proposal
Received: from SUMEX-AIM by SU-AI with PUP; 08-Apr-83 08:17 PST
Date: Fri 8 Apr 83 08:17:05-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: Re: Addition to ARPA joint proposal
To: kahn%USC-ISI@SUMEX-AIM.ARPA
cc: csd.betty@SU-SCORE.ARPA, csd.atkinson@SU-SCORE.ARPA, jmc%sail@SUMEX-AIM.ARPA,
rpg%sail@SUMEX-AIM.ARPA, dek%sail@SUMEX-AIM.ARPA
In-Reply-To: Your message of Thu 7 Apr 83 20:53:01-PST
I talked to Don Knuth this morning, and we agreed that the umbrella proposal
was broad enogh, and had a sufficient budget extension, to permit that
his proposed tasks to come in under it. This means that we can avoid revision
and only add his Vitae to the proposal.
I assume this has your concurrence.
Gio
-------
∂08-Apr-83 0847 RINDFLEISCH@SUMEX-AIM DARPA Equipment Rebudgetting
Received: from SUMEX-AIM by SU-AI with PUP; 08-Apr-83 08:47 PST
Date: Fri 8 Apr 83 08:48:31-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: DARPA Equipment Rebudgetting
To: CSL.JLH@SU-SCORE.ARPA, Ullman@SU-HNV.ARPA, RPG@SU-AI.ARPA, REG@SU-AI.ARPA,
Reid@SU-SHASTA.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA, Feigenbaum@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA,
Pattermann@SUMEX-AIM.ARPA, CSD.BScott@SU-SCORE.ARPA
Several of us met this morning to discuss the rebudgeting of the DARPA
Equipment Modernization contract as required by DSSW. We agreed to the
following guidelines:
1) We will retain the general funding breakdown between AI, CSL, and
CSD-CF as in the original proposal.
2) We will budget about $1.6M in the 1st contract year, $1.1M in the
2nd, and the remainder, $680K, in the 3rd. This is consistent with
DARPA funding commitments although I don't have their exact figures.
3) $70K originally budgeted for 2 VAX 11/750's for HPP will be
reallocated for Lisp workstations. HPP bought 2 VAX's out of its
own workstation money to enable delivery by DEC because of the equip
contract delay and needs to buy the workstations now. CSL will
budget for the remaining 8 or possibly reallocate money to buy two
complete sets of 5.
So now we need up-to-date budget estimates, backed up with vendor
quotations, for the various items in the proposal. The "piece of the
2080" in year 3 should be quoted and documented in terms of this year's
2060 prices to keep the auditors happy. Following are the assignments
for parts of the budget:
Rindfleisch: Assemble overall budget
Rindfleisch and Gabriel: Lisp workstations
Reid and Gorin: Central file server and laser printers
Gorin: Ethernet equipment, large time-shared mainframe, maintenance and
support staff.
Hennessy: SUN's, 11/750's, tape drive, disk drive, local SUN file
servers, and McSUN clusters (LOTS stations)
We will meet next Thursday (4/14) at 11:00 to go over the estimates and
fit things into the year by year breakdown.
Thanks, Tom R.
-------
∂08-Apr-83 0911 CSD.BSCOTT@SU-SCORE Re: Addition to ARPA joint proposal
Received: from SUMEX-AIM by SU-AI with PUP; 08-Apr-83 09:10 PST
Mail-from: SU-NET-TEMP host SCORE rcvd at 8-Apr-83 0851-PST
Date: Fri 8 Apr 83 08:39:53-PST
From: Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>
Subject: Re: Addition to ARPA joint proposal
To: Wiederhold@SRI-AI.ARPA, csd.betty@SU-SCORE.ARPA, csd.atkinson@SU-SCORE.ARPA
cc: jmc%sail@SUMEX-AIM.ARPA, rpg%sail@SUMEX-AIM.ARPA, dek%sail@SUMEX-AIM.ARPA,
CSD.BSCOTT@SU-SCORE.ARPA
In-Reply-To: Your message of Thu 7 Apr 83 20:53:29-PST
We can certainly add it,Gio, but I guess I have to say I am a little
annoyed at this point (if I am allowed to be annoyed). We thought the
proposal was final--SPO has already checked out the budget, and we are
ready to duplicate it for processing.
But I'll do it.
Betty
-------
∂08-Apr-83 1117 BUCHANAN@SUMEX-AIM search comm. mtg today 3:30, chm's conf. room
Received: from SUMEX-AIM by SU-AI with PUP; 08-Apr-83 11:12 PST
Date: Fri 8 Apr 83 11:10:29-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: search comm. mtg today 3:30, chm's conf. room
To: search-committee: ;
should be short, but we need to review progress to date. Time is
running out if we are going to make credible offers.
bgb
-------
∂08-Apr-83 1128 BUCHANAN@SUMEX-AIM Room Change for 3:30 mtg (sorry) -- Room 301 instead of Chm's Conf.Rm
Received: from SUMEX-AIM by SU-AI with PUP; 08-Apr-83 11:28 PST
Date: Fri 8 Apr 83 11:27:00-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: Room Change for 3:30 mtg (sorry) -- Room 301 instead of Chm's Conf.Rm
To: search-committee: ;
-------
∂08-Apr-83 1342 RPG Tasking
To: JMC@SU-AI, GIO@SU-AI, DEK@SU-AI, oliger@SU-NAVAJO, DCL@SU-AI,
FWH@SU-AI, ZM@SU-AI, TOB@SU-AI
CC: BS@SU-AI
Now that the umbrella portion of the proposal has been sent it is time to
submit the tasking portions of it. For this we need to specify tasks as
milestones in the traditional fashion. In addition a specific 3-year
budget with standard particulars must be sent. Gio and I are aiming
at sending as many of them as are ready on wednesday, April 13. If you
will not be ready by then please send them as soon as possible yourselves
to ARPA.
-rpg-
∂09-Apr-83 1542 AAAI@SRI-AI MIT Offer
Received: from SU-SCORE by SU-AI with PUP; 09-Apr-83 15:42 PST
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Sat 9 Apr 83 15:35:14-PST
Date: Sat 9 Apr 83 15:26:57-PST
From: Louis G. Robinson <AAAI%SRI-AI@SU-SCORE.ARPA>
Subject: MIT Offer
To: Nilsson%SRI-AI@SU-SCORE.ARPA, aaai%SRI-AI@SU-SCORE.ARPA
cc: AAAI-Executive-Committee: ;
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Nils,
I am circulating the attached copy of a letter from MIT
Press to the AAAI Executive Council: an offer from them
to present the $1,000 Publisher's Prize at next year's
AAAI conference. Looks good to me. I imagine folks
will get back to me if there are any objections; otherwise
I should be getting back to Satlow with our response within
a week or so.
Lou
THE MIT PRESS
Massachusetts Institute of Technology
28 Carleton Street
Cambridge, Massachusetts 02142
617-253-1623, 253-1605
5 April 1983
Louis G. Robinson
American Association for
Artificial Intelligence
445 Burgess Drive
Menlo Park, CA 94025
Dear Lou:
I noted on your announcement for AAAI 83 that a $1,000 Publishers
Prize will be awarded for the best paper. We were asked to sponsor
the prize last year, along with other publishers, but we were unable
to make a committment. Obviously you found a sponsor for this
worthy project. We would like to sponsor the 84 prize and would
prefer to give you a firm committment at this early stage. Would
this be possible on your part? Please let me know.
Sincerely,
Frank Satlow
Executive Editor
FS/mh
cc: Marvin Minsky
Frank Urbanowski
-------
∂11-Apr-83 0748 CSD.BSCOTT@SU-SCORE Re: Tasking
Received: from SU-SCORE by SU-AI with PUP; 11-Apr-83 07:41 PST
Date: Mon 11 Apr 83 07:39:23-PST
From: Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>
Subject: Re: Tasking
To: RPG@SU-AI.ARPA, JMC@SU-AI.ARPA, GIO@SU-AI.ARPA, DEK@SU-AI.ARPA,
oliger@SU-NAVAJO.ARPA, DCL@SU-AI.ARPA, FWH@SU-AI.ARPA, ZM@SU-AI.ARPA,
TOB@SU-AI.ARPA
cc: BS@SU-AI.ARPA, CSD.BSCOTT@SU-SCORE.ARPA
In-Reply-To: Your message of Fri 8 Apr 83 13:42:00-PST
Note that all tasking portions of the ARPA proposals are to be sent as
a regular proposal, i.e., submitted to me for processing through Stanford.
Betty
-------
∂11-Apr-83 1237 CLT on TTY75 (at TV-143) 1237
hi its semi sunny here
∂11-Apr-83 1633 RINDFLEISCH@SUMEX-AIM Re: DARPA Equipment Rebudgetting
Received: from SUMEX-AIM by SU-AI with PUP; 11-Apr-83 16:33 PST
Date: Fri 8 Apr 83 22:51:46-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Re: DARPA Equipment Rebudgetting
To: CSL.JLH@SU-SCORE.ARPA, Ullman@SU-HNV.ARPA, RPG@SU-AI.ARPA, REG@SU-AI.ARPA,
Reid@SU-SHASTA.ARPA
cc: Feigenbaum@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA, Pattermann@SUMEX-AIM.ARPA,
CSD.BScott@SU-SCORE.ARPA, Rindfleisch@SUMEX-AIM.ARPA
In-Reply-To: Your message of Fri 8 Apr 83 08:48:29-PST
In my last msg I left the "Systems Group 11/780" off of John Hennessy's
list of action items. Also Ron Ohlander called this afternoon and needs
information about the difference between prices quoted to us with
"university" and other discounts versus GSA contract prices (or list
prices for vendors w/o GSA contracts). He has to convince the
bureaucracy that Stanford should buy all this stuff (we get it cheaper)
as opposed to having the govt furnish it. So, when you get quotes on
items, ask for GSA schedule prices as well.
Tom R.
-------
∂11-Apr-83 2359 RV AI Qual
I'm interested in finding out more about the AI qual (remember, I
talked to you a few weeks ago?). Have you and the committee decided
when it might be, its format, and a reading list for it? The first and
last items would be especially useful to me.
Thanks,
R.V.
∂12-Apr-83 1645 NAN@SU-SCORE pmessage
Received: from SU-SCORE by SU-AI with PUP; 12-Apr-83 16:40 PST
Date: Tue 12 Apr 83 16:35:31-PST
From: Nancy Dorio <Nan@SU-SCORE.ARPA>
Subject: pmessage
To: jmc@SU-AI.ARPA
Stanford-Phone: (415) 497-2273
Stuart Schwartzstein called. Please return the call at 617 492-2116
nan
-------
∂12-Apr-83 2005 WIEDERHOLD@SUMEX-AIM Re: ARPA Umbrella
Received: from SUMEX-AIM by SU-AI with PUP; 12-Apr-83 20:05 PST
Date: Tue 12 Apr 83 18:58:26-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: Re: ARPA Umbrella
To: CSD.BSCOTT@SU-SCORE.ARPA
cc: jmc@SU-AI.ARPA
In-Reply-To: Your message of Tue 12 Apr 83 09:00:05-PST
There will be no changes now.
Any specifics can go into the individula task proposals. I discussed that with
Luckham on friday. I did not realize that he had in effect put a hold on the
proposal. Sorry . Gio
-------
∂12-Apr-83 2005 WIEDERHOLD@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 12-Apr-83 20:05 PST
Date: Tue 12 Apr 83 18:52:43-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
To: JMC@SU-AI.ARPA
cc: DEK@SU-AI.ARPA, DKE@SU-AI.ARPA, RPG@SU-AI.ARPA, GIO@SU-AI.ARPA
In-Reply-To: Your message of Mon 11 Apr 83 12:29:00-PST
he is, with minimal rewrite.
Don's name was mentioned in the Golub section, and our tasks were broad
enough. Umbrella proposal is leaving the university. Gio
-------
∂12-Apr-83 2016 CSD.RUSSELL@SU-SCORE ai qual
Received: from SU-SCORE by SU-AI with PUP; 12-Apr-83 20:16 PST
Date: Tue 12 Apr 83 17:16:22-PST
From: Stuart J. Russell <CSD.RUSSELL@SU-SCORE.ARPA>
Subject: ai qual
To: jmc@SU-AI.ARPA
Dear Professor McCarthy,
I am interested in the possibility of taking the AI qual this
quarter. I must apologise for the late response, I am not a frequent
BBOARD reader.
Stuart Russell
-------
∂12-Apr-83 2024 RPG
To: GIO@SU-AI, JMC@SU-AI, DKE@SU-AI
∂12-Apr-83 2005 WIEDERHOLD@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 12-Apr-83 20:05 PST
Date: Tue 12 Apr 83 18:52:43-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
To: JMC@SU-AI.ARPA
cc: DEK@SU-AI.ARPA, DKE@SU-AI.ARPA, RPG@SU-AI.ARPA, GIO@SU-AI.ARPA
In-Reply-To: Your message of Mon 11 Apr 83 12:29:00-PST
he is, with minimal rewrite.
Don's name was mentioned in the Golub section, and our tasks were broad
enough. Umbrella proposal is leaving the university. Gio
-------
Since there were some other delays with the umbrella, I added
Don's section to it. I changed the title of Golub's part to:
Research in the Analysis of Algorithms and Architectures
and added Don's section as:
Research in the Analysis of Algorithms.
I think Knuth deserves that title, and Golub's section seemed to
be restricted to the combination of topics. The introduction
combines Golub's and Knuth's areas.
Since Luckham still isn't happy with it, I can change things up
until 11am tomorrow morning.
-rpg-
∂12-Apr-83 2113 WIEDERHOLD@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 12-Apr-83 21:13 PST
Date: Tue 12 Apr 83 21:08:21-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
To: RPG@SU-AI.ARPA
cc: GIO@SU-AI.ARPA, JMC@SU-AI.ARPA, DKE@SU-AI.ARPA
In-Reply-To: Your message of Tue 12 Apr 83 20:24:00-PST
Thanks. I was not aware of your update.
I do not understand luckhams problems withnthe umbrella now.
Gio
-------
∂14-Apr-83 1000 BUCHANAN@SUMEX-AIM next comm. mtg
Received: from SUMEX-AIM by SU-AI with PUP; 14-Apr-83 10:00 PST
Date: Thu 14 Apr 83 09:33:50-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: next comm. mtg
To: search-committee: ;
Let's postpone the search comm. mtg tentatively scheduled for this Friday
and conduct our business by message. Katevenis from Berkeley looks
like our best bet and we should make an offer. If another candidate
looks as good or better, we can almost certainly use an EE slot (and,
I presume, make a courtesy apptmt in CS). We have the following
candidates still expecting interviews: Clement (this week),
Novak (to be arranged), and Buckley (Apr. 29). Do you know of more?
bgb
-------
∂14-Apr-83 1002 BUCHANAN@SUMEX-AIM Michael Carey
Received: from SUMEX-AIM by SU-AI with PUP; 14-Apr-83 10:01 PST
Date: Thu 14 Apr 83 09:45:34-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: Michael Carey
To: search-committee: ;
Carey is scheduled to give a seminar on the 22nd. However, he is about
to accept another job offer. If he is number one on anyone's list, then
we should try (today!) to get him to hold off. I am still inclined to
go with Katevenis. Any dissent?
bgb
-------
∂14-Apr-83 1414 CSD.BSCOTT@SU-SCORE Re: honor yes, advantage no
Received: from SU-SCORE by SU-AI with PUP; 14-Apr-83 14:13 PST
Date: Thu 14 Apr 83 12:46:00-PST
From: Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>
Subject: Re: honor yes, advantage no
To: JMC@SU-AI.ARPA
cc: CSD.BSCOTT@SU-SCORE.ARPA
In-Reply-To: Your message of Wed 13 Apr 83 15:54:00-PST
Thank you, John. -- Betty
-------
∂14-Apr-83 1448 NAN@SU-SCORE pmessage
Received: from SU-SCORE by SU-AI with PUP; 14-Apr-83 14:47 PST
Date: Thu 14 Apr 83 14:41:55-PST
From: Nancy Dorio <Nan@SU-SCORE.ARPA>
Subject: pmessage
To: jmc@SU-AI.ARPA
Stanford-Phone: (415) 497-2273
Jack Minker called. Please return the call at 301 229-6480. he will be there
until 7:15 (this is his home number). I am to tell you to call back operator
#7.
Nan
-------
∂14-Apr-83 1650 AAAI@SRI-AI.ARPA AAAI Membership
Received: from SU-DSN by SU-AI with PUP; 14-Apr-83 16:50 PST
Received: From SRI-AI by SU-DSN.ARPA; Thu Apr 14 16:47:29 1983
Date: Thu 14 Apr 83 16:41:28-PST
From: Louis G. Robinson <AAAI%SRI-AI@SU-DSN.ARPA>
Subject: AAAI Membership
To: AAAI-Executive-Committee: ;
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Folks,
As of this afternoon, membership in AAAI is now at
2,302
Lou
-------
∂15-Apr-83 1307 DFH
If you don't object, I plan to leave early,
at 4 pm today. --Diana
∂15-Apr-83 1415 CSD.ULLMAN@SU-SCORE hrumpf
Received: from SU-SCORE by SU-AI with PUP; 15-Apr-83 14:15 PST
Date: Fri 15 Apr 83 14:17:06-PST
From: Jeffrey D. Ullman <CSD.ULLMAN@SU-SCORE.ARPA>
Subject: hrumpf
To: jmc@SU-AI.ARPA
I think I ought to tell you that I object to the outburst at the meeting
yesterday morning. I am working my butt off at an unrewarding job
trying to see, among other things, that the department doesn't go bankrupt.
It is work that as a very senior member of this department, might most
naturally have fallen to you. I don't pretend to know all about all
these issues, and I might even be completely wrong about disk space.
However, I think I deserve minimal courtesy when I am trying to
fulfill my responsibilities, even when I don;t agree with you.
-------
∂15-Apr-83 1455 RINDFLEISCH@SUMEX-AIM Draft Equipment Budget
Received: from SUMEX-AIM by SU-AI with PUP; 15-Apr-83 14:54 PST
Date: Fri 15 Apr 83 14:55:25-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Draft Equipment Budget
To: Ullman@SU-HNV.ARPA, JMC@SU-AI.ARPA, CSL.JLH@SU-SCORE.ARPA,
Feigenbaum@SUMEX-AIM.ARPA, RPG@SU-AI.ARPA, REG@SU-AI.ARPA, Reid@SU-SHASTA.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA
Following is the budget sketch we discussed at our meeting Thursday
morning. It has lots of things that need refining but is intended as
a cut at a year by year breakdown.
Year 1 Year 2 Year 3
Lisp workstations
HPP $ 430 $ 180 $
Formal Reasoning 210 150
Central File Server 185 50
CSL 11/780 274
Large time-shared 2060 450
11/750 workstations (6) 303
SUN workstations (30) 195 130 64
Local net file servers (3) 27 54
SUN terminal clusters 100
Laser printers (4) 60 60
Ethernet TIPS and gateways 55 55 55
Maintenance 45
CSD-CF Staff (inc ben) 40.0 43.2 46.7 8% inflation
Indirect Costs 58.6 29.8 32.2
------- ------- -------
$1708.6 $1026.0 $ 647.9
Total $3382.5 (needs to come down to $3379.7)
The action items are:
1) Ralph: large time-shared mainframe, central file server, refine SUN
clusters, Ethernet TIPs and gateways, maintenance, and staff.
2) Brian: Laser printers
3) Iterate on yearly breakdown and check with Adams/Ohlander on funding
constraints.
4) Assemble paper quotations on estimates.
Tom R.
-------
∂15-Apr-83 2053 WIEDERHOLD@SUMEX-AIM Re: Proposal
Received: from SUMEX-AIM by SU-AI with PUP; 15-Apr-83 20:53 PST
Date: Fri 15 Apr 83 20:53:50-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: Re: Proposal
To: RPG@SU-AI.ARPA, dcl@SU-AI.ARPA, jmc@SU-AI.ARPA, zm@SU-AI.ARPA, tob@SU-AI.ARPA,
dek@SU-AI.ARPA
cc: GIO@SU-AI.ARPA, BS@SU-AI.ARPA
In-Reply-To: Your message of Thu 14 Apr 83 00:04:00-PST
I left the (relatively small) task text with for betty tonight.
I suggest that any completed task proposals be bundled, but I dont want
to delay my task proposal more than a day, since we are running behind
schedule now. Gio
-------
∂16-Apr-83 1111 CLT boris
at the Kennedy Center Washington DC on April 20th
∂16-Apr-83 1335 CSD.BSCOTT@SU-SCORE Boeing
Received: from SU-SCORE by SU-AI with PUP; 16-Apr-83 13:35 PST
Date: Sat 16 Apr 83 13:36:11-PST
From: Betty Scott <CSD.BSCOTT@SU-SCORE.ARPA>
Subject: Boeing
To: JMC@SU-AI.ARPA
cc: CSD.BScott@SU-SCORE.ARPA
John, the file I have on the Boeing agreement is still pending. What did
you decide to do about this?
Betty
-------
∂17-Apr-83 0242 Lowell Wood <LLW at S1-A> Monday Meetings
Received: from S1-A by SU-AI with NCP/FTP; 17 Apr 83 02:39:28 PST; for: jmc
Date: 17 Apr 1983 0241-PST
From: Lowell Wood <LLW at S1-A>
Subject: Monday Meetings
To: jmc at SU-AI
CC: LLW at S1-A, rpg at SU-AI
∂16-Apr-83 1543 John McCarthy <JMC@SU-AI> visitors on Monday
Received: from SU-AI by S1-A with NCP/FTP; 16 Apr 83 15:43:12 PST
Date: 16 Apr 83 1540 PST
From: John McCarthy <JMC@SU-AI>
Subject: visitors on Monday
To: llw@S1-A
Joe Weening and Mark Todorovich are coming with me and Dick Gabriel on Monday.
Tom McWilliams will take care of Mark who is interested in S-1 class
computers and is a senior at Caltech who might be a potential employee.
I have suggested that Jeff Rubin talk to Joe about system performance and
its measurement. Dick Gabriel and I will be glad to talk about LISP
multi-processing if you are available, otherwise we'll talk to each other
about it.
[John: I'll look forward to seeing all of you. I mentioned to Dick that
I might not be around on Monday, due to being summoned out-of-town. It
now appears that I've evaded these summons. Lowell]
∂17-Apr-83 2238 JACK MINKER <minker.umcp-cs@UDel-TCP> Lerner's 70th Birthday
Received: from UDEL-TCP by SU-AI with NCP/FTP; 17 Apr 83 22:30:20 PST; for: JMC
Date: 14 Apr 83 17:00:57 EST (Thu)
From: JACK MINKER <minker.umcp-cs@UDel-TCP>
Subject: Lerner's 70th Birthday
To: JMC at Su-Ai
Via: UMCP-CS; 14 Apr 83 23:30-EST
Dear John,
Dorothy Hirsch of the Committee of Concerned
Scientists was in Israel and met with Sonya Lerner,
daughter of Alexander Lerner. She told Dorothy that
September 6 of this year is Lerner's 70th birthday.
It would be nice if we could do something to honor
him. At the AAAI meeting we could hold a special
session in Honor of Lerner where we would invite say you,
Simon, Newell and Minsky to deliver scientific talks
in his honor. To do this would need the approval of the
AAAI executive board. If you agree with this idea,
I could wrie to the executive board, or you could to
make this request. Morris Pripstein spoke to Lou
Robinson about this and Lou would be pleased to
cooperate if the executive board gave the approval.
It would also be nice if we could do something
like this at IJCAI. This is more difficult and might
involve a special session outside of the conference
as we did in Boston.
I would appreciate hearing from you on this
matter at your earliest convenience. Since the AAAI
meeting is still not locked in concrete, we can
possibly get a time slot if the executive board
agrees.
You can reach me by phone at:
(301) 229-6480 Home
or
(301) 454-4251 Office
I will be home this weekend so that if you have
some free time we can talk.
Best regards,
Jack Minker
∂17-Apr-83 2249 LGC Data Opportunity Reminder
This is just to jog your memory concerning a msg I sent you while you were
away, with the header:
∂04-Apr-83 1938 LGC Opportunity for data-acquisition .
You may not have had the time to consider it seriously when you first
saw it, and it might be worth taking a second look to see if you can
provide us with some useful data. If you need another copy of the msg,
I can provide one. -- Lew
∂18-Apr-83 1145 DFH Govt. publications
Soviet Military Power is in stock in SF office
and costs $6.50. The SF office is out of the
Statistical Abstract now, but should have it in
2 weeks--they will not take advance orders, so
will check again then. --Diana
∂18-Apr-83 1308 ALS
John
Could you supply me with the exact date of the early Dartmouth conference
and a complete list of the pasticipants? I intend to say something about it
in a talk that I am giving at the IJCAI meeting this summer.
The Dartmouth Summer Research Project on Artificial Intelligence occupied
the summer of 1956. However, only a few people were there the whole
time - perhaps Minsky and me, perhaps only me. Some came for only a few
days. I can't imagine any way
to get a complete list. Here is a partial list, and you might try to supplement
it by asking Minsky.
John McCarthy
Marvin Minsky
Oliver Selfridge
Nathaniel Rochester
Claude Shannon
Raymond Solomonoff
Julian Bigelow
Arthur Samuel
Alex Bernstein
Allen Newell
Herbert Simon
∂18-Apr-83 1744 KAHN at USC-ISI Re: dinner on Wednesday?
Received: from USC-ECL by SU-AI with NCP/FTP; 18 Apr 83 17:44:15 PST
Received: from USC-ISI by USC-ECL; Mon 18 Apr 83 17:42:54-PST
Date: 18 Apr 1983 1742-PST
Sender: KAHN at USC-ISI
Subject: Re: dinner on Wednesday?
From: KAHN at USC-ISI
To: JMC%SU-AI at USC-ECL
Cc: kahn%USC-ISI at USC-ECL
Message-ID: <[USC-ISI]18-Apr-83 17:42:35.KAHN>
In-Reply-To: Your message of 18 Apr 83 0020 PST
John,
lets plan on dinner wed. how about meeting here at
arpa about 6pm?
Do you think it would be worth inviting ron ohlander or
do you prefer it if just the two of us get together then?
bob
I would be glad to have Ron too.
∂19-Apr-83 1016 BUCHANAN@SUMEX-AIM Gael Buckley's papers
Received: from SUMEX-AIM by SU-AI with PUP; 19-Apr-83 10:12 PST
Date: Tue 19 Apr 83 09:21:58-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: Gael Buckley's papers
To: search-committee: ;
I have put the papers we requested from Buckley in her file along with
a brief explanation of who was responsible for which parts of the work,
which she sent.
Please mark your calendars for her visit Friday, May 29. She wants
meetings arranged with "established graduate students and recently
hired junior faculty". Eleyse will arrange apptmts.
bgb
-------
∂19-Apr-83 1441 ALS
Do I remember correctly that Church and Kleene were both in Dartmouth for
part of the time? I am reasonably sure that I heard one of them speak but
I have the two slightly confused in my mind.
Herb Gelernter was there for sure and I have a feeling that Selfridge was
also there.
You are right about Gelernter and Selfridge, but I am pretty sure
that neither Church nor Kleene was there. I my state of mind at that
time, I wouldn't have invited either.
∂19-Apr-83 1627 DFH
To: JMC, JJW, YOM, RPG, CG, JK, CLT
Is anyone interested in fact-checking a 45-page
manuscript of an AI book for children for $150?
Or do you know of someone who might be? This is
for Franklin-Watts Publishing Co. If so, let me
know. -- Diana
∂19-Apr-83 1916 KAHN at USC-ISI
Received: from USC-ECL by SU-AI with NCP/FTP; 19 Apr 83 19:16:38 PST
Received: from USC-ISI by USC-ECL; Tue 19 Apr 83 17:13:40-PST
Date: 19 Apr 1983 1711-PST
Sender: KAHN at USC-ISI
From: KAHN at USC-ISI
To: JMC%SU-AI at USC-ECL
Cc: kahn%USC-ISI at USC-ECL
Message-ID: <[USC-ISI]19-Apr-83 17:11:35.KAHN>
In-Reply-To: Your message of 18 Apr 83 2245 PST
WEll, i invited him, but hes got a conflict
although he would like to be there.
it'll have to be another time for him. see you
tomorrow about 6 at arpa.
bob
∂19-Apr-83 2015 KAHN@USC-ISI Re: Addition to ARPA joint proposal
Received: from SUMEX-AIM by SU-AI with PUP; 19-Apr-83 20:15 PST
Received: from USC-ISI by SUMEX-AIM.ARPA with TCP; Tue 19 Apr 83 20:07:04-PST
Date: 19 Apr 1983 1828-PST
Sender: KAHN at USC-ISI
Subject: Re: Addition to ARPA joint proposal
From: KAHN at USC-ISI
To: WIEDERHOLD at SUMEX-AIM
Cc: kahn at USC-ISI, csd.betty at SU-SCORE
Cc: csd.atkinson at SU-SCORE, jmc%sail at SUMEX-AIM
Cc: rpg%sail at SUMEX-AIM, dek%sail at SUMEX-AIM
Message-ID: <[USC-ISI]19-Apr-83 18:28:11.KAHN>
In-Reply-To: Your message of Fri 8 Apr 83 08:17:05-PST
gio,
ive just gotten back from travel and havent yet read over
dons note to me on the proposal, but i wondered what you
meant by the umbrella proposal had a sufficient budget
extension? does that mean don doesnt want additional
fudns and that you will provide it? or what?
would appreciate a clarification.
thanks
bob
∂19-Apr-83 2342 RPG
∂12-Apr-83 1437 JMC via Ethernet SU-SCORE Common Lisp
There was a discussion at the PI's meeting about it that went
rather well. There was some sentiment, initiated by Feigenbaum
citing Gordon Bell, that it would soon be time
to discuss ANSIizing Common Lisp. It was agreed that this was
probably premature and anyway depended on the Common Lisp group.
Ron Ohlander hoped the group would meet again in the not too
distant future, and would like to know if there is any present
plan to meet. Is there? There is favorable sentiment towards
Common Lisp even from people whose present use emphasizes
Interlisp. The only doubts about DARPA's intention to
emphasize Common Lisp were expressed by Tony Hearn. There was
general agreement with DARPA's present intention to
support directly financially only the 68000 implementation.
The blue and yellow pages ideas received nods but probably not
very widespread understanding. The discussion was regarded as
successful by all.
We (the Common Lisp Group) discussed ANSIizing it and decided it was the
wrong time.
∂20-Apr-83 0620 BUCHANAN@SUMEX-AIM correction: Buckley is coming Apr.29 (not May 29)
Received: from SUMEX-AIM by SU-AI with PUP; 20-Apr-83 06:20 PST
Date: Wed 20 Apr 83 06:13:25-PST
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: correction: Buckley is coming Apr.29 (not May 29)
To: search-committee: ;
-------
∂20-Apr-83 1238 YOM
Prof. McCarthy ,
I wanted to discuss the summer and related issues with you, but couldn't
catch you before you left. I will be away next week, going to STOC (Symp.
on the Theory Of Computing) held in Boston. In general, apart from a
conference on the east coast and a trip home, I intend to be here most of
the summer. I'll see you after I return, and discuss these matters in more
detail.
Yoram.
∂21-Apr-83 1022 WIEDERHOLD@SUMEX-AIM Re: Addition to ARPA joint proposal
Received: from SUMEX-AIM by SU-AI with PUP; 21-Apr-83 10:20 PST
Date: Wed 20 Apr 83 17:40:27-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: Re: Addition to ARPA joint proposal
To: KAHN%USC-ISI@SUMEX-AIM.ARPA
cc: csd.betty@SU-SCORE.ARPA, csd.atkinson@SU-SCORE.ARPA, jmc%sail@SUMEX-AIM.ARPA,
rpg%sail@SUMEX-AIM.ARPA, dek%sail@SUMEX-AIM.ARPA, dek@SU-AI.ARPA
In-Reply-To: Your message of Tue 19 Apr 83 18:28:00-PST
WE were advised to keep rthe umbrella budget request generous to allow
for extensions to tasks which may be defined and approved at a later time.
This budget, approved by the university and on its way to you, is hence
generous, although certainly not unlimited in that sense.
In one area it is especially generous. Since it still includes all of
Zohar Manna's desires, while his actual budget for his task was substantially
cut back, there is an expectation that our tasks will be at least $100000
less this year than the umbrella budget provides for.
It is this cushion that appears to be adequate for Don's task.
Don's task will still, of course, first have to be submitted and approved
by you. And when approved, it will require allocation of real money
to the task within the umbrella.
We are all glad that don is joining us, and hope that there will also
been intellectual interaction.
Gio
-------
∂21-Apr-83 1041 WIEDERHOLD@SUMEX-AIM Re: Addition to ARPA joint proposal
Received: from SUMEX-AIM by SU-AI with PUP; 21-Apr-83 10:40 PST
Date: Wed 20 Apr 83 23:32:58-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: Re: Addition to ARPA joint proposal
To: KAHN%USC-ISI@SUMEX-AIM.ARPA
cc: csd.betty@SU-SCORE.ARPA, csd.atkinson@SU-SCORE.ARPA, jmc%sail@SUMEX-AIM.ARPA,
rpg%sail@SUMEX-AIM.ARPA, dek%sail@SUMEX-AIM.ARPA, dek@SU-AI.ARPA
In-Reply-To: Your message of Tue 19 Apr 83 18:28:00-PST
( this is a second mailing, I have some evidence that mail composed
earlier today was lost. Please excuse the duplication if any. Gio)
20-Apr-83 17:41:09-PST,1038;000000000001
Date: Wed 20 Apr 83 17:41:08-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM>
Subject: resp.
WE were advised to keep rthe umbrella budget request generous to allow
for extensions to tasks which may be defined and approved at a later time.
This budget, approved by the university and on its way to you, is hence
generous, although certainly not unlimited in that sense.
In one area it is especially generous. Since it still includes all of
Zohar Manna's desires, while his actual budget for his task was substantially
cut back, there is an expectation that our tasks will be at least $100000
less this year than the umbrella budget provides for.
It is this cushion that appears to be adequate for Don's task.
Don's task will still, of course, first have to be submitted and approved
by you. And when approved, it will require allocation of real money
to the task within the umbrella.
We are all glad that don is joining us, and hope that there will also
been intellectual interaction.
Gio
(CC to DEK@sail)
-------
-------
∂22-Apr-83 1003 RINDFLEISCH@SUMEX-AIM [OHLANDER at USC-ISI: Re: Equipment Contract Budget]
Received: from SUMEX-AIM by SU-AI with PUP; 22-Apr-83 09:59 PST
Date: Fri 22 Apr 83 08:29:01-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: [OHLANDER at USC-ISI: Re: Equipment Contract Budget]
To: Ullman@SU-HNV.ARPA, Feigenbaum@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA,
CSL.JLH@SU-SCORE.ARPA, RPG@SU-AI.ARPA, REG@SU-AI.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA, Pattermann@SUMEX-AIM.ARPA
I would be quite unhappy about having to spend all the money at once, because
I think we don't have enough experience with the work station philosophy
and with the particular proposed work stations - whether LISP machines,
SUNs or VAXes. Also neither the 2080 nor the S-1 are ready to be
bought. Also we don't have enough experience with expanded file
size to support or refute my contention that we will require very
much larger file sizes than presently contemplated.
Ohlander's latest about the spreadout of the money seems ok to me.
Please forward this to other people concerned, because your list of
recipients doesn't work from SAIL.
The following is more info from Ohlander about DARPA money flow for the
equipment contract. Do any of you have additional qualms about spending
everything in one shot beyond what I expressed below?
Tom R.
---------------
1) 22-Apr OHLANDER at USC-ISI Re: Equipment Contract Budget
2) 22-Apr To: OHLANDER@USC-ISI Re: Equipment Contract Budget
Message 1 -- ************************
Return-Path: OHLANDER@USC-ISI
Received: from USC-ISI by SUMEX-AIM.ARPA with TCP; Fri 22 Apr 83 06:33:15-PST
Date: 22 Apr 1983 0632-PST
Sender: OHLANDER at USC-ISI
Subject: Re: Equipment Contract Budget
From: OHLANDER at USC-ISI
To: Rindfleisch at SUMEX-AIM
Cc: Adams at USC-ISI, Ullman%SU-HNV at SUMEX-AIM
Message-ID: <[USC-ISI]22-Apr-83 06:32:22.OHLANDER>
In-Reply-To: Your message of Wed 20 Apr 83 11:54:00-PST
Tom,
The loading is not too large. In fact my conversations with
DSSW indicate that they would prefer to negotiate a one-year contract
for the purchase of the equipment and delete the maintenace requirement.
If that happens, you can spend all of the money as sonn as you get it.
There are two years of money put out already ($1.85M) and the rest
would come in October of this year. The maintenance costs would be added on
to one of the research contracts and the equipment contract would be
decreased by that amount.
Ron
Message 2 -- ************************
Date: Fri 22 Apr 83 08:22:13-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Re: Equipment Contract Budget
To: OHLANDER@USC-ISI
cc: Adams@USC-ISI, Ullman%SU-HNV@SUMEX-AIM, Rindfleisch
In-Reply-To: Your message of Fri 22 Apr 83 06:32:00-PST
Hmmmm... Thanks for the information, Ron. The only reservation I have
right off about allocating all the money in one year is the timing of
hardware futures. We want to get some hands-on experience with the new
workstations (3600's, Dandelions, and Dorados) in-house for evaluation
before we decide on the final mix to be purchased. We can probably get
some 3600's and Dandelions by late this summer but I'm not sure how soon
we can get a Dorado out of the Xerox toy shop. Also there is the matter
of the "2080" announcement timing. We will get back to you with our
preferred strategy.
Regards, Tom R.
-------
-------
∂22-Apr-83 1332 ADMIN.LIBRARY@SCORE Lisp-Anwendungsgebiete, Grundbegriffe, Geschichte
Received: from SU-SCORE by SU-AI with PUP; 22-Apr-83 13:18 PST
Date: Fri 22 Apr 83 13:17:28-PST
From: C.S./Math Library <ADMIN.LIBRARY@SU-SCORE.ARPA>
Subject: Lisp-Anwendungsgebiete, Grundbegriffe, Geschichte
To: mccarthy@SU-SCORE.ARPA
Lisp...by Herbert Stoyan has been received in the Math/CS Library. Right now
it is in the office but it should be on the new books shelf next week.
H. Llull
-------
Thanks for the information about Stoyan's book. I assume that this is
the Dresden published edition, and it's interesting that the East Germans
continue to make this available. I have a copy Stoyan sent me.
∂22-Apr-83 1335 MRC@SCORE Re: TIP phone numbers
Received: from SU-SCORE by SU-AI with PUP; 22-Apr-83 13:35 PST
Date: Fri 22 Apr 83 13:38:23-PST
From: Mark Crispin <MRC@SU-SCORE.ARPA>
Subject: Re: TIP phone numbers
To: JMC@SAIL
Postal-Address: 725 Mariposa Ave. #103; Mountain View, CA 94041
Phone: (415) 497-1407 (Stanford); (415) 968-1052 (residence)
In-Reply-To: Your message of Fri 22 Apr 83 12:16:00-PST
There is no TIP in NYC, although NYU has an ARPA node there. The way
to get TIP, etc. access is to find out who the liaison is for that
particular TIP/whatever. Then you call that person and talk him/her
into giving you the number. For example, the liaison for the SU-TIP
is Ralph Gorin (actually the new name for TIP is TAC...just to confuse
people).
I recommend contacting the NIC to find out what nodes are in NYC and
who the contacts are. NIC@NIC ought to give you this information.
-- Mark --
-------
∂22-Apr-83 1339 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract Budget
Received: from SUMEX-AIM by SU-AI with PUP; 22-Apr-83 13:39 PST
Return-Path: OHLANDER@USC-ISI
Received: from USC-ISI by SUMEX-AIM.ARPA with TCP; Fri 22 Apr 83 11:45:09-PST
Date: 22 Apr 1983 1142-PST
Sender: OHLANDER at USC-ISI
Subject: Re: Equipment Contract Budget
From: OHLANDER at USC-ISI
To: Rindfleisch at SUMEX-AIM
Cc: Adams at USC-ISI, Ullman%SU-HNV at SUMEX-AIM
Message-ID: <[USC-ISI]22-Apr-83 11:42:14.OHLANDER>
In-Reply-To: Your message of Fri 22 Apr 83 08:22:14-PST
ReSent-date: Fri 22 Apr 83 13:39:57-PST
ReSent-from: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
ReSent-to: Feigenbaum@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA, CSL.JLH@SU-SCORE.ARPA,
RPG@SU-AI.ARPA, REG@SU-AI.ARPA
Tom,
Disregard my previous information. The contract will have
to run at least two years. There is already $1.615M positioned
at DSSW right now for forwarding to Stanford as soon as the
contract is signed. An additional $950K will be provided in
October of this year. The final $814K will come in October of
1984. You should make your plans accordingly. Maintenaance
money can be provided for two years. Does this pretty much meet
with your expectations.
Ron
∂22-Apr-83 1356 RPG Trip(s)
To: JMC@SU-AI, llw@S1-A, tm@S1-A
I have had a paper accepted at IJCAI on S-1 Lisp.
Since I have never been to Europe, I'd like to go there to
deliver it. Also, I would like to attend AAAI in Washington,
which is 2 weeks after IJCAI. I understand it is very difficult to
get foreign travel from LLL, and it is probably pretty hard for Stanford
too. There are several possibilities. LLL can pay round-trip day
coach to NY, and Stanford could provide the remainder. Since AAAI
is on the way back, and since I am booking charters now, I think the
total cost would not be too high. Another possibility is to consider
the two conferences to be one in NY and one in DC with a roundtrip in
both cases. This covers the circle route easily too.
Any ideas? IJCAI is in West Germany, August 8-12. AAAI is
August 22-26.
-rpg-
I'll try to get DARPA approval for the foreign travel. I don't think
it's impossible.
∂22-Apr-83 1459 AAAI%SRI-AI.ARPA@SCORE [Louis Robinson <AAAI-OFFICE@SUMEX-AIM.ARPA>: [Michael Genesereth <GENESERETH@SUMEX-AIM.ARPA>: request]]
Received: from SU-SCORE by SU-AI with PUP; 22-Apr-83 14:55 PST
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Fri 22 Apr 83 14:44:46-PST
Date: Fri 22 Apr 83 14:31:21-PST
From: Louis G. Robinson <AAAI%SRI-AI@SU-SCORE.ARPA>
Subject: [Louis Robinson <AAAI-OFFICE@SUMEX-AIM.ARPA>: [Michael Genesereth <GENESERETH@SUMEX-AIM.ARPA>: request]]
To: AAAI-Executive-Committee: ;
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Return-Path: AAAI-OFFICE@SUMEX-AIM.ARPA
Received: from SUMEX-AIM.ARPA by SRI-AI.ARPA with TCP; Fri 22 Apr 83 14:19:47-PST
Date: Fri 22 Apr 83 14:14:20-PST
From: Louis Robinson <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: [Michael Genesereth <GENESERETH@SUMEX-AIM.ARPA>: request]
To: aaai@SRI-AI.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Mail-From: GENESERETH created at 22-Apr-83 13:47:47
Date: Fri 22 Apr 83 13:47:44-PST
From: Michael Genesereth <GENESERETH@SUMEX-AIM.ARPA>
Subject: request
To: aaai@SUMEX-AIM.ARPA
Claudia or Lou,
Could you please mail the following letter to the exec committee with
subject "Status of AAAI-83 Program". Thanks.
mrg
Executive Councilors,
All of the papers for AAAI-83 are now in. This year, we
received a total of around 250 papers, in comparison with about 400
last year. Obviously, IJCAI-83 has had an effect, but the number of
submissions is still substantial. You'll find a breakdown of the
submissions by category below.
You will also find a list of program committee members. The
committee was organized somewhat later this year than in past years.
I just wasn't sure how many papers we would get due to IJCAI, and
in fact a week before the deadline we had only 16 papers. However,
most everybody I asked was extremely accommodating, and I think the
committee is a good one.
We plan to meet at Stanford on the weekend of May 7 and 8 to
make the final decisions on the papers and arrange the program. If
any of you have suggestions for panels, debates, or invited talks,
please try to send them to me before then.
mrg
Breakdown of AAAI-83 Submissions by Category
Knowledge Representation 33
Problem Solving and Inference 32
Search 12
Learning 18
Support Hardware and Software 12
Philosophy and Methodology 13
Natural Language 29
Speech Processing 3
Vision and Robotics 28
Cognitive Modelling 16
Expert Systems 48
AAAI-83 Program Committee
Bob Balzer ISI
Tom Binford Stanford
Ron Brachman Fairchild
Mike Brady MIT
Bruce Buchanan Stanford
Barbara Grosz SRI
Ben Kuipers Tufts
Wendy Lehnert Massachusetts
Doug Lenat Stanford
John McDermott CMU
Tom Mitchell Rutgers
Bob Moore SRI
Gordon Novak Texas and Stanford
Ray Reiter UBC
Howie Shrobe MIT and Symbolics
Kurt vanLehn Xerox
Dave Waltz Illinois
Bonnie Webber Pennsylvania
-------
-------
-------
∂22-Apr-83 1525 RINDFLEISCH@SUMEX-AIM Lisp Workstations
Received: from SUMEX-AIM by SU-AI with PUP; 22-Apr-83 15:25 PST
Date: Fri 22 Apr 83 15:25:36-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Lisp Workstations
To: JMC@SU-AI.ARPA, RPG@SU-AI.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA, Feigenbaum@SUMEX-AIM.ARPA
John/Dick,
Following is the budget breakdown I'm using for Lisp workstations in
the DARPA Equipment contract. I am budgeting everything in year 1,
using Ed Feigenbaum's strategy that we should make our hunger for
equipment manifest by consuming what DARPA offers as fast as it comes.
In fact the first year runs until June 1984 and even then we do not have
to spend everything if there is a good reason to wait. It can be pushed
into the second year. Let me know if you want to make any changes.
Tom R.
HPP Workstations:
1 ea. Symbolics LM-2 trade-in residual $ 25,830
3 ea. Symbolics 3600, 512K memory @68,800 206,400
1 ea. Symbolics CD-1000 color display subsyst 29,900
1 ea. Symbolics TD20 streamer tape drive 3,900
1 ea. Symbolics AM1200 service modem 2,000
8 ea. Xerox 1108 (Dandelion), 29 Mbyte disk
@26,720 213,760
1 ea. Xerox 1132 (Dorado) 120,435
1 ea. 10 Mbit/sec UNIBUS Ethernet interface
(Symbolics EN11) 6,900
Installation, cabling, and spares 3,000
--------
$612,125
Formal Reasoning Workstations:
5 ea. Symbolics 3600, 512K memory @68,800 344,000
1 ea. Symbolics TD20 streamer tape drive 3,900
1 ea. Symbolics AM1200 service modem 2,000
1 ea. 10 Mbit/sec UNIBUS Ethernet interface
(Symbolics EN11) 6,900
Installation, cabling, and spares 3,000
--------
$359,800
Note, I budgeted 2 sets of Symbolics tape/modem/Ethernet interface
equipment because we will have 2 separate sites (MJH and Welch Road).
-------
∂22-Apr-83 1528 RINDFLEISCH@SUMEX-AIM Still haven't received your earlier msg...
Received: from SUMEX-AIM by SU-AI with PUP; 22-Apr-83 15:28 PST
Date: Fri 22 Apr 83 15:29:28-PST
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Still haven't received your earlier msg...
To: JMC@SU-AI.ARPA
-------
∂22-Apr-83 1609 PB library data base
Here is a response I recently got from Harry Llull. I haven't had a chance
to talk to him in person yet; this is a response to a message I sent him.
It appears that RLG (whatever that stands for) wants to recoup some of their
costs. It still may be possible to do a one time only search to get our own
data base, if they don't put a prohibition on machine readable data capture.
Of course if that $60/hr is for 30 cps, that may be quite expensive.
I'll talk to some of the principals soon, and get the full story.
--peter
--------
∂19-Apr-83 1048 ADMIN.LIBRARY%SU-SCORE.ARPA@SCORE RLIN
Received: from SU-SCORE by SU-AI with PUP; 19-Apr-83 10:48 PST
Date: Tue 19 Apr 83 10:49:00-PST
From: C.S./Math Library <ADMIN.LIBRARY@SU-SCORE.ARPA>
Subject: RLIN
To: pb@SAIL
Peter,
I talked with the Library Systems Office concerning your suggestion. The
libraries are in negotiation now with RLG to obtain access to the tapes
in order to mount them for an online catalog. There has been no agreement
at this point and would not be a good time for a department to begin
negotiating on their own. However if McCarthy or the department would like
to set up a search only account you may go through LaVonne Gallo or Sarah
Thomas who is manager of library coordination at RLG 328-0920.
The charge is $60/hr connect time with a two hour per month minimum and a
$190 startup fee. With prepayment the connect time charge drops to $55
per hour.
Harry
-------
∂22-Apr-83 1630 Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Received: from USC-ECL by SU-AI with NCP/FTP; 22 Apr 83 16:30:46 PST
Received: from SUMEX-AIM by USC-ECL; Fri 22 Apr 83 16:27:45-PST
Date: Fri 22 Apr 83 16:19:28-PST
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
To: JMC%SU-AI@USC-ECL.ARPA
In-Reply-To: Your message of Fri 22 Apr 83 15:34:00-PST
John,
minor correction: I wasn't advocating SPENDING all the money; I was
advocating GETTING all the money as soon as we can. If we decide not to
spend it we can hold on to it through the end of the contract period.
And then if there is money left in the pot, we can ask for a "no cost
extension".
I am very optimistic about being to get MORE for equipment in 1984,
if we can show that we "need" it. We ALWAYS need it.
Ed
-------
∂22-Apr-83 2015 CSD.ROWE@SCORE Re: Omni
Received: from SU-SCORE by SU-AI with PUP; 22-Apr-83 20:15 PST
Date: Fri 22 Apr 83 16:56:54-PST
From: Neil Rowe <CSD.ROWE@SU-SCORE.ARPA>
Subject: Re: Omni
To: JMC@SAIL
cc: NCR@SAIL
In-Reply-To: Your message of Fri 22 Apr 83 12:24:00-PST
sorry, I haven't got a copy -- just saw it in a library.
-------
∂23-Apr-83 0730 Woody Bledsoe <ATP.Bledsoe@UTEXAS-20>
Received: from USC-ECL by SU-AI with NCP/FTP; 23 Apr 83 07:08:51 PST
Received: from UTEXAS-20 by USC-ECL; Sat 23 Apr 83 07:05:31-PST
Date: 23 Apr 1983 0905-CST
From: Woody Bledsoe <ATP.Bledsoe@UTEXAS-20>
To: JMC%SU-AI@USC-ECL
cc: ATP.Bledsoe@UTEXAS-20
In-Reply-To: Your message of 22-Apr-83 1232-CST
Thanks John for accepting our centennial invitation. I hope we can add a few
dollars (other than the $500. plus expenses) to make it more fair for you.
We will be in touch with you about it several weeks before the event. Woody
-------
∂23-Apr-83 1402 KAHN@USC-ISI Re: Addition to ARPA joint proposal
Received: from SUMEX-AIM by SU-AI with PUP; 23-Apr-83 14:01 PST
Received: from USC-ISI by SUMEX-AIM.ARPA with TCP; Sat 23 Apr 83 13:57:06-PST
Date: 23 Apr 1983 1352-PST
Sender: KAHN at USC-ISI
Subject: Re: Addition to ARPA joint proposal
From: KAHN at USC-ISI
To: WIEDERHOLD at SUMEX-AIM
Cc: KAHN at USC-ISI, csd.betty at SU-SCORE
Cc: csd.atkinson at SU-SCORE, jmc%sail at SUMEX-AIM
Cc: rpg%sail at SUMEX-AIM, dek%sail at SUMEX-AIM
Cc: dek%SU-AI at SUMEX-AIM
Message-ID: <[USC-ISI]23-Apr-83 13:52:16.KAHN>
In-Reply-To: Your message of Wed 20 Apr 83 17:40:27-PST
Gio,
Thanks. I had forgotten that this was to be the
test case of the new method to get a jump
on the contracting process. Without that context,
it was awfully hard to parse the earlier msgs.
thanks very much for clarifying this for me. I should
have remembered.
bob
∂23-Apr-83 2351 Lowell Wood <LLW at S1-A> Trips
Received: from S1-A by SU-AI with NCP/FTP; 23 Apr 83 23:50:54 PST; for: jmc
Date: 23 Apr 1983 2345-PST
From: Lowell Wood <LLW at S1-A>
Subject: Trips
To: rpg at SU-AI
CC: LLW at S1-A, jmc at SU-AI, TM at S1-A
∂22-Apr-83 1354 Dick Gabriel <RPG@SU-AI> Trip(s)
Received: from SU-AI by S1-A with NCP/FTP; 22 Apr 83 13:54:40 PST
Date: 22 Apr 83 1356 PST
From: Dick Gabriel <RPG@SU-AI>
Subject: Trip(s)
To: JMC@SU-AI, llw@S1-A, tm@S1-A
I have had a paper accepted at IJCAI on S-1 Lisp.
Since I have never been to Europe, I'd like to go there to
deliver it. Also, I would like to attend AAAI in Washington,
which is 2 weeks after IJCAI. I understand it is very difficult to
get foreign travel from LLL, and it is probably pretty hard for Stanford
too. There are several possibilities. LLL can pay round-trip day
coach to NY, and Stanford could provide the remainder. Since AAAI
is on the way back, and since I am booking charters now, I think the
total cost would not be too high. Another possibility is to consider
the two conferences to be one in NY and one in DC with a roundtrip in
both cases. This covers the circle route easily too.
Any ideas? IJCAI is in West Germany, August 8-12. AAAI is
August 22-26.
-rpg-
[Dick; We can (and cheerfully will) send you to AAAI. As you noted, it's
*much* harder to send you abroad, because the Lab's funds for foreign
travel are `colored' uniquely, and are usually spoken for at least a
half-year in advance, on an allocation basis. Could Stanford pick up the
NY to Germany to DC fare and in-Germany costs, with the Lab handling all
the domestic costs? The `two conferences' gambit won't work, as the Lab
wants all travel receipts, and demands to have you certify all reported
travel dates-and-times as the actual ones, with criminal penalties for
false statements (since the disbursement of Federal funds is involved).
Since the IJCAI paper is reporting on Lab work, I'm pretty sure that we
can pay you salary while you're at IJCAI, and travelling to and from there
(as well as to AAAI, of course). Keep me posted on what you're working out
on this. Lowell]
∂24-Apr-83 0004 Lowell Wood <LLW at S1-A> IPTO S-1 Support
Received: from S1-A by SU-AI with NCP/FTP; 24 Apr 83 00:02:39 PST; for: jmc
Date: 23 Apr 1983 2357-PST
From: Lowell Wood <LLW at S1-A>
Subject: IPTO S-1 Support
To: jmc at SU-AI
CC: LLW at S1-A
∂21-Apr-83 2356 John McCarthy <JMC@SU-AI> IPTO support of S-1
Received: from SU-AI by S1-A with NCP/FTP; 21 Apr 83 23:55:57 PST
Date: 21 Apr 83 2353 PST
From: John McCarthy <JMC@SU-AI>
Subject: IPTO support of S-1
To: LLW@SU-AI
I had dinner with Bob Kahn last night, and he brought up that he wanted
to support S-1 but on a smaller scale than the Navy and DoE and wanted
to make a distinctive contribution. Since we were talking about LISP
I suggested that he support making MACSYMA available in Common Lisp
for the S-1, and he liked that. Trying to phone you I got Mike
Farmwald who remarked that that would be nice but it might be rather
trivial in cost, because Symbolics was putting Macsyma into Zetalisp
which is very close. Perhaps Bob Kahn deserves the honor of a larger
contribution and I trust that my off-hand remarks won't deprive him
of that honor.
[John: After years of scrambling for pennies, I'm disinclined to turn down
just about any offer of S-1 support. I'll definitely give thought as to
what Bob might fund that would be `distinctive,' and will assuredly come
up with something whose cost is not so insultingly small as was Mike's
suggestion. Actually concluding any support arrangements with DARPA won't
be completely uncomplicated, however, as my management takes a dim view of
any Lab program having too many sponsors and also is a bit turned off at
the present time by DARPA management's views-and-behavior re funding of
Lab projects, computer-related and otherwise (`intolerably flaky and
micromanagerial' was one of the more kind characterizations recently).
Thanks very much for the information, in any case. Lowell]
∂24-Apr-83 1603 CSD.MCGRATH@SCORE [BUNDY@RUTGERS: New Lectureships]
Received: from SU-SCORE by SU-AI with PUP; 24-Apr-83 15:44 PDT
Date: Sun 24 Apr 83 15:47:13-PDT
From: Jim McGrath <csd.mcgrath@SCORE>
Subject: [BUNDY@RUTGERS: New Lectureships]
To: jmc@SAIL
Know any interested people?
---------------
Return-Path: BUNDY@RUTGERS
Received: from RUTGERS by SU-SCORE.ARPA with TCP; Sun 24 Apr 83 08:09:03-PDT
Date: 24 Apr 1983 1058-EDT
From: BUNDY@RUTGERS
Subject: New Lectureships
To: acw@MIT-AI, adis@MIT-AI, amsler@UTEXAS-20, bak@MIT-AI, barrow@SRI-KL,
beppe@MIT-AI, blue@MIT-AI, borning@WASHINGTON, brady@MIT-AI, bud@MIT-MC,
carl@MIT-AI, csd.mcgrath@SU-SCORE, dan@MIT-ML, dekleer@PARC-MAXC, ddm@MIT-AI,
djc@MIT-DMS, duffey@MIT-AI, ffm@MIT-MC, geoff@MIT-AI, gpb@MIT-ML,
harold@MIT-ML, hdt@MIT-AI, henry@MIT-AI, hinton@CMU-CS-C, jerryb@MIT-AI,
kdf@MIT-AI, kwc@MIT-AI, lcampbell@DEC-MARLBORO, maddox@MIT-ML, miller@MIT-AI,
milne@MIT-AI, pcsid@MIT-MC, pereira@SRI-AI, pratt@MIT-AI, reiner@CMU-CS-A,
shardy@SRI-KL, steels@MIT-AI, steveh@MIT-MC, warren@SRI-AI, weise@MIT-ML,
wjl@MIT-ML
cc: bundy@RUTGERS
As an indirect result of the Japanese 5th Generation
plans, and the interest in expert systems etc, AI is taking
off again in the UK. There have been too many inititives
for me to list them all, but the latest manifestation
is the awarding of new lectureships at many UK universities
for AI related activities. At Edinburgh we have 3 new posts.
One is an industrially sponsored post, to which we have just appointed,
and the other two are part of the Dept of Education IT initiative.
One of these is for computational linguistics, and the other
for IKBS (a new UK buzzword, meaning whatever you like, but
probably applied AI, with a leaning towards expert systems.)
Interested people should apply to Janet Lee, Dept of AI,
Univ. of Edin, Forrest Hill, Edin.
Alan
-------
-------
∂24-Apr-83 1656 CLT
when you put the roast in, take the foil off the meat, but leave the
squash wrapped.
It will be as you have decreed.
∂25-Apr-83 0328 host USC-ECL
Received: from USC-ECL by SU-AI with NCP/FTP; 25 Apr 83 03:28:21 PDT
Received: from SU-SCORE by USC-ECL; Mon 25 Apr 83 00:30:33-PDT
2α20⊂ 94⊂→≠⊗⊂_→_→a⊂!:)247∀*42&207∀≡()''" )" ( ε)26<7⊂()'')" ( *≥⊂$⊂(0:943 ⊗⊂&67(05⊂!⊂≤_→≠(472⊂∀≠_→$⊗ε)52≥⊂()''"4⊂⊂⊂+_⊃*7⊂()'')" ( εε()''"4⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂&20⊂→≠ 9⊂_→⊂⊂⊂⊂⊂⊂⊂+7:_≥⊂$_ε*70*74εε⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂⊂ 6944⊗+1:7()''⊗ε"02⊂)⊂→≠⊂ 9⊂≤→→→→$≥→"*ε#97⊂!:)247∀*42&20∂
M⊂I↓MTM
=I∀¬IA∧4)IAI↓MTM
=I∀4)M]4(4)]A%4)4)4(4)QM
=I∀4)5A
I5⊃@αuable.
Fernando Pereira and Chuck will both be responsible for the editorial
work.
Some of the people that answered the initial bboard message stated
that they would be interested in general Prolog and logic programming
information, but not in the nitty-gritty details of Prolog systems and
their bugs. Therefore, we have decided to setup two mailing lists,
PROLOG and PROLOG-HACKERS: PROLOG will provide a digest of articles of
general interest; PROLOG-HACKERS will provide a direct mailing for the
nasty stuff. Finally, PROLOG-REQUEST should be used for distribution
requests. In a nutshell
Mail to for
------- ---
PROLOG@SCORE sending articles of general interest
PROLOG-HACKERS@SCORE sending articles of limited interest
PROLOG-REQUEST@SCORE getting in and out of the mailing
lists and other bureaucracy
At the moment, everyone that replied to the bboard announcement is in
the PROLOG list. If you want to be in PROLOG-HACKERS as well, send in
a request.
Please do not send the same message to PROLOG and PROLOG-HACKERS. We
will move messages between lists if appropriate.
An archive of the USEnet prolog interest group interchange can be
FTP'd from SCORE using standard anonymous login convention. The
pathname is
[SU-SCORE]PS:<PROLOG>UPIG.ARCHIVE
We intend to continue distributing net.lang.prolog articles, but of
course we hope the main source of material will be YOU, the people in
this list.
Looking forward to your contributions,
Fernando Pereira
Chuck Restivo
[ PS: Forthcoming attraction: the Prolog and logic programming
bibliography I have collected in the last 6 years. - FP ]
------------------------------
End of PROLOG Digest
********************
∂25-Apr-83 1100 JK
To: JMC, JK, CLT, RPG, JJW
ON DEFINITIONS AND HIGH ORDER MATCHING IN EKL
I have implemented a (complete) high order matching algorithm
as a part of the EKL rewriting system. It is based on the following
fact: When specialised to matching (i.e., one-sided unification),
the Huet algorithm for high order unification terminates.
Note that the general case is undecidable.
It does not seem to slow down the system: In fact, the current matcher
runs about 25% faster on my test files than the one I had in the fall.
I am enclosing some examples on how to use this facility for lisp definitions.
The DEFINE command in EKL requires a proof of the existence of an object
before a definition is accepted. For instance, in the first example
the wff
∃COLL.∀X Y.(ATOM X⊃COLL(X)=LIST(X))∧COLL(X.Y)=COLL(X)*COLL(Y)
has to rewrite to true.
Note the subtleties involved in this particular match:
(1) The variable PARS in sexpinductiondef is of type ground*; a list
type. It will have to matched to ().
(2) The unifiable variables in sexpinductiondef are ATOMCASE,DEFSEXP
and PARS,X,Y. The latter group occurs past an existential variable:
Since we have high order matching, we can freely use functional interpretations
for variables of this type
(3) The existential variable COLL, of type GROUND→GROUND, has to be matched
against the existential variable FUN, of type GROUND*→GROUND. This is permissible
since COLL is a purely functional variable (we use extensionality here).
(get-proofs lispax)
(proof collect)
(show sexpinductiondef)
;labels: SEXPINDUCTIONDEF
33. (AXIOM
|∀ATOMCASE DEFSEXP.(∃FUN.(∀PARS X Y.(ATOM X⊃FUN(X,PARS)=ATOMCASE(X,PARS))∧
FUN(X.Y,PARS)=
DEFSEXP(X,Y,FUN(X,PARS),FUN(Y,PARS),
PARS)))|)
;5s cpu
(define coll
|∀X Y.(ATOM X⊃COLL(X)=LIST(X))∧COLL(X.Y)=COLL(X)*COLL(Y)|
(use sexpinductiondef))
(show listinductiondef)
;labels: LISTINDUCTIONDEF
31. (AXIOM
|∀NILCASE DEF.(∃FUN.(∀PARS X U.FUN(NIL,PARS)=NILCASE(PARS)∧
FUN(X.U,PARS)=DEF(X,U,FUN(U,PARS),PARS)))|)
;2s cpu
(define appen
|(∀x u v.(appen(nil,v)=v∧appen(x.u,v)=x.appen(u,v)))|
(use listinductiondef))
∂25-Apr-83 1159 GOTELLI@SCORE Re: new account
Received: from SU-SCORE by SU-AI with PUP; 25-Apr-83 11:46 PDT
Date: Mon 25 Apr 83 11:46:37-PDT
From: Lynn Gotelli <GOTELLI@SCORE>
Subject: Re: new account
To: JMC@SAIL
In-Reply-To: Your message of Fri 22 Apr 83 11:58:00-PST
Thank you.
-------
∂25-Apr-83 1259 ZAVEN@SCORE DARPA, RADC
Received: from SU-SCORE by SU-AI with PUP; 25-Apr-83 12:45 PDT
Date: Mon 25 Apr 83 12:44:08-PDT
From: Lawrence Markosian <ZAVEN@SCORE>
Subject: DARPA, RADC
To: jmc@SAIL
Judith Daley of ARPA, whom I visited on Friday, told me that in
addition to the general problem at RADC which we discussed, there
is this: some time ago DARPA funded a project thru RADC and
the contract which resulted (between RADC & the supplier) was
"nothing like" what DAPRA had sought to fund and so they don't
deal with RADC anymore.
Lawrence Markosian
-------
∂25-Apr-83 1432 RPG Buying all the 3600's the first year
To: rindfleisch@SUMEX-AIM, JMC@SU-AI, REG@SU-AI, EAF@SU-AI
Enough people whom I really trust have expressed satisfaction with the
3600 and with the Lisp workstation concept that I do not feel too
bad about buying all of them the first year. Given that I don't see
an alternative for the Formal Reasoning group in the near future
for a quality, large-address-space Lisp, I'm not inclined to wait
and see too long. We have work that needs to get done now.
I have heard, though, that there is some chance that NIL on Vaxes
is rapidly moving towards Common Lisp and is quite usable. I will
be visiting DEC and MIT later this week and will test this rumor
first-hand. If it turns out to be true, then I might agree with John.
-rpg-
∂25-Apr-83 1609 CSD.FOYSTER@SCORE Denelcor HEP
Received: from SU-SCORE by SU-AI with PUP; 25-Apr-83 16:09 PDT
Date: Mon 25 Apr 83 16:11:45-PDT
From: Gordon T. Foyster <CSD.FOYSTER@SCORE>
Subject: Denelcor HEP
To: jmc@SAIL
cc: csd.foyster@SCORE
The sites that have the Denelcor HEP MIMD computer are:
Ballistics Research Laboratory - 4 processors
Cornell University - 1 processor
Georgia Tech. - 1 processor
Washington State University has done some studies of MIMD algorithms
on the machine - at Denelcor in Denver I guess.
Steve Lundstrom - SLNDSTRM@SHASTA has more information on the computer.
Gordon Foyster
-------
∂25-Apr-83 1653 DFH Chicago travel expenses
In starting to do the expense report for this
conference, I notice there is no charge slip
or receipt for the hotel bill. Looking over
the correspondence and the hotel confirmation,
it seems that it my have been covered by the
museum. Do you have any recollection of actually
paying them? I can call to double-check.
--Diana
Please double-check, because I don't remember.
∂26-Apr-83 1433 DFH phone call from Prof. Schwartz (sp?)
He is at Nils Nisson's office at SRI this
afternoon and asks that you call him there.
859-2311.
--Diana
∂27-Apr-83 0408 KDF%MIT-OZ@MIT-MC metaphors
Received: from USC-ECL by SU-AI with NCP/FTP; 27 Apr 83 04:08:49 PDT
Received: from MIT-MC by USC-ECL; Wed 27 Apr 83 04:07:54-PDT
Date: Wed, 27 Apr 1983 06:59 EDT
From: KDF%MIT-OZ@MIT-MC
To: John McCarthy <JMC%SU-AI@USC-ECL>
Cc: phil-sci%mit-oz%MIT-MC@USC-ECL
Subject: metaphors
In-reply-to: Msg of 27 Apr 83 0257 PDT from John McCarthy <JMC%SU-AI at USC-ECL>
Yes. Sterile.
∂27-Apr-83 0616 Communications Satellite <COMSAT @ MIT-MC> Msg of Wednesday, 27 April 1983 06:56 EDT
Received: from USC-ECL by SU-AI with NCP/FTP; 27 Apr 83 06:16:43 PDT
Received: from MIT-MC by USC-ECL; Wed 27 Apr 83 06:12:07-PDT
Date: 27 April 1983 09:12 EDT
From: Communications Satellite <COMSAT @ MIT-MC>
Subject: Msg of Wednesday, 27 April 1983 06:56 EDT
To: JMC%SU-AI @ USC-ECL
FAILED: frd at SU-AI; Host appears to be permanently down or not accepting mail.
Failed message follows:
-------
Mail-from: ARPANET site SU-AI rcvd at 27-Apr-83 0301-PDT
Date: 27 Apr 83 0257 PDT
From: John McCarthy <JMC%SU-AI@USC-ECL>
Subject: metaphors
To: phil-sci%mit-oz%MIT-MC@USC-ECL
The whole discussion of the scientific community metaphor for mind
and the associated collection of metaphors reinforces my conviction
that metaphors are not science but are a mere decoration. No
argument that depends on metaphors for its effect can be regarded
as convincing. In short the last two week's discussion seems empty
to me. Does anyone else think so?
∂27-Apr-83 0839 GENESERETH@SUMEX-AIM Invitation
Received: from SUMEX-AIM by SU-AI with PUP; 27-Apr-83 08:39 PDT
Date: Tue 26 Apr 83 18:13:20-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM.ARPA>
Subject: Invitation
To: Bobrow%PARC-MAXC@SUMEX-AIM.ARPA, Brown%PARC-MAXC@SUMEX-AIM.ARPA,
BEngelmore%SRI-KL@SUMEX-AIM.ARPA, Feigenbaum@SUMEX-AIM.ARPA,
RHayes-Roth%SRI-KL@SUMEX-AIM.ARPA, hart%SRI-KL@SUMEX-AIM.ARPA,
nilsson%SRI-KL@SUMEX-AIM.ARPA, jmc@SU-AI.ARPA, tw@SU-AI.ARPA
cc: genesereth@SUMEX-AIM.ARPA
Gentlemen,
The program committee meeting for this year's AAAI conference will be
held at Stanford on the 7th and 8th of May. In the interests of
getting all of the socializing out of the way before the meeting
begins, I have asked the program committee members to my house (1242
Forest Avenue in Palo Alto) for a reception at 7 on Friday the 6th. I
would very much appreciate it if you could join us. It would be a
nice gesture for the committee members, and it's a prime opportunity
to make suggestions about the conference and air any concerns you
might have. Please let me know if you can make it.
mrg
AAAI-83 Program Committee
Bob Balzer ISI
Tom Binford Stanford
Ron Brachman Fairchild
Mike Brady MIT
Bruce Buchanan Stanford
Barbara Grosz SRI
Ben Kuipers Tufts
Wendy Lehnert Massachusetts
Doug Lenat Stanford
John McDermott CMU
Tom Mitchell Rutgers
Bob Moore SRI
Gordon Novak Texas and Stanford
Ray Reiter UBC
Howie Shrobe MIT and Symbolics
Kurt vanLehn Xerox
Dave Waltz Illinois
Bonnie Webber Pennsylvania
-------
∂27-Apr-83 1006 FEIGENBAUM@SUMEX-AIM BBC's HORIZON program on AI
Received: from SUMEX-AIM by SU-AI with PUP; 27-Apr-83 10:06 PDT
Date: Wed 27 Apr 83 09:58:13-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: BBC's HORIZON program on AI
To: lenat@SUMEX-AIM.ARPA, nilsson%SRI-AI@SUMEX-AIM.ARPA, dertouzos%MIT-MC@SUMEX-AIM.ARPA,
myers@SUMEX-AIM.ARPA, hpp-exec@SUMEX-AIM.ARPA
cc: jmc@SU-AI.ARPA, tw@SU-AI.ARPA, bundy%RUTGERS@SUMEX-AIM.ARPA, kahn%USC-ISI@SUMEX-AIM.ARPA
I received a videotape of the BBC Horizon (3/23/83) program on AI,
in which some of us participated. I couldn't "see" it because of
compatibility problems with American TV standards. But I could hear
it (audio was fine; it doesn't seem to be worth seeing anyway, but that
might be possible with video output to video monitor, i.e. avoiding
the nasty RF; if any one wants to pursue that, let me know).
As our British friends have rumored, the program ranged from OK to
awful, but the package makes a very bad impression. It's a throwback
to the media treatment of ten years ago. Feigenbaum (2 sentences; OK);
Lenat (2 sentences;OK); Nilsson ( 4 or 5 sentences; OK); a little
Jack Myers (OK); lots of Weizenbaum and ELIZA (demo of ELIZA; trailer
using ELIZA); lots and lots and lots of Ed Fredkin saying his normal
wild things (see chapter in Pam McCorduck's book Machines Who Think).
Also: "The race is on...Fifth Generation...Michie: where is the money?...
Mike Dertouzos "shocked" at Japanese plan...etc. etc."
But mostly it's the polar-opposite duo, Fredkin and Weizenbaum that
are featured, from beginning to end, with Ed's remarks lending
credibility to Joe's criticisms regarding social responsibility.
If anyone wants a copy of the audio, let me know and I'll take the
total order over to a tape copying place, have them produced, and
send you the bill. If anyone wants to fool with the video, let me know.
I'll play the tape at Stanford in some public way soon (probably arrange
a bag-lunch sometime).
(Note to Alan Bundy: do you have any news on British reaction to that program
that you can share with your American friends?)
-------
∂27-Apr-83 1244 GAVAN%MIT-OZ@MIT-MC metaphors
Received: from USC-ECL by SU-AI with NCP/FTP; 27 Apr 83 12:44:27 PDT
Received: from MIT-MC by USC-ECL; Wed 27 Apr 83 12:41:25-PDT
Date: Wed, 27 Apr 1983 09:07 EDT
From: GAVAN%MIT-OZ@MIT-MC
To: John McCarthy <JMC%SU-AI@USC-ECL>
Cc: phil-sci%mit-oz%MIT-MC@USC-ECL
Subject: metaphors
In-reply-to: Msg of 27 Apr 83 0257 PDT from John McCarthy <JMC%SU-AI at USC-ECL>
From: John McCarthy <JMC%SU-AI at USC-ECL>
The whole discussion of the scientific community metaphor for mind
and the associated collection of metaphors reinforces my conviction
that metaphors are not science but are a mere decoration.
Please tell me what you think science is. Also, in your view, are any
of the models that scientists concoct (e.g. life worlds) in some sense
metaphorical? If not, what is their relation to the "real world"?
No argument that depends on metaphors for its effect can be regarded
as convincing.
It seems to me that metaphors aren't used with an eye toward making
arguments, but rather with an eye toward developing heuristics and
solving puzzles.
∂27-Apr-83 1305 GAVAN%MIT-OZ@MIT-ML metaphors
Received: from USC-ECL by SU-AI with NCP/FTP; 27 Apr 83 13:05:33 PDT
Received: from MIT-ML by USC-ECL; Wed 27 Apr 83 13:03:31-PDT
Date: Wed, 27 Apr 1983 14:49 EDT
From: GAVAN%MIT-OZ@MIT-ML
To: KDF%MIT-OZ@MIT-MC
Cc: John McCarthy <JMC%SU-AI@USC-ECL>
Subject: metaphors
In-reply-to: Msg of 27 Apr 1983 06:59-EDT from KDF%MIT-OZ at MIT-MC
Date: Wednesday, 27 April 1983 06:59-EDT
From: KDF%MIT-OZ at MIT-MC
To: John McCarthy <JMC%SU-AI at USC-ECL>
cc: phil-sci%mit-oz%MIT-MC at USC-ECL
Re: metaphors
Received: from MIT-MC by USC-ECL; Wed 27 Apr 83 04:07:54-PDT
Yes. Sterile.
This is a metaphor.
∂27-Apr-83 1316 DAM%MIT-OZ@MIT-ML metaphors
Received: from USC-ECL by SU-AI with NCP/FTP; 27 Apr 83 13:16:02 PDT
Received: from MIT-ML by USC-ECL; Wed 27 Apr 83 13:11:28-PDT
Date: Wed, 27 Apr 1983 15:03 EDT
From: DAM%MIT-OZ@MIT-ML
To: JMC%SU-AI@USC-ECL
cc: phil-sci%MIT-OZ@MIT-ML
Subject: metaphors
I sent this message earlier but it didn't seems to get through.
Date: 27 Apr 83 0257 PDT
From: John McCarthy <JMC%SU-AI at USC-ECL>
No argument that depends on metaphors for its effect can be regarded
as convincing. In short the last two week's discussion seems empty
to me. Does anyone else think so?
I tend to agree with the above in this case. Metaphors are useful
when a precise relationship exists between things. For example
Franklin considered electricity to be analogous to a fluid. It turns
out that electricity IS analogous to a fluid in a precise sense (the
dynamical equations have the same form). I think that socities and
individuals are analogous in the precise sense that a society is a
collection of interacting individuals. Thus one might better
understand societies by understanding individuals. However I think a
stronger claim is made in the "society of mind" metaphore. There is
supposedly some level of abstraction at which both societies and
individuals are collections of interacting agents. The nature of the
relationship between societal agents and agents within an individual
has never been clear to me. What is the proper analogy between
societies and individuals (is it a well defined analogy or just a
basis for fuzzy thinking)?
David Mc
∂27-Apr-83 1328 DAM%MIT-OZ@MIT-MC metaphors
Received: from USC-ECL by SU-AI with NCP/FTP; 27 Apr 83 13:25:36 PDT
Received: from MIT-MC by USC-ECL; Wed 27 Apr 83 13:23:08-PDT
Date: Wed, 27 Apr 1983 12:02 EDT
From: DAM%MIT-OZ@MIT-MC
To: JMC%SU-AI@USC-ECL
cc: phil-sci%MIT-OZ@MIT-MC
Subject: metaphors
Date: 27 Apr 83 0257 PDT
From: John McCarthy <JMC%SU-AI at USC-ECL>
No argument that depends on metaphors for its effect can be regarded
as convincing. In short the last two week's discussion seems empty
to me. Does anyone else think so?
I tend to agree with the above in this case. Metaphors (analogies)
are useful when a precise relationship exists between things. For
example when Franklin considered electricity to be analogous to a
fluid. It turns out that electricity IS analogous to a fluid in a
precise sense (the dynamical equations have the same form). I think
that socities and individuals are analogous in the precise sense that
a society is a collection of interacting individuals. Thus one might
better understand societies by understanding individuals. However I
think a stronger claim is made in the "society of mind" metaphore.
The claim is (I think) that both societies and minds are collections of
agents where there is some precise relationship (perhaps at an
abstract level) between the interactions of agents in a society and
the interaction of agents in a single mind. For this analogy to be
useful as a theory of mind the nature of the relationship between
the "agents" of society and the "agents" of a single mind should be
made more explicit.
David Mc
∂27-Apr-83 1328 RICKL%MIT-OZ@MIT-ML Re: metaphors
Received: from USC-ECL by SU-AI with NCP/FTP; 27 Apr 83 13:26:11 PDT
Received: from MIT-ML by USC-ECL; Wed 27 Apr 83 13:24:08-PDT
Date: 27 Apr 1983 1610-EDT
From: RICKL%MIT-OZ@MIT-ML
Subject: Re: metaphors
To: JMC%SU-AI@USC-ECL
cc: phil-sci%MIT-OZ@MIT-ML
In-Reply-To: Your message of 27-Apr-83 0257-EDT
Date: 27 Apr 83 0257 PDT
From: John McCarthy <JMC%SU-AI@USC-ECL>
Subject: metaphors
The whole discussion of the scientific community metaphor for mind
and the associated collection of metaphors reinforces my conviction
that metaphors are not science but are a mere decoration. No
argument that depends on metaphors for its effect can be regarded
as convincing. In short the last two week's discussion seems empty
to me. Does anyone else think so?
I think that you are trading upon the common distinction between the
context of discovery and the context of justification. I take it that
you would not deny that metaphorical thinking figures prominently in
scientific discovery --- history is too full of examples to the
contrary for such a denial to be coherent with observed fact. Thus
at the very least they provide a powerful influence ("heuristic", if you
will) which has frequently guided researchers in fruitful scientific
investigation.
If you then claim that some of the tasks of science are to provide as rigorous
as possible a description of the phenomena, to attempt to embed the
observed within a formal system if possible, to extend the range of
phenomena covered by such formal systems and connect them in systematic
fashion to other such formal systems, and to attempt to "verify" (which I
take to mean, investigate the accuracy and scope of --- Popper shows that
no theory is verifiable, only falsifiable, and all theories are trivially
falsified in virtue of being generalizations which abstract from their
"ceteris paribus" clauses) to "verify" knowledge and theories not already
well-accepted; then I agree. Metaphorical or analogical reasoning is
notoriously unreliable on its own, and requires a formal justification
before it can be believed in. On the other hand, when a practicing
scientist is reasoning about his/her daily work (perhaps thinking about a
proposed experiment), I do not believe that such reasoning is conducted
entirely within an "uninterpretted formalism" connected to the real world
by "bridge principles". In addition to being cumbersome, such formalisms
have only been advanced for some parts of physics, chemistry, and a
little bit of biology. Typically they are not put forward until the
field is already so well understood that it is essentially static, i.e.
no longer a topic of active scientific research. This would seem to
imply that scientific research does not utilize scientific reasoning,
which is absurd. Neurophysiology is an example of a field in which no
rigorous formalization is yet accepted, but in which research is progressing
rapidly and which we definitely want to call "science". The absurdity
results from imagining that logico-deductive systems encompass all of
scientific reasoning --- they are clearly *necessary* and *important*,
but also not *sufficient*.
-=*=- rick
-------
∂27-Apr-83 2016 JMC*
Look up Tommy Koh.
∂27-Apr-83 2055 Christopher C. Stacy <CSTACY @ MIT-MC>
Received: from USC-ECL by SU-AI with NCP/FTP; 27 Apr 83 20:55:34 PDT
Received: from MIT-MC by USC-ECL; Wed 27 Apr 83 19:04:01-PDT
Date: 27 April 1983 21:43 EDT
From: Christopher C. Stacy <CSTACY @ MIT-MC>
To: me @ SU-AI, roz @ SU-AI, lee @ SU-AI, jmc @ SU-AI
If you get this message, please reply to me so that I know our
mailer is doing routing for SU-AI properly. Thanks!
Chris
Yes, I am getting messages including those from the phil-sci list.
I think we will be on TCP in a few days.
∂28-Apr-83 0743 CSD.BSCOTT@SCORE
Received: from SU-SCORE by SU-AI with PUP; 28-Apr-83 07:42 PDT
Date: Thu 28 Apr 83 07:46:04-PDT
From: Betty Scott <CSD.BSCOTT@SCORE>
To: JMC@SAIL
cc: CSD.BSCOTT@SCORE
In-Reply-To: Your message of Wed 27 Apr 83 16:50:00-PDT
We need to prepare a promotion recommendation form--simple form. I
need to talk with you about your research associate salaries in general.
Will you have a few minutes sometime this afternoon?
Betty
-------
∂28-Apr-83 1042 BUCHANAN@SUMEX-AIM COMTEX introduction
Received: from SUMEX-AIM by SU-AI with PUP; 28-Apr-83 10:24 PDT
Date: Thu 28 Apr 83 09:34:46-PDT
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: COMTEX introduction
To: jmc@SU-AI.ARPA
John,
Les had a couple of small corrections. Do you have any? The
Comtex people are getting anxious. (Les also had some photos that
I'll send -- don't know if they will use them, but they would add
some interest.)
thanks,
bgb
-------
I have a fair number of corrections if I can find my copy.
∂29-Apr-83 1401 CSD.ULLMAN@SCORE Adcom
Received: from SU-SCORE by SU-AI with PUP; 29-Apr-83 13:30 PDT
Date: Fri 29 Apr 83 13:24:11-PDT
From: Jeffrey D. Ullman <CSD.ULLMAN@SCORE>
Subject: Adcom
To: jmc@SAIL
I know you just took on a big university responsibility,
but could I ask you to take on one more little one?
In Brian Reid's absence, I need a chairman of the admissions committee
to rule on the enevitable cases of complaints regarding our rejections.
I'm not suggesting that you hold anybody's hand, just be a voice
with authority telling them NO.
-------
I am willing to do it, but Vaughan Pratt has been paying much more
detailed attention to the candidates for admission than I have.
If he is willing, he would be better than I.
∂29-Apr-83 1600 CSD.ULLMAN@SCORE
Received: from SU-SCORE by SU-AI with PUP; 29-Apr-83 15:36 PDT
Date: Fri 29 Apr 83 15:34:30-PDT
From: Jeffrey D. Ullman <CSD.ULLMAN@SCORE>
To: JMC@SAIL
In-Reply-To: Your message of Fri 29 Apr 83 15:10:00-PDT
Vaughan is pretty much on leave this quarter.
I'd appreciate it if you would be the one.
-------
OK, I'll do it.
∂29-Apr-83 1600 BUCHANAN@SUMEX-AIM comtex intro
Received: from SUMEX-AIM by SU-AI with PUP; 29-Apr-83 15:22 PDT
Date: Fri 29 Apr 83 15:15:29-PDT
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM.ARPA>
Subject: comtex intro
To: les@SU-AI.ARPA, jmc@SU-AI.ARPA
Many thanks for the comments. I have incorporated them into a draft that
I hope to send out next weJohn, I softened the parts you marked
"humph", also.
Les, at your suggestion I am including names & last known addresses of
PhD's from SAIL. The CSD records have no address for Morgan Ohwovoriole
or Paul Martin -- can either of you help?
thanks,
bgb
-------
Paul Martin is at SRI.
∂29-Apr-83 1636 CSD.ULLMAN@SCORE
Received: from SU-SCORE by SU-AI with PUP; 29-Apr-83 16:31 PDT
Date: Fri 29 Apr 83 16:27:46-PDT
From: Jeffrey D. Ullman <CSD.ULLMAN@SCORE>
To: JMC@SAIL
In-Reply-To: Your message of Fri 29 Apr 83 16:13:00-PDT
Thanks.
-------
∂29-Apr-83 1719 MAILER failed mail returned
Because the destination ARPA host(s) are not running NCP,
the following message was undeliverable to recipient(s)
king@KESTREL:
∂28-Apr-83 1540 JMC skyhooks
To: king@KESTREL
There was a recent article by Hans Moravec (note spelling) in the L-5
Society magazine. It contains references. Unfortunately, I didn't
keep it. Moravec is HPM at SAIL and is either that or Moravec
at CMUA. He can give you references. SAIL hasn't yet switched
to TCP, so messages have to be forwarded through ECLC. However,
it has been announced that we will switch this weekend if the
remaining debugging goes well.
∂30-Apr-83 1128 CSD.BSCOTT@SCORE Diana Hall
Received: from SU-SCORE by SU-AI with PUP; 30-Apr-83 11:27 PDT
Date: Sat 30 Apr 83 11:31:08-PDT
From: Betty Scott <CSD.BSCOTT@SCORE>
Subject: Diana Hall
To: JMC@SAIL
cc: CSD.BScott@SCORE
I think Diana has done an outstanding job in learning financial and other
procedures in the short time she has been here. She is conscientious,
her work is accurate, and she shows considerable initiative in the
accomplishment of her tasks. It it's agreeable to you, I would like to
accelerate her progress, i.e., give her a 5% salary increase at the end
of six months (6/l/84) instead of waiting the usual year. Is this o.k?
Betty
-------
I am entirely satisfied and would perhaps advocate being even more
generous.
∂30-Apr-83 1258 JJW Slander?
To: JMC, ME
The last paragraph of this rather long message may be of interest and
perhaps deserves a reply.
∂30-Apr-83 1229 pratt@Navajo "Stanford mail"
Received: from SU-HNV by SU-AI with PUP; 30-Apr-83 12:29 PDT
Received: from Shasta by Diablo with PUP; Sat, 30 Apr 83 12:13 PDT
Received: from Navajo by Shasta with PUP; Sat, 30 Apr 83 12:15 PDT
Date: Sat, 30 Apr 83 12:15 PDT
From: Vaughan Pratt <pratt@Navajo>
Subject: "Stanford mail"
To: nethax@shasta
Cc: mailhax@diablo
(This is a reply to mail sent to nethax. Mailhax@diablo is a more
appropriate forum for this topic.)
Brian and I evidently stand on opposite sides of the fence when it comes
to choosing between global principles and local pragmatics. Once a day we
both walk through the gate and again take up positions on opposite sides.
When I proposed hacking up Navajo's name to be "navajo" to solve a problem
I was having with getting uucp replies to work properly, Brian objected
strenuously as a matter of principle to this (very simple) method of fitting
into the benighted case-sensitive uucp world (as did everyone else, and rightly
so, the principle is a fine one). Now Brian objects to Mark because Mark's
principles override his concerns for local problems. Talk about the pot
calling the kettle black!
After considering both the principles and the pragmatics of the case problem
I came to the conclusion that the simplest solution (straight name change)
was unacceptable. In this case principle won out over pragmatics. (Keeping
the case of Navajo capitalized the way it is is almost PURELY a matter of
principle, with maybe just a touch of esthetics thrown in.)
I have considerable respect for principle (today at least - I won't as soon as
I have squeezed past Brian through this here gate). It seems that Brian's
objection to Mark is that he sets global principles above local problems.
With objections like that who needs letters of recommendation? If Mark ever
does need such a letter from Brian he need merely get the parity
right for the day on which he requests it.
My interpretation of what Mark is up to is that he is trying to solve a local
problem with the least disruption to global principles. I consider this to
be what first-rate engineering is all about! If Mark can actually pull that
off then by that standard he's a good engineer.
The attitude of "the rest of the world be damned, we have our own problems to
worry about" leads to WAITS. WAITS was put together by some very smart
people, and I'm sure if Stanford put its mind to it it could pull off another
WAITS fairly easily. It might even get classified as research. However such
research is more often than not misguided when done in the spirit of
"the rest of the world be damned," particularly in such areas as the standards
to adhere to in mail delivery.
-v
At the time WAITS was developed, there was no rest of the world. There is
no system presently extant - not TOPS-10, not TENEX, not ITS that existed
at the time WAITS was developed. Nor was a hardware configuration suitable
for time-sharing available from one manufacturer. WAITS started with the
DEC time-sharing (I forget what it was called) but had to modify to get
a reliable disk service and to serve displays. D.E.C. later made
modifications, some of which paralleled those of WAITS, to get TOPS-10.
The people who said, "the rest of the world be damned " were ITS and
BBN. BBN copied their SDS-940 system with improvements onto the PDP-10.
SAIL lost through not pursuing time-sharing development money and not
writing papers about the features of WAITS. At no time was WAITS
considered a research project.
∂01-May-83 2317 FEIGENBAUM@SUMEX-AIM IBM
Received: from SUMEX-AIM by SU-AI with PUP; 01-May-83 23:17 PDT
Date: Sun 1 May 83 23:18:11-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
Subject: IBM
To: jmc@SAIL
John, after your phone call on Saturday, I remembered that I had heard the
same story about IBM's attitude ("when it clearly becomes important, we'll
do it") from Bob Kahn, in regard to the meeting that Kahn and Cooper had
with IBM, during the period when Cooper was making his various site visits
to various places (including ours). You might want to check this outwith
Bob Kahn, who will be on campus on May 13 for a meeting that Lederberg
is chairing on the military applications of AI (Defense Science Board).
Ed
-------
∂02-May-83 0918 NILSSON%SRI-AI.ARPA@SCORE Lou Robinson Resignation
Received: from SU-SCORE by SU-AI with PUP; 02-May-83 09:18 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Mon 2 May 83 09:21:21-PDT
Date: 2 May 1983 0914-PDT
From: Nilsson at SRI-AI
Subject: Lou Robinson Resignation
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-AI,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-AI,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-AI,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL
Lou Robinson asked me to forward the following note to the excom
announcing that he wishes to resign from his position as Executive
Director of AAAI. I have discussed this matter with Lou and have
decided to accept it. I have also had some discussions with Claudia
Mazzetti about her willingness and ability to pinch-hit for Lou
until the excom can come to a decision about a replacement for
Lou. Claudia and I are confident that she can manage the transition,
and Lou has generously offered to help as needed.
Lou has also indicated that he would be interested in helping AAAI in
various capacities involving public relations, publications and other
related matters. Don Walker and I will be talking with Lou about
such things and will probably be coming to the excom soon with some
specific proposals.
Here is Lou's note:
Dear AAAI Executive Committee member:
I would like to resign as Executive Director of the American
Association for Artificial Intelligence. While I think that my
professional interests and abilities were exactly what the AAAI
needed during its formative years, I believe that the Association now
needs a person with a full-time commitment to the job.
I would like my resignation to take effect July 1, 1983. In one week I
will most probably undergo surgery. I expect the recuperation period
to be completed by the end of May at which time I intend to travel with
my family throughout China.
I would very much like to continue to offer my services to the AI
community and believe I can best do so outside the context of the
AAAI Office. I will make every effort to make the necessary transition
as smooth as possible.
Thank you all very much for your help in establishing AAAI as a strong
national scientific association whose membership is now close to 2,500.
A special note of thanks to Don Walker for his invaluable assistance
during AAAI's formative stages and thereafter.
And I doubt if anyone has ever really appreciated the role my wife,
Ann Robinson, has played in establishing AAAI as a valuable, viable
national scientific society.
Cordially,
Lou Robinson
==============
I would appreciate hearing from any of you about specific problem
areas that you anticipate during this transition. We'll do our
best to cover them.
Sincerely, Nils
-------
∂02-May-83 1022 BOSACK@SCORE [Ed Franceschini <FRANCESCHINI.CMCL1@NYU.ARPA>: RE: We will have a user in your area....]
Received: from SU-SCORE by SU-AI with PUP; 02-May-83 10:22 PDT
Date: Mon 2 May 83 10:25:44-PDT
From: Len Bosack <BOSACK@SCORE>
Subject: [Ed Franceschini <FRANCESCHINI.CMCL1@NYU.ARPA>: RE: We will have a user in your area....]
To: JMC@SAIL
This is the reply I received from my query about remote NYC access. At $15
per hour it may be cheaper to pay for calls to the CORADCOM TAC (at Ft.
Monmouth, NJ -- just south of NYC). I will send some mail to the CORADCOM
liason asking for information.
Len
---------------
Return-Path: FRANCESCHINI.CMCL1@NYU.ARPA
Received: from nyu.ARPA by SU-SCORE.ARPA with TCP; Mon 2 May 83 07:37:35-PDT
Date: 2 May 83 10:29 EST
From: Ed Franceschini <FRANCESCHINI.CMCL1@NYU.ARPA>
To: BOSACK@SU-SCORE.ARPA
Subject: RE: We will have a user in your area....
Cc: Ed Franceschini <FRANCESCHINI.CMCL1@NYU.ARPA>
Message-ID: <12239A464.017B0023;1983@CMCL1.NYU.ARPA>
In-Reply-To: Message of 22-APR-1983 18:50 from Len Bosack <BOSACK@SU-SCORE.ARP
The only access point in N.Y. currentlt is through the NYU facility.
We have a charge of $15.00 per connect time hour.
If you are interested let me have your address and I will send you the
paperwork.
(This provides only USER TELNET service. We do not provide a mailbox or
file space for FTP, for outside users.)
-------
-------
∂02-May-83 1037 BOSACK@SCORE Your NYC User
Received: from SU-SCORE by SU-AI with PUP; 02-May-83 10:37 PDT
Date: Mon 2 May 83 10:38:47-PDT
From: Len Bosack <BOSACK@SCORE>
Subject: Your NYC User
To: JMC@SAIL
Do you have a name/address/phone for your remote user? Among other things,
I need the phone number s/he will use to access the net from to get phone
line charges to various access points in the East.
Len
-------
∂02-May-83 1032 DFH Statistical Abstract of US
Govt. bookstore now has this in. Price is $11.
Check to be made out to Supt. of Documents.
--Diana
I see that SCORE has become very prompt in forwarding mail.
∂02-May-83 1943 RMS%MIT-OZ%MIT-MC.ARPA@SCORE Abstract for talk "How Flavors Differ from Smalltalk"
Received: from SU-SCORE by SU-AI with PUP; 02-May-83 19:42 PDT
Received: from MIT-MC.ARPA by SU-SCORE.ARPA with TCP; Mon 2 May 83 19:46:00-PDT
Date: Monday, May 2, 1983 10:34PM-EDT
From: Richard M. Stallman <RMS%MIT-OZ@MIT-MC>
Subject: Abstract for talk "How Flavors Differ from Smalltalk"
To: mccarthy at SU-SCORE
Many languages nowadays provide techniques for
defining abstract objects. The main dimension in which
they differ is in their provisions for inheritance in
between types. I will discuss the inheritance features
of flavors in the MIT Lisp machine system and their impact
on implementation.
Please schedule the talk for Tuesday or Thursday or Friday.
Monday and Wednesday are taken by gamelan playing.
∂02-May-83 1949 SIEGMAN@SU-SIERRA Re: C.I.T. and lunch
Received: from SU-SIERRA by SU-AI with PUP; 02-May-83 19:49 PDT
Date: Mon 2 May 83 19:45:06-PDT
From: Tony Siegman <SIEGMAN@SU-SIERRA>
Subject: Re: C.I.T. and lunch
To: JMC@SU-AI
cc: SIEGMAN@SU-SIERRA
In-Reply-To: Your message of Mon 2 May 83 16:58:00-PDT
John, I would be glad to -- although (a) having done what I've done thus
far in response to an invitation from the Academic Council COC to be one
of a number of "consultants" to them on a revised charge for the C-CIT,
I don't intend to be further involved in any active way in subsequent
developments; and (b) I gather there may be some announcements some
time soon -- though I am not privy to any details -- concerning some
sort of major adjustments in information technology planning at
Stanford. I guess we can watch the Daily for details.
-------
How about Wedneday, Thursday or Friday of this week?
∂03-May-83 0134 MRC@SCORE Re: Abstract for talk "How Flavors Differ from Smalltalk"
Received: from SU-SCORE by SU-AI with PUP; 03-May-83 01:33 PDT
Date: Tue 3 May 83 01:37:10-PDT
From: Mark Crispin <MRC@SCORE>
Subject: Re: Abstract for talk "How Flavors Differ from Smalltalk"
To: JMC@SAIL
cc: JJW@SAIL, ME@SAIL
Postal-Address: 725 Mariposa Ave. #103; Mountain View, CA 94041
Phone: (415) 497-1407 (Stanford); (415) 968-1052 (residence)
In-Reply-To: Your message of Mon 2 May 83 19:48:00-PDT
When reloaded, Score gets its time from the Ethernet, which has become
about 3 minutes fast. SAIL maintains its own timebase. So it is quite
possible for mail to appear to break the time barrier.
One new thing is that Score is quite capable of passing a relayed message
with no delay, unless there is a backlog of other "new requests" or if the
message couldn't be immediately delivered. In other words, it is quite
possible for a message to spend only 10 seconds or so at Score. At least
that's the way the software is supposed to work.
-------
∂03-May-83 0757 SIEGMAN@SU-SIERRA
Received: from SU-SIERRA by SU-AI with PUP; 03-May-83 07:57 PDT
Date: Tue 3 May 83 07:57:31-PDT
From: Tony Siegman <SIEGMAN@SU-SIERRA>
To: JMC@SU-AI
cc: SIEGMAN@SU-SIERRA
In-Reply-To: Your message of Mon 2 May 83 20:42:00-PDT
Any day is open -- I'd chose Thursday, if that's all right with you.
Would you like to meet in the Faculty Club lobby, or somewhere else?
-------
I have made a Faculty Club reservation for two at 12:15 on Thursday.
∂03-May-83 1555 BSCOTT@SCORE
Received: from SU-SCORE by SU-AI with PUP; 03-May-83 15:55 PDT
Date: Tue 3 May 83 15:53:56-PDT
From: Betty Scott <BSCOTT@SCORE>
To: JMC@SAIL
cc: BSCOTT@SCORE
In-Reply-To: Your message of Tue 3 May 83 14:34:00-PDT
Thank you, John. Then what I propose is the 5% acceleration on 6/1, and a
6.3% increase on 9/1. This total is much more than most people will have
between now and 9/1, given that we have a 6.5% control.
Betty
-------
∂03-May-83 1610 RINDFLEISCH@SUMEX-AIM Equipment Contract Rebudgeting
Received: from SUMEX-AIM by SU-AI with PUP; 03-May-83 16:10 PDT
Date: Tue 3 May 83 16:10:30-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Equipment Contract Rebudgeting
To: Ohlander@USC-ISI.ARPA
cc: Rindfleisch@SUMEX-AIM, Kahn@USC-ISI.ARPA, Adams@USC-ISI.ARPA,
Feigenbaum@SUMEX-AIM, JMC@SAIL, CSL.JLH@SCORE, Ullman@Diablo,
CSD.BScott@SCORE
Ron,
I finally have finished updating the Equipment Modernization contract
budget and have express mailed 2 copies to you through the Sponsored
Projects Office here (c/o Bob Kahn). Please forward it as necessary
to the DSSW negotiator. The new budget is for 2 years with allocations
of $2,562,300 in year 1 and $817,299 in year 2 ($3,379,599 grand total).
I've also listed below the major equipment items, with discount
information, that you requested to justify Stanford doing the purchase
rather than the government. Omissions include items being built
in-house, incidental equipment (net connections, installation, etc.),
minor test equipment, and personnel/support/maintenance costs.
Let me know if you need more information or have further questions.
Tom R.
MAJOR EQUIPMENT SUMMARY:
Year 1 Year 2
Symbolics Lisp machines (1 LM-2 subject
to trade-in and 8 3600's) $617,930
Xerox Lisp machines (8 1108's and
1 1132) 334,195
Central file server (DEC VAX 11/750,
1 TU78 tape, and 9 RA81 disk
spindles) 228,895
Systems research machine (DEC VAX
11/780, TU78 tape, and 3 RA81 disk
spindles) $234,783
Time-shared system (DEC 2060 system,
41% budgeted, remaining funding to
come from other sources) 420,000
VAX workstations (7 DEC 11/750's and
3 RA81 disk spindles) 320,550
SUN workstations (27 SUN Microsystems
SUN-1 100U systems) 396,240
Local net file servers (3 SUN
Microsystems SUN-1 150U systems with
169 Mbyte disks/tapes) 100,005
Laser printers (4 Xerox 8045 print
servers) 123,176
DISCOUNT INFORMATION:
Symbolics -- For the number of model 3600 machines we expect to buy,
the 20% Stanford discount is equal to the GSA discount. This discount
applies only to the basic Lisp machine. Options (e.g., memory, color
display, etc.) are not discounted under either arrangement.
The LM-2 is an interim machine that has allowed us to make substantial
research progress until the main purchase of model 3600 machines can
be made. It has been on a no-cost loan for the past 8 months (cleared
earlier with Duane Adams). We will purchase the LM-2 and recover 80%
of its purchase cost as a trade-in against the purchase of 5 3600
machines. This arrangement is available only for Stanford and through
GSA.
Xerox -- No GSA contract exists for the Xerox 1100 series of equipment.
Stanford qualifies for both a volume and educational discount on the
planned equipment and we have incorporated these -- 16.5% on the 1108
systems, 7% on the 1132 systems, and 8.5% on the printer systems. We
have figured these discounts based on the expected total number of each
machine we expect Stanford to buy over the next year from all funding
sources (DARPA, NIH, industry). Thus, each will mutually benefit from
the purchase volume.
DEC -- DEC does have a GSA contract and we budget all DEC equipment from
that schedule with the exception of the VAX 11/750 workstations. The
GSA discount (15%) exceeds the standard Stanford discount (12%). The
11/750 workstations are budgeted according to the special high discount
(61%) available to DARPA contractors.
SUN Microsystems -- SMI does not have a GSA contract. It does give
Stanford a 15% educational discount on which the budget is based.
-------
∂03-May-83 1627 SIEGMAN@SU-SIERRA
Received: from SU-SIERRA by SU-AI with PUP; 03-May-83 16:27 PDT
Date: Tue 3 May 83 16:24:27-PDT
From: Tony Siegman <SIEGMAN@SU-SIERRA>
To: JMC@SU-AI
cc: SIEGMAN@SU-SIERRA
In-Reply-To: Your message of Tue 3 May 83 11:48:00-PDT
OK, thank you, see you there.
-------
∂03-May-83 1654 RINDFLEISCH@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 03-May-83 16:54 PDT
Date: Tue 3 May 83 16:55:22-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
To: JMC@SAIL
In-Reply-To: Your message of Tue 3 May 83 16:12:00-PDT
Thanks, John. I just hope all the work gets rewarded with a contract
soon.
Tom R.
-------
∂03-May-83 1702 MAILER failed mail returned
The following message has expired without successful delivery:
∂30-Apr-83 1611 JMC
To: RMS@SU-AI
test
∂04-May-83 0921 AAAI-OFFICE@SUMEX-AIM Re: Lou Robinson Resignation
Received: from SU-SCORE by SU-AI with PUP; 04-May-83 09:21 PDT
Received: from SUMEX-AIM by SCORE with Pup; Wed 4 May 83 09:14:31-PDT
Date: Wed 4 May 83 09:10:27-PDT
From: NAME AAAI-OFFICE <AAAI-OFFICE@SUMEX-AIM>
Subject: Re: Lou Robinson Resignation
To: Nilsson@SRI-AI.ARPA
cc: Amarel@RUTGERS.ARPA, bobrow@PARC-MAXC.ARPA, buchanan@SUMEX-AIM,
BEngelmore@SRI-KL.ARPA, LErman@SRI-KL.ARPA, csd.genesereth@SCORE,
grosz@SRI-AI.ARPA, hart@SRI-KL.ARPA, csd.mccarthy@SCORE, mcdermot@YALE.ARPA,
minsky@MIT-AI.ARPA, rich@MIT-AI.ARPA, stan@SRI-AI.ARPA, gjs@MIT-AI.ARPA,
tenenbaum@SRI-KL.ARPA, walker@SRI-AI.ARPA, dwaltz@BBNG.ARPA,
bonnie.upenn@UDEL-RELAY.ARPA, AAAI-OFFICE@SUMEX-AIM
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
In-Reply-To: Your message of Mon 2 May 83 09:14:00-PDT
Some individuals have expressed concern regarding my resignation and
this year's AAAI Conference August 22 to 26 in Washington,D.C.
I am confident that this year's conference will be the most successful
one yet. The main,program part of the conference will be decided this
weekend. Speakers for the tutorial program have already been contracted.
The entire tutorial program has been defined and scheduled.
A technology transfer symposium has similarly been defined and
scheduled. The exhibit program will be very solid. A conference
decorator has already been contracted to deal with such things as
exhibitor electrical, telephone and furniture needs.
The following exhibitors have already been solicted and have indicated
that they would be exhibiting at the conference:
Xerox, Symbolics, Apollo Computer, Lawrence Erlbaum, NCR,
Smart Systems, Cognitive Systems, DEC, John Wiley, Addison-
Wesley, AI & DS, LMI, Ablex, MIT Press, Scion Corp., Comtex,
Microbot, Texas Instruments and Intelligenetics.
Others have shown a strong interest but cannot be considered as
definite as the above.
All elements of catering and local arrangements have been carefully
coordinated with the Washington Hilton's professional conference
staff including the marketing manager, conference scheduler and the
head of public relations there. Conference publicity pieces have
already been developed and distributed. These include such things
as press releases and a poster. A handsome conference brochure --
which will include all conference registration forms -- is now at the
printer's and will soon be distributed.
The formula for publishing and distributing proceedings -- through
Los Altos Publisher William Kaufmann, Inc. -- is already
firmly in place. This formula was tested for the first time in
Pittsburgh last year and proved very successful. As you know,
all AAAI and IJCAI proceedings are now being distributed through
William Kaufmann, Inc.
All of the mechanisms for a successful conference are firmly in place
and I have no hesitation in predicting that this year's conference will
be very successful.
Claudia Mazzetti, my assistant, has been working very closely with me on
all AAAI elements, and Nils and I feel absolutely confident in her
ability to follow through on all of the above.
Lou Robinson
-------
∂04-May-83 1000 JMC*
renewal
∂04-May-83 1600 JMC*
Keyworth
∂04-May-83 2121 RPG*
Pinnacles climbing guide?
-rpg-
Chuck Richards, Pinnacles Climbing Guide, 1974 is in my Concorde bag.
∂04-May-83 2125 ULLMAN@SCORE CF
Received: from SU-SCORE by SU-AI with PUP; 04-May-83 21:25 PDT
Date: Wed 4 May 83 17:08:34-PDT
From: Jeffrey D. Ullman <ULLMAN@SCORE>
Subject: CF
To: jmc@SAIL
It seems to me that you are the chairman of our computer facilities
committee, and that you can therefore start a policy discussion
of net charges without any action by me. Go to it; it's critical.
-------
∂04-May-83 2126 PERILLO%SRI-NIC.ARPA@SCORE Re: user in NYC
Received: from SU-SCORE by SU-AI with PUP; 04-May-83 21:26 PDT
Received: from SRI-NIC by SU-SCORE.ARPA with TCP; Wed 4 May 83 18:08:19-PDT
Return-path: <PERILLO@SRI-NIC>
Mail-From: PERILLO created at 3-May-83 16:49:14
Date: 3 May 1983 1649-PDT
From: Francine Perillo <PERILLO at SRI-NIC>
Subject: Re: user in NYC
To: JMC at SU-AI, nic at SRI-NIC
cc: PERILLO at SRI-NIC
In-Reply-To: Your message of 22-Apr-83 1550-PST
Remailed-date: 4 May 1983 1728-PDT
Remailed-from: Francine Perillo <PERILLO at SRI-NIC>
Remailed-to: mccarthy at SU-SCORE
Hello John,
Unfortunately, there are no TIPs (or TACs, as they are now called)
in the NYC area. Best we can do is advise you of one in NJ and
another at Griffiss AFB in NY state. The NIC does not distribute
TAC number; in fact, we don't even have any to distribute. The
host liaison of each TAC has that responsibility as well as
discretion to distribute it.
The liaison are as follows:
John Tugman DRSEL-MS-PE@OFFICE-7
Commander
Army Communications Electronics Command
Fort Monmouth, NJ 07703
(201) 532-1185
Robert Walker WALKER@RADC-MULTICS
Rome Air Development Center
Griffiss AFB, New York 13441
(315) 330-2501
Finally, if you know anyone at NYU's Computer Center they might
offer you or your friend an account on their machine which is
connected to the ARPANET.
Good Luck,
Francine /NIC
-------
∂04-May-83 2127 jlh@Shasta McSuns
Received: from SU-SHASTA by SU-AI with PUP; 04-May-83 21:26 PDT
Date: Wednesday, 4 May 1983 20:26-PDT
To: jmc at Sail
Subject: McSuns
From: John Hennessy <jlh@Shasta>
Since SMI Sun's have increased in price (along with just about every
other workstation-oriented terminal), the McSun approach (i.e. moderate
quality high-res terminals) makes increasing sense.
The Teletype BLIT terminal developed at Bell Labs is selling for
approximately $3400. While it needs a Unix system for development
support (i.e. writing new programs), I believe it can be used on any
system with downloaded code. The people at Bell Labs like them and have
developed much of the Sun functionality on them (although they are
definitely harder to program and more limited).
Comments?
∂04-May-83 2134 PERILLO@SRI-NIC Re: user in NYC
Received: from RANDOM-PLACE by SU-AI with TCP/SMTP; 4 May 83 21:34:36 PDT
Date: 3 May 1983 1649-PDT
From: Francine Perillo <PERILLO at SRI-NIC>
Subject: Re: user in NYC
To: JMC at SU-AI, nic at SRI-NIC
cc: PERILLO at SRI-NIC
In-Reply-To: Your message of 22-Apr-83 1550-PST
Hello John,
Unfortunately, there are no TIPs (or TACs, as they are now called)
in the NYC area. Best we can do is advise you of one in NJ and
another at Griffiss AFB in NY state. The NIC does not distribute
TAC number; in fact, we don't even have any to distribute. The
host liaison of each TAC has that responsibility as well as
discretion to distribute it.
The liaison are as follows:
John Tugman DRSEL-MS-PE@OFFICE-7
Commander
Army Communications Electronics Command
Fort Monmouth, NJ 07703
(201) 532-1185
Robert Walker WALKER@RADC-MULTICS
Rome Air Development Center
Griffiss AFB, New York 13441
(315) 330-2501
Finally, if you know anyone at NYU's Computer Center they might
offer you or your friend an account on their machine which is
connected to the ARPANET.
Good Luck,
Francine /NIC
-------
∂04-May-83 2145 @USC-ECLC:LLW@S1-A Re-Transmitting. . .
Received: from RANDOM-PLACE by SU-AI with TCP/SMTP; 4 May 83 21:44:34 PDT
Mail-from: ARPANET site S1-A rcvd at 30-Apr-83 1407-PDT
Date: 30 Apr 1983 1406-PDT
From: Lowell Wood <LLW at S1-A>
Subject: Re-Transmitting. . .
To: jmc@sail at USC-ECLC
CC: LLW at S1-A
∂30-Apr-83 0123 LLW Wednesday Evening
To: jmc at SU-AI
CC: LLW at S1-A
∂29-Apr-83 1723 John McCarthy <JMC@SU-AI> dinner?
Received: from SU-AI by S1-A with NCP/FTP; 29 Apr 83 17:23:11 PDT
Date: 29 Apr 83 1722 PDT
From: John McCarthy <JMC@SU-AI>
Subject: dinner?
To: llw@S1-A
Are you interested in dinner before or after your panel with Panofsky.
There may be ARPAnet problems in communicating, because Marty plans
to switch us to TCP tonight.
Unsolicited advice:
Keep discussion with Panofsky friendly. Also imagine that there are
potential recruits for helping with your defense research in the audience.
I think the President's appeal for scientific help will have some effect.
[John: I got a note while I was out of my office today indicating that I
am expected to show up for some pre-hostilities dinner with the folks who
are staging this extravaganza on Wednesday. As soon as I can find out on
Monday whether it's obligatory or optional, I'll be back in touch
(somehow--perhaps we can persuade Marty to install TCP for us soon as it's
been determined to be at least semi-functional at SAIL). Thanks for the
unsolicited advice--it seems quite sound. It's hard to win many points
roughing up a Grand Old Man, and besides, Panofsky seems like a rather
decent guy--not even Edward can manage to get very annoyed with him. I
have every intention of quite baldly appealing for recruits--it's the main
reason I decided to squander an evening preaching to the infidels! Hope
to see you sometime Wednesday evening. Lowell]
∂04-May-83 2326 RMS%MIT-OZ%MIT-MC.ARPA@SCORE My abstract
Received: from SU-SCORE by SU-AI with PUP; 04-May-83 23:26 PDT
Received: from MIT-ML.ARPA by SU-SCORE.ARPA with TCP; Wed 4 May 83 23:12:24-PDT
Date: 5 May 1983 0205-EDT
From: RMS%MIT-OZ@MIT-MC
Subject: My abstract
To: mccarthy@SU-SCORE
Did you get it? Is it ok? Have you found a place
I can stay? When is the talk going to be?
Have you called Fateman or has he called you?
-------
The abstract is ok, and you will hear from RPG about time of talk and
possibly a place to stay. I haven't heard from Fateman.
Does the following mean that TCP incoming mail is working, and I can
have the M.I.T. people mail directly to SAIL?
∂05-May-83 0015 @USC-ECLC,@MIT-MC:GAVAN%MIT-OZ@MIT-MC temporary mail routing
Received: from RANDOM-PLACE by SU-AI with TCP/SMTP; 5 May 83 00:14:02 PDT
Received: from MIT-MC by USC-ECLC; Sun 1 May 83 19:15:54-PDT
Date: Sun, 1 May 1983 22:14 EDT
From: GAVAN%MIT-OZ@MIT-MC
To: JMC%SU-AI@USC-ECLC
Cc: jcma%MIT-OZ@MIT-MC, cstacy%MIT-OZ@MIT-MC
Subject: temporary mail routing
In-reply-to: Msg of 01 May 83 1200 PDT from John McCarthy <JMC at SU-AI>
From: John McCarthy <JMC at SU-AI>
SAIL is now running TCP, but mail doesn't work yet.
Mail for jmc-lists@sail should be routed through SCORE.
When our mail is working, I'll send another message.
Right now, we're sending to you through USC-ECLC. Let me know if this
message reaches you, so that we'll know that other messages are
reaching you. Do you prefer that we route through SCORE? Let us know
when SAIL's TCP is working correctly, and we'll switch.
Routing through ECLC is pointless now, because we no longer are on
NCP. This message was received by SMTP from ECLC, but Martin Frost says SMTP
is still flakey. However, I suppose you might as well start sending
the messages directly to SAIL, since this one worked. SCORE
forwards by Ethernet, so that doesn't depend on our SMTP working.
∂05-May-83 0026 ME mail
Incoming network mail under TCP (using the mail protocol called SMTP)
is indeed up and working, but I'm not going to announce it or encourage
people to use it until I fix a couple of minor things in it. Also,
outgoing SMTP mail is NOT up, so MAIL still routes messages through Score.
Probably both directions of SMTP will be up officially in a couple of days.
∂05-May-83 0245 @USC-ECLC,@MIT-MC:RJF@MIT-MC
Received: from RANDOM-PLACE by SU-AI with TCP/SMTP; 5 May 83 02:45:09 PDT
Received: from MIT-MC by USC-ECLC; Mon 2 May 83 23:46:20-PDT
Date: 3 May 1983 02:30 EDT
From: Richard J. Fateman <RJF @ MIT-MC>
To: JMC @ MIT-MC
RMS (Richard Stallman) sent me a note indicating you wnated to get
Macsyma. Certainly you should be able to get a copy for a DEC-10 or -20
for Stanford; I hope also for VAX systems, soon (cost $500 from Symbolics
for the latter; if Symbolics can't supply it, or the cost is too high,
UC Berkeley can send a tape.)
Progress on getting a "public domain" copy depends on how MIT stonewalls
the Dept of Energy.
Feel free to give me a call at 415 642-1879.
Richard Fateman
∂05-May-83 1307 DFH additional phone
LGC has requested a phone extension on his
desk. Since Paul Wieneke will also be using
the desk shortly, and the phone already in the
room is paid for by HPP, I think it would be best
just to get another phone line for them. If this is
OK with you, I will put in the order and charge it to
the ARPA account.
--Diana
OK about LGC phone.
∂05-May-83 1524 minker.umcp-cs@UDel-Relay Lerner's 7oth Birthday and IJCAI
Received: from RANDOM-PLACE by SU-AI with TCP/SMTP; 5 May 83 15:23:54 PDT
Date: 5 May 83 12:17:39 EDT (Thu)
From: JACK MINKER <minker.umcp-cs@UDel-Relay>
Return-Path: <minker.umcp-cs@UDel-Relay>
Subject: Lerner's 7oth Birthday and IJCAI
To: JMC@SU-AI, "uw-beaver!ubc-vision!Reiter"@LBL-UNIX
Cc: Amarel@Rutgers
Via: UMCP-CS; 5 May 83 18:12-EDT
Dear John and Ray,
This September Alexander Lerner will be 70 years old.
I would like to do something to honor his birthday both at the
AAAI and IJCAI. The AAAI Committee will, shortly, make a decision
on a recommendation that I made. Since you are on the Committee,
I hope that you will support it. As far as IJCAI is concerned,
Amarel has been unable to get the Committee to do anything
officially. This is not surprising. I would like to propose the
following and hope that I receive your concurrence.
There is an open evening at IJCAI. I would like to have
a meeting outside of IJCAI, to honor Lerner. It would be low key.
I would give a short description of his scientific accomplishments
and his current situation. This would be followed by some talks
on an AI matter. In particular, since we all have done
work in non-monotonic logic, we could have brief statements on what we
have each done, and a discussion of where non-monotonic logic fits
into the AI picture.
If the topic is acceptable to you, would you be willing to
participate? Would you rather have another topic? Would you rather
have a different way of honoring him? Would you rather not participate
at all?
If anything is to be done, it will have to be soon. I would like
to receive your comments.
Best regards,
Jack
mail minker.umcp-cs@UDel-Relay,"uw-beaver!ubc-vision!Reiter"@LBL-UNIX,amarel%rutgers
lerner
A small session on non-monotonic reasoning would be fine, especially
as I have something new to say. Having it in honor of Lerner is also
fine. I suggested to Saul that it might be worthwhile to have Lerner
added to some appropriate already planned IJCAI panel and invite him
to come. If they chanced to let him out, having an extra panel
member wouldn't be a serious embarassment. Given the probable outcome,
you can protest.
∂06-May-83 1519 RESTIVO@SCORE latest Prolog
Received: from SU-SCORE by SU-AI with PUP; 06-May-83 15:19 PDT
Date: Fri 6 May 83 14:55:22-PDT
From: Chuck Restivo ( The Moderator ) <RESTIVO@SCORE>
Subject: latest Prolog
To: jmc@SAIL
Date: Fri 6 May 83 14:31:40-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Prolog utilities
To: Restivo@SU-SCORE,Prolog-Hackers@SU-SCORE
I've just brought up at SRI the complete set of Edinburgh utilities.
This is an extensive collection, including an algebraic simplifier,
an arbitrary precision rational arithmetic package, a formatted write
procedure, and many other goodies. Most of the code is well documented,
and there are a few help files as well. I have to check with Edinburgh
whether I can let people have this stuff freely (I believe I can).
The utilities were written for DEC-10/20 Prolog, and therefore some
of them are only practical on a high-performance system (for your
information, the DEc-10/20 system does 40000 LIPS, has tail recursion
optimization and stack garbage collection, which allows determinate
computations to run with an efficiency comparable with that of similar
algorithms in MACLISP).
Fernando
-------
cwr - this version should be running at SCORE by the end of the weekend.
-------
∂06-May-83 1524 DFH air reservations
I made reservations as follows:
Thurs May 12 SFO/Boston United 94 leave 1:30pm ar.9:45pm
Sun May 15 Boston/SFO TWA 847 leave 6:30pm ar. 9:31 pm
There is a non-stop that leaves at 9 am Thurs morning and
arrives in Boston at 5:17 pm if you prefer. Let me know if
you need to change anything. Do you need a hotel reservation?
--Diana
∂06-May-83 1833 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 6 May 1983
Previous Balance 0.65
Monthly Interest at 1.5% 0.01
Current Charges 0.30 (vending machine)
-------
TOTAL AMOUNT DUE 0.96
Please deliver payments to Diana Hall, room 358, Jacks Hall.
Make checks payable to: STANFORD UNIVERSITY.
To ensure proper crediting, please include your Pony account name on your check.
Bills are payable upon presentation. Interest of 1.5% per month will be
charged on balances remaining unpaid 25 days after bill date above.
You haven't paid your Pony bill since 12/82.
Accounts with balances remaining unpaid for more than 55 days are
considered delinquent and are subject to reduction of credit limit.
Please pay your bill and keep your account current.
∂07-May-83 1244 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN essences
Received: from RANDOM-PLACE by SU-AI with TCP/SMTP; 7 May 83 12:43:47 PDT
Date: 7 May 1983 1321-EDT
From: RICKL%MIT-OZ%MIT-MC@SU-DSN
Subject: essences
To: dam%MIT-OZ%MIT-MC@SU-DSN, batali%MIT-OZ%MIT-MC@SU-DSN, jmc@SU-AI
cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
Date: Thu, 5 May 1983 13:49 EDT
From: DAM@MIT-OZ
Subject: Essential Properties
Well consider a
glass of water. Suppose that I do not know that water is H2O and
therefore I consider it to be a natural kind. I do know however that
the water has a certain mass and that this mass is an essential
property of the water in the glass. On the other hands the fact that
the water is in the glass is a contextual property (I could pour the
water into a different glass and it would still be the same water).
Thus even though I don't know the "real essence" of the water I can
distinguish essential from contextual properties of the water.
However, it is only by reference to a theory that you believe that the
mass is an essential property of the water in the glass. If you
believed a different theory you might also believe that weight was an
essential property --- e.g. like the pre-Newtonian mechanists, who
identified it with mass. Alternatively you might believe a theory, like
relativity, which says that mass is a contextual property depending on
velocity. Even in your symmetric sets, you identify those properties as
essential which your *theory of symmetric sets* tells you are essential,
even though there you are able to rely on an even stronger theory which
is capable of supporting necessary & sufficient conditions. This
theory-relative nature of essential properties, I take it, is part of
JMC's point:
Date: 06 May 83 1153 PDT
From: John McCarthy <JMC@SU-AI>
Subject: natural kinds
Essential properties on the other hand seem irrevocably
epistemological and linguistic; they relate to conventions
whereby people will maintain communication in the face of
hypothetical new information. They can be only partially
successful.
More than merely communication about hypothetical new information,
however, what we take to be essential properties determines the
conceptual boxes into which we will try to fit this new information.
The fact (well-established, I think) that this can only be partially
successful means that the conceptual boxes (theoretical terms) provided
to us by our theories will never (or rarely) map over unchanged as our
theories change and evolve; this linguistic change is well-supported
historically. I think that scientists are pretty much in the position
of BATALI's robot
From: BATALI@MIT-OZ
Subject: Essential Properties
On the other hand, it might be a reasonable committment for a robot to
make that objects have such essential properties, but that it could
never be sure what they are. In this way, the robot could understand
that whatever particular nominal essense it had been using could be
wrong, and it has to try again to approximate the "real essense."
recognizing that their theories could be wrong (though I think he is
distinguishing "nominal" & "real" essences in a way which JCM & DAM are
not) --- the linguistic/taxonomic component by which these theories
describe the world is inextricably bound up with the theories which
employ them, & the robot can also use the observation that some of the
terms/essences have been gotten wrong as an indication that some of the
theories which employ them may need to be revised, & vice versa.
-=*=- rick
-------
∂07-May-83 1247 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN
Received: from RANDOM-PLACE by SU-AI with TCP/SMTP; 7 May 83 12:46:56 PDT
Date: Fri, 6 May 1983 22:29 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: JMC@SU-AI
Cc: jcma%MIT-OZ%MIT-MC@SU-DSN, cstacy%MIT-OZ%MIT-MC@SU-DSN
In-reply-to: Msg of 05 May 83 0037 PDT from John McCarthy <JMC at SU-AI>
Date: 05 May 83 0037 PDT
From: John McCarthy <JMC at SU-AI>
To: gavan%oz%MIT-MC at SU-SCORE
Received: from SAIL by SCORE with Pup; Thu 5 May 83 00:40:45-PDT
Routing through ECLC is pointless now, because we no longer are on
NCP. This message was received by SMTP from ECLC, but Martin Frost says SMTP
is still flakey. However, I suppose you might as well start sending
the messages directly to SAIL, since this one worked. SCORE
forwards by Ethernet, so that doesn't depend on our SMTP working.
Mail should now reach you directly. Did this get through?
Thanks, Gavan. Your message dated Fri, 6 May 1983 22:59 EDT arrived
at 7 May 1983 12:46:56. It would be nice to know the cause of the
delay, but unless the delay is a known one at your end, it is better
to wait a week or two for further debugging of TCP here before chasing
problems.
∂07-May-83 1253 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN metaphor
Received: from RANDOM-PLACE by SU-AI with TCP/SMTP; 7 May 83 12:53:27 PDT
Date: Fri, 6 May 1983 23:36 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: metaphor
In-reply-to: Msg of 05 May 83 0110 PDT from John McCarthy <JMC at SU-AI>
From: John McCarthy <JMC at SU-AI>
I suppose that metaphors are sometimes of use in generating scientific
hypotheses, although I am not conscious of using them myself.
Well, I think they're always of use in hypothesis-formation. It's
possible, of course, that there might be some other algorithm. If you
know of another, I'd be happy to hear it.
However, the metaphors are not the hypotheses themselves. It seemed
to me that the discussion which I criticized presented the metaphors
themselves rather than going on to definite hypotheses. For example,
the "society of mind" metaphor went on airily without anyone noticing,
until Marvin pointed it out, that the society to which he was imagining
the mind analogous was more like an anthill than like human society.
Marvin's view was that the parts did not individually have minds or
intelligence. Well the anthill analogy is fine as far as it goes,
but to go beyond the analogy requires yet more specific hypotheses.
Well I had difficulty understanding how Marvin's description of the
society envisioned differs all that much from human society. Human
societies and anthill societies share many features, one of which is
elite domination of the mass (the iron law of oligarchy). What do you
mean by "minds" and "intelligence"? Certainly the parts must be smart
about something if they are to perform their jobs successfully. I
think it was also Marvin who once (correct me if I'm misrepresenting
you) characterized mind as a (group of) machine table states. If the
parts are machine table states (or substates) in what way are they not
intelligent?
Total precision is not quite the issue;
Thank you.
it seemed to me that the discussion was becoming ever more airy and
that the emphasis on metaphor was a big part of the problem.
Well this sometimes happens, but I don't think that was the thrust of
the discussion. I agree also that formulating hypotheses is bound to
be more fruitful than the "airy" bandying-about of metaphors. So
here's an hypothesis: Intelligent agents will need to be able to
formulate hypotheses in order to serve some goal more effectively and
efficiently. They will formulate hypotheses on the ground of
metaphorical matches. No match will ever be perfect (or precise).
Any such match will need to be debugged deductively.
I'm suggesting (for example) is that you really can't inductively
build a class structure without some constraint on the potential
candidates for membership in particular classes. I would contend that
this constraint can be characterized as metaphorical, in that the
candidate will resemble other members of the class in some respects
and will differ from them in other respects. The problem of abduction
(or, more "precisely", the problem of an abduction used as a
preliminary to induction in a class structure) is the problem of
deciding which are the candidates for induction. The candidates will
be METAPHORS in that they will share some (and not share other)
features which are (almost) always present among the members of the
class. There's a tension between the target and the base in a
metaphorical match that can be exploited (as a constraint) both in the
class and the candidate. The candidates' shared features (properties,
relations) with members of the class count as evidence for its
inclusion. Features of the candidates not shared by the class members
mitigate against inclusion. They might also motivate the
reformulation of the universal (or the class and its relationships to
other classes). More can be done, of course, such as searching for
compatibilities and non-compatibilities (with the behavior of class
members in particular contexts) of the candidates' unique features and
their implications (extracted from the universals of the features
themselves).
All this is to say that some features must be considered more
"essential" than others. Nevertheless, there are NO purely essential
features. I think this is what Wittgenstein means by "family
resemblance."
Metaphors are imprecise. Discussions about them can seem airy.
Nevertheless, they're an important part of reasoning. Intelligent
programs will need to be able to make metaphors. So they'll have to
be engineered. So it makes eminent sense to discuss them.
∂07-May-83 1651 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN
Received: from MIT-MC by SU-AI with TCP/SMTP; 7 May 83 16:51:35 PDT
Date: Sat, 7 May 1983 19:48 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
In-reply-to: Msg of 07 May 83 1317 PDT from John McCarthy <JMC at SU-AI>
Date: 07 May 83 1317 PDT
From: John McCarthy <JMC at SU-AI>
To: gavan%oz%MIT-ML at SU-SCORE
Received: from SAIL by SCORE with Pup; Sat 7 May 83 13:21:43-PDT
Thanks, Gavan. Your message dated Fri, 6 May 1983 22:59 EDT arrived
at 7 May 1983 12:46:56. It would be nice to know the cause of the
delay, but unless the delay is a known one at your end, it is better
to wait a week or two for further debugging of TCP here before chasing
problems.
As you can see, the random bits are still being appended to your mail.
The delay may indeed be at our end, due to the cruftiness of the TOPS-20
mailer, which does not compare with the smarts of ITS's COMSAT.
∂07-May-83 1716 @MIT-MC:BATALI%MIT-OZ%MIT-MC@SU-DSN Unnatural kinds
Received: from MIT-MC by SU-AI with TCP/SMTP; 7 May 83 17:15:56 PDT
Date: Sat, 7 May 1983 20:08 EDT
From: BATALI%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Cc: batali%oz%MIT-ML@SU-SCORE, dam%oz%MIT-ML@SU-SCORE,
phil-sci%oz%MIT-ML@SU-SCORE, rickl%oz%MIT-ML@SU-SCORE
Subject: Unnatural kinds
In-reply-to: Msg of 06 May 83 1153 PDT from John McCarthy <JMC at SU-AI>
From: John McCarthy <JMC at SU-AI>
I think the best way of thinking about natural kinds is as
a fact of physics.
It is a fact about the world that cows aren't
continuous with horses, and 1 isn't continuous with 2. We can
imagine another world in which we would have to establish an
arbitrary boundary between cows and horses, just as, in our
world, we would need to be arbitrary in order to distinguish
hills from mountains. Perhaps there is a planet somewhere in
with terrain in which there is a sharp distinction.
It is a fact about US that we have an easy time distinguishing cows
from horses. And the reason that we have such an easy time doing so
may have to do with (ultimately physics) facts about cows and horses.
I don't see how this implies that there are any necessary and/or
sufficient conditions for cow or horse hood. I don't think that
"discontinuity" is a good way to characterize natural kinds either,
because it requires some notion of the measure along which the
discontinuity falls. There are lots of measures along which horses
and cows are continuous -- electric charge, tendency to be called
"Bessy" and so on. It seems that the intuitions about discontinuity
assumes a measure like "ability for humans to distinguish".
The key epistemological property of natural kinds is that
the kind has properties that we don't a priori know about and
can subsequently learn. Amusingly, even when objects are not
natural kinds, the concepts and words for them often are.
Whether something is natural kind (if there are such things) or not,
it is possible that we don't know all there is about it. Mathematical
objects are a good example. A mathematicial typically defines a class
of objects precisely and formally. It may still take a lot of work to
discover facts about the class of objects. Or a non-math example:
let us take the class of all white sneakers. Certainly this is not a
natural kind (?). There is a number R such that R is the percent of
the sneakers with a hole over the left big toe. Clearly it will take
a lot of work to determine the value of R.
All in all, I think that "natural kinds" are natural more to the
degree that we "naturally" distunguish them and use concepts refering
to them, than to any facts about physics.
natural kinds
I'll try one more argument about natural kinds being objective - at
least more so than essential properties. Can you imagine a computer
program or some other intelligent animal to which there would be
a clear distinction between hills and mountains? There are multiple
objective differences between cows and horses, and different experience
will determine which are taken as defining. A race that classified
solely on the basis of mass and which therefore didn't make
a sharp distinction would have to be feeble minded. There are physical
distinctions and the genetic distinction that cows only come from
cows and horses from horses.
It seems to me that taking a subjective position on cows and horses
leads to taking a subjective position on almost all distinctions and
therefore two an impoverished science and an impoverished robotics.
∂07-May-83 1733 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN metaphor
Received: from MIT-MC by SU-AI with TCP/SMTP; 7 May 83 17:30:35 PDT
Date: 7 May 1983 2021-EDT
From: RICKL%MIT-OZ%MIT-MC@SU-DSN
Subject: metaphor
To: kdf%MIT-OZ%MIT-MC@SU-DSN, levitt%MIT-OZ%MIT-MC@SU-DSN, jmc@SU-AI,
isaacson%USC-ISI@SU-DSN
cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
From: KDF@MIT-OZ
Subject: Isolation and approximation
From: LEVITT@MIT-OZ
.... simply because we hope something so
familiar will be applicable, and we don't know what else to do yet.
We pull simple linear and second-order models out of a Bag of
"general-purpose" descriptive Tricks.
The space of all equations in all possible parameters
is so large that we should venture into it only when we have isolated
a small subset of it as likely to be relevant. Just how to do that,
of course, is still shrouded in mystery.
The other is, any complex-but-narrow program might be a worthwhile
part of a model organized as a bag of tricks, and a mind or scientist
that tries different ones.
Right. Eventually we may be able to put some more structure on the
bag, to say just what techniques can be applied when. By analogy with
low-level vision....
Thank you both for providing an excellent practical illustration of the use
of analogy and metaphor in research and discovery. Here the analogical base
is low-level vision, the metaphorical base a Bag of Tricks with structure,
and the target our understanding of Science. When we don't know what else
to do yet, surely one powerful mechanism in the Bag of Tricks (or perhaps
more formally, "collection of discovery heuristics") is to hope some familiar
situation will be applicable. Considering the small subset of things *known*
to be relevant in the base may suggest a small subset of things *likely* to
be relevant in the target, as well as giving a first cut at a possible
taxonomy, likely "essential features", and expectations of domain behavior.
Furthermore, any important area of human cognition which is still shrouded
in mystery is surely of interest to us as artificial intelligence researchers.
Date: 05 May 83 0110 PDT
From: John McCarthy <JMC@SU-AI>
Subject: metaphor
the "society of mind" metaphor went on airily without anyone noticing,
until Marvin pointed it out, that the society to which he was imagining
the mind analogous was more like an anthill than like human society.
Those parts of the society of minds which depend critically upon the
intelligence of the agents being no greater than ants will probably
not be applicable. Those parts of the theory of water dealing with
solubility don't apply to electrical theory, either, because chemicals
don't dissolve in beakers of electrons --- but as DAM pointed out, the
differential equations of water flow and electricity have the same form.
Those parts of the society of minds which emphasize Rational Action (agent
as doer) instead of Rational Belief (agent as thinker) probably won't
apply either, as MONTALVO.HP-HULK (name?) pointed out recently.
Those parts of the society of minds which **don't** depend critically
upon ant-like agents, but instead depend more on the activity of parallel
knowledge-structuring agents, *are* suggested as *likely* to be relevant.
Subject: Re: developmental view of science
From: ISAACSON at USC-ISI
To: RICKL%MIT-OZ at MIT-ML
I leafed through your notes which are quite extensive. It will
take me some time to digest the stuff. One point, though, I wish to
comment on right now.
RE margin note <p. 72, C-Lines>, margin note <p. 85, par 1>,
and part of your message to DAM & Phil-Sci dated 28 Apr 83,
relating to scientific journal systems and Minsky's C-lines.
A few weeks ago, Derek de Solla Price of Yale gave here a talk on
"Citation Mapping, Computers and Artificial Intelligence". [He
is said to hold degrees in experimental physics and history of
science (Cambridge). He is Avalon Professor of the History of
Science at Yale.] Although I'd discount his views on computers
and AI, he presented some very interesting data from his
researches in Library Science. He discovered graph-structures
(networks), based on *real* data, that essentially support what
you say about the scientific journal and archiving system. You
may wish to contact him for his data and, perhaps, some reprints.
These margin notes are in oz:ps:<rickl>margin.txt, and I can send them
to people who live in net-land.
From: John McCarthy <JMC@SU-AI>
However, the metaphors are not the hypotheses themselves. It seemed
to me that the discussion which I criticized presented the metaphors
themselves rather than going on to definite hypotheses....
the anthill analogy is fine as far as it goes,
but to go beyond the analogy requires yet more specific hypotheses.
I agree --- I think that one of the main uses of metaphor is as a guide
to future research directions, or in attempting to understand a
problematic domain. N.B. there is a *big* difference between metaphor
and analogy --- JCMa has been working on this and perhaps he will say
something about it --- but I presently take "analogy" to imply a much
tighter coupling between base and target than analogy; thus I think that
much scientific reasoning can be analogical without being imprecise, and
also that analogical reasoning *does* *not* *exclude* deductive
reasoning (I have said that before); I think that the discussion on
society of minds and science was "metaphorical", which is however
***not*** the same as useless.
I actually thought (and still do) that I ***was*** proposing more
specific hypotheses in my original communication to the net (15 April).
In the interests of steering the discussion back in a direction you
would consider more fruitful, in virtue of embodying more specific
hypotheses for understanding science, I will reprint them here (they
have grown from 10 points to 11, but the thrust is the same). Of
course, they represent *preliminary* *hypotheses* about a *very* poorly
understood domain.
People who remember can stop reading here. People who don't should
also see DAM's response (15 April).
-=*=- rick
================================================================
I want to propose a computational interpretation something like the
following:
1) "paradigm" be given a technical definition in terms of a
small related set of concrete scientific achievements, canonical
experiments, and practical illustrations;
2) scientific reasoning occurs by applying these to novel
situations, by constructing an analogical mapping between pieces of the
novel situation and examples from the paradigm;
3) scientific language is a construct for describing scientific
reasoning; a term derives meaning from its place and function within
the structure of (1-2); the mapping (2) only then goes to the "real
world" with a term;
4) the paradigm is annotated to constrain and guide the matcher,
including particularly "cause" relationships;
5) "normal science" consists in extending the annotation to apply
to related situations, or more conveniently (perhaps lower computational
cost?);
6) "normal science" experiments consist in determining which of
the few possible matches to the novel situation, given the paradigm, is
in fact the case in our world (e.g., mapping "cause" relationships);
7) the number of possible matches to experimentally discriminate
among provides one criterion to guide proposed research directions:
known result (only one possible match); promising research area (a few);
too hard, poorly understood (too many); "unscientific" (none);
8) "revolutionary science" is the process of finding a better (=
more power, lower computational cost) paradigm; "sparseness" argues for
a revolutionary rather than incremental change;
9) axiomatization and construction of formal logico-deductive
systems, when it occurs, is always after the fact (i.e., after the
domain is already firmly accepted "science"), and merely formalizes what
is by then accepted knowledge;
10) induction is rarely applied directly to the results of
experiments --- usually they are designed to answer particular questions
about the paradigm's match to a particular novel situation; however,
the annotation (3-5) *is* frequently the subject of induction;
11) science education is a necessary prerequisite to science
practice, being the structured presentation of a series of paradigms to
a student capable of learning precedents by example.
I have been thinking of a developmental view of Science which is along
the lines of Kuhn's treatment --- i.e. that scientific development
occurs in stages, guided by paradigms which embody specific examples of
concrete scientific achievement. I believe much scientific reasoning,
particularly search and discovery, is of the form of Winston-like
analogical reasoning; and that there are specific constructs in
scientific paradigms (short-hand for a cluster of concepts, largely
embodied in examples) to support this.
-------
∂07-May-83 2001 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN failed mail return
Received: from MIT-MC by SU-AI with TCP/SMTP; 7 May 83 20:01:28 PDT
Date: Sat, 7 May 1983 23:00 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Cc: jcma%MIT-OZ%MIT-MC@SU-DSN
Subject: failed mail return
In-reply-to: Msg of 06 May 83 1331 PDT from John McCarthy <JMC at SU-AI>
From: John McCarthy <JMC at SU-AI>
When I MAIL to phil-sci, I get failed messages back aimed at recipients
I never heard of. I presume this is because of wrong entries on the
phil-sci mailing list. Can I avoid this by only copying phil-sci?
I've found two bad entries in the list and have corrected them.
Please let us know whenever you receive undeliverable phil-sci mail.
Thanks.
∂07-May-83 2211 @MIT-MC:KDF%MIT-OZ%MIT-MC@SU-DSN visit
Received: from MIT-MC by SU-AI with TCP/SMTP; 7 May 83 22:11:41 PDT
Date: Sun, 8 May 1983 01:07 EDT
From: KDF%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Subject: visit
In-reply-to: Msg of 07 May 83 1014 PDT from John McCarthy <JMC at SU-AI>
Sounds like fun. I'm in afternoons and evenings, due to
machine availability.
Ken
∂08-May-83 0247 RMS BBOARD message
I was just editing a bboard message to raise the issue
of sharing programs, and whether Stanford ought to do so.
It is in [1,rms]msg.txt. I do not know how to install
that as a bboard message. Can yo do that for me?
My disk booted ok, but I haven't played with it.
I spent the time writing this message and talking
to some students instead.
∂08-May-83 1528 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN Re: natural kinds
Received: from MIT-MC by SU-AI with TCP/SMTP; 8 May 83 15:27:47 PDT
Date: 8 May 1983 1816-EDT
From: RICKL%MIT-OZ%MIT-MC@SU-DSN
Subject: Re: natural kinds
To: Batali%MIT-OZ%MIT-MC@SU-DSN, batali%MIT-OZ%MIT-MC@SU-DSN, jmc@SU-AI,
dam%MIT-OZ%MIT-MC@SU-DSN
cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
In-Reply-To: Your message of 8-May-83 1509-EDT
Date: Sunday, 8 May 1983, 14:59-EDT
From: John Batali <Batali%MIT-OZ@MIT-MC>
Subject: natural kinds
I accept the
fact that there are objective distinctions between cows and horses.
What we could tell it [the robot]
that would be of some help would be that cows and horses
are "discontinuous" in the way that hills and mountains are not. This
would allow it to assume that pro-cow evidence is anti-horse evidence
and thus give it a way to form a differential recognition procedure for
large herbivore mammals.
If we agree that cows and horses are objectively discontinuous, then
cows and pigs are discontinuous in the same way that cows and horses are;
so are cows and rabbits, cows and mountains, and in fact cows and anything
else. Pro-cow evidence is anti-rabbit evidence just as much as anti-horse
evidence, etc. But if cows are discontinuous with everything else then
they fall into their own little clump separated by an (objective)
discontinuous gap from everything else. Do we run into any problems if
we take this to be the defining "essential property" of natural kinds?
For example, mountains don't fall into a clump with a discontinuous gap,
because there are hills arbitrarily close; but rabbits would for the
same reasons cows do.
This works very well for static theories but encounters problems as
soon as theories undergo change. An example: before Copernicus every
educated person knew that there were three natural kinds in the
heavens, based on objective discontinuities --- (1) the planets, which
moved in regular patterns and included the sun and moon; (2) the
comets, which moved, but erratically; and (3) the stars, which were
fixed and included everything else. After Galileo, every educated
person knew that the three natural kinds in the heavens were --- (1)
the planets, which no longer included the sun or the moon but now
included the earth, and also the comets as a deviant case; (2) the stars,
which now included the sun; and (3) a new category, satellites, which
included the moon and also some heavenly bodies not previously known.
These were also based on objective differences, but in virtue of a
more powerful theory the objective differences were more fundamental.
The point is that there are several different ways that things can be
objectively partitioned into discontinuous clumps, and we strive to
discover and use the most powerful and fundamental partitions.
Date: 07 May 83 2342 PDT
From: John McCarthy <JMC@SU-AI>
Subject: natural kinds
It seems to me that taking a subjective position on cows and horses
leads to taking a subjective position on almost all distinctions and
therefore two an impoverished science and an impoverished robotics.
Please, this is not taking a subjective position, unless you believe that
science is subjective. If you believe that science is objective then this
is an objective position. All it says is that a theory cannot take its
taxonomy for granted, that part of science *really* *is* deciding how the
world should be (or should have been) rationally carved up, and that a
previous theory's attempt may need to be revised or discarded. If one
prefers to think of this process as objectively correcting the errors of
the past, fine. We then all agree, of course, that the future may
objectively correct the errors which the present made when objectively
correcting the errors of the past.
-=*=- rick
-------
∂08-May-83 1728 Mailer@SCORE Message of 8-May-83 10:49:32
Received: from SU-SCORE by SU-AI with PUP; 08-May-83 17:28 PDT
Date: Sun 8 May 83 17:19:09-PDT
From: The Mailer Daemon <Mailer@SCORE>
To: JMC@SAIL
Subject: Message of 8-May-83 10:49:32
Message undelivered after 0 days -- will try for another 3 days:
RMS@MIT-AI.ARPA: Cannot connect to host.
------------
Received: from SAIL by SCORE with Pup; Sun 8 May 83 10:49:31-PDT
Date: 08 May 83 1041 PDT
From: John McCarthy <JMC@SU-AI>
Subject: mailing to bulletin boards
To: RMS%SU-AI@SU-SCORE
-------
∂08-May-83 1816 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN natural kinds
Received: from MIT-MC by SU-AI with TCP/SMTP; 8 May 83 18:15:56 PDT
Date: Sun, 8 May 1983 21:10 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: natural kinds
In-reply-to: Msg of 07 May 83 2342 PDT from John McCarthy <JMC at SU-AI>
From: John McCarthy <JMC at SU-AI>
It seems to me that taking a subjective position on cows and horses
leads to taking a subjective position on almost all distinctions and
therefore two an impoverished science and an impoverished robotics.
It seems to me that you've skipped a few steps in your reasoning here.
How do you get from subjectivism to impoverished science and robotics?
That is, how is it that the former leads to the latter? Also, what is
it that makes a distinction objective rather than subjective? How do
we know, other than by the existence of intersubjective agreement,
that a particular distinction is objective? Or is objectivity defined
as intersubjective agreement? If not, what is it?
∂08-May-83 2341 @MIT-MC:MINSKY%MIT-OZ%MIT-MC@SU-DSN visit
Received: from MIT-MC by SU-AI with TCP/SMTP; 8 May 83 23:40:57 PDT
Date: Mon, 9 May 1983 02:23 EDT
From: MINSKY%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Cc: minsky%oz%MIT-ML@SU-SCORE
Subject: visit
In-reply-to: Msg of 08 May 83 2108 PDT from John McCarthy <JMC at SU-AI>
Believe so. Will be definite as soon as I check with G. to be sure
nothing drastic like fumigation, etc.
∂09-May-83 0850 OHLANDER@USC-ISI Stanford Proposal.
Received: from USC-ISI by SU-AI with TCP/SMTP; 9 May 83 08:40:15 PDT
Date: 9 May 1983 0840-PDT
Sender: OHLANDER at USC-ISI
Subject: Stanford Proposal.
From: OHLANDER at USC-ISI
To: BSCOTT at SU-SCORE
Cc: TOB at SU-AI, JMC at SU-AI, ZM at SU-AI, RPG at SU-AI, WIEDERHOLD at SUMEX-AIM, LUCKHAM at SU-SCORE
Message-ID: <[USC-ISI] 9-May-83 08:40:19.OHLANDER>
Betty,
As I told you on the phone, we have considerable problems
with the current proposal. What has been provided to us is
nothing like that which we had talked about on at least three
occasions. I am providing two things below. One of them is
Gio's first cut at an SOW for the proposal. The second item is
some inputs from Berkeley. The stuff from Berkeley is fairly
redundant in the scope and SOW sections, but it provides a
reasonable example of the kind of statements that are needed.
The way you should structure the SOW is to provide about 4-5
pages, divided into sections according to the major research
areas. Each section should lead off with a statement that sets
the context for that section, e.g. "Stanford will carry out
research in the area of computer vision to address the problems
of photointerpretation and robotic sensing. Examples of the
kinds of research that will be undertaken include:" Following
this statement should come a listing, by number, of the specific
kinds of research that you are proposing to undertake. Each item
should be only one or two sentences. You want to leave the
coverage broad enough to cover work that might be undertaken at a
later date, even though it will not currently be funded. The level
of detail provided by Gio in his previous note, as shown, below is
not too bad while that provided by Binford is too detailed. He
should consolidate some of his items.
Stanford should strive to get this SOW as soon as possible. Send
a draft to me over the net with copies to Adams and Machado.
Regards, Ron
9-May-83 08:18:45-PDT,6910;000000000001
Mail-From: SMTP created at 26-Feb-83 23:55:33
Return-path: WIEDERHOLD@SUMEX-AIM.ARPA
Received: FROM SUMEX-AIM BY USC-ISI.ARPA WITH TCP ; 26 Feb 83 23:53:05 PST
Date: Sat 26 Feb 83 23:46:33-PST
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM.ARPA>
Subject: Proposed SCOPE for 1983-1986 ARPA proposal
To: adams@USC-ISI.ARPA, ohlander@USC-ISI.ARPA, mach@USC-ECLB.ARPA
cc: csd.armer@SU-SCORE.ARPA, csd.betty@SU-SCORE.ARPA, jmc@SU-AI.ARPA, zm@SU-AI.ARPA,
csl.dcl@SU-SCORE.ARPA, tob@SU-SCORE.ARPA, wiederhold@SUMEX-AIM.ARPA
Please comment on the appropriateness of the Scope as follows.
Thank you - we will continue to prepare the full Contract proposal
as well as individual task requests.
Gio Wiederhold for McCarthy, Luckham, Manna, Binford, and himself.
Stanford proposes to perform Research in the following Areas:
A. Basic Research in Artificial Intelligence
--------------------------------------------
1. Formalization of facts about the common sense world including laws
determining the effects of actions including mental actions.
2. Formalization (both procedural and declarative) of the heuristic knowledge
used in common sense reasoning and problem solving.
3. Formalization of common sense reasoning including non-monotonic reasoning.
4. Representation of facts in the memory of a computer. Connections between
AI representation work and database research.
5. Languages for communications among computers and between computers and
people, especially semantic aspects of such languages.
6. Techniques for computer reasoning and computer-assisted human reasoning.
7. AI Programming languages including LISP, especially Common Lisp, but also
research aimed at new languages.
8. Techniques of automatically specializing programs.
9. Programs that take advice from users and give advice in return, and
programs that decide what to do by formal reasoning.
B. Basic research in Software Engineering and Program Verification
------------------------------------------------------------------
1. Languages for formal specification of systems designs including
both software and hardware.
2. Languages for formal documentation and annotation of programs.
3. Techniques for specifying and analysing systems designs.
4. Techniques for analyzing sequential and concurrent programs, including
consistency analysis, analysis of runtime behavior, analysis of communication
in concurrent programs.
5. Concurrent programming methodology.
6. Formalization of programming knowledge.
7. Techniques for systematic development of programs, and for reasoning about
programs and programming.
8. Techniques for computer-supported reasoning.
9. Design and implementation of tools for advanced programming environments
including tools for error detection and consistency analysis, and for
knowledge based programming assistance.
10. Techniques for interactive and automatic program synthesis.
11. Theorem proving techniques.
12. Logical tools for program verification and synthesis
C. DATA MANAGEMENT
------------------
We propose to investigate advanced techniques to improve database access and
management. Within this objective we will analyze the semantics inherent in
the data and in the operations persformed on data. Where the boundaries of
algorithmic approached are reached we will develop heuristic techniques. We
will stress methods which are domain-independent so that the technology can
be transferred to a wide variety of applications.
Specifically we propose to:
1. Specify and use data semantics. We will use a categorization of data
semantics to establish those semantics which can aid in designing and
distributing well-structured databases, and those which may be of operational
utility.
2. Address ambiguity, view and performance problems in database update.
3. Extend and develop semantic aids to the anlaysis of database contents.
4. Develop communication and audit-trail concepts in design databases.
5. Investigate algorithms for maintaining data on optical disks.
6. Acquire and maintain data resourses and database management systems to
support this research.
D. Image Understanding and Robotics
-----------------------------------
This project focusses on designing and implementing Intelligent and Expert
Systems for interpretation and planning actions. Its elements include:
1. Expert systems for Image Understanding and expert systems for
Manufacturing; in particular the ACRONYM system and its planned successor;
2. Subsystems for interaction with the user:
2A. Rule-based, intelligent interfaces including geometric editing,
computer graphics, incorporation of natural language, Speech I/O
and mechanical devices;
3. Subsystems for modeling the world:
3A. Geometric modeling and representation of the physical world;
3B. Learning and abstraction in building and using databases for
the physical world;
4. Subsystems for planning, interpretation and decision making;
4A. Representation of information as constraint neetworks;
4B. Geometric reasoning utilizing mathematical representations in
a rule base, including constraint resolution, domain-specific constraints
with specialized procedures for resolution of constraints; geometric
reasonong incorporating symbolic statistical distributions and
maximum-likelihood methods with subspace factorization;
4C. Problem formulation in geometric reasoning;
4D. Reasonong with space/time and sequences of events and actions;
4E. Interpretation of images by matching with generic and specific models;
4F. Planning of robot actions;
4G. navigation, path-finding, and motion control of mobile robots;
map-making and world modeling;
5. Subsystems for interaction with the world:
5A. Analysis of images and sequences of images to obtain image features
including regions, region boundaries, and junctions of boundaries;
5B. Segmentation and aggregation of image features to analyze texture
regions, canonical groupings, and figure-ground discrimination;
5C. VLSI implementation of image algorithms and architecture studies;
5D. Inference rules for interpretation of image structures as spatial
structures;
5E. Multi-sensor integration with collateral information and knowledge
bases;
5F. Stereo mapping, motion parallax and object motion in spatial
interpretation;
5G. Use of shadows combined with general and special cues to spatial
structure in image analysis;
6. Robotics:
6A. Design of robot hands and arms;
6B. Force control in assembly;
6C. Control of cooperative action of multiple fingers and arms;
6D. Programming systems for robots.
-------
Berkeley's Input:
2. Scope
We propose two related lines of research whose overall
objective is to expand the ways in which computers enhance
human productivity. One line of research investigates how to
improve the coupling between a user and the computer by
developing better man-machine interfaces in general and
enhanced programming environments for the computer profes-
sional in particular. The other line of research concerns
how to develop higher performance computer systems and com-
puter systems which take advantage of the increasing
viability of distributed computing.
The University of California shall carry out a research
program in man-machine interfaces and programming environ-
ments with a goal of providing efficient utilization of
workstations and shared computer resources by human beings.
Examples of the kinds of important problems in this area
which may be investigated include: systems which model 3-
dimensional mechanical parts and permit a non-specialist
user to construct quickly an image of the part, preview it
on a variety of terminals, and create hardcopy engineering
drawings from it; relational data base systems for engineer-
ing data which have data types and access methods suitable
for VLSI and other engineering data; schemes for allowing
non-programmers to construct sophisticated data access
modules for data base systems which can display information
including graphics on bit-mapped displays; intelligent
natural language interfaces; an interactive graphics editor
for typesetting mathematical equations, tables, and figures;
interactive source language editors and translator front
ends which deal with static semantics, version histories,
consistency of code and documentation, etc; and use of a
general-purpose relational data base system to store
detailed program information and answer general questions
about programs, versions, configurations, and abstractions.
The University of California shall also carry out a
research program in high performance and distributed
computing which is aimed at developing new techniques in
these areas. Examples of the kinds of problems which may be
explored include: high performance computer architectures
that use various combinations of cache memories, pipelines,
memory system interfaces, and I/O architectures; algorithms
for high performance digital circuits such as adders, multi-
pliers, and floating point units; schemes for controlling
the flow of parallel computations including both data-flow
and control-flow architectures and both centralized and dis-
tributed control; algorithms that exploit parallelism and
pipelining; investigation of the functionality of the 4.2BSD
kernel for distributed applications; user programs which
better exploit the 4.2BSD distributed computing environment;
the performance of the 4.2BSD system for distributed appli-
cations; the application of scheduling algorithms to prob-
lems of distributed computing; encryption algorithms and
protocols for secure computation, protection of data, and
automated transactions; and algorithms to maximize the per-
formance of a distributed data base system during query pro-
cessing and crash recovery.
3. Statement of Work
We will carry out research in man / machine interfaces
and programming environments with the goals of: increasing
the productivity of the user of computers; easing the use of
computers for non-specialists; providing flexibility to
adapt the work environment to new approaches and design
methodologies; creating machine-independent and retargetable
tools; facilitating the development and maintenance of large
software systems; finding novel ways to use graphics in the
interaction with computers; improving access to information
with the use of expert systems and knowledge bases; and
exploring the feasibility of natural-language interfaces.
We will carry out research in high-performance and dis-
tributed computing with the goals of: understanding how to
exploit parallelism, asynchronism and distributed computing
effectively; discovering new parallel and distributed algo-
rithms; finding ways to express algorithms so that inherent
parallelism can be discovered and exploited; developing
architectures for execution of parallel and distributed
algorithms; studying various organizations of control in
parallel machines; experimenting with benchmarking methodol-
ogy for evaluating parallel and distributed systems; deriv-
ing performance metrics for parallel and distributed
machines; determining application-oriented criteria and
methods for optimal work-sharing between workstations and a
host; finding mechanisms for load balancing in distributed
systems; solving problems inherent to distributed computing
such as communication, synchronization, file and database
placement, and updates of multiple copies; and exploring
approaches to fault tolerance and crash recovery.
∂09-May-83 1017 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN visit and seminar
Received: from MIT-MC by SU-AI with TCP/SMTP; 9 May 83 10:16:22 PDT
Date: Mon, 9 May 1983 13:12 EDT
From: DAM%MIT-OZ%MIT-MC@SU-DSN
To: JMC@SU-AI
Subject: visit and seminar
In-reply-to: Msg of 07 May 83 1014 PDT from John McCarthy <JMC at SU-AI>
Sounds good, I will review circumscription and
reserve the afternoon.
David Mc
∂09-May-83 0921 YEARWOOD@SCORE Student Support for Summer 83
Received: from SU-SCORE by SU-AI with PUP; 09-May-83 09:16 PDT
Date: Mon 9 May 83 09:19:59-PDT
From: Marlene Yearwood <YEARWOOD@SCORE>
Subject: Student Support for Summer 83
To: jmc@SAIL
cc: hall@SCORE
Stanford-Phone: (415) 497-2266
Let me share with you the list of RAs that our records show you are currently
supporting and when that list expires. If you want to extend one of these
appointments, or if you want to cancel an appointment that is continuing into
Summer let us know by Monday, May 16th. If this list does not match your
perception of reality, please let us know. Again, we ask that you give this
your immediate attention as we are trying to set up support for students on
time this quarter, and we must have all of the information by May 16th.
On your account 2FCZ427 (NSF) you have:
CAROLYN TALCOTT at 50% from Jan-01-1983 through Jun-30-1983
On your account 2FCZ451 (NSF) you have:
JOSEPH WEENING at 10% from Apr-01-1983 through Jun-30-1983
On your account 2FCZ762 (ONR) you have:
YORAM MOSES at 50% from Apr-01-1983 through Jun-30-1983
Marlie
-------
∂09-May-83 1136 ME
To: YM, JMC
You're getting multiple copies of a reminder because I'm debugging
the rmind phantom. Don't worry, the copies will stop sometime.
∂09-May-83 1404 DFH flight times
To: JMC, CLT
Wed 5/11 San Jose / Los Angeles lv. 8:25am ar. 9:29 am Air Cal 362
Thurs 5/12 Los Ageles / Boston lv. 12 noon ar. 8:08 pm American 12
Sun 5/15 Boston/San Francisco lv. 6:30 pm ar. 9:31pm TWA 847
Those reservations are entirely satisfactory.
∂09-May-83 1532 minker.umcp-cs@UDel-Relay Special Session on Non-Monotonic Logic
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 9 May 83 15:31:29 PDT
Date: 9 May 83 17:49:23 EDT (Mon)
From: JACK MINKER <minker.umcp-cs@UDel-Relay>
Return-Path: <minker.umcp-cs@UDel-Relay>
Subject: Special Session on Non-Monotonic Logic
To: Bundy@RUTGERS
Cc: JMC@SU-AI, Moore@SRI-AI, "uw-beaver!ubc-vision!Reiter"@LBL-UNIX,
Reddy@CMU-CS-A, PJH@SU-AI, ATP.Bledsoe@UTexas-20, Walker@SRI-AI,
Wrightson@RUTGERS, Siekmann@RUTGERS, Amarel@RUTGERS
Via: UMCP-CS; 9 May 83 18:12-EDT
Dear Alan,
Saul Amarel sent me a Net message on 7 May 1983 and
sent a copy to you and the members of the IJCAI organizing
committee. He has informed me that the IJCAI conference committee
has approved my proposal for a special session on the subject
of non-monotonic logic. The session is intended to honor the
70 birthday of the noted Soviet cyberneticist, Dr. Alexander
Lerner.
Last Friday I sent a Net message to Ray Reiter and John
McCarthy asking that they participate in the session with me. All
three of us have done work in non-monotonic logic. Through Saul
I understand that Reiter has agreed to be a part of the session.
Since Bob Moore has done work in this field and has informed
Saul that he would be pleased to be part of the session, I will
be in touch with him as soon as possible to invite him to participate.
I would appreciate it if you would mention the session
in the programme brochure as Saul has suggested. I understand that
the Wednesday evening of the conference has not been scheduled.
It would be good to have the session on that evening. However, I am
open to your suggestions.
I will be in direct contact with Reiter, McCarthy, and Moore
in the next few days and will let you know where things stand.
Thanks for your assistance on this matter.
Sincerely yours,
Jack Minker
∂09-May-83 1800 JMC*
Golodner
∂09-May-83 1840 JMC*
shop
∂09-May-83 1926 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN visit
Received: from MIT-MC by SU-AI with TCP/SMTP; 9 May 83 19:26:32 PDT
Date: Mon, 9 May 1983 20:54 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Subject: visit
In-reply-to: Msg of 08 May 83 2351 PDT from John McCarthy <JMC at SU-AI>
From: John McCarthy <JMC at SU-AI>
As you may know, I am coming to Boston for the time-sharing history
session this coming weekend. I will actually arrive Thursday night
and will probably stay at Minsky's. Where and when do I actually
appear for the videotaping?
I have no idea, but Richard Solomon does, no doubt. I will forward this
message to him, and, hopefully, he'll respond promptly. He usually reads
his mail every day, even if he's halfway across the world. FYI his net
address is Solomon.Datanet@MIT-Multics. As you may remember, Multics is
case-sensitive, so be careful of your capitalizations in messages to him.
You might want to cc: to Pool.Datanet@MIT-Multics in case Richard is out
of pocket. In any event, if you don't hear from him within 24 hours, let
me know and I'll cajole the right people.
I will give an AI seminar on Friday about circumscription and
the blocks world.
So I've heard. I'll probably be there, unless it's at some ungodly
hour in the middle of the night, like 11am.
Many thanks for refixing my address in phil-sci. Incidentally,
a guest account at OZ was arranged for me, but it happened just
about the time we went off ARPAnet and now it has disappeared.
If it isn't too much trouble to revive it, I would be grateful.
No problem. I'll send you a message later this evening with the specifics
of the account (user-name, password, etc).
∂09-May-83 2040 GENESERETH@SUMEX-AIM Re: Keyworth
Received: from SUMEX-AIM by SU-AI with PUP; 09-May-83 20:40 PDT
Date: Mon 9 May 83 20:41:34-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
Subject: Re: Keyworth
To: JMC@SAIL
In-Reply-To: Your message of Mon 9 May 83 13:22:00-PDT
OK, thanks, JOhn. I'll be waiting for your go ahead.
mrg
-------
∂09-May-83 2102 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN [JMC: visit ]
Received: from MIT-MC by SU-AI with TCP/SMTP; 9 May 83 21:02:40 PDT
Date: Mon, 9 May 1983 20:58 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: Solomon.Datanet%MIT-Multics@SU-DSN, Pool.Datanet%MIT-Multics@SU-DSN
cc: JMC@SU-AI
Subject: [JMC: visit ]
I just received the following message from John McCarthy. Someone
should send him net mail detailing the taping schedule. Just send
mail from Multics to JMC@SU-AI.
Date: 08 May 83 2351 PDT
From: John McCarthy <JMC at SU-AI>
To: gavan%oz%MIT-ML at SU-SCORE
Re: visit
Received: from SAIL by SCORE with Pup; Sun 8 May 83 23:59:05-PDT
As you may know, I am coming to Boston for the time-sharing history
session this coming weekend. I will actually arrive Thursday night
and will probably stay at Minsky's. Where and when do I actually
appear for the videotaping? ....
∂09-May-83 2120 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN account
Received: from MIT-MC by SU-AI with TCP/SMTP; 9 May 83 21:16:22 PDT
Date: Mon, 9 May 1983 21:20 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Subject: account
In-reply-to: Msg of 08 May 83 2231 PDT from John McCarthy <JMC at SU-AI>
You now have an account on OZ. You are JMC. Your password is
"object". Feel free to change it. When the daemon that recognizes
new users fires and (subsequently) asks you what group you are in, you
should probably put yourself in the K group (K is for Kollaborating
researcher), since various other system daemons will not be as nasty
to you as they would otherwise. Once you've logged in for the first
time, I'll check all the bits to ensure that they've been set
optimally.
BTW, would it be possible for you to reciprocate with a SAIL account?
I've used it before, so I'm familiar with its operation, and know how
to avoid leaving jobs hanging, etc. I would like to use (at off-peak
hours, mostly) the NS program as a source of data for some
natural-language research I'm conducting.
∂09-May-83 2254 @MIT-MC:MINSKY%MIT-OZ%MIT-MC@SU-DSN natural kinds
Received: from MIT-MC by SU-AI with TCP/SMTP; 9 May 83 22:53:59 PDT
Date: Tue, 10 May 1983 01:43 EDT
From: MINSKY%MIT-OZ%MIT-MC@SU-DSN
To: RICKL%MIT-OZ%MIT-MC@SU-DSN, minsky%MIT-OZ%MIT-MC@SU-DSN
Cc: Batali%MIT-OZ%MIT-MC@SU-DSN, batali%MIT-OZ%MIT-MC@SU-DSN,
dam%MIT-OZ%MIT-MC@SU-DSN, jmc@SU-AI, phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: natural kinds
I am bothered by this talk of natural kinds. Some sentences began
this way (I'm just using some of RICKL's because I'm replying to one
of his messages):
If we agree that cows and horses are objectively
discontinuous, then ...
But if cows are discontinuous with everything else then ...
And so on. Then RICKL asks: "Do we run into any problems if we take
this to be the defining "essential property" of natural kinds?
Well, it seems to me that for most cognitive purposes it doesn't make
any difference whether a group is two disjoint clusters, or a single
dumbbell cluster with a rather thin neck. You'll still need two
different symbols for them, both in your head and in your theories -
just as, if there were a few rabbit-ducks or hosre-cows, you'd still
need prototype-frame-clusters for rabbits, ducks, horses and cows,
since unified theories of them wouldn't say most of the things you
want to say. Just as we think of solids and liquids - and will
continue to - even with all those jellies, jams, goos, pastes, slimes,
glops, smears, and what have you.
So I think the idea of "natural kinds" can't have much to do with
disjointness. I'm not convinced that it is a good, stimulating
concept. What I think IS a good idea is that, given certain pairs
(WORLD, MIND) then certain "events" in that WORLD will induce (often,
dependning on lots of other things) certain SYMBOLS, or stereotypes,
etc., in that kind of MIND.
I claim that this is a good idea, and that there should be a word for it.
You have my permission to call these "natural kinds mod(MIND)" for
that WORLD.
AS for the question of whether certain worlds - say the one we're in -
have "events" that must induce symbols in any ("intelligent") MIND
- so that "natural kind" can stand apart from the psychology involved,
I'm not sure I'd care to get into that, since you'd have to agree on
which kinds of minds are worthy of the name, even assuming that we
agree about all sorts of facts about our world, like that it contains
the sorts of objects that we commonsensically agree it contains.
Marvin, you would find some (perhaps one) of the philosophy papers on
natural kinds interesting. Unfortunately, I don't have the reference, but
I'll bring it if I can find it; it's relatively recent. I was the one who
raised the issue in the phil-sci discussion - contrasting natural kinds
with essential properties. The idea is this:
1. Previously philosophers supposed that concepts had definitions
that determined (for example) what you meant by a lemon.
2. But then suppose some geneticist comes up with a blue lemon.
If yellow was part of the definition of a lemon, how do we account
for the fact that people will accept blue lemons as lemons even
if they don't know genetics.
3. The idea of natural kinds is that people in fact don't recognize
things by definitions or even prototypes. Consider a child. It
hears that we are going to the store to buy a lemon. It decides
that there is something called lemons, and it is ready to learn
more. It sees this small yellow fruit in the store, and says to
itself, "That little yellow fruit is a lemon". It can then
recognize other lemons. However, it is prepared to learn that
some lemons are blue and would be prepared to learn that some
lemons are as large as grapefruits.
4. What makes this work, since it seems that it could be stretched
into calling anything a lemon?
5. It is a fact about the world that many kinds of objects are
discretely separated from other objects. In principle, there
could be a continuous range between lemons and grapefruit or
between horses and cows, but there isn't. Of course, sometimes
there is as between hills and mountains. When there is no
bridge, there are fewer problems, and this situation is so
common that we have evolved and our language has evolved to
take advantage of it.
6. The point at issue in the phil-sci discussion is that natural
kinds are natural; they aren't a linguistic convention. The
world helps us out in this way, and Martians would be equally
pleased to empirically discover that there are no bridge fruit
between lemons and grapefruit or bridge animals between horses
and cows.
7. Robots should also be designed to take advantage of this
phenomenon.
∂09-May-83 2355 @MIT-MC:MINSKY%MIT-OZ%MIT-MC@SU-DSN
Received: from MIT-MC by SU-AI with TCP/SMTP; 9 May 83 23:54:58 PDT
Date: Tue, 10 May 1983 02:50 EDT
From: MINSKY%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Cc: minsky%oz%MIT-ML@SU-DSN
In-reply-to: Msg of 09 May 83 2312 PDT from John McCarthy <JMC at SU-AI>
I'm afraid I still think this is trivial, for several reasons. First,
as I said, I don't think it makes much difference whether there is a
thin bridge or not - or even a think one, for the reasons I gave.
I don't think there are significaltly fewer problems, psychologically -
and that there only seem fewer to those philosophers.
1. I, too regard as wrong to suppose that all concepts have definitions.
I hope you've read "Learning Meaning". It may happen, if a concept
proceeds smoothly (perhaps by accident) to never need an
"accumulation" that it may apepar to have a definition - but that's a
psychological accident.
2. In Learning Meaning, I explain that when a blue lemon comes up, you
may be able to then make "lemon" into an accumulation.
Psychologically, this will be OK if you are in a fruit context that
already regards color as not so essential - no so default-rigid in the
frame from which the "lemon" concept was partly inherited from. If
there was no solid such precorsor, then you'll just replace "yellow"
by "blue or yellow" or by "any color". My paper (and Winston's
thesis) discusses why one might choose one or the other of these
strategies under different circumstances.
3. If you read Learning meaning, you'll see that this is far beyond
the stereotype of "prototype" that you appear to believe I'm stuck in.
A meaning is a large web of different structures, with different ways
to move around depending on the intentional context.
4. Under extreme conditions, one might allow all parts of the
representation to drift and get anything to be called "lemon". But
that won;t happen if there is some intelligence and a big network
about the world; instead the child will decide that there's something
wrong. "What makes this work?", you ask. It works, not because the
"natural kinds" are discrete, but because we can reason only by
"triangulating" a continuous space by a discrete structure supplied by
the mind - like homology theory. Thus I think you and the
philosophers have things PRECISELY WRONG. We could still have minds
almost as good as ours in a continuous world, provided that the laws
of nature were also continuous!
5. I maintain that the problems just aren't significantly simpler!
6. I don't say it is linguistic at all. I say it is a product of the
ability to symbolize, without which thinking is unlikely. Look at
your own example - color. They're continuous, but it would be useless
to perceive them as such. And it does little harm so far as I can
see, to do otherwise. It is better to think of "red fruit" and "green
fruit" than "fruit".
7. When there is a large, discrete separation, robots will recognize
this by observing that there aren't any caveats or conditions or
appendages attached to a representation. Fine. Also inevitable.
No special provision need be made! I usually will just emerge from
the non-necessity to make an accumulation or difference-network.
So long as you believe that knowledge can be perfectly modular, and
resist the need for "propositions" to be embedded in networks with
comments and heuristics concerning their use, you'll be fooled by
these attractive but psychologically pathological examples of special
clarity. I value your other technical copntributions, like
work on counterfactuals and factoring machines, etc., but this
"natural kind" is attractive to you for a wrong reason, and deflects
away from the important problems.
∂10-May-83 0035 ISAACSON@USC-ISI Control-Z: Inquir at OZ
Received: from SU-DSN by SU-AI with PUP; 10-May-83 00:34 PDT
Received: From USC-ISI by SU-DSN.ARPA; Tue May 10 00:38:07 1983
Date: 10 May 1983 0034-PDT
Sender: ISAACSON at USC-ISI
Subject: Control-Z: Inquir at OZ
From: ISAACSON at USC-ISI
To: JMC%SU-AI at SU-DSN
Cc: isaacson at USC-ISI
Message-ID: <[USC-ISI]10-May-83 00:34:50.ISAACSON>
Gavan tells me that you have trouble with Control-Z to end your
entry in your OZ new-user Inquir-entry run. Not that I know I
to remedy that, but I had the same problem when I first tried to
do the same. To the best of my recollection, I fooled it
somehow, but forgot how. I suppose, if all else fails, you may
ask your sponsor to enter the information for you. -- JDI
∂10-May-83 0047 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN [JMC: <ctrl>z ]
Received: from MIT-MC by SU-AI with TCP/SMTP; 10 May 83 00:47:48 PDT
Date: Tue, 10 May 1983 03:43 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: cstacy%MIT-OZ%MIT-MC@SU-DSN
cc: JMC@SU-AI
Subject: [JMC: <ctrl>z ]
Chris: JMC is having trouble getting ↑z through a
SAIL->ARPAnet->MC->CHAOSnet->OZ connection.
He can't finish running inquir because of this.
Any suggestions?
Date: 09 May 83 2157 PDT
From: John McCarthy <JMC at SU-AI>
To: gavan%oz%MIT-MC at SU-SCORE
Re: <ctrl>z
Received: from SAIL by SCORE with Pup; Mon 9 May 83 22:03:08-PDT
My head is stuck half way through this new user hole. The program
wants <ctrl>z, but I don't know how to sent it through MC.
It is now 0058 your time.
∂10-May-83 0108 RESTIVO@SCORE PROLOG Digest V1 #2
Received: from SU-SCORE by SU-AI with PUP; 10-May-83 01:07 PDT
Date: Monday, May 9, 1983 3:56PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: 712 Partridge Av, Menlo Park, CA 94025
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #2
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Sunday, 8 May 1983 Volume 1 : Issue 2
Today's Topics:
Administrivia - Bad Digest Fixed and Redistributed,
Queries - Performance & Autonomous Control,
Implementations - Personal Computers,
Applications - Knowledge Acquisition, Programming - Objects
----------------------------------------------------------------------
Date: 9 May 1983 0741-PDT (Monday)
From: abbott at AEROSPACE (Russ Abbott)
Subject: PROLOG Digest V1 #2
Correction:
The end of my message was cut off and (presumably) the start of
someone else's. I have no paper in IJCAI-83.
Russ
[ Sorry, folks. The previous transmission of PROLOG Digest V1 #2 was
messed up, so we are re-transmitting this new version of it, which
has been scanned with a fine-toothed comb and should be immaculate.
-- Chuck ]
------------------------------
Date: Friday, 6 May 1983 10:10:16 EDT
From: Brad.Allen@CMU-RI-ISL
Subject: Inferences/second
A lot of the Japanese presentation of the 5G machine's capabilities
are couched in terms of how fast the machine will be in "logical
inferences per second". Two questions: 1) How reasonable a measure do
you think this is, and 2) roughly how fast are current and projected
near-term implementations of PROLOG in LIPS?
------------------------------
Date: 26 Apr 1983 0216-EDT
From: NUDEL.CL@RUTGERS (NUDEL.CL)
Subject: Autonomous Control Strategy
If anyone out there has the time and knowledge, I would be
particularly interested to know what is the current state-of-the-art
in giving PROLOG what Kuwalski has called "autonomous control strategy
optimization". My Ph.D. research develops heuristics for good search
orderings in the Constraint Satisfaction (Consistent Labeling,
Discrete Relaxation, Constraint Propagation) problem. I am thinking
of applying this to PROLOG.
Mail to NUDEL@RUTGERS (not NUDEL.CL@RUTGERS as in the header above).
Bernard
------------------------------
Date: Fri, 6 May 83 10:24:05 EDT
From: "John B. Black" <Black@YALE.ARPA>
Subject: Prolog for PCs and CAI
I would like to find out what versions of prolog exist for personal
computers and if one can actually do anything useful with them. Has
anybody out there actually used microprolog for anything? If so what
was your experience?
I am particularly interested in whether there is a version of
microprolog for the regular IBM-PC. I know there is version from
Logic Programming Associates for the IBM-PC with a Z80 card, but is
there a version for the 8088?
I am interested in these things because I am about to start a project
trying out Prolog for Intelligent Computer Assisted Instruction, and I
want to transfer later to a PC. Is anybody else out there doing CAI
in Prolog?
------------------------------
Date: Fri 6 May 83 17:56:21-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Another Prolog for personal computers
An English company, Expert Systems Ltd., has rewritten Chris Mellish's
PDP-11 Prolog for Z80-based systems (and possibly for others). This
system has the advantage of being compatible with the Clocksin and
Mellish book, unlike MicroProlog. I don't have an address for them,
but I understand a US company is planning to start selling it soon.
Expert Systems sells equivalent systems for PDP-11s and VAXes (No,
this is no advertisement, I have no connection with Expert Systems...)
Fernando Pereira
------------------------------
Date: 6 May 1983 0803-PDT (Friday)
From: abbott at AEROSPACE (Russ Abbott)
Subject: Using Prolog for Knowledge Acquisition
I'd be interested in discussing the use of Prolog for knowledge
acquisition/representation with anyone working in the area. I'm
working on a "knowledge acquisition system" for use in the analysis of
specifications in what we call the "documentation workbench" project.
The long term goal is to be able to take a requirements document and
help the author check it for consistency, completeness, etc. The
following discusses (very informally) my current approach and state.
Progress on Automatic Analysis of Requirements Specifications.
Following is a brief discussion of the approach toward which I have
been evolving in the documentation workbench project.
Automatic parsing of natural language is feasible but difficult.
Basic parsing is straightforward. Realistic, sophisticated parsing
cannot be done independently of semantic analysis. In addition, the
parsing of unconstrained natural language is technically detailed,
requiring significant amounts of information about grammatical
structures, in addition to the semantic information. Finally, even if
one could parse successfully, that would be only the first step. What
one really wants to be able to do is to analyze the information
contained in the parsed specification, i.e., process the semantic
information. Therefore I have decided not to focus on parsing but to
leave that for a later stage of the project.
This suggests that the appropriate initial point of view is to imagine
that a specification is written in a quasi-formal language, one whose
structure is understood a priori by the specification analysis system,
but one that is unconstrained semantically. Therefore the current
research is in developing such a language. Given such a language, and
that is certainly not a given, one could then do two things.
1. Generate the natural language version of the specification.
This means that one can take the formal language in which the
specification information is given and, as a by product of
analyzing it, produce an English language version of the same
information. It is much simpler to translate a formal language
into a natural language than the reverse, and parts of this
translation are already working. As an example, the current
version of the system is able to translate Prolog rules such as:
a(x) :- b(X)
which is read in Prolog as:
If predicate b is true of variable X then one may conclude
that predicate A is true of variable X.
into the English statement
If X is a "b" then X is an "a."
(There is an additional assumption used in this translation: a
one-place predicate about which one has no other information is
really a type.)
When this aspect of the processing is more polished, one could
imagine that specifications might very well be written
originally in such a formal language, and that the English
version will be generated automatically. The advantage is that
one would be sure of what one is actually saying. The
disadvantage is that one would have to write in this formal
language. But when the formal language is completely developed,
it might not seem like quite such an unreasonable burden. (See
below for examples.)
2. Analyze the information presented in the quasi-formal language,
suggest possible contradictions, and answer questions about it.
At the simplest level (and the value of this should not be
discounted), since the input language is quasi-formal, one can
do the same kind of analysis on it that one can do on any formal
language, e.g., check for syntactic conflicts, generate cross
reference lists, etc. For example, one of the features of the
language that I'm developing is the ability to use types in a
natural way (in effect any common noun is a type), so one can
analyze the input language for type conflicts.
In some sense it seems impossible to develop a useful formal language
that has no semantics, since there seems no way to deal with a
language whose semantics are not known. And of course that is true.
Therefore what we really have to develop is a language whose semantics
permits one to express any information, i.e., a language for knowledge
acquisition. This, of course, is a very difficult, but current,
research area. The approach we are taking is to develop a language
based on what might be called typed predicate calculus, or a
combination of predicate calculus and relational data bases. That is,
the semantics of the language provides for the declaration and
definition of both
1. Types and relations (i.e., predicates) and
2. Deduction rules, e.g., If Predicate←A(X, Y, Z) and
Predicate←B(W, X, Y) then Predicate←C(W, X, Z).
Prolog provides the basic framework for expressing the deduction
rules. But Prolog does not have any built in means for expressing
types. They must implemented (as was suggested above) as one place
predicates.
As an example of the current state of the definition of this (quasi-
formal) language consider the following:
organization$shall←be←interoperable←with$organization$for$mission(
'SOC', 'STC', Mission)
:-
'DOD←satellite←operation'(Mission) and
mission$is←assigned←by$organization(Mission, 'STC').
This is a prolog rule that says in effect:
SOC shall be interoperable with STC for any DOD satellite
operation mission assigned by the STC.
As far as prolog is concerned it is of the form:
predicate←a(A, B, C) :- predicate←b(C) and predicate←c(C, B).
Prolog takes each of the long predicate names simply as a name and
makes nothing more of them. The processing I have added so far parses
the predicate names into a name along with type declarations. For
example, the last line of the above rule is interpreted both as a
declaration:
is←assigned←by(M : Mission; O : Organization)
as well as a use of this implicitly declared predicate, with Mission,
a variable in the rule, as the actual parameter replacing the formal
parameter M, and hence of type Mission, and STC, a constant in the
rule, as the actual parameter replacing the formal parameter O, and
hence of type Organization.
So far this seems like a promising approach to the quasi-formal
language, and I am proceeding along these lines.
I am currently able to accept statements of the above form, to
translate them into more-or-less English, to store the information,
and respond to very limited questions about the information.
------------------------------
Date: Fri 6 May 83 17:50:20-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Quasiformal languages
LESK is such a language, developed by Doug Skuce of the CS Dept. of
the University of Ottawa, Canada. He has implemented it in Prolog.
The language allows the definition of classes (types), isa
relationships, and complex part-whole relationships, and has a formal
semantics (it's just logic in disguise). It has a nice English-like
flavor. A reference is
"Expressing Qualitative Biomedical Knowledge Exactly Using the
Language LESK", D. S. Skuce, Comput. Biol. Med., vol. 15, no. 1, pp.
57-69, 1982.
Fernando
------------------------------
Date: Friday, 6 May 1983 10:02:24 EDT
From: Brad.Allen@CMU-RI-ISL
Subject: Object-oriented programming in LPGs
I heard a talk here at CMU yesterday by David Rosenthal of Edinburgh
on a graphics front-end for PROLOG which he is developing. During the
talk, he gave an interesting account of how concurrent PROLOG can be
used to implement Smalltalk-like objects. I'd be interested in hearing
more about this technique, and in general, anything about object-
oriented programming in logic programming languages. My particular
interest lies in the integration of schema-based knowledge
representations with LPGs, a topic about which I have written a paper
to be given at IJCAI-83.
------------------------------
End of PROLOG Digest
********************
∂10-May-83 0111 @LBL-CSAM.ARPA:uw-beaver!ubc-vision!reiter@LBL-CSAM Lerner
Received: from LBL-CSAM by SU-AI with TCP/SMTP; 10 May 83 01:11:29 PDT
Date: 10 May 83 01:13:25 PDT (Tue)
From: uw-beaver!ubc-vision!reiter@LBL-CSAM
Subject: Lerner
Return-Path: <uw-beaver!ubc-vision!reiter@LBL-CSAM>
Message-Id: <8305100813.AA15071@LBL-CSAM.ARPA>
Received: by LBL-CSAM.ARPA (3.320/3.21)
id AA15071; 10 May 83 01:13:25 PDT (Tue)
To: uw-beaver!lbl-csam!JMC@SU-AI, uw-beaver!lbl-csam!minker.UMCP-CS@UDEL-RELAY
Dear Jack,
Yes, I'm agreeable on participating in an IJCAI special session in honour
of Lerner. The subject of such a session at AAAI came up at the program
committee meeting this last weekend. The following motion was unanimously
approved by the program committee and transmitted to the executive
committee for their approval:
Jack Minker has made the following request: that there be a privately
sponsored evening specialist technical session in honor of Alexander
Lerner's 70th birthday. The committee unanimously concurs and suggests
that it be held on Tuesday night, so as not to conflict with the main
program, and that it be announced in the printed program.
I don't think you should count on me to participate since I am not
planning to attend the AAAI meeting.
I shall be at Maryland June 13-16, conferring with Bary and John. Will
you be around then? I'd welcome a chance to see you.
Best wishes, Ray.
∂10-May-83 0326 @USC-ECL,@MIT-ML:GAVAN%MIT-OZ%MIT-ML@SU-DSN natural kinds
Received: from USC-ECL by SU-AI with TCP/SMTP; 10 May 83 03:24:15 PDT
Received: from MIT-ML by USC-ECL; Mon 9 May 83 19:13:49-PDT
Date: Mon, 9 May 1983 22:07 EDT
From: GAVAN%MIT-OZ%MIT-ML@SU-DSN
To: John McCarthy <JMC@SU-AI>
Cc: phil-sci%MIT-MC@SU-DSN
Subject: natural kinds
In-reply-to: Msg of 08 May 83 2231 PDT from John McCarthy <JMC at SU-AI>
From: John McCarthy <JMC at SU-AI>
It seems that Gavan is correct that our acceptance of an objective
distinction between cows and horses is related to a correspondence
theory of truth. As to faith, I haven't figured out how to answer
someone who asserts that X is a matter of faith. However, if it is
a matter of faith, rather than say a Pascal type wager, we can build
a science on it if we want, and Gavan can only grumble about it.
Gavan doesn't always grumble about it. But he does raise a suspicious
eyebrow anytime any subject makes a claim to objectivity.
My remark about an impoverished science comes to this. To many of
us the fact that the cow-horse distinction is different from the
hill-mountain distinction is important for the design of robots.
I don't dispute that it would be necessary to design the robot so that
it can differentiate between discrete and continuous distributions.
Hopefully, though, the robot will be able to recognize when the speech
community in which it operates has intersubjectively validated a
change in the distribution. That is, I hope that when the geneticists
develop a continuum of species between cows and horses the robot will
be able to update its internal tables like we will be able to do.
Similarly, I hope that, when the community agrees to call every mound
less than X meters high a hill and every mound greater than X meters
high a mountain, the robot will be able to adjust.
A theory that cannot distinguish or even dislikes distinguishing
between these distinctions is impoverished.
I don't object to making the distinction. I just object to the claim
that such distinctions can be made objectively.
Saying otherwise should be supported by examples where (say) replacing
"objective" by "intersubjective" has led to scientific success.
I suppose that given enough time in the library I could drag out
countless examples. For now, let me just say that whenever a young
scientist breaks out of the straight-jacket of normal science and
makes a scientific revolution, he/she has recognized that the laws,
observations, etc., expounded upon and PRESENTED AS OBJECTIVE by
his/her professors are not objective at all but are rather the
subjective OPINIONS of a group of people (the scientific community).
If this scientist were to believe that these laws and observations are
objective, then he/she would have no reason to doubt them.
Finally, it is true that genetic engineers may someday produce a
continuum between cows and horses especially if Gavan becomes head
of NSF. That supports the point that natural kinds are a fact about
the world and not just a fact about concepts.
What world? If you cannot prove the existence of the world, then you
can't prove there are any "facts" about it. I, for one, believe in
the existence of a world "out there", but the only world I know
anything about is the one inside my head. I just have faith in the
belief that there really is a world "out there" somewhere. I have
many opinions about this world, but no facts. I have no facts, and I
cannot be objective, precisely because my concepts (necessarily)
constrain my percepts. Perception is fallible. See Wittgenstein on
the rabbit and the duck problem.
paradigms
Once upon a time Thomas Kuhn wrote a book called "The Structure of
Scientific Revolutions" that introduced the notion of "paradigm
shift" with such examples as the transition from medieval to the
Galileo-Newton view of mechanics and the transition from classical
to quantum mechanical views. Now every PhD thesis proposes its
own paradigm shift, and Vaughan Pratt has five new paradigms in
one report dated 1982. It's clear that the technology of making paradigms
has entered an era of expansion like that in semi-conductor
memories. However, note that Pratt has five rather than four
paradigms. My proposed meta-paradigm is that the number of new
paradigms in each paper can be expected to increase by a factor
of five every three years. We will therefore reach the mega-paradigm
paper in the year 2026.
∂10-May-83 1041 DCL Umbrella
To: csd.bscott@SU-SCORE, JMC@SU-AI
Betty,
my understanding of Ron Ohlander's message is that:
1. We need to include 2 or 3 page SOW in a prominent position
in the original umbrella proposal; each task will have a subsection
of the SOW. This is the thing that will form the basis for a contract;
it has to be "legally obligating" but also "general".
2. We should NETmail the SOW first for approval before sending the
ammended proposal.
3. Things are time critical if we want our money October.
4. There is some fuzziness about how the separate task proposals
should be formatted to refer to the umbrella, but this is not
our immediate concern.
- David
∂10-May-83 1100 JMC*
Mult-proc. lisp to Allen%bbnf,Don Allen
∂10-May-83 1104 RPG Varia concerning Ohlander
To: BS@SU-AI, TOB@SU-AI, JMC@SU-AI, ZM@SU-AI, GIO@SU-AI
We showed both Machado and Ohlander the draft of the proposal when they
were here. Machado said in front of Ohlander: ``This looks like the beginning
of a good proposal here. Add some initial text and link the sections to
that initial text.'' I can put working on this proposal more on my agenda, but I
cannot see that I can spend any time on it before June.
-rpg-
∂10-May-83 1154 DFH Ohlander
Ron Ohlander called Betty Scott. He will be
in his office today until 1 pm our time,
202-694-5051. Tomorrow he will be there until
3:30, but don't think that is very convenient
for you. Please call him.
--Diana
fc 2840, tok
taipei, 2994
sing, 3894, tok taipei, sing.
osaka-taipei 524
1965-1218, return 1st vs.
Pan Am 19 1:45 - 5:25pm on 22nd at Osaka, Saturday May 21
Osaka - Taipei, Cathay 521 10:30 -12:10, Sunday June 5
Taipei - Singapore, 9am, Singapore 9:00 - 1:05pm business
Singapore - SFO, Pan Am, 7:20am - 10am, June 15.
Taipei - SFO, sunday, 5:50pm - 2:05pm, June 12
∂10-May-83 1309 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN paradigms
Received: from MIT-MC by SU-AI with TCP/SMTP; 10 May 83 13:08:59 PDT
Date: Tue, 10 May 1983 15:58 EDT
From: DAM%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Subject: paradigms
In-reply-to: Msg of 10 May 83 1044 PDT from John McCarthy <JMC at SU-AI>
I have been wondering what Vaughan Pratt is up to lately.
Could you give me a reference to recent papers (or an ARPA address,
his ARPANET directory entry is way out of date).
David Mc
∂10-May-83 1442 TOB
∂10-May-83 1554 @MIT-MULTICS.ARPA:JCMa@MIT-OZ my mail address
Received: from MIT-MULTICS by SU-AI with TCP/SMTP; 10 May 83 15:53:40 PDT
Date: Friday, 6 May 1983, 20:16-EDT
From: JCMa@MIT-OZ
Subject: my mail address
To: JMC@SU-AI
In-reply-to: The message of 6 May 83 14:54-EDT from John McCarthy <JMC at SU-AI>
Supersedes: The message of 6 May 83 20:15-EDT from John C. Mallery <JCMa at MIT-OZ>
Return-path: <JMC@SU-AI>
Received: from SAIL by SCORE with Pup; Fri 6 May 83 12:00:17-PDT
Date: 06 May 83 1154 PDT
From: John McCarthy <JMC@SU-AI>
Subject: my mail address
To: phil-sci-request%oz%MIT-ML@SU-SCORE
Our TCP seems to be working well enough so that it is no longer
necessary to go through another computer, and my address should
be jmc-lists@su-ai with whatever decoration is required to get
out of oz.
You mailer, or some forwrding mailer, seems to have added some random
characters to your message. If this gets to you, we can change you
address as desired.
Your message arrived ok, and my request to use JMC-LISTS@SU-AI for the
phil-sci list seems to have been implemented.
∂10-May-83 1657 ZM DARPA proposal
∂10-May-83 1738 RITA@SCORE Package for you
Received: from SU-SCORE by SU-AI with PUP; 10-May-83 17:38 PDT
Date: Tue 10 May 83 17:35:19-PDT
From: Rita Leibovitz <RITA@SCORE>
Subject: Package for you
To: jmc@SAIL
cc: sharon@SCORE
Stanford-Phone: (415) 497-4365
John, Sharon Bergman has asked me to mail you a msg. that you have a package
from Federal Express and it is in the Reception Area to pick-up.
Thanks.
rita and sharon bergman
-------
∂10-May-83 1814 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN paradigms
Received: from MIT-MC by SU-AI with TCP/SMTP; 10 May 83 18:14:19 PDT
Date: Tue, 10 May 1983 21:00 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: paradigms
From: John McCarthy <JMC at SU-AI>
Once upon a time Thomas Kuhn wrote a book called "The Structure of
Scientific Revolutions" that introduced the notion of "paradigm
shift" with such examples as the transition from medieval to the
Galileo-Newton view of mechanics and the transition from classical
to quantum mechanical views. Now every PhD thesis proposes its
own paradigm shift . . .
Right. Kuhn's book has touched off a new rage in pseudo-scientific
fashion -- "Claim You've Made a New Paradigm." Now of course most of
these claims are foolish. But claiming a paradigm shift does not a
paradigm shift make. If I recall my Kuhn correctly (more or less), an
actual paradigm shift requires that a group of scientists (linguistic
community) start engaging in normal science within this paradigm.
Whether or not most claims to paradigm shift actually result in a
paradigm shift is irrelevant to the point to which I believe you're
referring. No paradigm shift could ever be made by any scientist who
assumes that the knowledge (laws and observations) of his colleagues
and elders is objective.
∂10-May-83 1848 FWH
To: JMC
CC: DCL
B. Research and Development for Advanced Programming Environments
-----------------------------------------------------------------
Stanford will investigate advanced techniques for production and maintenance
of software. This research has the goal of developing automated aids for all
stages of the production of software systems (requirements, design,
implementation, and maintenance). New high level languages will be designed.
Tools providing automated support for these languages will be developed and
techniques for integrating them into programming environments will be studied.
Special emphasis is placed on the production of new complex systems that
utilize concurrent and distributed processing. The research will focus
primarily on software, but applicability to hardware design will also be
studied.
Specific projects will be undertaken within each of the following tasks:
1. Design new high level languages for formulation of systems requirements,
design specifications, formal annotation of implementations, and
documentation. Such languages are to be machine processable.
2. Develop techniques and guidelines for specifying systems designs and
implementations in these languages.
3. Design, implement, and test tools supporting software production in these
languages. Such support tools will include tools for (i) testing and
validation of requirements, design specifications, and implementations, (ii)
compilation of specifications for rapid prototyping, and (iii) formal
consistency analysis.
4. Develop techniques and tools for analyzing the parallel activity in
systems; this includes consistency analysis, analysis of runtime behavior, and
analysis of communication errors among parallel threads of control.
5. Study the formalization of programming knowledge and its codification in
the broad spectrum languages (task 1) for use in automated knowledge-based
programming assistance.
6. Develop techniques for computer-supported reasoning as required for
advanced automated tools under tasks 1 through 5 above.
7. Design of advanced programming environments that integrate tools developed
under tasks 1 through 6 above.
∂10-May-83 1848 FWH SOW
You just got our SOW
- David
∂10-May-83 1849 FWH SOW
John,
Please let me have a copy of the final combined SOW that you
send out to Ohlander etc.
This is DCL in disguise
- David
∂10-May-83 2250 NOVAK@SUMEX-AIM Petition
Received: from SUMEX-AIM by SU-AI with PUP; 10-May-83 22:49 PDT
Date: Tue 10 May 83 22:50:38-PDT
From: Gordon Novak <NOVAK@SUMEX-AIM>
Subject: Petition
To: jmc@SAIL
Three cheers for your petition re Hoover! I wish to sign.
-- Gordon
-------
I shall be out of town so please phone Prof. Alphonse Juilland.
∂11-May-83 0224 POURNE@MIT-MC test
Received: from MIT-MC by SU-AI with TCP/SMTP; 11 May 83 02:24:13 PDT
Date: 11 May 1983 05:26 EDT
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: test
To: POURNELLE @ MIT-MC, JMC @ SU-AI
I'm told that we can now get through to SAIL without kludges. T
his is a test to see. Meanwhile, I have got a numnber of
letters going to Bugliarello. Who knows...
Best wishes,
Your message sent at 5:26 arrived at 2:24.
∂11-May-83 1105 Mailer@SCORE Message of 8-May-83 10:49:32
Received: from SU-SCORE by SU-AI with PUP; 11-May-83 11:05 PDT
Date: Wed 11 May 83 11:08:09-PDT
From: The Mailer Daemon <Mailer@SCORE>
To: JMC@SAIL
Subject: Message of 8-May-83 10:49:32
Message undeliverable and dequeued after 3 days:
RMS@MIT-AI.ARPA: Cannot connect to host.
------------
Received: from SAIL by SCORE with Pup; Sun 8 May 83 10:49:31-PDT
Date: 08 May 83 1041 PDT
From: John McCarthy <JMC@SU-AI>
Subject: mailing to bulletin boards
To: RMS%SU-AI@SU-SCORE
You can mail from E by <ctrl><meta>XMAIL SU-BBOARDS<cr>.
It is worthwile to make the first line of your message a subject
in which case the command should be <ctrl><meta>XMAIL SU-BBOARDS/SU<cr>.
This will mail the current page to all bulletin boards - currently
several tens and twenties and vaxen.
-------
∂11-May-83 1109 @MIT-MC:LEVITT%MIT-OZ%MIT-MC@SU-DSN paradigms
Received: from MIT-MC by SU-AI with TCP/SMTP; 11 May 83 11:08:56 PDT
Date: Wed, 11 May 1983 13:19 EDT
From: LEVITT%MIT-OZ%MIT-MC@SU-DSN
To: GAVAN%MIT-OZ%MIT-MC@SU-DSN
Cc: John McCarthy <JMC@SU-AI>, phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: paradigms
In-reply-to: Msg of 10 May 1983 21:00-EDT from GAVAN
From: GAVAN
To: John McCarthy <JMC at SU-AI>
cc: phil-sci
Re: paradigms
...
Whether or not most claims to paradigm shift actually result in a
paradigm shift is irrelevant to the point to which I believe you're
referring. No paradigm shift could ever be made by any scientist who
assumes that the knowledge (laws and observations) of his colleagues
and elders is objective.
I hate to get into this, but obviously Einstein's discovery was a
reformulation of "objective", annoyingly repeatable observations that
"ether" and other theories failed to explain. By clumping theories
with observations, you're obscuring your point.
∂11-May-83 1333 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN natural kinds
Received: from MIT-MC by SU-AI with TCP/SMTP; 11 May 83 13:33:23 PDT
Date: 11 May 1983 1517-EDT
From: RICKL%MIT-OZ%MIT-MC@SU-DSN
Subject: natural kinds
To: minsky%MIT-OZ%MIT-MC@SU-DSN
cc: batali%MIT-OZ%MIT-MC@SU-DSN, dam%MIT-OZ%MIT-MC@SU-DSN, jmc@SU-AI,
phil-sci%MIT-OZ%MIT-MC@SU-DSN
From: MINSKY@MIT-OZ
Subject: natural kinds
Some sentences ... began this way:
If we agree that cows and horses are objectively
discontinuous, then ...
But if cows are discontinuous with everything else then ...
And so on. Then RICKL asks: "Do we run into any problems if we take
this to be the defining "essential property" of natural kinds?
As you recall, I then went on to point out in detail one of the problems
which such a position encounters, the gist of which argument was that it
is one's *theories* (conceptions) about the world which determine what
counts as objectively discontinuous, hence that what you take to be
"natural kinds" changes as your view of the world does. This is why the
above were phrased "*If* we agree...", "But *if* cows...", "... *if* we
take...", etc. However, in light of your & DAM's examples I am not sure
that even a static theory-relative notion of "discontinuity" works
completely, even in the restricted world of science.
Just as we think of solids and liquids - and will
continue to - even with all those jellies, jams, goos, pastes, slimes,
glops, smears, and what have you.
Date: Tue, 10 May 1983 12:37 EDT
From: DAM@MIT-OZ
The best example I can think of is a sailors description of wind....
Another example would be the famous case of the eskimo's vocabulary
for types of snow.
and it is very clear that we do have theories of liquids (van der Waal's
forces), solids, meteorology, and that Eskimos could probably develop
theories of snow. It is possible that liquids and solids could be
disambiguated --- e.g. crystalline structure for solids --- but this
is not certain, & probably hopeless for weather. Perhaps another
example would be geology, and gradations of rocks.
Hmmm....
From: MINSKY
What I think IS a good idea is that, given certain pairs
(WORLD, MIND) then certain "events" in that WORLD will induce (often,
dependning on lots of other things) certain SYMBOLS, or stereotypes,
etc., in that kind of MIND.
You have my permission to call these "natural kinds mod(MIND)" for
that WORLD.
I am very interested in your idea of the distinction between "natural
kinds mod(MIND)" and just "SYMBOL". Unless you think that *every*
symbol represents a "natural kind mod(MIND)", I infer from your message
that certain "events" induce what you call "natural kinds mod(MIND)" and
certain other "events" don't. A lot of the "natural kinds" discussion
on the net has been how to distinguish these. Could you clarify further?
-=*=- rick
-------
∂11-May-83 1613 @MIT-MC:MINSKY%MIT-OZ%MIT-MC@SU-DSN natural kinds
Received: from MIT-MC by SU-AI with TCP/SMTP; 11 May 83 16:12:21 PDT
Date: Wed, 11 May 1983 19:02 EDT
From: MINSKY%MIT-OZ%MIT-MC@SU-DSN
To: RICKL%MIT-OZ%MIT-MC@SU-DSN
Cc: batali%MIT-OZ%MIT-MC@SU-DSN, dam%MIT-OZ%MIT-MC@SU-DSN, jmc@SU-AI,
phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: natural kinds
In-reply-to: Msg of 11 May 1983 15:17-EDT from RICKL
I am working on an essay about "the a priori" which will include
sparseness theory and some other stuff. The following reply
to a note from JMC explains some of my views:
I think you're all wasting valuable time rehashing unsuccessful
philosophical ideas like natural kind. The psychology of it all just
seems miserably antedated. (Except RICKL and LEVITT and BATALI,
usually.) Gee. If you'd only appreciate the pwoer of new AI ideas,
you could solve these problems instead of trying to make its predecessor's
failures seem less miserable.
Date: Tuesday, 10 May 1983 02:50-EDT
From: MINSKY
To: John McCarthy <JMC%SU-AI at SU-DSN>
cc: minsky%oz at MIT-ML
I'm afraid I still think this is trivial, for several reasons. First,
as I said, I don't think it makes much difference whether there is a
thin bridge or not - or even a think one, for the reasons I gave.
I don't think there are significaltly fewer problems, psychologically -
and that there only seem fewer to those philosophers.
1. I, too regard as wrong to suppose that all concepts have definitions.
I hope you've read "Learning Meaning". It may happen, if a concept
proceeds smoothly (perhaps by accident) to never need an
"accumulation" that it may apepar to have a definition - but that's a
psychological accident.
2. In Learning Meaning, I explain that when a blue lemon comes up, you
may be able to then make "lemon" into an accumulation.
Psychologically, this will be OK if you are in a fruit context that
already regards color as not so essential - no so default-rigid in the
frame from which the "lemon" concept was partly inherited from. If
there was no solid such precorsor, then you'll just replace "yellow"
by "blue or yellow" or by "any color". My paper (and Winston's
thesis) discusses why one might choose one or the other of these
strategies under different circumstances.
3. If you read Learning meaning, you'll see that this is far beyond
the stereotype of "prototype" that you appear to believe I'm stuck in.
A meaning is a large web of different structures, with different ways
to move around depending on the intentional context.
4. Under extreme conditions, one might allow all parts of the
representation to drift and get anything to be called "lemon". But
that won;t happen if there is some intelligence and a big network
about the world; instead the child will decide that there's something
wrong. "What makes this work?", you ask. It works, not because the
"natural kinds" are discrete, but because we can reason only by
"triangulating" a continuous space by a discrete structure supplied by
the mind - like homology theory. Thus I think you and the
philosophers have things PRECISELY WRONG. We could still have minds
almost as good as ours in a continuous world, provided that the laws
of nature were also continuous!
5. I maintain that the problems just aren't significantly simpler!
6. I don't say it is linguistic at all. I say it is a product of the
ability to symbolize, without which thinking is unlikely. Look at
your own example - color. They're continuous, but it would be useless
to perceive them as such. And it does little harm so far as I can
see, to do otherwise. It is better to think of "red fruit" and "green
fruit" than "fruit".
7. When there is a large, discrete separation, robots will recognize
this by observing that there aren't any caveats or conditions or
appendages attached to a representation. Fine. Also inevitable.
No special provision need be made! I usually will just emerge from
the non-necessity to make an accumulation or difference-network.
So long as you believe that knowledge can be perfectly modular, and
resist the need for "propositions" to be embedded in networks with
comments and heuristics concerning their use, you'll be fooled by
these attractive but psychologically pathological examples of special
clarity. I value your other technical copntributions, like
work on counterfactuals and factoring machines, etc., but this
"natural kind" is attractive to you for a wrong reason, and deflects
away from the important problems.
∂11-May-83 2015 Hewitt%MIT-OZ@MIT-MC your seminar
Received: from SU-DSN by SU-AI with PUP; 11-May-83 20:15 PDT
Received: From MIT-MC by SU-DSN.ARPA; Wed May 11 20:18:33 1983
Date: Wednesday, 11 May 1983, 23:15-EDT
From: Carl Hewitt <Hewitt%MIT-OZ@MIT-MC>
Subject: your seminar
To: jmc%SU-AI@SU-DSN
Cc: Hewitt%MIT-OZ@MIT-MC, cobb%MIT-OZ@MIT-MC
John,
We have arranged for your seminar this Friday at 2:00PM.
Would you be interested in participating
a Chinese Banquet that evening?
Cheers,
Carl
∂11-May-83 2023 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN paradigms
Received: from MIT-MC by SU-AI with TCP/SMTP; 11 May 83 20:23:18 PDT
Date: Wed, 11 May 1983 22:13 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: LEVITT%MIT-OZ%MIT-MC@SU-DSN
Cc: John McCarthy <JMC@SU-AI>, phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: paradigms
In-reply-to: Msg of 11 May 1983 13:19-EDT from LEVITT
From: LEVITT
I hate to get into this, but obviously Einstein's discovery was a
reformulation of "objective", annoyingly repeatable observations that
"ether" and other theories failed to explain. By clumping theories
with observations, you're obscuring your point.
I understand what you're saying, and I agree that this would tend to
obscure the point. But the point still stands. There are also
innumerable cases where observations were dismissed because they did
not conform to "objective" theories. Galileo comes to mind.
A PHIL-SCI participant told me (personally) last night that the
objectivists (as I'll call them) cling adamently to their objectivism
because they would like to bring to bear (in the development of a
robot) some particularly elegant and powerful mathematical formalisms
which they believe would be precluded by a subjectivist metaphysics.
Is this correct? Which formalisms? Why would subjectivism preclude
them?
∂12-May-83 0220 GOLUB@SCORE Re: DARPA scope document
Received: from SU-SCORE by SU-AI with PUP; 12-May-83 02:20 PDT
Date: Thu 12 May 83 02:23:38-PDT
From: Gene Golub <GOLUB@SCORE>
Subject: Re: DARPA scope document
To: JMC@SAIL
In-Reply-To: Your message of Tue 10 May 83 16:55:00-PDT
Thanks John for taking care of this.
GENE
-------
∂12-May-83 0515 Solomon.Datanet@MIT-MULTICS.ARPA memo for weekend seminar
Received: from MIT-MULTICS by SU-AI with TCP/SMTP; 12 May 83 05:10:03 PDT
Date: 11 May 1983 21:28 edt
From: Solomon at MIT-MULTICS (Richard Jay Solomon)
Subject: memo for weekend seminar
To: jmc at SU-AI
cc: Solomon.Datanet at MIT-MULTICS (hold.sv)
Acknowledge-To: Solomon.Datanet at MIT-MULTICS
HISTORY OF THE COMPUTER: Timesharing Origins Page 1
INTRODUCTION TO THE COMPUTER HISTORY PROJECT
The discussion on the origins of timesharing in Cambridge and its
subsequent role on digital computer developments, which we are to hold
on May 14-15, is one of a series of videotaped sessions on the
history of the computer which have been funded at MIT by the Alfred P.
Sloan Foundation.
This is not a TV program. There is no script. The sessions are to be
informal discussions among people who worked together during various
critical phases of this history. Its purpose is to prepare an
archive in order to aid future historians.
The videotaped discussion is intended to be relaxed. (1) There is no
audience, other than the video technicians and related staff of the
MIT Research Program on Communications Policy. If there is anything
you wish to say off the record, just indicate this and we will stop the
recording, or even erase comments you made previously. You can feel
free to interrupt at will, leave the room as you need to, and change the
subject if you think the questions are not to the point.
The purpose of this memo is to stimulate discussion. Please do not
feel bound by our outline or topics. The series as a whole has two
purposes:
1. To recapture for an archival historical record,
while the pioneers are still with us, the mood and
spirit of the time during which they worked. History
is always a rewriting of the past from the perspective
of the present. However, to the extent that it is
possible, we wish to have our discussants tell us what
they were thinking about at the time; how problems and
difficulties had been perceived then; and what their
expectations and hopes for the technologies had been.
To this end, we have brought together small groups of
people during these videotapings who have worked with
each other in the past so that they can discuss and
recollect what they had done and thought on a daily
basis.
2. As the unifying intellectual theme, we want to
trace what streams of ideas came together in the
various early efforts that we are reviewing. It is
recognized that the computer is not a single invention,
←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←← (1) The
studio will be set up as if it were a living room. Informality
extends to your clothing. Though lighting will be kept at a minimum,
the studio may get warm after a number of hours. Wear whatever you
feel comfortable in. Colored shirts are best; and horizontal,
finely-striped material should be avoided.
May 14-15, 1983 MIT Research Program on Communications Policy ! HISTORY OF THE COMPUTER: Timesharing Origins Page 2
but grew from diverse sources, often from things
brought together for purposes different from that of
computing. So as we look at any point in the past we
would like to know:
o What were your conceptions of what a computer was, or
should be, and what a computer would or could do? How
did the evolution of machines, computing, or
electronics, during the days before the first
timesharing machines had been demonstrated, affect your
conception of what was happening and what was needed to
fulfill your notions about computers and computing?
o What intellectual influences and previous experiences
do you think had been seminal in each new effort toward
the modern computer and timesharing?
o What combinations of previously unrelated skills and
application of technology had been coming together to
make new and useful tools for the evolving computer
community?
o What pivotal innovations had been added to the previous
store of knowledge and how did this change perhaps
earlier perspectives on what was needed to make
interactive computing practical?
In regard to the pieces that came together, certain hardware and
software advances were, of course, necessary for timesharing and
interactive computing to be conceived and developed. While it is
valuable to discuss how and why this came about from your
perspective, it seems to us other, more fundamental concepts about
convergence of technologies also deserve some attention, for example:
o What was the extent that these designs and conceptions
were driven by need or driven by available technology?
How dominant were the requirements of sponsors or
particular technical problems for which prototype
elements had been originally intended? How much was
there a "Mount Everest phenomenon" -- one believed that
a problem could now be conquered, so conquered it must
be.
o Related is the question of convergence of differing
technological disciplines: How did these traditions
begin to merge into the interactive computer; a
personal tool, perhaps best expressed by today's
microcomputers? How did you see the relation of the
novel concepts that you were working with to the
batch-processors and number crunchers of the day?
May 14-15, 1983 MIT Research Program on Communications Policy ! HISTORY OF THE COMPUTER: Timesharing Origins Page 3
One issue that inevitably arises in all discussions of the
history of technology, but which is not our central concern, is that
of exact priorities and sequences. It is an important issue on which
those who participate may certainly from time to time want to express
their convictions, but we do not believe that the video medium is the
best for establishing precise facts that need documentation. The
documented written record is far better for that. Videotape may be
able to better capture feelings and impressions about what
important players had been doing. We want you to try to recall how you
saw things. We are not asking each of you to be an historical
encyclopaedia.
With that in mind, we will present some thoughts and questions about
the evolution of timesharing in Cambridge, during the period of the
late 1950's to the beginnings of Project MAC and commercial
timesharing projects around 1963, in order to help focus the
discussion:
Origins of Timesharing in Cambridge
A simple description of events may aid in structuring the time frame,
though details abound in a number of books and articles:
Digital computers, by the late 1950's had moved from a series of
experimental machines, to a set of mostly commercially-produced
devices following what has loosely been known as the "von
Neumann" architecture. At the same time, we had the introduction of
high-level programming languages, which permitted a better
utilization of programmers' time for certain classes of problems. But
access to machines, though now demanded by a larger number of users, had
become somewhat remote -- by today's standards, at least.
There were a few attempts to make computers more "accessible". Some
applications, such as the SAGE air defense system, implied realtime
access by users, but the programming tasks were not necessarily for
the immediate user to worry about. And for varied reasons, certain
machines, such as TX-0, became available for a small number of
researchers to use directly, in their own time frame.
It was during this period that the concept of giving programmers direct
access to computing, via interactive input/output devices, began to
surface.
Our panelists know the history best; they have written about their
experiences and have delivered lectures on the subject. It would be
presumptuous for this memo to attempt to more than cursorily outline
what happened; and in the discussion, we expect that each panelist's
memory will be a better "jog" for the others than anything that can be
written here.
We want to discuss how people worked together; what they thought
May 14-15, 1983 MIT Research Program on Communications Policy ! HISTORY OF THE COMPUTER: Timesharing Origins Page 4
they were doing; what they thought others, including the "outside
world", were doing; where they thought their contributions were going;
what they needed, and what they wanted to accomplish these goals; levels
of satisfaction, inspiration and frustration in their work; how
problems were approached, solved, or abandoned; and specific examples --
and anecdotes -- which illustrate why and how certain philosophies
were followed or not followed, and why basic decisions were made to
pursue lines of research and development as it related timesharing
developments, particularly in the Cambridge computer community during
the early days.
We hope to distill how the evolution of ideas and cross-
fertilization works between individuals, what constitutes useful
breeding grounds for novel concepts, how one recognizes an idea, when
a technical insight becomes applicable, and how one decides to abandon a
specific route for alternatives.
The evolution of timesharing should be considered in these terms. Many
things happened during those days, but the following list appears to
us to be the important ones to cover in the videotaped discussion.
However we will leave it to the panelists to choose what they think
are the most important to be discussed, especially if such items
or concepts have been left off the list:
1) The background of people is important. We would
like to have you discuss your own background; how you
became interested in technology, mathematics,
electronics, engineering, computing, etc.; what you may
have done which inspired these avocations in grade
school, high school, and before you had completed your
college-level studies; how these interests evolved
until, and after you joined key projects; what
professional interests were affected afterwards; and
what you knew of others who made significant
contributions at the time.
2) We need a definition of timesharing. What do you
consider to be the significant elements? Shared
resources? Interactive programming or computation?
Telecommunications? A special type of machine
architecture? Specific applications?
3) How did the original problems related to timesharing
affect perceptions as to what had to be done after
things were completed. What was known of other
efforts, contemporary and historical to solve similar
problems. Did new problem-oriented goals later affect
designs? What types of machines and services may have
anticipated the development of specific timesharing and
interactive computing components?
4) Which components for the evolution of computer
May 14-15, 1983 MIT Research Program on Communications Policy ! HISTORY OF THE COMPUTER: Timesharing Origins Page 5
timesharing were critical and required novel and
imaginative application? How was that approached?
Some components, from today's point of view, seem to
have been very important to the ultimate success of
timesharing and interactive machines. These appear to
be expanded memory, faster cpu's, expanded mass
storage, better input/output devices for the
interactive user, improved telecommunications links.
How much did existing machines and systems have to be
modified for timesharing and what were the
modifications?
What other hardware problems did you have and how did
the engineering solutions affected the architecture of
later machines and systems?
When and how did the concepts of an operating system to
control resources among users begin to evolve? What
about security, privacy, and reliability questions in
multiple access? How did these differ from single-user
and batch processes as perceived at the time? What
about charging for use? How may have these problems and
concepts have actually to the concept and demonstration
of timesharing and interactive computing?
5) Describe your interactions in the working
environment. Where and when did you talk about plans,
problems, and solutions? Over lunch at the F&T? In
classrooms? In your homes?
6) Retrospective questions are very subjective, but
instructive to future generations. So, in retrospect,
and for varying time cycles after specific ideas
related to interactive computing and timesharing were
tested or completed: what worked and what did not work
in terms of technology, process, environment, funding,
and people? What did you perceive was happening in
electronics related to computers and computing, at the
time and immediately afterwards, a few years later when
more advanced devices were in the works, a decade later
when timesharing was generally accepted for certain
applications? What of your experiences, and that of the
computer community, helped, hindered and led? Would
there have been better ways to learn about the problems
and opportunities, or worse ways?
7) Along with the machines and computer developments
that have been well documented, work was going forward
in cryptography and other fields which may still remain
secret. Without probing classified activities, we
would like to enquire how aware you were of such
May 14-15, 1983 MIT Research Program on Communications Policy ! HISTORY OF THE COMPUTER: Timesharing Origins Page 6
activities, and indeed, of their very existence. Were
there individuals who served as a bridge? Who were
they? What contributions, if any, did they make to the
development of interactive and timeshared computers in
the civilian world? Or did secrecy make such work a
dead end?
8) Where did the resources to develop various
components of timesharing and interactive computing
come from? How much was internally generated and how
much from government and industry? Was this
satisfactory?
9) How much did computing theory affect your thinking
about the electronics or mechanism of timeshared,
interactive computers, and if so, in what ways?
10) What other seminal technologies or theories were
important to digital computer developments at that time
and how did you and others perceive their evolution and
utility? As examples, developments were proceeding
apace in integrated circuits, highspeed cpu's, magnetic
storage devices, and in different types of software and
applications. Did the work in artificial intelligence,
list processing languages and realtime interpretive
languages have anything to do with ideas related to
timesharing?
11) On even the first prototypical interactive
machines, the initial excitment of getting one's hands
on such a tool may have been similar to today's
generation when they first get their own microcomputer.
How did this affect your impression of what you could
do and how your life might change because of computers?
Did you try out new and perhaps trivial ideas, and
generally attempt serendipitous experiments, even
games, with the machine? At what stage did you begin
to think about users as being other than technical
professionals? When did you start thinking about
devices that would be used by laymen?
12) What else would you like to say to future
generations of computing professionals and to
historians?
May 14-15, 1983 MIT Research Program on Communications Policy !
∂12-May-83 0728 @MIT-MC:RICKL%MIT-OZ%MIT-MC@SU-DSN counterfactuals & etc.
Received: from MIT-MC by SU-AI with TCP/SMTP; 12 May 83 07:28:35 PDT
Date: 12 May 1983 1023-EDT
From: RICKL%MIT-OZ%MIT-MC@SU-DSN
Subject: counterfactuals & etc.
To: jmc@SU-AI
cc: rickl%MIT-OZ%MIT-MC@SU-DSN
marvin's last message referred to some work you had
done on counterfactuals --- could you please provide
me with a pointer to it?
unfortunately i will probably not be able to meet
you when you are out here --- at 2:00 friday i will
be in a philosophy of science seminar with tom
kuhn, it turns out, somehow things always seem
to conflict --- but i will be out in the golden
state (bay area even) this summer & will look you
up unless you go elsewhere.
-=*=- rick
-------
∂12-May-83 1554 RV qual details
As I recall, we arranged for my qual to be on the 17th.
Have you figured out what time it will be given and who
will be on the committee? That information would help
me greatly.
Thanks,
Rick Vistnes
∂12-May-83 1829 NILSSON%SRI-AI.ARPA@SCORE Online abstracts and articles
Received: from SU-SCORE by SU-AI with PUP; 12-May-83 18:29 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Thu 12 May 83 18:33:09-PDT
Date: 12 May 1983 1823-PDT
From: Nilsson at SRI-AI
Subject: Online abstracts and articles
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-MC,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL
A few months ago I circulated a rough proposal for AAAI getting involved
in making abstracts and (eventually) articles available online. I did get
some feedback about the idea that convinced me that there a number of
things to think about in order to get started. What I would now like
to ask is for suggestions about who would like to be the "chief worrier
and implementer" of this idea. Perhaps one of you. Perhaps someone you
know who would like to do this and would be good at it. If we could
identify such a person in the next few weeks, that person (together with,
perhaps, Lou Robinson, who has expressed interest in the matter) could
get started with their preliminary thinking and then come to the excom
mtg in August with a specific proposal. Excom mtgs being what they
are, we will probably say (if we like the idea), "Great! Why don't you
go off and do it?" So, do I have a volunteer? --Nils
-------
∂12-May-83 1903 NILSSON%SRI-AI.ARPA@SCORE Online abstracts and articles
Received: from SU-SCORE by SU-AI with PUP; 12-May-83 19:02 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Thu 12 May 83 19:00:34-PDT
Date: 12 May 1983 1823-PDT
From: Nilsson at SRI-AI
Subject: Online abstracts and articles
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-MC,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL
A few months ago I circulated a rough proposal for AAAI getting involved
in making abstracts and (eventually) articles available online. I did get
some feedback about the idea that convinced me that there a number of
things to think about in order to get started. What I would now like
to ask is for suggestions about who would like to be the "chief worrier
and implementer" of this idea. Perhaps one of you. Perhaps someone you
know who would like to do this and would be good at it. If we could
identify such a person in the next few weeks, that person (together with,
perhaps, Lou Robinson, who has expressed interest in the matter) could
get started with their preliminary thinking and then come to the excom
mtg in August with a specific proposal. Excom mtgs being what they
are, we will probably say (if we like the idea), "Great! Why don't you
go off and do it?" So, do I have a volunteer? --Nils
-------
∂12-May-83 1903 NILSSON%SRI-AI.ARPA@SCORE Agenda Item
Received: from SU-SCORE by SU-AI with PUP; 12-May-83 19:03 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Thu 12 May 83 19:00:46-PDT
Date: 12 May 1983 1833-PDT
From: Nilsson at SRI-AI
Subject: Agenda Item
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-MC,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL
Here's an item I'll put on the agenda for our excom mtg in August.
People might like to think a bit about it ahead of time. Rick
Hayes-Roth has agreed to discuss the matter at the excom mtg. -Nils
7-May-83 07:20:43-PDT,2378;000000000001
Return-Path: RHAYES-ROTH@SRI-KL.ARPA
Received: from SRI-KL.ARPA by SRI-AI.ARPA with TCP; Sat 7 May 83 07:20:29-PDT
Date: Sat 7 May 83 07:14:32-PDT
From: RHAYES-ROTH@SRI-KL.ARPA
Subject: AAAI Exec Mtg Agenda Item--Professional Society
To: reddy@CMU-CS-A.ARPA, nilsson@SRI-KL.ARPA, aaai@SRI-KL.ARPA
cc: bengelmore@SRI-KL.ARPA
Raj & Nils:
I'd like to suggest that you bring before the AAAI Exec
Committee a serious proposal regarding widening the AAAI to
make it both a professional society and make it serve the needs
of practicing AI workers. The current association has conveyed
to its members and outsiders who come to its national conference
an image of academic narrowness. While the academic work of
the field's leaders is surely critical to the future of American
AI, a perception that the AAAI is purely of academic interest
will surely limit the association's growth, significance, and
appeal.
Specifically, I believe the Association should encourage
the development of several parallel lines internally: research,
education, applications, and commercial standards. These activity
and interest areas would parallel those pursued by other successful
organizations, such as the IEEE and ACM. The first steps needed,
it seems to me, include: (1) having the current executive committee
adopt a resolution that affirms the AAAI's interest in the broad set
of issues confronting the AI community; (2) adapting future AAAI
conference programs to reflect this diversity of audiences; (3) seeking
to broaden the executive committee to include members with interests
spanning these diverse areas; and (4) publicizing the widened and
increased professional orientation of the society as part of an effort
to recruit more industrially and commercially oriented members.
Since coming to Teknowledge, I have heard frequent criticisms
of AAAI by industrial clients who tried it and found it irrelevant to
their concerns. In my opinion, technical societies that appear irrelevant
cannot achieve their potential. Other competing societies will arise
and, because they will serve more interests they will eventually
dominate the first. I think the AAAI has reached a critical decision
point in this regard.
Please let me know if I can be of any help in clarifying issues
or developing plans.
Cheers,
Rick
-------
-------
∂13-May-83 0816 @MIT-MC:GAVAN%MIT-OZ%MIT-MC@SU-DSN [CSTACY: forwarded]
Received: from MIT-MC by SU-AI with TCP/SMTP; 13 May 83 08:16:21 PDT
Date: Fri, 13 May 1983 11:13 EDT
From: GAVAN%MIT-OZ%MIT-MC@SU-DSN
To: jmc@su-ai
Subject: [CSTACY: forwarded]
Date: 12 May 1983 10:32 EDT
From: Christopher C. Stacy <CSTACY at MIT-MC>
To: GAVAN
All he has to do is type the "send a ↑Z command through" athis TELNET
program. I dont remember how to do this on WAITS, but I forwarded
his message to their main system hacker.
∂13-May-83 1001 Solomon.Datanet@MIT-MULTICS.ARPA Saturday and Sunday
Received: from MIT-MULTICS by SU-AI with TCP/SMTP; 13 May 83 10:00:56 PDT
Date: 13 May 1983 12:57 edt
From: Solomon at MIT-MULTICS (Richard Jay Solomon)
Subject: Saturday and Sunday
To: jmc at SU-AI
cc: Solomon.Datanet at MIT-MULTICS (hold.sv)
Acknowledge-To: Solomon.Datanet at MIT-MULTICS
We meet in Bldg 9 off Mass Avenue on the Third Floor video studios at 11
AM both Sat and Sunday. We expect to go through 5PM, but will probably
break up earlier onsunday. Wear comfortable clothes, it gets a bit warm.
If Bldg 9is locked on Saturday, just go through 77 Mass avenue. The
first floor ends up onthe 2nd of Bldg 9.
Call me at 413 267 5171 if you have any questions. I will try to get you
at the Parker House.
Richard Solomon.
∂13-May-83 1037 DBL ai qual
John,
Is the AI Qual still this Tuesday, May 17?
If so, what times (recall that 10:30-12:15 and 4-5 are bad for me)?
Do the students know where and when?
If not: when is it?
Doug
∂13-May-83 1108 DFH AI Qual
To: feigenbaum@SU-SCORE, lenat@SU-SCORE, JMC@SU-AI
To: Doug Lenat, Ed Feigenbaum
cc: JMC
From: Diana Hall (dfh@su-ai), secretary to John McCarthy
Subject: AI Qual
This is just to remind you of the AI Qual which is to be held
on Tuesday, May 17. I am going to tell the three students involved
to meet you in room 352 at 9:30 am. The room is reserved until
1:30 pm. I realize that this is probably not enough time, but it
was the longest stretch I could reserve in one room. If necessary,
you could switch to a different room after 1:30.
∂13-May-83 1349 DFH AI Qual
I reserved room 352 from 9:30 am - 1:30 pm on
May 17, and I sent messages regarding this to
Lenat, Feigenbaum, and the three students
involved. One of the students, Stuart Russell,
had not been informed of the date prior to
my telling him, however, and wants to make
arrangements with you to take it at a later
time. He will probably contact you about this
on Monday.
∂13-May-83 1447 DFH Ben Grosof
Could Ben Grosof's summer support be charged
to Jussi's account? There seems to be a lot of
SRA money there. -- Diana
∂13-May-83 1518 minker.umcp-cs@UDel-Relay Special Session on Non-Monotonic Logic
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 13 May 83 15:18:00 PDT
Date: 13 May 83 14:01:27 EDT (Fri)
From: JACK MINKER <minker.umcp-cs@UDel-Relay>
Return-Path: <minker.umcp-cs@UDel-Relay>
Subject: Special Session on Non-Monotonic Logic
To: Bundy@RUTGERS
Cc: "uw-beaver!ubc-vision!Reiter"@LBL-UNIX, Moore@SRI-AI, McDermott@YALE,
JMC@SU-AI, Amarel@RUTGERS
Via: UMCP-CS; 13 May 83 18:04-EDT
Dear Alan,
The special session that I am organizing should be entitled:
Non-Monotonic Logic - In Honor of Alexander Lerner's
70th Birthday
Participants: John McCarthy
Drew McDermott
Jack Minker, Chairman
Robert Moore
Raymond Reiter
All of the above have agreed to participate in the special
session. I have not been able to reach John McCarthy, but from a
previous conversation and from Saul Amarel and Bob Moore, I have no
doubt that he will participate. John is out of town and won't be
back until Monday. It would be nice to include all of the participants
in the programme brochure.
As far as the time for the special session, I would like to
leave this up to your discretion. When I originally proposed the
session I was unaware that other special sessions were being organized.
I suggested the Wednesday evening since I did not want to conflict with
your program. It would be best to schedule the session any time after
Bob Moore gives his scheduled talk at IJCAI and so that it does not
conflict with speaking commitments of the other participants. Moore's
talk will be on non-monotonic logic and hence, it would be best to
schedule the special session any time after his session.
Thank you for your cooperation. I especially appreciate your
very prompt responses to my messages.
Best regards,
Jack
∂13-May-83 1614 @MIT-MC:LEVITT%MIT-OZ%MIT-MC@SU-DSN oops -- corrected; back to data types
Received: from MIT-MC by SU-AI with TCP/SMTP; 13 May 83 16:14:28 PDT
Date: Fri, 13 May 1983 19:03 EDT
From: LEVITT%MIT-OZ%MIT-MC@SU-DSN
To: RICKL%MIT-OZ%MIT-MC@SU-DSN, jmc%MIT-OZ%MIT-MC@SU-DSN
Cc: MINSKY%MIT-OZ%MIT-MC@SU-DSN, phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: oops -- corrected; back to data types
In-reply-to: Msg of 13 May 1983 11:02-EDT from RICKL
The whole point of EVERY PHRASE was that "AI ideas" argue for different
perspectives on a problem, and the whole point of the whole message was
that AI and philosophy (and the other disciplines I mentioned) are different
perspectives on the same problems.
I completely missed this point (as did Minsky), and I'm grateful you
clarified it. I'm sorry that the only realistic interpretation I
could generate was so unflattering, and that I hurt you with my
dramatic language. (Actually, I was trying to be responsible; as the
great linguists have said, needlessly "bad" words remain dangerous
because they're not used widely enough. I wish more people would help
me defuse them.)
I didn't follow because in my mind the analogy is so remote. The idea
of multiple viewpoints is itself very significant, but I don't couple
it with any idea of *democracy* about viewpoints. "Equal time"
regulations make TV *less* informative than simple First Amendment
rights permit, because every interesting new idea must be cut short
and rebutted with a old idea, considered "the other side" by a very
narrow audience. And multiple representation is an incredible waste
if you have no ways of deciding when to use which viewpoint, and which
should be given very low priority or thrown away.
I think JMC and DAM have made this point -- that while analogy and
conjecture are important, they're weak unless you have a theory of how
to make relevant ones. (The same is true of inference, but DAM and
JMC seem to believe logic's a more promising area to study. I'm not
convinced.) I misunderstood you because while the form of your
analogy was perfect, your evidence suggested that only AI has been
productive.
(JMC gave a talk about conjecture today at MIT AI. Any thoughts?)
I don't want our misunderstanding to obscure my real point -- that
"natural kinds" sound like a very primitive idea of a abstract
datatype, an area which is under active experimental study and which I
suspect is more pregnant than viewing common sense as an extension to
logic. JMC, how do you think "objects" and "data types" fit with your
"circumscription" ideas?
∂13-May-83 1701 DFH Computer Facilities Meeting
I have not had a great deal of success on getting
a time that suits everyone. Thursday at 2 I can
get Ullman, Mayr,Gorin, and Mogul. Wed at 10 I
can get Mayr, Novak, Reid, Cheriton, possibly
Oliger (Gorin says he can come, but Ullman says
they are both in a meeting that conflichts.)
Feigenbaum has been away. Friday seems to be
bad because it is the day of the Systems Qual.
∂13-May-83 1707 MAILER failed mail returned
The following message has expired without successful delivery to recipient(s)
RMS@MIT-OZ:
∂10-May-83 1754 JMC
To: RMS@SU-AI
Your manuals are on my file cabinet.
∂14-May-83 0201 ME mail
To: JMC@SU-AI, RMS@SU-AI
∂13-May-83 1652 JMC via mail rms
To: rms@MIT-MC, ME@SU-AI
mail to rms here is forwarded to rms at the now non-existent ai.
I believe MC is the correct place to forward it.
ME - It is forwarded there as of earlier today. AI forwardings were
changed to OZ, but that wasn't working because OZ isn't on the ARPAnet
yet, so they were changed to MC.
∂14-May-83 0238 @MIT-MC:HDT%MIT-OZ%MIT-MC@SU-DSN via mail rms
Received: from MIT-MC by SU-AI with TCP/SMTP; 14 May 83 02:37:47 PDT
Date: 14 May 1983 05:35 EDT (Sat)
From: Howard D. Trachtman <HDT%MIT-OZ%MIT-MC@SU-DSN>
To: John McCarthy <JMC@SU-AI>
Subject: via mail rms
In-reply-to: Msg of 13 May 83 1652 PDT from John McCarthy <JMC at SU-AI>
Date: 13 May 83 1652 PDT
From: John McCarthy <JMC at SU-AI>
To: rms at MIT-MC, ME at SU-AI
Re: via mail rms
mail to rms here is forwarded to rms at the now non-existent ai.
I believe MC is the correct place to forward it.
rms's mail probably should be forwarded to mc, as mc will forward it to oz
which is where he reads it.
ps: I enjoyed your lecture here today (well, it was yesterday by now).
∂14-May-83 2122 DEK DARPA scope document
To: JMC@SU-AI
CC: BS@SU-AI
Your bold asserions (abstracted from my more hedgy prose)
are fine and no changes are needed. Thanks much.
∂14-May-83 2320 Kanef%HP-HULK.HP-Labs%RAND-RELAY.ARPA@SCORE Circumscription paper
Received: from SU-SCORE by SU-AI with PUP; 14-May-83 23:20 PDT
Received: from rand-relay.ARPA by SU-SCORE.ARPA with TCP; Sat 14 May 83 23:24:06-PDT
Date: 14 May 1983 2118-PDT
From: Kanef.HP-HULK@Rand-Relay (Bob Kanefsky)
Return-Path: <Kanef%HP-HULK.HP-Labs@Rand-Relay>
Subject: Circumscription paper
Received: by HP-VENUS via CHAOSNET; 14 May 1983 21:18:06-PDT
To: Kanef%JMC.HP-LABS@Rand-Relay, Kanef%%.HP-LABS@Rand-Relay, SAIL@SU-SCORE
Reply-To: Kanef.HP-Hulk%Rand-Relay@Score
Message-Id: <421820287.10139.hplabs@HP-VENUS>
Via: HP-Labs; 14 May 83 23:11-PDT
I'm writing a paper on presuppositions for a linguistics class, and in it
I mention briefly your notion of circumscription. I want to make sure I
cite your paper ("Circumscription -- A Form of Non-Monotonic Reasoning")
correctly. I only have the copy you handed out in class, so I don't know
if it was ever published. If it was published, could I trouble you for
the publishing data? If not, in what year should I say it was written?
(Currently I'm saying "I don't know if it was ever published, so I
assume it was not". I wonder if anyone will catch the humor in that.)
Thanks.
--Kanef
-------
The paper was published in Artificial Intelligence, April 1980.
∂15-May-83 0747 FEIGENBAUM@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 15-May-83 07:46 PDT
Date: Sun 15 May 83 07:47:46-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
To: DFH@SAIL, csd.ullman@SCORE, csd.Feigenbaum@SCORE, csd.mayr@SCORE,
Computer Facilities Committee Members@SAIL
cc: jmc@SAIL
In-Reply-To: Your message of Wed 11 May 83 11:39:00-PDT
There's an important guest speaker speaking at 1:15 to 2:15 on May 19
so neither I nor my repreentative can be there at that time. I'll get
back to you with other suggested times when some other balls drop out of
the air. (The speaker is Randy Davis from MIT).
Ed Feigenbaum
-------
∂16-May-83 0900 JMC*
∂16-May-83 0927 BSCOTT@SU-SCORE.ARPA ARPA Umbrella
Received: from SU-SCORE by SU-AI with TCP/SMTP; 16 May 83 09:27:32 PDT
Date: Mon 16 May 83 09:30:51-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: ARPA Umbrella
To: DCL@SU-AI.ARPA
cc: JMC@SU-AI.ARPA, ZM@SU-AI.ARPA, TOB@SU-AI.ARPA, Wiederhold@SUMEX-AIM.ARPA,
BScott@SU-SCORE.ARPA
David,
In response to your inquiry, Ron Ohlander stopped by my office briefly on
Friday. He had not had an opportunity to go over the revised scope in-
formation which John sent; he hoped to do this on Sunday, and said he
would send comments today.
Betty
-------
∂16-May-83 0947 DFH
To: JMC
Re: Facilities Committee Meeting
Did you note Feigenbaum's message about the Thurs 2 pm time conflicting
with a speaker? How much time do you need for the meeting? If we made
it 2:30-4, would that be sufficient? (CIS groundbreaking ceremony is
at 4). --Diana
∂16-May-83 1048 DFH Meeting
To: "@FACCOM.[1,DFH]"@SU-AI
To: Facilities Committee
From: Diana Hall (dfh@su-ai)
Subject: Meeting
The meeting will be held Thursday, May 19, from 2:30 - 4:00 pm in the
Chairman's conference room. I am aware that there are a number of you
who cannot make this time, but it has proved impossible to find a time
agreeable to everyone involved.
∂16-May-83 1237 ME leased line
To: JMC@SU-AI
CC: LB@SU-AI
∂16-May-83 1109 JMC line inoperative
The line from my house seems to be inoperative. Is that
part of the rewiring?
ME - Yes. If we're lucky, it'll be working by tonight.
∂16-May-83 2009 JMC*
passport
∂16-May-83 2205 RPG Dialups???
To: REG@SU-AI, ME@SU-AI, LB@SU-AI
CC: JMC@SU-AI
I find the following maddening:
1) the dialups went down with almost no notice,
2) the notice was given via the wrong medium (BBOARD),
3) SAIL dialups seem to be the last ones to come up, since I think
all others are up, and SAIL's are still down.
My work schedule has been ruined by this escapade. In the future warn us.
Bring up some resources from each machine before all from one.
-rpg-
∂17-May-83 1833 ME leased line
Your leased line should now be working. Please let me and LB know whether
or not it is.
∂17-May-83 2139 KARP@SUMEX-AIM Facilities committee meeting Thursday
Received: from SUMEX-AIM by SU-AI with PUP; 17-May-83 21:39 PDT
Date: Tue 17 May 83 21:40:45-PDT
From: Peter Karp <KARP@SUMEX-AIM>
Subject: Facilities committee meeting Thursday
To: jmc@SAIL
Both Ken Brooks and I are interested in being student members on the facilities
committee. I would like to check out the meeting Thursday, and I imagine Ken
will also. Would you tell me when and where the meeting is? Thanks,
Peter
-------
∂17-May-83 2249 RPG
∂17-May-83 2224 JMC common.msg
I think it's time to start a new common.msg; it now takes between
eight and ten seconds of computer time to get E to look at the last page.
Done.
∂18-May-83 0742 OHLANDER@USC-ISI Stanford SOW
Received: from USC-ISI by SU-AI with TCP/SMTP; 18 May 83 07:41:56 PDT
Date: 18 May 1983 0740-PDT
Sender: OHLANDER at USC-ISI
Subject: Stanford SOW
From: OHLANDER at USC-ISI
To: JMC at SU-AI
Cc: BScott at SU-SCORE
Message-ID: <[USC-ISI]18-May-83 07:40:56.OHLANDER>
John,
I have performed considerable work on the draft SOW and I now
believe it to be in reasonable shape. It should be added (modulo any minor
changes that you might want to make) to the proposal. You will note that I
decided to consolidate Luckham's and Manna's work afterall.
There are still a few other problems with the proposal. Some of them are
minor. We would like to have all of the names of the coprincipal investigators
removed from the cover and we would like to have a title for the proposal.
Another more serious problem regards the budget. It has to be spelled out
in detail. I will talk to Betty Scott offline about this.
Duane Adams is going to be at Stanford Tomorrow. I believe he plans to talk
to you about these issues.
Regards, Ron
Statement of Work
Stanford University proposes to conduct a broad program of research in the
field of computer science. The specific research will include: fundamentals of
artificial intelligence; advanced programming techniques and environments; data
management; image understanding and robotics; analysis of algorithms; and
relations between algorithms and architectures. The following specific task
areas are proposed:
1. Basic Research in Artificial Intelligence
Stanford University proposes to conduct research in aspects of basic artificial
intelligence technology with the objectives of making significant advances in
machine reasoning capabilities, automatic speedup of programs, and artificial
intelligence language design and implementation. Examples of the kinds of
tasks that will be undertaken in this area include: development of
formalizations of common sense reasoning and about facts and knowledge used in
the common sense world; development of languages for communications among
computers and between computers and people; development of systems for computer
reasoning and computer-assisted human reasoning and systems that take advice
from users and give advice in return; improvement of artificial intelligence
languages and also performance of research aimed at new languages; and
experimentation with techniques of automatically specializing programs to make
them run faster.
2. Research in Advanced Programming Techniques and Environments
Stanford University also proposes to carry out research in the development of
advanced programming environments and software production with the goals of
providing much better programming tools than currently exist. Examples of the
kinds of tasks that will be undertaken in this area include: design and
implementation of new high level languages for formulation of systems
requirements, design specifications, formal annotation of implementations, and
documentation; design and implementation of an advanced programming environment
that supports software production in these languages; development of techniques
and tools for analyzing the parallel activity in systems and for verifying and
synthesizing concurrent programs; study of the formalization and codification
of programming knowledge and introduce machine reasoning techniques for
software production applications; design of a high-level logic programming
language using new deductive techniques; development of tools for rigorously
establishing properties of software and hardware systems (using logic of
programming such as dynamic logic and temporal logic); and investigation of
program manipulation techniques in such areas as maintenance, transformation,
and optimization.
3. Data Management
Stanford University also proposes to carry out a research project in data
management with a goal of improving data management capabilities by the
development of better algorithms and the incorporation of machine intelligence.
Examples of the kinds of tasks that will be undertaken in this area include:
investigation of advanced techniques to improve database access, storage,
update, and management; development of new techniques for the use of databases
in VLSI design; investigation of techniques for maintaining reliability and
integrity in distributed data base systems; and exploration of methods of
inferring new knowledge from data.
4. Image Understanding and Robotics
Stanford University also proposes to carry out research in image understanding
and robotics with the goals of achieving automatic understanding of images and
sensing and intelligent planning by robotic devices. Examples of the kinds of
tasks that will be undertaken in this area include: analysis, design and
implementation of intelligent systems for interpretation and planning actions;
study of planning, navigation and path-finding, motion control map-making and
world modeling for mobile robots; study, design and development of mechanisms
for interpretation of images; and investigation of architecture of image
algorithms and their implementation in VLSI.
5. Analysis of Algorithms
Stanford University also proposes to conduct research in the analysis of
algorithms with the objective of achieving dramatic speedups in certain classes
of algorithms. Examples of the kinds of tasks that will be undertaken in this
area include: development of new computer algorithms for a class of practical
problems in order to explore general issues of efficiency of algorithms;
development of the mathematics required to determine efficiency of algorithms;
and extension of programming methodologies so that such algorithms can be
implemented more quickly and reliably than with present techniques.
6. Relations between Algorithms and Architectures
Stanford University also proposes to carry out research in examining the
relations between algorithms and architectures in order to gain an appreciation
of the processing power required for certain classes of algorithms. Examples
of the kinds of tasks that will be undertaken in this area include;
determination of lower bounds on the optimal ratio of processor power to
communication capability for representative problems; relation of algorithmic
requirements in terms of topology, data rates, and processor capability to
physical and technological restrictions; investigation of inherent tradeoffs in
convergence rates and complexity with concurrency; and investigation of the
possibility of a metalanguage for the description of application problems in
terms of standard computational processes which express both the complexity of
the computation and its topology and which allows accurate simulation of the
process for various architectures.
Ohlander has tinkered further and is now satisfied with what follows.
Betty Scott thinks she can do everything that remains. However, since
I am leaving for Japan and Taiwan for 3 weeks on Saturday, I suggest
you co-ordinate with her. Welcome back, and I hope you had an
interesting trip.
∂18-May-83 0900 JMC*
Carol Lynch
∂18-May-83 1000 JMC*
2810 shares
∂18-May-83 1000 JMC*
John Dupre
∂18-May-83 1132 BSCOTT@SU-SCORE.ARPA Re: SOW
Received: from SU-SCORE by SU-AI with TCP/SMTP; 18 May 83 11:32:27 PDT
Date: Wed 18 May 83 11:35:53-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Re: SOW
To: JMC@SU-AI.ARPA
cc: BSCOTT@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Wed 18 May 83 08:55:00-PDT
I will take care of it, John.
The budget problems Ron mentions, John, have to do with furnishing more
detail on certain budget categories, e.g., travel. I hope I can handle the
questions without bothering anyone.
Betty
-------
∂18-May-83 1259 GENESERETH@SUMEX-AIM Re: Keyworth is out
Received: from SUMEX-AIM by SU-AI with PUP; 18-May-83 12:59 PDT
Date: Wed 18 May 83 13:00:16-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
Subject: Re: Keyworth is out
To: JMC@SAIL
In-Reply-To: Your message of Wed 18 May 83 12:24:00-PDT
John,
Too bad about Keyworth. Frank Press is a good suggestion, though I don't
know if he knows anything at all about ai. I'll think it over and ask
around, and let you know what happens. Thanks for your efforts.
mrg
-------
Another possibility would be to ask if there is another official from
Keyworth's office who would be appropriate. You could call his executive
secretary Carol Lynch 202 456-7116 and ask.
∂18-May-83 1441 NILSSON%SRI-AI.ARPA@SU-SCORE.ARPA Letter
Received: from SU-SCORE by SU-AI with TCP/SMTP; 18 May 83 14:41:10 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Wed 18 May 83 14:32:23-PDT
Date: 18 May 1983 1423-PDT
From: Nilsson at SRI-AI
Subject: Letter
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-MC,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL
Here's a copy of a letter I just sent to Lou Robinson:
Mr. Louis Robinson
American Association for Artificial Intelligence
445 Burgess Drive
Menlo Park, CA 94025
Dear Lou:
I have received your letter of May 2, 1983 in which you state that
you would like to resign your position as Executive Director of AAAI
effective July 1, 1983. On behalf of the Executive Committee of
AAAI, I am hereby accepting your resignation and want to convey our
very best wishes for success in your future activities. The ExCom
joins me in thanking you for your skillful and industrious efforts on
behalf of AAAI during its important formative years.
We are particularly appreciative of your offer to help make the
transition as smooth as possible, and we will be counting on you for
help and advice especially before and during our upcoming conference
this summer. As we have discussed on various occasions, I think there
are many ways in which you can continue to serve AAAI--especially in
the area of publications and publicity.
Don Walker, the AAAI Secretary-Treasurer, will be staying in
touch with you concerning compensation matters during these last few
months. As you, Don and I have already observed, some portion of the
compensation depends on exhibit and tutorial activity at the
conference.
Again, thanks for all that you have done in the past few years. We
look forward to the possibility that your association with AAAI will
continue.
Sincerely,
Nils J. Nilsson
President
cc: Don Walker
-------
∂18-May-83 2242 WIEDERHOLD@SUMEX-AIM Re: DARPA scope document
Received: from SUMEX-AIM by SU-AI with PUP; 18-May-83 22:42 PDT
Date: Wed 18 May 83 22:13:39-PDT
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM>
Subject: Re: DARPA scope document
To: JMC@SAIL
cc: DEK@SAIL, golub@SCORE, GIO@SAIL, bscott@SCORE, ZM@SAIL
In-Reply-To: Your message of Tue 10 May 83 16:55:00-PDT
Thanks for helping us all out in our absence. Should I do something now
that i am back? Gio
-------
∂19-May-83 0006 LLW@S1-A Second Choice
Received: from S1-A by SU-AI with TCP/SMTP; 19 May 83 00:05:52 PDT
Date: 19 May 83 0007 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Second Choice
To: jmc@SU-AI
CC: LLW@S1-A
∂18-May-83 2314 JMC@SU-AI Keyworth alternative
Received: from SU-AI by S1-A with TCP/SMTP; 18 May 83 23:14:35 PDT
Date: 18 May 83 2311 PDT
From: John McCarthy <JMC@SU-AI>
Subject: Keyworth alternative
To: llw@S1-A
Keyworth has reserved August for vacation. Is there someone else in that
office worth asking to speak at AAAI?
[John: Perhaps Doug Pewitt, his Assistant Director for General Science,
who is the OSTP guy in charge of masterminding the Government's response
to the Japanese initiative (among many other matters). Doug is smart,
opinionated, Washington-street-wise, and probably a live-wire speaker
(though I've never heard him give a formal address); he was Deputy
Director of Energy Research in DoE prior to going to OSTP. I'd certainly
recommend him as an alternate to Jay, though he necessarily will speak
with considerably less authority. Lowell]
∂19-May-83 0112 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #3
Received: from SU-SCORE by SU-AI with TCP/SMTP; 19 May 83 01:10:51 PDT
Date: Wednesday, May 18, 1983 10:38PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: 712 Partridge Av, Menlo Park, CA 94025
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #3
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Thursday, 19 May 1983 Volume 1 : Issue 3
Today's Topics:
Implementations - Performance
Applications - Objects
Publications - Recent Technical Reports
Call for papers - 1984 American Logic Programming Conference
----------------------------------------------------------------------
Date: Thu 12 May 83 22:59:59-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: LIPS
The LIPS (logical inferences per sec.) measure for Prolog (and maybe
other logic programming systems) is not as useless as it might appear
at first sight. Of course, resolving a goal against a clause head
takes a different amount of work for different goals and clauses, but
a similar observation could be made about the MIPS measure for
conventional machines. The speed of the concatenate loop
conc([],L,L).
conc([X|L1],L2,[X|L3]) :- conc(L1,L2,L3).
appears to be a remarkably good indicator of the speed of a Prolog
implementation for large "pure" Prolog programs (ie. Horn clauses+cut
but no evaluable predicates except maybe arithmetic). For example,
compiled Prolog on a DEC 2060 runs at 43000 LIPS with this estimate,
and (interpreted) C-Prolog on a VAX 11/780 runs at 1500 LIPS. Prolog
compilers for the VAX and similar machines are starting to be
developed, and at least one is expected to reach 15000 LIPS on a VAX
780 (it will be quite a while before these are incorporated into full
Prolog systems). The first Prolog machine prototype from Japan (the
Psi machine from ICOT) is expected to reach 40000 LIPS.
Extensive use of evaluable predicates may invalidate the measure to a
large extent (but then, we aren't talking about *logic* programs
anymore, and "logical inference" is no longer the main operation).
-- Fernando Pereira
------------------------------
Date: Tue 10 May 83 10:37:51-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Object-oriented programming in LPGs
I've seen papers on two techniques for embedding an object-oriented
style into LPGs: Ehud Shapiro's Concurrent Prolog and Bill Kornfeld's
unification-with-one-way-equality. I think neither of the two papers
have been widely distributed. Ehud's address is
Ehud Shapiro
Dept. of Applied Mathematics
Weizmann Institute of Science
Rehovot 76100, Israel
Bill Kornfeld is on this list, so I will let him elaborate on his
technique.
-- Fernando Pereira
------------------------------
Date: Sat, 14 May 1983 21:38 EDT
From: BAK%MIT-OZ@MIT-MC
Subject: Object-oriented programming in Prolog
Fernando referred to both my work and that of Ehud (as, indeed, we
both refer to it) as "object-oriented programming." Actually the
problems we are each trying to solve are quite different. Shapiro is
concerned in his work with problems of synchronization, multi-
processing, and scheduling as would be needed, for example, by an
operating system. My work is instead concerned with extensible
datatypes and class-structuring of data as would be found in languages
like Smalltalk. (I think the term "object-oriented," as most people
use it, would not properly apply to what Ehud is doing, but that is
not an important point).
The technique, basically, is to extend unification with assertions
about equality. Class structuring is gotten by equating the term
representing the subclass with a term representing the superclass.
For example if an "equilateral←triangle" is a kind of
"regular←polygon" we might represent this by the equality assertion:
equals(equilateral←triangle(L),regular←polygon(3,L)).
where regular←polygon is a binary term that takes a number of sides
and a side length, and equilateral←triangle is monadic and takes just
a side.
What this buys us is that many relations can be defined in terms of
the superclass and then they will be inherited by the subclass. For
example, we might have:
area(regular←polygon(N,L],A) :- ... .
Then if we tried to prove area(equilateral←triangle(15),A), the
equilateral←triangle term would unify with the regular←polygon term
resulting in N being 3 and L 15. Equality assertions seem to give you
all the flexibility of class structuring as found in Smalltalk-like
languages.
There is a "one-way chaining" rule in the procedural semantics of
equals assertions that is critical to its performance. When unifying
the terms WITHIN equals assertions, you can recursively look for new
equals assertions only for the second pair of terms. It is then
possible at compile time to compute the possible chains of functors
through which an equality proof is possible. This information can be
stored in a hash table and used to check whether a proof via equals
assertions is possible. If it is not possible, failure happens in a
(small) constant amount of time. When use of equals assertions is
restricted to class structuring this will in practice give constant
time failure almost all the time when failure would actually happen.
I wrote a paper on this which will appear in IJCAI this summer.
People are welcome to FTP it. The pathname on MIT-OZ is
PS:<BAK>IJCAI.TEX for the source and PS:<BAK>IJCAI.PRS for a press
file. I also made a few extra hardcopies that I could mail.
-- Bill K.
------------------------------
Date: 15 May 1983 20:46:53-PDT (Sunday)
From: Adrian Walker <ADRIAN.IBM-SJ@Rand-Relay>
Subject: Prolog mailing lists
Here is some material for the Prolog mailing lists:
Reports available from IBM T.J. Watson Research Center, Distribution
Services, Post Office Box 218, Yorktown Heights, New York 10598.
Automatic Generation Of Explanations Of Results From
Knowledge Bases. Report RJ 3481. Adrian Walker.
Prolog/Ex1, An Inference Engine Which Explains Both Yes
and No Answers. Report RJ 3771. Adrian Walker.
Report available from Adrian Walker, Department K51, IBM Research
Laboratory, 5600 Cottle Road, San Jose, CA 95193. (Adrian @ IBM-SJ).
Data bases, Expert Systems, and Prolog. Report RJ 3870.
Adrian Walker.
Report available from Department of Computer Science, New York
University, 251 Mercer Street, New York, NY 10012.
Syllog: a knowledge based data management system. Report
No. 034, Department of Computer Science, New York University.
Adrian Walker.
Best wishes,
Adrian
------------------------------
Date: Mon 16 May 83 15:50:49-PDT
From: Ken Laws <Laws@SRI-AI.ARPA>
Subject: Recent Technical Reports
I have extracted the following titles on logic and logic programming
from a list of recent technical reports on file at the Stanford
Math/CS library.
-- Ken Laws
Treleaven, P.C. Gouveia Lima, I.*Japan's Fifth Generation
Computer Systems.* Newcastle upon Tyne U. Computing Lab.*
No. 176.*1982.
Shapiro, E.Y.* Alternation and the Computational Complexity
of Logic Programs.* Yale U. Comp.Sci.Dept.*Res.Rpt. 239.*
1982.
Orlowska, E.* Logic of Vague Concepts: Applications of Rough
Sets.* Polish Academy of Sciences. Inst. of Comp.Sci.*ICS
PAS rpt. no. 474.*1982.
Bergstra, J. Chmielinska, A. Tiuryn, J.*" Hoare's Logic Is
Not Complete When It Could Be".* M.I.T. Lab. for
Comp.Sci.*TM-226.*1982.
Bergstra, J.A. Tucker, J.V.* Hoare's Logic For Programming
Languages With Two Data Types.* Mathematisch Centrum.*IW
207/82.*1982.
Stabler, E.P. Jr.* Database and Theorem Prover Designs For
Question Answering Systems.* Western Ontario U. Cognitive
Science Centre.*COGMEM 12.*1982.
Ballantyne, M. Bledsoe, W.W. Doyle, J. Moore, R.C. Pattis,
R. Rosenschein, S.J.* Automatic deduction (Chapter XII of
Volume III of the Handbook of Artificial Intelligence,
edited by Paul R. Cohen and Edward A. Feigenbaum).*
Stanford U. Comp.Sci.Dept.*STAN-CS-82-937; Stanford U.
Comp.Sci.Dept. Heuristic Programming Project.*HPP-82-019.*
1982. 64p.
Kautz, H.A.*A First-Order Dynamic Logic For Planning.*
Toronto U. Comp. Systems Res. Group.*CSRG-144.*1982.
Boehm, H.-J.*A Logic For Expressions With Side-Effects.*
Cornell U. Comp.Sci.Dept.*Tech.Rpt. 81-478.*1981.
Itai, A. Makowsky, J.*On the Complexity of Herbrand's
Theorem.* Technion - Israel Inst. of Tech.
Comp.Sci.Dept.*Tech.Rpt. 243.*1982.
Makowsky, J.A. Tiomkin, M.L.*An Array Assignment For
Propositional Dynamic Logic.* Technion - Israel Inst. of
Tech. Comp.Sci.Dept.*Tech.Rpt. 234.*1982.
Boyer, R.S. Moore, J.S.*A Mechanical Proof Of the
Unsolvability Of the Halting Problem.* Texas U. Computing
Sci. and Comp.Appl.Inst. Certifiable Minicomputer
Project.*ICSCA-CMP-28.*1982.
Manna, Z. Pneuli, A.*How to Cook a Temporal Proof System
For Your Pet Language.* Stanford U. Comp.Sci.Dept.*
STAN-CS-82-954.*1982. 14p.
Bowen, D.L. (ed.)* DECsystem-10 Prolog User's Manual.*
Edinburgh U. A.I. Dept.*Occasional Paper 027.*1982.
Bundy, A. Welham, B.*Utility Procedures In Prolog.*
Edinburgh U. A.I. Dept.*Occasional Paper 009.*1977.
Byrd, L. (ed.)*User's guide to EMAS Prolog.* Edinburgh U.
A.I. Dept.*Occasional Paper 026.*1981.
Sterling, L. Bundy, A.* Meta Level Inference and Program
Verification.* Edinburgh U. A.I. Dept.*Res. Paper 168.*
1982.
Demopoulos, W.*The Rejection Of Truth Conditional Semantics
by Putnam and Dummett.* Western Ontario U. Cognitive
Science Centre.*COGMEM 06.*1982.
Kozen, D.C.*Results On the Propositional Mu-Calculus.*
Aarhus U. Comp.Sci.Dept.*DAIMI PB-146.*1982.
Mosses, P.* Abstract Semantic Algebras!* Aarhus U.
Comp.Sci.Dept.*DAIMI PB-145.*1982.
Griswold, R.E.*The Control of Searching and Backtracking In
String Pattern Matching.* Arizona U. Comp.Sci.Dept.*TR
82-20.*1982.
Hagiya, M.*A Proof Description Language and Its Reduction
System.* Tokyo U. Info.Sci.Dept.*Tech.Rpt. 82-03.*1982.
Goto, E. Soma, T. Inada, N. Ida, T. Idesawa, M. Hiraki, K.
Suzuki, M. Shimizu, K. Philipov, B.*Design of a Lisp
Machine - FLATS.* Tokyo U. Info.Sci.Dept.*Tech.Rpt.
82-09.*1982.
Sakamura, K. Ishikawa, C.* High Level Machine Design By
Dynamic Tuning.* Tokyo U. Info.Sci.Dept.*Tech.Rpt.
82-07.*1982.
Sato, M.*Algebraic Structure of Symbolic Expressions.*
Tokyo U. Info.Sci.Dept.*Tech.Rpt. 82-05.*
------------------------------
Date: 15 May 1983 16:39:44-EDT (Sunday)
From: Doug DeGroot <DEGROOT.YKTVMV.IBM@Rand-Relay>
Subject: Call for papers - 1984 American Logic Programming Conference
CALL FOR PAPERS
The 1984 International Symposium on
LOGIC PROGRAMMING
Atlantic City, New Jersey February 6-9, 1984
Sponsored by the IEEE Computer Society
and its Technical Committee on Computer Languages
The symposium will consider fundamental principles and important
innovations in the design, definition, and implementation of logic
programming systems and applications. Of special interest are papers
related to parallel processing. Other topics of interest include (but
are not limited to): distributed control schemes, FGCS, novel
implementation techniques, performance issues, expert systems, natural
language processing and systems programming.
Please send ten copies of an 8- to 20-page, double spaced, typed
manuscript, including a 200-250 word abstract and figures to:
Doug DeGroot
Program Chairman
IBM Thomas J. Watson Research Center
P.O. Box 218
Yorktown Heights, New York 10598
Program Committee
-----------------
Jacques Cohen (Brandeis) Fernando Pereira (SRI International) Doug
DeGroot (IBM Yorktown) Alan Robinson (Syracuse University) Don
Dwiggins (Logicon) Joe Urban (Univ. Southwestern Louisiana) Bob Keller
(University Utah) Adrian Walker (IBM San Jose) Jan Komorowski
(Harvard) David Warren (SRI International) Michael McCord (IBM
Yorktown) Jim Weiner (Univ. New Hampshire)
Walter Wilson (IBM DSD Poughkeepsie)
Summaries should explain what is new or interesting about the work and
what has been accomplished. It is important to include specific
findings or results, and specific comparisons with relevant previous
work. The committee will consider appropriateness, clarity,
originality, significance, and overall quality of each manuscript.
Manuscripts whose length exceeds 20 double spaced, typed pages may
receive less careful scrutiny than the work merits.
If submissions warrant, the program committee will compose a four day
program.
September 1, 1983 is the deadline for the submission of manuscripts.
Authors will be notified of acceptance or rejection by October 30,
1983. The accepted papers must be typed on special forms and received
by the program chairman at the above address by December 15, 1983.
Authors of accepted papers will be expected to sign a copyright
release form.
Proceedings will be distributed at the symposium and will be
subsequently available for purchase from IEEE Computer Society.
Conference Chairman Technical Chairman Publicity Chairman Joe Urban
Doug DeGroot David Warren Univ. SWern Lousiana IBM Research SRI
International CS Dept. P. O. Box 218 333 Ravenwood Avenue P. O. Box
44330 Yorktown Hts., NY 10598 Menlo Park, CA 94025 Lafayette, LA 70504
(914)945-3497 (415)859-2128 (318)231-6304
------------------------------
End of PROLOG Digest
********************
∂19-May-83 1002 NOVAK@SUMEX-AIM Why MCC chose Austin
Received: from SUMEX-AIM by SU-AI with PUP; 19-May-83 10:01 PDT
Date: Thu 19 May 83 10:02:55-PDT
From: Gordon Novak <NOVAK@SUMEX-AIM>
Subject: Why MCC chose Austin
To: Feigenbaum@SUMEX-AIM, Buchanan@SUMEX-AIM, Rindfleisch@SUMEX-AIM, JMC@SAIL,
ULLMAN@SCORE
18-May-83 11:13:28-CDT,000001028;000000000001
Date: 18 May 1983 1113-CDT
From: CS.DALE@UTEXAS-20
Subject: UT Support for Computer Sciences
To: CS.Novak
cc: CS.DALE
Gordon:
For your information, the University commitment to computer sciences and
EE, made public today in relation to the MCC location in Austin is as
follows:
1. Establishment of $15 million in faculty endowments
($5 million currently in place).
2. Creation of 30 new faculty positions over the next
3 years.
3. Establishment of graduate fellowships at the level
of $750,000 per year.
4. $1 million per year for research support (equipment
maintenance, support personnel, and other operating
expenses).
5. Additional $5 million over the next two years for
research equipment.
The commitment states that UT is committed to the development of world
class research and education programs in CS and EE and intends to bring
these areas rapidly to the forefront where they will be comparable to MIT,
Stanford, CMU, and Berkeley.
-------
-------
∂19-May-83 1022 DMC AI qual
To: JMC@SU-AI
CC: DFH@SU-AI, DBL@SU-AI, russell@SU-SCORE
Dear Professor Mc Carthy,
I was wondering when the AI qual is rescheduled for.
Thanks.
- Dave Chelberg (DMC @SAIL)
Please try to reschedule the AI qual for late June.
∂19-May-83 1928 @MIT-MC:LEVITT%MIT-OZ%MIT-MC@SU-DSN no learning systems yet??
Received: from MIT-MC by SU-AI with TCP/SMTP; 19 May 83 19:28:00 PDT
Date: Thu, 19 May 1983 22:21 EDT
From: LEVITT%MIT-OZ%MIT-MC@SU-DSN
To: John McCarthy <JMC@SU-AI>
Cc: phil-sci%oz%MIT-ML@SU-DSN
Subject: no learning systems yet??
In-reply-to: Msg of 19 May 83 1718 PDT from John McCarthy <JMC at SU-AI>
From: John McCarthy <JMC at SU-AI>
I don't understand the part about "acknowledges the duality". This doesn't
mean that one can be defined in terms of the other.
I wasn't saying it can, but if I understand what you meant by
"duality", you don't think datatype issues are "orthogonal" to the
discovery process, as DAM claimed.
Maybe our disagreement is that you think much of the discovery process
is *independent* of the representation system we are building the new
concepts out of. I doubt this; but am I getting warmer?
I also don't understand
about not offering related discoveries. I have argued that a child
learns a name and presumes (often correctly) that it names an entity
about which much more is to be learned. ... I do acknowledge
that I don't know that anyone has done it yet.
I don't understand -- doesn't Winston's learning system do exactly
this? Are Kripke and Putnam aware of this work, or do they have
independent, more plausible scenarios?
no learning systems yet?
I have to think about whether Winston's system admits natural kinds,
but at first glance it seems so. Kripke and Putnam aren't interested
in programming. They don't have AI type scenarios, but do have human
type scenarios. Kripke's "Naming and Necessity" is full of scenarios.
Indeed philosophers rely excessively on scenarios.
∂19-May-83 2316 WIEDERHOLD@SUMEX-AIM Re: Stanford Proposal.
Received: from SUMEX-AIM by SU-AI with PUP; 19-May-83 23:16 PDT
Date: Thu 19 May 83 23:17:40-PDT
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM>
Subject: Re: Stanford Proposal.
To: OHLANDER@USC-ISI.ARPA
cc: BSCOTT@SCORE, TOB@SAIL, JMC@SAIL, ZM@SAIL, RPG@SAIL, LUCKHAM@SCORE
In-Reply-To: Your message of Mon 9 May 83 08:40:00-PDT
Thank you very much for the large amount of effort you did to make our proposal
conform. We'll keep it as a model for the future.
I will bereachable at Stanford with little delay during the summer and beyond.
Please let me know of something further should be done. Gio
-------
∂20-May-83 0030 ARK@SU-SCORE.ARPA SCORE Modem Service
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 May 83 00:30:11 PDT
Mail-From: ARK created at 20-May-83 00:13:50
Date: Fri 20 May 83 00:13:50-PDT
From: Arthur Keller <ARK@SU-SCORE.ARPA>
Subject: SCORE Modem Service
To: SU-BBoards@SU-SCORE.ARPA
ReSent-date: Fri 20 May 83 00:25:41-PDT
ReSent-from: Arthur Keller <ARK@SU-SCORE.ARPA>
ReSent-to: Ullman@SU-SCORE.ARPA, REG@SU-AI.ARPA, Bosack@SU-SCORE.ARPA,
JMC@SU-AI.ARPA
This past weekend, CSD Computer Facilities removed the modems connected
to SCORE that had been sold to EE with Sierra and replaced them with
cheaper modems. Users were not warned of this unavailability of
service in advance. Along with this process, they decided to rewire
the modem room.
The cheaper modems, however, are not functionally equivalent to the original
modems. That is, they only support the Bell 300 and Bell 1200 protocols and
not the Vadic 1200 protocol. CSD-CF made no attempt to inform users that
this change was going to take place. While they advocated that users purchase
the Vadic 3451 modems which support all three protocols, users were not
adequately warned against purchase of the considerably cheaper Vadic 3455
modems which support only the Vadic 1200 protocol. Users were not informed
that this protocol would not be supported at some future point. Rather, it is
reasonable to expect that CSD-CF would continue to support this service as they
had instituted it without warning us that it was temporary.
I deplore this lack of consideration on the part of the CSD-CF administration.
I request that CSD-CF reinstate support for the Vadic 1200 protocol for
direct connection to SCORE. I further request that CSD-CF administrative
personnel endeavor to inform users of their intentions when it affects them.
I apologize for the duplication of messages caused by the difficulties of using
SCORE when TELNETed from SAIL (which, by the way, continues to support the
Vadic 1200 protocol).
Arthur
-------
∂20-May-83 0455 OHLANDER@USC-ISI Re: Stanford Proposal.
Received: from USC-ISI by SU-AI with TCP/SMTP; 20 May 83 04:55:32 PDT
Date: 20 May 1983 0452-PDT
Sender: OHLANDER at USC-ISI
Subject: Re: Stanford Proposal.
From: OHLANDER at USC-ISI
To: WIEDERHOLD at SUMEX-AIM
Cc: BSCOTT at SCORE, TOB at SAIL, JMC at SAIL, ZM at SAIL
Cc: RPG at SAIL, LUCKHAM at SCORE
Message-ID: <[USC-ISI]20-May-83 04:52:17.OHLANDER>
In-Reply-To: Your message of Thu 19 May 83 23:17:40-PDT
Gio,
One thing more should be included in the proposal. I need a
list of all equipment currently owned by the government and used
by any of the researchers. A statement should be made requesting
that authorization be granted to continue to use that equipment
under the new contract. Include this information in a separate
section.
Ron
∂20-May-83 0814 DIFFERDING@SUMEX-AIM Reducing number of Altos
Received: from SUMEX-AIM by SU-AI with PUP; 20-May-83 08:14 PDT
Date: Fri 20 May 83 08:15:32-PDT
From: Joan Differding <DIFFERDING@SUMEX-AIM>
Subject: Reducing number of Altos
To: JMC@SAIL
cc: REG@SAIL
In response to the message on the bboard: Sounds like a great idea
to me. I have often had to wait hours at a time to get onto an Alto
in order to use it graphics programs while people are using them to
type papers. It's really just as easy to use EMACS and SCRIBE on
SCORE.
Joan Differding
-------
∂20-May-83 0828 TVR
To: JMC@SU-AI
CC: REG@SU-AI, Other-SU-BBoards@SU-AI
∂20-May-83 0116 JMC reducing number of Altos
To: su-bboards@SU-AI
It has been proposed to reduce number of Altos to those required for
services not available on other machines. What would be the reaction
to reducing them so much that Bravo was infeasible? There would be
more terminals in offices some more maintainable replacements in
the terminal rooms. Please react to JMC (chairman facilities committee)
by June 15 with cc REG@sail.
TVR - At this point, there are not enough public SUNs available to
replace them. One of their common uses is to provide multiple-window
TELNET use which does not have an equivalent besides the SUNs. Also,
we still don't have a BRAVO replacement that has been mentioned many
times in the past. The one attempt for the SUN that i was aware of
never became available before its author graduated.
Even if we can get more SUN's from SMI, it is rumored that they apparently
are no longer willing to deliver 3Mb Ethernet interfaces. We may have to
either build them ourselves, or prematurely add 10Mb Ethernet to MJH. If
we do the latter, we will probably have to do a fair amount to the gateway,
to make it handle the differet protocols, and to speed it up, as most 10Mb
SUN traffic would have to go through it.
∂20-May-83 0934 NOVAK@SU-SCORE.ARPA Altos
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 May 83 09:28:59 PDT
Date: Fri 20 May 83 09:32:14-PDT
From: Gordon Novak <NOVAK@SU-SCORE.ARPA>
Subject: Altos
To: JMC@SU-AI.ARPA, REG@SU-AI.ARPA
I have observed that a lot of the use of the HPP Alto is non-CSD people
using Bravo to write term papers, announcements of religious meetings,
and even commercial advertising flyers. (The latter was a staff person
from EE, whom I chased off.)
-- Gordon
-------
∂20-May-83 0943 WINSLETT@SU-SCORE.ARPA alto reduction
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 May 83 09:42:53 PDT
Date: Fri 20 May 83 09:45:59-PDT
From: Marianne Winslett <WINSLETT@SU-SCORE.ARPA>
Subject: alto reduction
To: jmc@SU-SCORE.ARPA
cc: reg@SU-AI.ARPA
I, for one, would be very unhappy if the altos could not run bravo.
--Marianne Winslett
-------
∂20-May-83 0948 trickey@Diablo Altos
Received: from SU-HNV by SU-AI with PUP; 20-May-83 09:47 PDT
Date: Fri, 20 May 83 09:20 PDT
From: Howard Trickey <trickey@Diablo>
Subject: Altos
To: jmc@sail
Cc: reg@sail
I don't care about Bravo, but I do use Draw and Redraw for all figures
in my papers, so I would like to have access to an Alto for, say,
three days a month (average total time). I hope someone is working
on a replacement for these programs to run on Suns.
-- Howard Trickey
∂20-May-83 1012 RPG travel strategy
To: JMC@SU-AI, llw@S1-A
I am making travel plans to Germany. The question I have is:
do you expect that when accounting time comes, the actual air fares are
looked at or will it be `regular coach fare?' The reason is that if
regular coach is assumed, I can try to get a low fare, trading lack of
comfort and flakey schedules for money, spending what's left over on
accomodations. If the actual tickets are inspected, then I'll just
book something standard (like TWA).
-rpg-
∂20-May-83 1142 NOVAK@SUMEX-AIM [NAME AAAI-OFFICE <AAAI-OFFICE@SUMEX-AIM.ARPA>: Potential Exhibitors for AAAI-83]
Received: from SUMEX-AIM by SU-AI with PUP; 20-May-83 11:42 PDT
Date: Fri 20 May 83 11:43:23-PDT
From: Gordon Novak <NOVAK@SUMEX-AIM>
Subject: [NAME AAAI-OFFICE <AAAI-OFFICE@SUMEX-AIM.ARPA>: Potential Exhibitors for AAAI-83]
To: JMC@SAIL
I expect that you know the Lisp software companies better than I.
-- Gordon
---------------
Mail-From: AAAI-OFFICE created at 19-May-83 11:28:00
Date: Thu 19 May 83 11:27:59-PDT
From: NAME AAAI-OFFICE <AAAI-OFFICE@SUMEX-AIM.ARPA>
Subject: Potential Exhibitors for AAAI-83
To: Novak@SUMEX-AIM.ARPA
cc: aaai-office@SUMEX-AIM.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Dear Gordon,
It was a pleasure meeting you at the Program Committee meeting. I was
wondering if you help me; I'm looking to identify the names and addresses
of software manufacturers that support the different LISP machines. I
remeber hearing the names of several start-up companies that provide
software support to LMI, Symbolics, DECetc. I'm looking to invite these
companies to have exhibits at our national conference. Also, govt and
universities can have demos of their different AI systems at the conference.
You might think about the possibility of having a demo of GLISP at the
conference. What do you think?
I hope you can help me with this request. Thanks for any assistance you
may provide.
Claudia Mazzetti
-------
-------
Gordon, I don't have these names, and I'm leaving tomorrow for 3 weeks.
Dick Gabriel might be able to help.
∂20-May-83 1147 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN representing sequences by sets
Received: from MIT-MC by SU-AI with TCP/SMTP; 20 May 83 11:47:23 PDT
Date: Fri, 20 May 1983 14:43 EDT
From: DAM%MIT-OZ%MIT-MC@SU-DSN
To: JMC@SU-AI
Subject: representing sequences by sets
Date: 19 May 83 1337 PDT
From: John McCarthy <JMC at SU-AI>
I am puzzled by its first sentence.
Every text on set theory mentions the Wiener-Kuratowski (early 1920s)
representation of ordered pairs by sets and then goes on to show how
this allows the representation of sequences.
I have changed the first sentence of the abstract. What I really
meant to say was that it is impossible to represent a finite list
OF ATOMS (I call them points) by using a finite set OF ATOMS (points).
Of course I am familiar with the standard construction of sequences
from sets (It is used in my paper).
David Mc
∂20-May-83 1217 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN role of logic in AI
Received: from MIT-MC by SU-AI with TCP/SMTP; 20 May 83 12:16:53 PDT
Date: Fri, 20 May 1983 14:52 EDT
From: DAM%MIT-OZ%MIT-MC@SU-DSN
To: JMC@SU-AI
cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: role of logic in AI
Date: 18 May 83 0031 PDT
From: John McCarthy <JMC at SU-AI>
If DAM disagrees with these views, I would like to read specific
criticisms.
I do disagree with many of the views expressed in your message
on first order logic (I also don't like the common practice of using
the terms "logic" and "first order logic" synonymously). However
I do not have time to provide specific critisisms now. I will give
a detailed response on Monday.
David Mc
∂20-May-83 1332 YOUM@SU-SIERRA Removal of ALTO
Received: from SU-SIERRA by SU-AI with PUP; 20-May-83 13:31 PDT
Date: Fri 20 May 83 13:31:47-PDT
From: Youm Huh <YOUM@SU-SIERRA>
Subject: Removal of ALTO
To: jmc@SU-AI
cc: su-bboards@SU-SIERRA
One of the most important features of ALTO that we appreciate is its
drawing capability such as draw, sil, etc.
-------
∂20-May-83 1341 SHARON@SU-SCORE.ARPA tickets
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 May 83 13:41:51 PDT
Date: Fri 20 May 83 13:45:23-PDT
From: Sharon Bergman <SHARON@SU-SCORE.ARPA>
Subject: tickets
To: jmc@SU-SCORE.ARPA
Dr. McCarthy: Hi. Dina Bolla Travel Agency just delivered some
tickets. I'll leave them in your mail slot.
Sharon
-------
I have the tickets, but in the future if you can catch them before they
run off have them take them up one more flight to me or to Diana Hall.
∂20-May-83 1447 BSCOTT@SU-SCORE.ARPA Layoff
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 May 83 14:47:14 PDT
Date: Fri 20 May 83 14:50:28-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Layoff
To: JMC@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
John,
Here is some information for your layoff letter; some of it is bureaucratic,
but necessary so the University personnel people say. Also, I was wrong
about the amount of notice--had forgotten that a three-month notice is re-
quired. It is o.k. to give him six months if you wish.
The letter should include some version of the following:
This letter is to confirm our discussion (or to notify if you haven't
talked with him) regarding the elimination of your position, effective
←←←←←←←←←←←←←←. You should consider this a permanent layoff in accordance
with University guidelines, and is necessary because←←←←←←←←←←←←←←←←←←←←
(programmatic or funding or ?? problems).
Should you desire to continue employment at Stanford University you should
contact Tom Castillo in Pel; his extension is 7-3314. The Benefits section of Personnel can give you information regarding continuation of your benefits;
this extension is 7-1777.
In the event you have not secured another position with Stanford by
←←←←←←←←←←←←←(date of layoff), you will be entitled to severance pay less
any deductions required by law. The amount of severance pay has not been
calculated; we will let you know about this sometime next week. You
should also know that if you are reemployed by Stanford after the layoff
date, and you are in receipt of severance pay, you may be required to pay back
all or a portion of it.
We regret the necessity for this layoff and wish to thank you for your
years of service to Computer Science. Please let me know if I can be of
assistance to you in securing another position.
------------------
If you wish John, leave out all the bureaucracy concerning University rules
and regulations, and tell him I will write to him next week about benefits,
etc.
Betty
-------
∂20-May-83 1448 BSCOTT@SU-SCORE.ARPA layoff letter
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 May 83 14:47:58 PDT
Date: Fri 20 May 83 14:51:25-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: layoff letter
To: JMC@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
Tom Castillo is in Personnel--don't know what happened to that word in the
previous message. --Betty
-------
∂20-May-83 1600 SHARON@SU-SCORE.ARPA visitor
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 May 83 15:59:53 PDT
Date: Fri 20 May 83 16:03:28-PDT
From: Sharon Bergman <SHARON@SU-SCORE.ARPA>
Subject: visitor
To: jmc@SU-SCORE.ARPA
Someone from Dr. Suppes office just called and said that there is
a Brazilian visitor seeing him now over in Ventura Hall and that
you might want to talk to him.
Sharon
-------
∂20-May-83 2206 shore@Shasta Altos
Received: from SU-SHASTA by SU-AI with PUP; 20-May-83 22:06 PDT
Date: Friday, 20 May 1983 22:10-PDT
To: JMC at SAIL
Cc: REG at SAIL
Subject: Altos
From: Andrew Shore <shore@Shasta>
Indeed there are many services not available elsewhere (yet or in abundance).
Multi-window telnet (talk), EtherWatch, PupWatch, Markup, Draw,
PrePress, PressEdit, Icarus, Bravo, Fred, Sil, ReadPress, ....
These may not be used often, but are indespensible when needed. The
Altos must be kept around and kept servicable until a viable
replacement exists for each and every one of these functions (and
probably more I haven't thought of. If it is necessary to reduce the
size of the Alto "pool" in order to guarentee a certain critical (> 5)
mass of them ALWAYS to be working, than so be it. Restricting Bravo
use seems reasonable if it means that there will be Altos guarenteed
available for the other functions. What we really need is more SUNs
with the above software (or something BETTER) available on them.
∂21-May-83 0107 WIEDERHOLD@SUMEX-AIM [OHLANDER at USC-ISI: Re: Stanford Proposal.]
Received: from SUMEX-AIM by SU-AI with PUP; 21-May-83 01:07 PDT
Date: Sat 21 May 83 01:09:02-PDT
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM>
Subject: [OHLANDER at USC-ISI: Re: Stanford Proposal.]
To: bscott@SCORE
cc: jmc@SAIL
do you have this information?
I can provide it for the terminals i have purchased, but dont know what it implies
in terms of makor equipment. thanks gio
---------------
Return-Path: OHLANDER@USC-ISI
Received: from USC-ISI by SUMEX-AIM.ARPA with TCP; Fri 20 May 83 04:56:01-PDT
Date: 20 May 1983 0452-PDT
Sender: OHLANDER at USC-ISI
Subject: Re: Stanford Proposal.
From: OHLANDER at USC-ISI
To: WIEDERHOLD at SUMEX-AIM
Cc: BSCOTT at SCORE, TOB at SAIL, JMC at SAIL, ZM at SAIL
Cc: RPG at SAIL, LUCKHAM at SCORE
Message-ID: <[USC-ISI]20-May-83 04:52:17.OHLANDER>
In-Reply-To: Your message of Thu 19 May 83 23:17:40-PDT
Gio,
One thing more should be included in the proposal. I need a
list of all equipment currently owned by the government and used
by any of the researchers. A statement should be made requesting
that authorization be granted to continue to use that equipment
under the new contract. Include this information in a separate
section.
Ron
-------
∂21-May-83 1223 DIETTERICH@SUMEX-AIM Re: reducing number of Altos
Received: from SUMEX-AIM by SU-AI with PUP; 21-May-83 12:22 PDT
Date: Sat 21 May 83 12:24:14-PDT
From: Tom Dietterich <DIETTERICH@SUMEX-AIM>
Subject: Re: reducing number of Altos
To: jmc@SAIL, reg@SAIL
cc: DIETTERICH@SUMEX-AIM
In-Reply-To: Your message of Fri 20 May 83 01:16:00-PDT
JMC,
I'm confused by your bboard message requesting comments about the
possibility of reducing the number of Altos to those required for
services not available on other machines. In particular, I'm confused
by your question "What would be the reaction to reducing them so much
that Bravo was infeasible?" In what way would Bravo be infeasible?
Would the Altos be modified in some way (e.g., remove one disk or some
memory) so that Bravo didn't run any more? Or would it just be that
with only, say, 2 Altos working, priority would be given to non-Bravo
tasks?
Given that there are only (to my knowledge) four working Altos
in MJH (one in 450, one in 460, one in 408, and one in the HPP
terminal room), this "proposal" has already been virtually carried
out. I guess the proposal now amounts to removing the non-functioning
Altos from the terminal rooms and offices and replacing them with
SUNs? This sounds like a good long-term plan, but in the short term,
the non-functioning Altos do serve as good network terminals, and they
are in nearly-constant use. As has been pointed out in previous
messages, they also have the advantage of being able to talk to the
3Mb ethernet.
I have no objection to reducing the number of Altos to the point where
bravo is not useful as a practical text editor. Indeed, given that
there are so few fully-functional Altos, I think we have reached the
point where Bravo use is anti-social (at least for large documents).
This past week, I've been putting together camera-ready copy using
Scribe manuscript files with figures inserted using Draw. I need to
use the Alto programs Draw, ReDraw, Pressedit, and Ftp to accomplish
this. Several times I had to wait for Bravo users to finish. They
tend to tie up the Altos for hours at a time. However, I don't think
we should ban the use of bravo entirely. You often need to use bravo
to change the file "user.cm" to set up fonts for Markup and Draw.
Also, some people have many old bravo files that they might want to
modify slightly and then print. It would be nice to keep bravo around
for such purposes. Perhaps the solution is to impose a time limit on
Alto use: Each Alto session should last no more than 30 minutes if
someone else is waiting. Perhaps we could have sign-up sheets next
the Altos too.
Many of us have looked forward to the time when we would have more
Lisp machines than Altos. Ironically, that time has arrived already,
and at very little cost--just neglect.
--Tom
-------
∂23-May-83 0009 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #4
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 May 83 00:09:25 PDT
Date: Sunday, May 22, 1983 9:10PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584, Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #4
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Monday, 23 May 1983 Volume 1 : Issue 4
Today's Topics:
Publications - Logic Programming Bibliography,
Applications - Objects
----------------------------------------------------------------------
Date: Sun 22 May 83 16:30:07-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Prolog and Logic Programming Bibliography
From my master bibliography file, I have extracted a bibliography on
Prolog and logic programming. While I haven't made any attempt at
completeness or uniformity of coverage (I wouldn't have the time or
patience), I have tried to annotate each entry with a few keywords. If
you have bibliography entries that I have not included, I would be
glad to receive them (particularly if in Scribe format!).
Because the bibliography is over 35000 characters long, it is not
being directly included here, but instead it will be stored at an
FTP-accesible place to be announced.
-- Fernando Pereira
[ The bibliography is kept at [SCORE]PS:<PROLOG>Prolog-Bib.Doc and can
be FTP'd. SU-SCORE supports the anonymous login convention.
For those who receive the digest but cannot reply or use the FTP
I have a limited number of hard copies that could be mailed.
--Chuck ]
------------------------------
Date: Thu 19 May 83 01:43:12-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Object-oriented styles
I agree that Concurrent Prolog doesn't cover the class structure
aspect of object-oriented programming. On the other hand, Concurrent
Prolog incorporates message-passing and state change (or object
regeneration) aspects which are often considered part of the
object-oriented programming paradigm. In particular, message-based
window systems can be described very nicely in Concurrent Prolog.
Curiously, from an operational point of view both approaches (Bill's
and Ehud's) provide means for delaying the execution of goals. I feel
there is some connection here waiting to be uncovered ...
-- Fernando
------------------------------
Date: Friday, 20 May 1983, 14:56-EDT
From: Henry Lieberman <Henry%MIT-OZ@MIT-MC>
Subject: Object oriented programming in Prolog
A problem with object-oriented programming in Prolog is the
representation of default information. Let's say we want to say that
"birds can fly". In an object-oriented language, we have an object
for BIRD which responds to a message TRANSPORT-MODE, returning FLY,
whereas a TURTLE object might respond CRAWL or a FISH would say SWIM.
To have a bird which is an exception such as an OSTRICH, we create a
new object that inherits from BIRD. All messages not explicitly
handled by OSTRICH will be handled by BIRD. We define a message
handler for OSTRICH which "intercepts" the TRANSPORT-MODE message and
replies WALK instead of FLY. This is called "shadowing" of messages
and is a powerful and fundamental technique.
How would this work in Prolog? Let's try
Bird (X) := Transport-Mode (X, Fly) ;; Birds can fly Bird (Ostrich) ;;
An ostrich is a bird Transport-Mode (Ostrich, Walk) ;; Ostriches
transport by walking
Oops, now ostriches can both walk and fly! How about
Bird (X) := Transport-Mode (X, Fly) Or Ostrich (X)
But this is clumsy since our "general" rule had to mention specific
exceptions which doesn't usually happen in object oriented languages.
------------------------------
End of PROLOG Digest
********************
∂23-May-83 0815 BSCOTT@SU-SCORE.ARPA Re: phone conversation
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 May 83 08:14:58 PDT
Date: Mon 23 May 83 08:18:23-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Re: phone conversation
To: JMC@SU-AI.ARPA
cc: BSCOTT@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Fri 20 May 83 16:52:00-PDT
Thanks for your message. I will not discuss it with anyone.
Betty
-------
∂23-May-83 0846 REG
To: SU-BBoards@SU-SCORE, ARK@SU-SCORE
CC: Ullman@SU-SCORE, JMC@SU-AI, Bosack@SU-SCORE
It seems to me that the burden of Arthur Keller's complaint [Bulletin Boards
May 20] is four fold.
First, Arthur says that we gave no warning. On May 12 notice was sent to
SU-BBOARDS notifying users of the planned downtime for conversion & wiring
changes.
Second, that ``users were not informed that this [i.e., the VA3400]
protocol would not be supported at some future point... It is reasonable
to expect that CSD-CF would continue to support this service as they had
instituted it without warning us that it was temporary.'' Surely
Arthur does not lament our lack of support for the B5500, the 7090,
the PDP-1, the Librascope, the robot cart, the former Score computer, etc.
Everything in this business is temporary.
Third, that we should give ``due consideration'' to those who willfully
ignored our recommendation and bought the wrong modem. We have
given due consideration to all users of the facility, including those who
wish to pay less for better service: the new modems are less expensive
and more reliable than the ones they replace. In essence Arthur wants
CF to invest thousands of dollars to buy unreliable modems so that
his two modems can be supported.
Fourth, that the rewiring of the modem room should not have been attempted
at this time. It is true that the rewiring was more traumatic that we
had intended; I apologize. However, it was apparent in December that we
would no be able to support the various new demands for telephone access
(by Whitney and other systems, e.g., expansion of Score) without such a
rewiring.
CF will not reinstate support of the Vadic 3400 protocol for Score.
Among the reasons are the expense, the space required, reliability
problems, and the effort needed. As has been noted, SAIL supports
these modems at the present time (though at some future time these
may be removed from SAIL).
Ralph Gorin
∂23-May-83 1035 DIETTERICH@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 23-May-83 10:35 PDT
Date: Mon 23 May 83 10:36:50-PDT
From: Tom Dietterich <DIETTERICH@SUMEX-AIM>
To: REG@SAIL, ARK@SCORE
cc: Ullman@SCORE, JMC@SAIL, Bosack@SCORE, DIETTERICH@SUMEX-AIM
In-Reply-To: Your message of Mon 23 May 83 08:46:00-PDT
Arthur is not alone in objecting to the lack of Vadic support for
SCORE. I also am no longer able to use SCORE from home. Although
Ralph has presented good financial reasons for dropping Vadic support,
I wonder if lost revenues to CSD-CF will offset any savings obtained
from the change.
Arthur was also objecting to the general way in which the changeover
was accomplished. The CSD-CF administration continues to give the
impression of making important decisions without consulting its users
or inviting comments. Contrary to Ralph's claims in his reply to
Arthur, the May 12 message in no way announced that the Vadic support
was going to be dropped. And the message gave less than 72 hours
notice about the change. That message wasn't any kind of notification
or request for comments--it was a decree. And it is this dictatorial
style of operation that has embittered many CSD-CF users, including
myself, toward the present administration.
The appropriate way to have proceeded with the Vadic changeover would
have been to describe the alternatives available along with their
consequences in a message to the CSD-CF community. This could have
been followed by a few weeks commentary period after which the
decision could have been made and announced. The changeover could
then be announced to occur about a month later. Such a schedule would
have allowed people to make other arrangements (e.g., purchasing new
modems), so that service would not be interrupted.
The present managerial style of CSD-CF is not in the long term
financial (or educational, or research) interests of the CSD-CF or the
Computer Science Department.
--Tom Dietterich
-------
∂23-May-83 1038 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN role of logic in AI
Received: from MIT-MC by SU-AI with TCP/SMTP; 23 May 83 10:38:00 PDT
Date: Mon, 23 May 1983 13:08 EDT
From: DAM%MIT-OZ%MIT-MC@SU-DSN
To: JMC@SU-AI
cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: role of logic in AI
The arguments against first order logic are quite subtle,
but I think they are extremely important and damning. Because of
the subtlety and technical nature of the arguments I will not
sart by presenting the arguments directly but will instead state
point by point where I disagree with JMC's message about logic.
In a later message I will present details arguments for my position.
Date: 18 May 83 0031 PDT
From: John McCarthy <JMC at SU-AI>
1. First order logic is logically universal in that any other logic
can be formalized withing first order logic by introducing appropriate
entities as objects. Specifically, second order logic is so
formalizable, and set theory formalized in first order logic is more
powerful than any high order logic. "More powerful" must be taken in
a technical sense that involves similar restrictions on what can be
objects in order for this statement to make sense.
I'm not sure what "more powerful" means in the above claim but
under the standard definitions I have seen first oder logic is far
weaker than, say, second order logic. In a later message I will give
a precise definition of what it means for a statement to be
expressible in a logic and show how there are many common notions
(such as the natural numbers and the notion of a transitive closure)
which are not expressible in first order logic. Set theory is a very
special case that deserves close examination. There are two ways to
view set theory. First set theory can be considered to be a set of
first order axioms. Under this view set theory CAN NOT express the
simple notions of the natural numbers or transitive closures. On the
other hand set theory can be considered to be a special logic (not
first order) where the sentences of set theory are interpreted over AN
INTENDED MODEL which is "the" universe of all sets. Under this
interpretation set theory is just about the most powerful logic going.
3. In order to justify using a more elaborate logic than first order
logic, whether it be higher order or modal, there has to be a good
reason. Many of the candidate modal logics fail, because they don't
admit enough meta-reasoning within the logic for practical purposes,
so that certain first order formalisms are actually more powerful. My
paper on First Order Theories of Individual Concepts and Propositions
has examples of this.
There seems to be an implicit assumption in the above that
powerful logics must be "elaborate" and therefore the use of a more
powerful logic must be "justified". However there are powerful logics
which are simpler than first order logic. For example consider
statements of the form x=y+z where x, y, and z are taken to be
variables ranging over real numbers. This statement is very simple
but can not be expressed in first order logic because the real numbers
can not be specified (axiomatized) in first order logic.
4. Logic can be used in AI in four ways.
I am not sure sure whether a distinction is being made here between
logic in general, and first order logic in particular.
a. The systems most committed to logic do their computation by
reasoning in logical systems. I don't know any way to do this very
efficiently, but I also know no proof that it is impossible. (I see
it as very important to formulate precise assertions about the
computational power or lack of it in various formalisms).
I think that there is an important distinction beteween
inference and other cognitive computational mechanisms (perception,
motor control, goal and subgoal formation, scientific hypothesis
generation, classical induction generating assumptions of the form
"all swans are white", and undoubtedly many more). Thus I see NO
MOTIVATION for trying to make all computation look like inference.
The distinction between inference and other computational mechanisms
is orthogonal to the debate about the power of first order logic as
one particular paradigm for inference.
d. As Newell emphasized in his 1980 AAAI presidential
address, and as I emphasize in my Ascribing Mental Qualities to
Machines, even if a system uses neither sentences for representing
information or deduction as computation, it may often be
conveniently described at a "logical level". Daniel Dennett makes
similar points in his Herbert Spencer lectures when he talks
about "the intentional stance".
Consider a program which takes two binary bit strings which
represent numbers and produces a third binary bit string representing
the sum of those two numbers. This program can be specified in terms
of the numbers represented by the bit strings. This specification can
be written in some formal language (though not first order logic since
the numbers aren't specifiable (axiomatizable) in first order logic.
However I see no reason to think of addition as inference (I could
think of an addition procedure as a Turing machine too, but whats the
point?). There are OTHER computations which ARE best thought of as
inferentcial processes.
I choose to emphasize the epistemological part of the problem
in my own work, because I do that much better than writing programs.
I think this is a fine position to take. I also emphasise
epistemological problems, though I also enjoy writing inference
procedures. However I think it is exactly first order logic's
epistemological weakness that is its biggest problem.
∂23-May-83 1712 CAB Altos
To: JMC
CC: REG, DEK, CAB
Regarding the phasing out, by plan or attrition, of the Altos--
I am organizing an international conference on computer-assisted type design,
at Stanford, August 1 - 7, this summer, 1983. We plan to use the Altos to
demonstrate the FRED, DRAW, and PrePress programs. Though relatively antiquated,
these are significant programs.
Thus, we need the Altos in August.(Though not after.)
But of course we need WORKING Altos.
What maintenance assistance is available from Xerox? Any? Some? None?
If currently none, shall I ask another division of Xerox to contribute some?
--Chuck Bigelow
∂23-May-83 1724 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract
Received: from SUMEX-AIM by SU-AI with PUP; 23-May-83 17:24 PDT
Date: Mon 23 May 83 17:25:43-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Re: Equipment Contract
To: Feigenbaum@SUMEX-AIM, JMC@SAIL, CSL.JLH@SCORE, Ullman@Diablo, RPG@SAIL,
REG@SAIL
cc: CSD.BScott@SCORE, Rindfleisch@SUMEX-AIM
In-Reply-To: Your message of Tue 3 May 83 16:10:36-PDT
Hurray!!! We heard from DSSW today that they are ready to issue our
equipment contract effective June 1. The first increment of funding
will be $1.6M through December. The remaining year 1 funding will come
from FY 84 money after October. We need to decide how to divvy up this
early money.
The HPP would like to move out asap on the Lisp workstations. Do others
of you have similar urgent needs or, on the contrary, preferences to
delay some purchases until the second increment comes? We will need
to plan this in detail very soon.
Tom R.
-------
∂23-May-83 1729 @MIT-MC:KDF%MIT-OZ%MIT-MC@SU-DSN role of logic in AI
Received: from MIT-MC by SU-AI with TCP/SMTP; 23 May 83 17:29:03 PDT
Date: Mon, 23 May 1983 20:16 EDT
From: KDF%MIT-OZ%MIT-MC@SU-DSN
To: DAM%MIT-OZ%MIT-MC@SU-DSN
Cc: JMC@SU-AI, phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: role of logic in AI
In-reply-to: Msg of 23 May 1983 13:08-EDT from DAM
I can't resist any longer.
1. Logic as a descriptive tool: Because there are many
different computational interpretations of logic, expressing a theory
in terms of something like FOPC is better for communicating ideas
because it presupposes less about the way the information is to be
used than handing someone a program does. In theory, grungy issues
like the descriptions of datastructures are no longer an impediment to
understanding what is going on.
I submit that, in fact, the case is far from settled. While
many people profess interest in Hayes' axioms for liquids or
McDermott's axioms for time, I have met few people who have actually
understood them line for line. By contrast, most people who read
Allen's papers on time quickly understand what is going on. In papers
on QP theory, I've found almost no one reads the axioms (including
people who have asked for them in the first place!) and the really
subtle questions people ask are from considering the sugared notation,
for which a precise but informal (thank you for the distinction, DAM)
semantics is provided. In fairness, there are times when one really
needs the mental hyigene of writing down axioms. However, note that
in the physical sciences axioms always came AFTER a long phase of
ontological development and discovery of laws. And presenting ideas
solely in terms of axioms can lead to unnecessary turgidity.
2. What kind of logic? I take substituting reification for
higher-order or modal logics with a grain of salt, for two reasons.
First, I spent alot of time looking over Moore's axioms for knowledge
and action (for those who aren't familar with it, he presents some
modal axioms for knowing, believing, etc., but translates them to a
first-order model to draw conclusions. There is an SRI technical
report about it). Most steps in the proof consisted of the
translation to and from the model language. If we wrote an inference
engine on modal logics the proof could become much shorter (hence
findable, given most domain-independent heuristics for controlling
theorem provers). Second, I spent several days once trying to write a
full, formal, axiomization of QP theory. It got ugly - before long I
was quantifying over sets of statements with every turn, gettting
axioms that were half a page long that could be said more clearly in
an English sentence. Such encodings can be like like writing
Quicksort in set theory - it can be done (indeed, there are people in
the Logic Programming community doing so), but it will take thousands
of steps to do what a programmer using a more appropriate language
will do in a page.
3. Logic as computational tool: All of the "pure" logic
systems I know of are theorem provers, and there are some reasonable
characterizations of their computational properties - they are
exponential, and for most sets of axioms will not terminate if the
assertion to be generated is unprovable. There appear to be no
good domain-independent ways to control instantiation and search
to mitigate the problem.
Aside from the "exponential disaster", there are some classes
of deductions which are so common that it is worth having specialized
inferential mechanisms for them. Spatial reasoning is one example
(see the Metric Diagram/Place Vocabulary stuff in my MS), temporal
reasoning is surely another (Allen's theory). The price one pays for
specialized representations is that inferences not in the class it is
designed for can be difficult or impossible. I don't find this
disturbing, though. I see no evidence that people know all
consequence of their beliefs, nor that it would be a very useful thing
for the minds we build to do so.
If carefully limited, logic can be a very useful
implementation tool. DAM has illustrated how a class of propositional
deductions can be viewed as constraint propagation, and while
incomplete, the computattion does terminate (and usually rather
quickly, I might add). Most of the work that happens in my code, both
rules and programs, is performed by asserting some implication or
other logical statement and letting the TMS figure out whether or not
it is true (a typical Process Instance expands into 25 or so clauses,
for example).
4. Epistemological vs. Heuristic distinction: While I have no
objection to people using pencil and paper rather than more expensive
aids, I'm not sure I really buy this distinction. "Heuristic",
certainly in the early days of AI, seemed to mean "whatever worked".
I think Marr's three levels of explanation are a better cut; we would
have epistemological, computational, and heuristic theories. Unlike
most AI heuristics, computational theories state explicitly what facts
of the world the rely upon to work. Note that Marr's framework is
organized around explaining some particular task; the issue of what
languages to express these theories in is somewhat orthogonal. I
think working solely on epistemological problems, as JMC does, can
certainly be reasonable. For myself, I just don't trust my intuitions
enough to do without some mechanical assistence (as Bacon suggested,
nature is likely to be more subtle than our arguments).
∂23-May-83 2329 RMS%MIT-MC.ARPA@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 May 83 23:27:40 PDT
Received: from MIT-MC.ARPA by SU-SCORE.ARPA with TCP; Mon 23 May 83 23:31:12-PDT
Date: 24 May 1983 02:29 EDT
From: Richard M. Stallman <RMS @ MIT-MC>
To: mccarthy @ SU-SCORE
You offered to pay for my car rental.
Should I mail you a copy of the receipt?
It's too bad you didn't manage to dig out my modal logic stuff,
but if you still want to, I will try to get the files restored.
∂24-May-83 1049 KARP@SUMEX-AIM Facilities committee
Received: from SUMEX-AIM by SU-AI with PUP; 24-May-83 10:49 PDT
Date: Tue 24 May 83 10:50:29-PDT
From: Peter Karp <KARP@SUMEX-AIM>
Subject: Facilities committee
To: jmc@SAIL
I believe I am now the student member of the facilities committee. Thus
please add me to the appropriate mailing list. Thanks,
Peter
-------
∂24-May-83 1057 BSCOTT@SU-SCORE.ARPA ARPA Contract MDA903-80-C-0102
Received: from SU-SCORE by SU-AI with TCP/SMTP; 24 May 83 10:57:23 PDT
Date: Tue 24 May 83 10:50:20-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: ARPA Contract MDA903-80-C-0102
To: JMC@SU-AI.ARPA, TOB@SU-AI.ARPA, Lantz@SU-HNV.ARPA
cc: DFH@SU-AI.ARPA, MAS@SU-AI.ARPA, Kovach@SU-SIERRA.ARPA, BScott@SU-SCORE.ARPA
This contract period was 10-1-79 through 3-31-83, and a final report on
the research is due. I would appreciate your having your secretaries
coordinate the report effort for you, and then give me the reports to
send to ARPA.
Betty
-------
∂24-May-83 1658 PACK@SU-SCORE.ARPA Philosophy-of-science
Received: from SU-SCORE by SU-AI with TCP/SMTP; 24 May 83 16:58:22 PDT
Date: Tue 24 May 83 17:01:55-PDT
From: Leslie E. Pack <PACK@SU-SCORE.ARPA>
Subject: Philosophy-of-science
To: jmc@SU-AI.ARPA
I just had my name put on the philosophy of science mailing list. The
first message I got was an argument against something you submitted on
the use of first order logic in AI. I am interested in reading it--is
it accessible on line somewhere?
I am also interested in reading more of the literature regarding
epistemological considerations in AI. Could you perhaps give me some
pointers to things to read?
Thanks,
Leslie Pack
I don't know what to recommend beyond the references I cited in
my course. You might look at Newell's 1980 presidential address
and papers by Bob Moore. If you are interested in specific topics
come see me. Also if you want to do some (for the present unpaid)
work in the areas.
-------
Plan:
Howdy.
If I'm not here, I'm probably at
HOME : 213 Laurel Ave.
Menlo Park 94025
321-9963
LOTS-X : R.Rasty
STAR VAX : Pack (imaginative, ain't it?)
"Recently I was with a group of mathematicians and philosophers. One
philosopher asked me whether I believed man was a machine. I replied,
'Do you really think it makes any difference?' He most earnestly
replied, 'Of course! to me it is the most important question in
philosophy.'
I had the following afterthoughts: I imagine that if my friend finally
came to the conclusion that he were a machine, he would be infinitely
crestfallen. I think he would think: 'My God! How horrible! I am
only a machine!' But if I should find out I were a machine, my
attitude would be totally different. I would say: 'How amazing! I
never before realized that machines could be so marvelous!'"
- Raymond Smullyan
This Book Needs No Title
∂25-May-83 1452 MDP@SU-SCORE.ARPA PROLOG Digest V1 #5
Received: from SU-SCORE by SU-AI with TCP/SMTP; 25 May 83 14:51:24 PDT
Date: Tuesday, May 24, 1983 11:34PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #5
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Wednesday, 25 May 1983 Volume 1 : Issue 5
Today's Topics:
Applications - Objects
----------------------------------------------------------------------
Date: Monday, 23 May 1983 11:57-PDT
From: narain at rand-unix
Subject: Object Oriented Programming in Prolog
If you implement inheritance properly in Prolog, then representation
of default information is as simple as in other object-oriented
languages. This has of course been done by Kenneth Kahn in his paper
"Actors in Prolog".
His mechanism for searching for the proper response to a message can
be improved considerabbly and I have reimplemented his ideas to make
the resulting system an order of magnitude more efficient, both in
time and in space.
Another quick and dirty alternative to making "clumsy" exceptions to
general rules is:
transport←mode(X,walk):-ostrich(X),!. transport←mode(X,fly):-bird(X).
The "cut" will make sure that the second rule is invoked IF and only
IF ostrich(X) fails in the first rule. So the transport mode of any
bird is "fly", unless X is an ostrich in which case it is "walk",
which is what is desired.
-- Sanjai Narain
------------------------------
Date: Monday, 23 May 1983, 23:07-EDT
From: Henry Lieberman <Henry%MIT-OZ@MIT-MC>
Subject: Reply to Narain
Sanjai,
I concede the "quick and dirty" use of cut works after a fashion, but
aside from matters of taste, it criticially requires that all the
"exceptions" appear BEFORE the general rule. This makes it more
difficult to incrementally add new knowledge to already existing
systems, a hallmark of object-oriented programming.
Ken Kahn's solution in Intermission is much better, but it amounts to
an embedding of an object-oriented language in Prolog rather than a
solution directly in a natural Prolog style. You can, of course,
implement actors in Prolog, just like you can implement actors in
machine language. Whereas in raw Prolog you might naturally say
Transport-Mode (X, Y), in Intermission you have to say Send (X,
Transport-Mode, Y) where the Send interpreter does the work.
Nevertheless, I'd be very interested in hearing about your
improvements to Ken Kahn's implementation scheme.
-- Henry
------------------------------
Date: 23 May 1983 0806-PDT (Monday)
From: Abbott at AEROSPACE (Russ Abbott)
Subject: Object Oriented Programming in Prolog
Object oriented programming is, indeed, a powerful approach to program
design. To map object oriented programming directly onto Prolog, one
could, as suggested, define a one-place predicate for each type and
make other predicates depend on it. For example,
% bird(X): X is a bird. transport←mode(X, fly) :- % If X is a bird
then X's
bird(X). % transport←mode is fly.
The question raised concerns type specializations. One could define a
specialization as follows, e.g., for the specialization ostrich.
% ostrich(X): X is an ostrich. bird(X) :- % Every ostrich is a bird.
ostrich(X). transport←mode(X, walk) :- % If X is an ostrich
then X's
ostrich(X), % transport←mode is walk.
!.
It is possible in Prolog to ensure that a specialization's predicates
will appear earlier in the database than those of its generalization.
In that way the specialization's rules will be considered first.
For example, let us assume that:
ostrich(a). If one then asks:
transport←mode(a, M). the response will be:
transport←mode(A, walk), since the ostrich transport←mode rule
appears before the bird transport←mode rule in the database.
One may request other answers. But since backtracking is not
permitted through the "cut" (!), Prolog will never find any other
solutions. If one left the cut out, then indeed Prolog will see that:
bird(X) :- % Every ostrich is a bird.
ostrich(X).
and hence that transport←mode(A, fly) also holds.
If one wanted to be a little less ad hoc about it, one could
explicitly include in every predicate a type test. Thus, every
predicate would include in its name the types of its parameters.
(This is something like "overloading" in Ada.) So in addition to the
predicate
transport←mode
there would be (at least) two additional predicates stored in the
database:
bird$transport←mode and ostrich$transport←mode
One could then define a general transport←mode predicate as in the
following Prolog pseudo code.
transport←mode(X, Y) :-
Type(X),
not specialization←predicate←defined(Type, transport←mode, X),
Type$transport←mode(X, Y).
(Note this is not directly legal Prolog. It is pseudo code because I
have left out the steps that manipulate predicate names. In
particular, Type and Type$transport←mode are predicate variables that
would have to be "built." In particular, the predicate name
Type$transport←mode would be built from the name of the most
specialized type that X satisfied.)
Thus
specialization←predicate←defined(Type, Predicate, X) is
defined to hold exactly in those cases in which:
1) There is a specialization (ostrich) of the Type (bird) that
holds on X and
2) Predicate (in our case transport←mode) is defined
explicitly for
that type. (In our case ostrich$transport←mode is
defined.)
specialization←predicate←defined(Type, Predicate, X) could be
defined by the following Prolog pseudo code. (Again I have left out
the term building steps).
specialization←predicate←defined(Type, Predicate, X) :-
specialization(Type, Specialization),
Specialization(X),
specialization←predicate←exists(Specialization, predicate,
Specialization←Predicate),
clause(Specialization←Predicate(X), ←).
where,
specialization(Type, Specialization) holds if the type
Specialization is a specialization of the type Type. Of course that
information would have to be provided explicitly at some point, but
that is to be expected. The system could then add the rule
Type(X) :-
Specialization(X). and
specialization←predicate←exists(Specialization, Predicate,
Specialization←Predicate),
takes a type (Specialization) and a predicate (Predicate) and returns
the name (Specialization←Predicate) of the predicate that would apply
explicitly to the specialization.
clause(Specialization←Predicate(X), ←) indicates whether or
not such a predicate is defined.
In general, any <predicate> for any <type> would be transformed into
something like the following.
type$predicate(X, Y) :-
type(X),
not specialization←predicate←defined(type, predicate, X),
< ... definition of Y>.
It would take a little more work in setting up the database, but that
could be automated by a preprocessor, also written in Prolog. One
must expect to have to do some work. After all, the object oriented
languages have knowledge of objects and types built into them by their
implementers. In Prolog one must express that knowledge explictly.
But it seems that to do so is not such a burden.
I am in the process of building a system with features such as this
and will be able to let you know in a little while how well it works.
------------------------------
End of PROLOG Digest
********************
-------
∂27-May-83 1459 RPG Common Lisp
To: ohlander@USC-ISI
CC: rpg-q@SU-AI
CC: JMC@SU-AI
Griss and I are developing a proposal for a high-performance, portable
Common Lisp implementation, as you suggested to us earlier this spring.
It is nearly complete and I will ship you a draft early next week.
An interesting aspect that just arose is that Symbolics has approached
me asking for a co-operative effort in this area between Stanford and
Symbolics in Palo Alto. Do you have any warnings or advice about the
limits of such a collaboration?
I have discussed the following plan with them: Stanford pays part
of the salary of several individuals who are also partly employed by
Symbolics. These people help build the public domain compiler-compiler,
runtime system, and general methodology. Symbolics is allowed to distribute
the compiler-compiler and associated items. They are allowed to claim ownership
of specific compilers produced by the compiler-compiler (presumably those that
compile to machines that they build now or may build in the future.)
Let me know any thoughts you have on the subject.
-rpg-
∂29-May-83 0026 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #6
Received: from SU-SCORE by SU-AI with TCP/SMTP; 29 May 83 00:24:11 PDT
Date: Saturday, May 28, 1983 6:08PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #6
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Sunday, 29 May 1983 Volume 1 : Issue 6
Today's Topics:
Queries - Prolog Availability &
Expert Systems & Declaring Predicates Transitive
----------------------------------------------------------------------
Date: Thu 26 May 83 09:08:43-MDT
From: John Dawson <BDM.DAWSON@SANDIA.ARPA>
Subject: Search for PROLOG source
To whomever can help me:
Greetings;
I am trying to find a vendor to supply PROLOG to run on a VAX
11/780 (under VMS) or, more desirably, a DEC-20 (with TOPS-20
operating system). I have heard rumors that it is available from the
University of Edinburgh for 150 pounds (British). Can anyone confirm
or deny and/or offer any info to assist my quest?
Muchas gracias,
Reply to: John Dawson
BDM.DAWSON@SANDIA.ARPA
Neuvo Mexico, USA
------------------------------
Date: Thu 26 May 83 12:12:30-PDT
From: David Warren <WARREN@SRI-AI>
Subject: Prolog for VAX and DEC-20
AVAILABILITY OF PROLOG FOR VAX AND DEC-10/-20
C-Prolog for VAXes under Unix or VMS is available from:
EDCAAD
Dept. of Architecture
University of Edinburgh
20 Chambers Street
Edinburgh EH1 1GZ
Scotland
Tel: 031-667-1011 ext 4598.
The license fee for noncommercial use is currently 100 pounds.
For information about the availability of Prolog for DEC-10s and -20s
under Tops-10 or Tops-20, contact Fernando Pereira or David Warren at
SRI International:
PEREIRA@SRI-AI (415) 859-5494
WARREN@SRI-AI (415) 859-2128
--David Warren
------------------------------
Date: 27 May 1983 0923-PDT (Friday)
From: Abbott at AEROSPACE (Russ Abbott)
Subject: Alternate systems for Expert Systems
Does anyone know of any comparisons between Prolog and the popular
systems for building expert systems (such as KAS/Prospector, EMYCIN,
Rosie, etc.)? I'd be interested in answers to questions such as:
Does any system have a useful construct that is either not available
or difficult to build in the others? For example, many of the
explicitly knowledge based systems have a "probabilistic inference
engine." Prolog does not. Does that turn out to be important?
What development/debugging tools are available and how useful are
they?
What limitations are there?
How fast do they run (on what computer)?
How portable are they (over what range of what computers)?
------------------------------
Date: 27 May 1983 1008-PDT (Friday)
From: Abbott at AEROSPACE (Russ Abbott)
Subject: Declaring predicates transitive
Does anyone know of a good way to write a predicate in Prolog that
declares some other predicate transitive? For example:
transitive(R).
should declare that R is a transitive relation. This should have the
effect of ensuring that
R(A, C) :-
R(A, B),
R(B, C).
One would like to do all the usual things, such as:
R(a, b). and get either "yes" or "no."
R(a, X). and get an X, and then others on backtracking.
R(X, a). and get an X, and then others on backtracking.
R(X, Y). and get an X, Y pair, and then others on
backtracking.
One might imagine writing:
transitive(R) :-
assert(R(A, C) :- R(A, B), R(B, C)).
One problem is that if R is reflexive, i.e., R(X, X), then
R(a, c) :- R(a, a), R(a, c).
and one gets into an infinite recursion on R(a, c).
Even if R is not reflexive, if one asks:
R(a, X)
one gets
R(a, X) :- R(a, B), R(B, X)
and
R(a, B)
is essentially the same as
R(a, X)
leading to another infinite recursion.
Also, if one asks:
R(X, c)
the same kind of problem arises. Since
R(X, c) :- R(X, B), R(B, c).
in attempting to satisfy R(X, B), one gets
R(X, B) :- R(X, B←1), R(B←1, B)
and another infinite recursion.
It seems that one wants (at least) the following:
1) To keep track of which elements have been reached so far and not to
use them again.
2) To distinguish between variable and atomic parameters, so that the
first call to R in the transitive rule does not itself call the
transitive rule but attempts to exactly take one R step.
3) To distinguish between variable and atomic parameters in order to
determine in which "direction" to go.
Both of the two approaches that I have considered actually test to see
whether the transitive rule is being applied. In one a flag (ugh!) is
set and unset the database. (It must even be unset during backtrack-
ing.) In the other, before one attempts R(X, Y) one looks at the
body of the about-to-be-tried R clause to see if it is the transi-
tive clause. Both of these seem ugly.
Although I haven't tried it, I suppose it would work to define a new
predicate such as:
transitive←R(X, Z) :-
transitive←R(X, Z, [X, Z]).
transitive←R(X, Z, Used) :-
R(X, Z).
transitive←R(X, Z, Used) :-
R(X, Y),
not member(Y, Used),
transitive←R(Y, Z, [Y | Used]).
The problem with this approach is that it doesn't solve the problem I
want to solve. I want to use the original predicate name R and have it
act as if it is transitive after I declare it to be.
As a test case, imagine that R is less←than←or←equal←to and that =< is
not defined. Define less←than←or←equal←to(X, Y) as:
less←than←or←equal←to(X, X).
less←than←or←equal←to(0, 1).
less←than←or←equal←to(X, Y) :-
integer(X),
Y is X + 1.
less←than←or←equal←to(X, Y) :-
integer(Y),
X is Y - 1.
------------------------------
End of PROLOG Digest
********************
∂29-May-83 1853 ROODE@SRI-NIC [ROODE at SRI-NIC (David Roode): x.25 interfaces]
Received: from SRI-NIC by SU-AI with TCP/SMTP; 29 May 83 18:52:48 PDT
Date: 29 May 1983 1442-PDT
From: ROODE at SRI-NIC (David Roode)
Subject: [ROODE at SRI-NIC (David Roode): x.25 interfaces]
To: BillW at SRI-KL, McLure at SRI-UNIX, rcf at SRI-NIC,
Roode.UCI at RAND-RELAY, Try.UCI at RAND-RELAY, Sarvela at SRI-KL,
Yduj.HPLABS at RAND-RELAY, HSS at SRI-KL, JQJ at SU-SCORE,
JPBION at SU-SCORE, Chernikoff at SRI-TSC, Lauren at UCLA-LOCUS,
Ray at SRI-KL, Vivian at SRI-KL, JMC at SU-AI, King at KESTREL,
Ellen at MIT-MC, Satz at SRI-TSC, Pourne at MIT-MC, JSol at USC-ECLC,
Pleasant at RUTGERS, Zellich at OFFICE-3
Location: EJ296 Phone: (415) 859-2774
Would you think people would like a public electronic mail host, say a Unix
system with good mail utilities capable of supporting 15-20 users,
which charged something like:
local nationwide
5pm-1am $1.50 $2.50
1am-7am $1.00 $2.00 per hour
7am-9am $1.50 $6.50
9am-5pm $3.00 $8.00
and also offered Bulletin boards, program libraries, some games, etc.?
Why do OnTyme, Telemail and so forth charge so much? Is it a case
of charging all the traffic can bear? For one thing, some of these services
do not quite have the idea down right, because they staff toll free phone-in
reference services, where they should be encouraging people to do this
type of interaction on-line.
---------------
Date: 29 May 1983 1415-PDT
From: ROODE at SRI-NIC (David Roode)
Subject: x.25 interfaces
To: klh, feinler
cc: roode
Location: EJ296 Phone: (415) 859-2774
I was just thinking how it would be good to open a service like "The
Source" or "Compuserve Information Services" but only charge $1 per
hour for connect time (excluding communications charges) in non
daytime hours, instead of $5-$8 (which includes public network
communications) like these services do. (In the daytime, the rates
are $20-$25).
Communication-wise, Tymnet sells connect time nationwide in what is
called "Leisure Tyme" for $.90 per hour, except your systems users
have to spend at least $5000 that way per month. The first so many
packets per hour are included, and it is close to enough at 300 baud
(though probably not at 1200) to run without excess packet charges.
Telenet has a comparable service, so it ought to be possible to
accomplish nationwide access for a surcharge on the order of $1 per
hour, which local users could escape.
.
.
.
-------
-------
David: The key thing is the personnel budget. Tymshare charges a lot,
because only 12.5 percent of their costs are for running the machines.
The rest are for handholding, sales, etc. I don't know exactly what is
included in what, and my information is five years old.
∂31-May-83 0640 OHLANDER@USC-ISI Re: Common Lisp
Received: from USC-ISI by SU-AI with TCP/SMTP; 31 May 83 06:40:12 PDT
Date: 31 May 1983 0639-PDT
Sender: OHLANDER at USC-ISI
Subject: Re: Common Lisp
From: OHLANDER at USC-ISI
To: RPG at SU-AI
Cc: rpg-q at SU-AI, JMC at SU-AI
Message-ID: <[USC-ISI]31-May-83 06:39:40.OHLANDER>
In-Reply-To: Your message of 27 May 83 1459 PDT
Dick,
I am not sure why Symbolics wants to get involved with this. dd
Don't they already have compiler?
Ron
∂31-May-83 1302 @SU-SCORE.ARPA:LFW@MIT-ML
Received: from SU-SCORE by SU-AI with TCP/SMTP; 31 May 83 13:01:52 PDT
Received: from MIT-ML by SU-SCORE.ARPA with TCP; Tue 31 May 83 12:43:44-PDT
Date: 31 May 1983 15:10 EDT
From: Laurence F. Wood <LFW @ MIT-ML>
To: PROLOG-HACKERS @ SU-SCORE
YES!! I would like to get prolog hacker info.
∂31-May-83 1514 TGD Minski's Learning Meaning
To: JMC@SU-AI
CC: TGD@SU-AI
Do you have a copy of Minski's paper on "learning meaning" that I could
borrow and photocopy? I saw it mentioned on the phil-sci mailing
liist.
--Tom
It's Minsky, and I'll bring in my copy if you haven't already found one.
∂31-May-83 1657 NOVAK@SU-SCORE.ARPA Alto use
Received: from SU-SCORE by SU-AI with TCP/SMTP; 31 May 83 16:56:26 PDT
Date: Tue 31 May 83 16:59:48-PDT
From: Gordon Novak <NOVAK@SU-SCORE.ARPA>
Subject: Alto use
To: jmc@SU-AI.ARPA
Yesterday I needed an Alto, and found 4 of them up -- all occupied by
business students writing term papers. If the dept. posted signs that
the Altos were for CS use only, it could get by with fewer.
-------
∂31-May-83 2011 @MIT-MC:BLOCK%MIT-OZ%MIT-MC@SU-DSN Re: role of logic in AI
Received: from MIT-MC by SU-AI with TCP/SMTP; 31 May 83 20:11:24 PDT
Date: 31 May 1983 2306-EDT
From: BLOCK%MIT-OZ%MIT-MC@SU-DSN
Subject: Re: role of logic in AI
To: JMC@SU-AI
In-Reply-To: Your message of 18-May-83 0031-EDT
I notice you refer to your "Ascribbing Mental Qualities to Machines".
How can I get a copy?
Ned Block (ned@MIT-COGS@MIT-MC)
-------
I just saw your message after return from trip. She'll send you a
paper copy of "Ascribing ... " if you send my secretary DFH@SU-AI
a U.S. mail address".
∂31-May-83 2324 RPG Common Lisp
To: ohlander@USC-ISI
CC: JMC@SU-AI, rpg-q@SU-AI
They have a compiler at the moment for their various machines (LM-2, 3600).
Talking to Marc LeBrun - who was asked to talk to me about a portable Common Lisp
by Henry Baker - I get the impression that there are two things that Symbolics
would like to get out of it.
First, they want to appear to the world as concerned about the common good.
Thus, they want to show that they are willing to put some investment into
the community for PR purposes.
Second, being a company, they want to make a profit from the portable
Common Lisp. I think there are three basic means for profit. They can
sell portable Common Lisp. This is out, since a collaboration with an
ARPA-funded institution cannot result in a profit-making piece of
software. They can sell Common Lisp systems ported using the developed
technology. I doubt that even the most successful portable Common Lisp
project could expect to be able to develop a portability technology that
could get up a Common Lisp on an arbitrary machine in less than 6 months.
Perhaps they want to sell that 6 months' worth of effort for various
machines. Finally, they can sell other software they develop for the
portable Common Lisp, giving away the Common Lisp system that the
technology produces.
I also suppose they wouldn't mind getting some of the compiler technology that
this project would produce.
I will talk to them more. I think that they will be fairly honest about their
interests.
-rpg-
∂03-Jun-83 0024 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #7
Received: from SU-SCORE by SU-AI with TCP/SMTP; 3 Jun 83 00:24:33 PDT
Date: Thursday, June 2, 1983 4:15PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #7
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Friday, 3 Jun 1983 Volume 1 : Issue 7
Today's Topics:
Representation - Declaring Predicates Transitive,
Applications - Objects
----------------------------------------------------------------------
Date: 30 May 1983 1543-EDT (Monday)
From: Vijay.Saraswat@CMU-CS-A
Subject: Declaring Predicates Transitive
The following solution is adequate for all the 4 cases mentioned by
Abbott:
transitive(R,Elem):-
atom(R), nonvar(Elem),
Rxy=..[R,X,Y], Rxz=..[R,X,Z], Rzy=..[R,Z,Y],
assert((Rxy:- nonvar(X),Rxz, X \==Z, Rzy)),
assert((Rxy:-nonvar(Y),Rzy, Y \==Z, Rxz)),
assert((Rxy:- X is Elem, Rxy)).
Elem is any element in the domain over which R is defined.
-- V. Saraswat.
------------------------------
Date: Tuesday, 31 May 1983 14:12-PDT
From: Narain at Rand-Unix
Subject: Reply to Henry Liebermann
1) I agree that actors may be implemented in any language. The
advantage of of choosing Prolog is however that you can use its most
powerful features like pattern matching, automatic backtracking,
deductive facility and well understood theoretical foundations.
Because of these features, an object oriented interpreter in Prolog
has a very simple definition.
Both styles "Object Oriented" and "natural Prolog" are very powerful
and the nice thing is to be able to use both in the same system.
2) An improvement in Kahn's "Actors in Prolog".
In Kahn's paper, actor behaviors are stored as clauses under just one
single predicate "sent", each of which has the form:
sent(Actor,Message,Result):-<body having other calls to sent>.
To compute the result "R" of passing message "M" to actor "A", one
simply makes the call:
:-sent(A,M,R)
Clearly, as more objects are added and more behaviors defined, any
call to "sent" will have to search through all clauses under "sent".
This effectively means that the knowledge base of all actors is being
searched, even when many actors may be totally irrelevant to the one
that received the message.
A fairly straightforward improvement over this search strategy is
given. Essentially, the search for a matching procedure is made on
nodes only along one branch of the inheritance tree instead of on all
the nodes of the inheritance tree. The idea is simple and no doubt
exists in other object oriented implementations. It does at Rand
where we have developed a language called ROSS for doing
knowledge-based simulations.
When an actor receives a message, it determines its immediate ancestor
and examines its knowledge base for a matching procedure. If it finds
a match it executes it, otherwise it goes on to the next ancestor in
the inheritance tree.
The convention adopted here is that all clauses in an <actor-name>'s
knowledge base are stored under the predicate <actor-name>←behavior.
Thus to search the knowledge base of <actor-name> we only need to
examine clauses under <actor-name>←behavior, and not all clauses under
"sent".
The modified interpreter is:
sent(Actor, Msg, Result):-
ancestor(Actor, Actor←anc),
procedure(Actor←anc, Actor, Msg, Result, Procedure),
call(Procedure),
!.
/* The "cut" is necessary only if we wish to constrain the response
to be the first one found. It is strictly unnecessary otherwise.
When "Actor" receives "Msg", the result is "Result" if "Actor←anc"
is an ancestor of "Actor", and "Procedure" is a call to
<Actor←anc>←behavior, and "Procedure" is successfully called.
Kahn's actor classes or generic actors are essentially "records"
and are implemented using Prolog lists. A generic actor is just
a list of partially instantiated elements where each element
represents one field of the "record". An instantiation of this
generic actor is just a term that matches the generic actor (a list).
Predicate "procedure" defines for each generic actor (its first
argument), the call to clauses inside its knowledge base, and binds
"Procedure" to this call. "Myself" is the actor that was originally
sent the message.
Thus: */
procedure([list], Myself, Msg, Result, list←behavior(Myself, Msg,
Result)). procedure([nlist,←,←,←],
Myself,Msg,Result,nlist←behavior(Myself,Msg, Result)).
procedure([slist|X],Myself,Msg,Result,slist←behavior(Myself,Msg,Result)).
procedure([],Myself,Msg,Result,null←list←behavior(Myself,Msg,Result)).
/* Definition of 'ancestor' */ ancestor(X,X). /* search begins with
the actor, so it is its own "ancestor" */ ancestor(X,Y):-parent(X,Y).
ancestor(X,Y):-parent(X,Z),ancestor(Z,Y).
/* Hierarchy in the context of lists */ parent([],[list]).
parent([nlist,←,←,←],[list]). parent([slist|X],[list]). /* a simple
list, where "slist" stands for "cons" */
/* Now "nlist←behavior", "list←behavior" etc. are defined as they
were originally by Kahn. Only the behaviors for the "print←elements"
message are outlined here, but other behaviors in Kahn's paper
including the Sieve of Eratosthenes alorithm were also programmed.
Note that behaviors are still defined with the same clarity, if the
naming convention is adhered to */
list←behavior(Myself,print←elements,←):-
sent(Myself,you←are←empty,true),!.
list←behavior(Myself,print←elements,←):-
write(' '),
sent(Myself, first, First),
write(First),
sent(Myself, rest, Rest),
sent(Rest,print←elements,←).
/* Special behavior for nlist */
nlist←behavior(Nlist,print←elements,←):-
sent(Nlist,length,Length),
Length > 5,
sent(Nlist, print←elements←with←dots, ←).
When an actor of the form [slist|X] receives a message to
"print←elements" it finds no behaviors in its own knowledge base. It
then determines that it is an instance of [list] and finds an
appropriate response under "list←behavior". It does not look under
"nlist←behavior".
On a DEC-20, the Sieve algorithm as programmed originally generated
only 11 primes (till 31) before running out of space. The algorithm
when reprogrammed as above generated 103 primes (till 563) before
running out of space, and ran three times as efficiently.
-- Sanjai Narain
------------------------------
Date: Thursday, 2 June 1983, 01:07-EDT
From: Henry Lieberman <Henry%MIT-OZ@MIT-MC>
Subject: Object-Oriented Programming in Prolog, reply to Narain
Sanjai,
Thank you for your informative reply. My intention was just to
distinguish the claim "Prolog is a good language for implementing
object-oriented programming languages", which your arguments support,
from the claim "Prolog IS a good object-oriented programming
language". The first claim I am sympathetic to, because of the
features you cite, but the second I am skeptical about, since all such
attempts I've seen involve adding additional interpretive mechanism
and changing programming style conventions rather drastically. I
believe Kahn's original intent in Intermission was to support the
first claim rather than the second.
The problem with the "hundred flowers" approach to supporting multiple
programming styles in the same language (efficiency aside) is
confusion for the poor user. In your scheme, a programmer might start
out thinking "I'm not going to use the hairy features of Actors, why
don't I just program directly in Prolog" and writes predicates like
Above(Block-1, Block-2). Then later in the evolution of the system,
the need for default information arises, so all previous code has to
be rewritten to say Sent(Block-1, Above, Block-2, True). This comes up
all the time in Lisp-embedded object-oriented programming [Flavors,
Loops, etc.] The point is if you're convinced enough about the virtues
of object-oriented programming, you might as well make your language
use it for everything.
I like your optimization to Kahn's message reception scheme, it works
out very nicely. Something I think Prolog-based object-oriented
languages might be promising for is in dealing with the so-called
"multiple-inheritance problem" which has been a difficult point in
other languages so far. Perhaps it would be possible in such a scheme
to just add some more clauses involving Ancestor and Parent relations
to resolve inheritance conflicts.
-- Henry
------------------------------
End of PROLOG Digest
********************
∂04-Jun-83 2200 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 4 June 1983
Previous Balance 0.96
Monthly Interest at 1.5% 0.01
-------
TOTAL AMOUNT DUE 0.97
Please deliver payments to Diana Hall, room 358, Jacks Hall.
Make checks payable to: STANFORD UNIVERSITY.
To ensure proper crediting, please include your Pony account name on your check.
Bills are payable upon presentation. Interest of 1.5% per month will be
charged on balances remaining unpaid 25 days after bill date above.
You haven't paid your Pony bill since 12/82.
Accounts with balances remaining unpaid for more than 55 days are
considered delinquent and are subject to reduction of credit limit.
Please pay your bill and keep your account current.
∂06-Jun-83 0041 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #8
Received: from SU-SCORE by SU-AI with TCP/SMTP; 6 Jun 83 00:41:32 PDT
Date: Sunday, June 5, 1983 5:34AM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #8
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Sunday, 5 Jun 1983 Volume 1 : Issue 8
Today's Topics:
Query - Equivalence & A Puzzle,
Representation - Declaring Predicates Transitive,
LP Library List Available!
----------------------------------------------------------------------
Date: 4 June 1983 1650-PDT (Saturday)
From: Abbott at AEROSPACE (Russ Abbott)
Subject: Equivalence
This may seem like a stupid question, but I'll risk looking stupid and
ask it anyway. How does one deal with equivalence in Prolog?
Imagine that one is building a system that interactively constructs a
knowledge base in the form of Prolog rules. While interacting with a
user one stores a number of predicates that are defined in terms of
some predicate p(X).
a(x) :- ..., p(x), ..., .
Also, one stores a number of predicates that are defined in terms of
some other predicate q(X).
b(x) :- ..., q(x), ..., .
Various definitions for p and q are also given by the user. The user
then tells you that p(X) means the same thing as q(X), i.e.,
p(X) :- q(X). and
q(X) :- p(X).
One can't put both of these rules into the database for fear of
infinite loops. Are there any clean ways out of this predicament?
------------------------------
Date: 2 Jun 83 04:52:02 EDT (Thu)
From: Bruce T. Smith <bts.unc@UDel-Relay>
Subject: Prolog puzzle
Starting with problem #26 in his book "What is the Name of this
Book?", Raymond Smullyan presents a number of variations on a
familiar logic puzzle. Here's his introduction to the section:
There is a wide variety of puzzles about an island in
which certain inhabitants called "knights" always tell
the truth, and others called "knaves" always lie. It
is assumed that every inhabitant is either a knight or
a knave.
All the puzzles involve meeting a small party of the island's
inhabitants. Sometimes a question must be asked, other times
information is volunteered. In all cases, the goal is to determine
which are knights and which are knaves. A sample of this type of
puzzle is the following:
#31. Again we have three people, A,B,C, each of whom is
either a knight or a knave. A and B make the following
statements:
A: All of us are knaves.
B: Exactly one of us is a knight.
What are A,B,C?
I've played around at solving these puzzles using Prolog, with
only mixed results. Here's the challenge for Prolog programmers:
Write a Prolog program which will solve this type of problem, where a
problem description is a list of names (e.g. A,B,C) and statements.
Restrict the statements the people may make, but at least let them
make claims about which groups they and their companions belong to.
This simplest puzzle may seem too easy, the set of possible
solutions is small enough to examine them all, for instance. The
puzzles in the book quickly increase in difficulty, however. One
simple variation is introduced in problem #39, with the addition of
"normals", who sometimes tell the truth and sometimes lie. (By the
end of the book, Smullyan is playing around with Godel's Theorem.)
I'd be interested in seeing how far people can get with knights and
knaves problems (or beyond), comments on solving this type of puzzle
(where the truth or falsity of statements must be determined before
they can be used), etc.
-- Bruce Smith, UNC-CH
duke!unc!bts (USENET)
bts.unc@udel-relay (other NETworks)
------------------------------
Date: 3 June 1983 0916-PDT (Friday)
From: Abbott at AEROSPACE (Russ Abbott)
Subject: Still trying to Declare Predicates Transitive
Vijay,
Thanks for your answer to my question concerning declaring predicates
transitive. I have two problems with your solution.
1) If there is a cycle longer than just one step, your X \== Z
and X \== Y tests won't catch it. But as I originally suggested
that's fairly easy to fix by keeping track of all the elements
seen so far.
2) Your solution depends implicitly on your added predicates never
being tried more than once. That is assured for the particular
example I gave because:
(a) Your added predicates come last in the database, and
(b) There are (implicilty) an unbounded number of values that are
greater than some given value. I.e., your solution never says
"no."
What would you do if I changed the example as follows. (It really
shouldn't matter what the R is that is being declared transitive.)
The following makes two modifications:
a) R(X, Y) if floor(X/3) and floor(Y/3) are equal.
E.g., 3, 4, and 5 are all in the same equivalence class.
b) The set of values is finite.
less←than←or←equal(X, Y) :-
integer(X),
0 is floor(X/10), % X is equivalent to Y if
1 is floor((X+1)/3) - floor(X/3), % floor(X/3) = floor(Y/3).
Y is X - 2. less←than←or←equal(X, Y) :-
integer(Y),
0 is floor(Y/10), % X is equivalent to Y if
1 is floor(Y/3) - floor((Y-1)/3), % floor(X/3) = floor(Y/3).
X is Y + 2. less←than←or←equal(0, 1).
less←than←or←equal(X,Y) :-
integer(X),
0 is floor(X/10),
Y is X + 1. less←than←or←equal(X, Y) :-
integer(Y),
0 is floor(Y/10),
X is Y - 1.
------------------------------
Date: Sat 4 Jun 83 23:50:34-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Prolog Library List Available!
The following describes the Prolog programs available on directory
<PROLOG> at SRI-AI. They are all accessible to an anonymous FTP login.
This file is PROLOG-LIBRARY.LIST in that directory.
The programs have been tested on DEC-10s and DEC-20s with the
Edinburgh Prolog system. They might need minor changes to run under
other Prolog systems, such as C-Prolog.
Copy and enjoy!
Fernando Pereira
======================================================================
* The utilities package from the Department of Artificial Intelligence,
University of Endinburgh. Contributors include Alan Bundy, Lawrence
Byrd and Richard O'Keefe.
util The top level file, which loads all the others
util.hlp A minimal (and outdated) help file
writef.pl Formatted write (writef)
trace.pl Tracing routines
readin.pl Read in a sentence
listro.pl List routines
setrou.pl Set routines
applic.pl Application routines
multil.pl Multi list routines
flagro.pl Flag handling
struct.pl Structure crunching
cmisce.pl Miscellaneous
long.pl Rational arithmatic package
tidy.pl Expression tidy/evaluator
invoca.pl Invocation routines
imisce.pl Miscellaneous (interpreted)
writef.hlp Documentation for the formatted write
* Pretty printer and utilities from Harry Barrow, Fairchild.
pp Pretty printer utils Utilities
* Tutorial programs and text by Ernie Davis and Udi Shapiro, Yale and
Weizmann Institute.
tutori.pl Programs
tutori.lpt Text
* Teach-yourself Prolog program, by William Wong, Rutgers.
To run it, load it into Prolog and then call the predicate 'hi'.
cai.pl
* Prolog cross-reference program from Edinburgh
xref.hlp Documentation (ignore the installation-specific stuff)
xref.pl The cross-referencer (compile it, otherwise it will
run out of space very quickly)
xref.mic MIC command file to create a runnable XREF image
(installation-specific)
xref.tec TEC124 macros to update Prolog sources with public
declarations.
* Richard O'Keefe's Prolog ToolKit.
Richard has started to develop an integrated Prolog toolkit. A nice
feature is a general help facility that allows interactive perusal
of keyword-indexed help files from inside Prolog. I have made some
minor modifications to make the toolkit behave better with TOPS-20
filenames, but there might be other installation dependencies. In
my version, all files mentioned in the code come from directory
<PROLOG>.
toolkit.hlp The basic help file for the ToolKit distribution
toolkit.pl Compiles and sets up the toolkit. It should be loaded
on top of the Edinburgh UTIL package (you can avoid
this if you scan the "imports" comments to find which
UTIL predicates are used, and load just those).
toolkit.exe Created by calling 'save←toolkit' after loading the above.
Due to excessive cleverness in the 'save' evaluable
predicate, saved images are installation-dependent,
so this is irrelevant outside SRI-AI; rebuild it locally
as described above.
helper.pl Prolog help facilities, hacked for TOPS-20.
helper-10.pl The original TOPS-10 version of the same.
helper.hlp Help for help.
pp.pl Program pretty-printer and browser (has a nice
partial match facility for predicate and functor
names). Do not confuse with pp., Harry Barrow's
term pretty-printer
pp.hlp Help for the same.
ixref.pl Interactive cross-referencer (based on XREF above,
but nicer).
ixref-10.pl The original version, with one Edinburgh-specific
file name.
ixref.def Initial knowledge for the cross-referencer.
ixref.hlp Help.
count.pl Counts the number of clauses in Prolog files (for the
"largest Prolog program" award).
count.hlp Help.
vcheck.pl Checks Prolog files for nonanonymous variables occurring
only once in a clause; invaluable!
vcheck.hlp Help
------------------------------
End of PROLOG Digest
********************
∂06-Jun-83 2224 ME uncooperative user
To: REG@SU-AI, LB@SU-AI, JMC@SU-AI, MRC@SU-AI
RMS has routinely maintained an obvious password on his Sail account when
it has existed for the last year or more. On various occasions I have
asked him not to do that, but he insists on making it possible for lots of
people to use his account. Apparently various other people do use it.
Several months ago, because of his unwillingness to abide by our policy of
not sharing accounts, I deleted all of his directories on Sail. Recently,
JMC created a directory for RMS to let him use Sail while he was in
California. And once again RMS has been maintaining an obvious password.
Last night I noticed someone logged in as RMS from SRI-UNIX, although
FINGER showed no one resembling RMS logged in at SRI-UNIX. So I changed
the password and again asked RMS not to let others use his account, as
that is our policy. He replied that his policy was not to help others
enforce computer security. Consequently, I have again deleted his files
and directory from Sail. In the process I noticed that he had an NS
notification request in which was outputting to one of his files. So I
cancelled that request and observed that it was one that would match all
stories. Thus his request would use up quite a bit of disk space if it
were't constantly edited to remove stories from the output file. In
deleting the NS request, I noticed that it had been entered today --
apparently last night, when "he" was logged in just before I changed his
password. Since we don't know if RMS himself was actually the one logged
in as him last night, we don't know who made the NS request, but it had
already taken up about 90 blocks of disk space. I think I know who
the person was at SRI-UNIX, since there were only a couple of people
logged in there at the time.
As far as I'm concerned, RMS is persona non grata at Sail.
∂07-Jun-83 0029 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #9
Received: from SU-SCORE by SU-AI with TCP/SMTP; 7 Jun 83 00:28:58 PDT
Date: Monday, June 6, 1983 10:32PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #9
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Tuesday, 7 Jun 1983 Volume 1 : Issue 9
Today's Topics:
Administrivia - Bad Formatting of Predicate Fixed,
Prolog Availability,
Implementations - Performance
----------------------------------------------------------------------
Date: 6 June 1983 0806-PDT (Monday)
From: Abbott at AEROSPACE (Russ Abbott)
Subject: Formatting
The formatting of my revised less←than←or←equal predicate for use in
declaring predicates transitive apparently got fouled up in transit.
It should be as follows.
[ My apologies to all. -- Chuck ]
less←than←or←equal(X, Y) :-
integer(X),
0 is floor(X/10), % X is equivalent to Y if
1 is floor((X+1)/3) - floor(X/3), % floor(X/3) = floor(Y/3).
Y is X - 2.
less←than←or←equal(X, Y) :-
integer(Y),
0 is floor(Y/10), % X is equivalent to Y if
1 is floor(Y/3) - floor((Y-1)/3), % floor(X/3) = floor(Y/3).
X is Y + 2.
less←than←or←equal(0, 1).
less←than←or←equal(X,Y) :-
integer(X),
0 is floor(X/10),
Y is X + 1.
less←than←or←equal(X, Y) :-
integer(Y),
0 is floor(Y/10),
X is Y - 1.
------------------------------
Date: 6 Jun 83 16:23:06 EDT
From: Sean McLinden <MCLINDEN@RUTGERS.ARPA>
Subject: C-Prolog for VAX running 4.xx bsd UNIX
Is there a version of Prolog available for Vaxes running Berkeley
UNIX? If there is, how would I go about getting a copy?
Thanks.
Sean McLinden
Decision Systems Lab
University of Pittsburgh
School of Medicine
------------------------------
Date: Mon 6 Jun 83 15:30:03-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Another Prolog in C
Besides C-Prolog, which follows the DEC-10/Clocksin & Mellish dialect,
there are a few other Prolog interpreters in C, none of which I have
used myself. One of them, UNH-Prolog, seems to be also compatible
with the DEC-10/Clocksin & Mellish dialect. It can be obtained for
$100 from
Prof. Jim
Dept. of Computer Science
University of New Hampshire
Kingsbury Hall
Durham, New Hampshire 03824
-- Fernando Pereira
------------------------------
Date: Mon 6 Jun 83 15:23:02-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Getting C-Prolog
HOW TO GET C-PROLOG
30 May 1983
C-Prolog is a Prolog interpreter written in C for VAXes, but it will
also run on MC68000 and Z8000 systems with enough address space. It
will run both under Unix and VMS. C-Prolog is almost entirely
compatible with DEC-10 Prolog and also very similar to the dialect of
Prolog described in the book "Programming in Prolog" by Bill Clocksin
and Chris Mellish. In contrast with most other Prolog systems, it
handles floating-point numbers.
Copies of C-Prolog can be obtained from
EdCAAD
Dept. of Architecture
University of Edinburgh
20 Chambers Street
Edinburgh EH1 1GZ
Scotland
(011-44-31)667-1011 x. 4598
There is a small handling charge for research and teaching uses, and a
distribution for commercial purposes is also available.
-- Fernando Pereira
------------------------------
Date: 1 Jun 83 10:59:05-PDT (Wed)
From: harpo!seismo!presby!burdvax!mckay @ Berkeley
Subject: More on LIPS and Benchmarking Problems
In articles and talks about logic programming implementations and in
this news group, the efficiency of a system is measured in
"resolutions per second" or "logical inferences per second" (aka
LIPS). These are usually mentioned in association with horn clause
systems of various flavors, E.g. DEC-10/20 Prolog, LOGLISP. By far the
fastest systems appear to be the DEC-10/20 Prolog systems which claim
rates of 20,000 to 40,000 "resolutions per second" for compiled
clauses.
A benchmark is important because of interactions of the specific
program being used and the definition of "LIPS". Consider the two
obvious definitions. First, the one suggested by the use of the
phrase "resolutions per second". This suggests one is measuring
successful unifications of a literal with the head of a clause from
some procedure as well as finding the appropriate clauses with which
to attempt unification. This means that the measurement includes many
unification attempts which may fail. It is extremely dependent on the
order of clauses within a procedure and the clarity of the predicate
involved. The measurement can be severely effected by the "shape" of
clauses or literals. A second definition to consider is attempted
unifications regardless of whether they fail or succeed. This would
equate LIPS with unification. One still has the problem with the
clarity of the literals involved but the problem with the order of
clauses has been minimized. While unification is a critical component
of a logic programming system, it by itself does not measure
"progress" of a computation.
All of this suggests that LIPS (whichever definition one uses) is
extremely application specific and, therefore, if one is quoting a
LIPS for a particular system one MUST state with what the measurement
was done, I.e. a plain LIPS figure is is not good enough, it must be
"LIPS with respect to X".
Therefore:
What is appropriate to measure for such systems?
What are reasonable benchmarks?
What have you used for benchmarks in the past?
What AND WHY did you choose the specific logic programs as benchmarks?
What measurements are there for the various systems?
------------------------------
Date: Sun 5 Jun 83 02:01:34-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: LIPS again
Although the author of the USENET article burdvax.781 is correct in
theory about all the tricky factors than enter in any LIPS evaluation,
in practice (8 years of it) the small set of benchmarks in David
Warren's "Implementing Prolog" technical reports (available from the
Dept. of Artificial Intelligence of Edinburgh University) seems to
provide a good estimate of the performance of a Prolog system for a
large variety of tasks.
The figures of 40000 LIPS for DEC-20 Prolog and 1500 LIPS for C-Prolog
on a VAX 780 come from timing "naive reverse" defined as follows:
nrev([],[]).
nrev([X|L],R) :- nrev(L,R0), conc(R0,[X],R).
conc([],L,L).
conc([X|L1],L2,[X|L3]) :- conc(L1,L2,L3).
Although this is very simple, it tests the shallow backtracking and
structure crunching which is charateristic of complex Prolog programs
such as compilers, language parsers, term rewrite systems (E.g.
algebraic simplifiers and equation solvers), theorem provers, etc. The
program doesn't test deep backtracking and cut, but these seem to be
comparatively fast in those systems that are good enough to be worth
our trouble measuring.
To get accurate figures for this kind of benchmark, we use a test
program which calls nrev once to allocate space, page in, etc., then
fails and goes into a fail loop calling nrev a number of times.
For the benefit of those who don't have access to the "Implementing
Prolog" papers, I intend to submit the collection of benchmarks and
results for DEC-20 Prolog on a 2060 and C-Prolog on a VAX 780 in the
near future. I hope this will provide a "de facto" standard for the
evaluation of Prolog systems.
-- Fernando Pereira
------------------------------
End of PROLOG Digest
********************
∂07-Jun-83 1453 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA ARPA proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 7 Jun 83 14:16:03 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 7 Jun 83 14:19:35-PDT
Date: Tue 7 Jun 83 14:06:17-PDT
From: Gio <Wiederhold%SRI-AI@SU-SCORE.ARPA>
Subject: ARPA proposal
To: rpg@SU-AI.ARPA, bscott@SU-SCORE.ARPA, jmc@SU-AI.ARPA, pickering%SRI-AI@SU-SCORE.ARPA
I have dictated a new introduction for the ARPA proposal.
Sections which can be used from the old one will have to be merged,
and BettY's equipment transfer and budget section inserted.
When Jayne has it typed in I will do the merge, using RPG's format.
Was there a chnage of title suggested - I do recall so, but dont know what the
change should be.
RPG - can you FTP me the file? or tell me where it is - I'll transfer it at night
because of the size.
Gio
-------
∂08-Jun-83 1353 BSCOTT@SU-SCORE.ARPA Re: ARPA proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 8 Jun 83 13:52:01 PDT
Date: Wed 8 Jun 83 13:16:45-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Re: ARPA proposal
To: Wiederhold@SRI-AI.ARPA, rpg@SU-AI.ARPA, jmc@SU-AI.ARPA,
pickering@SRI-AI.ARPA
cc: BSCOTT@SU-SCORE.ARPA
In-Reply-To: Message from "Gio <Wiederhold@SRI-AI.ARPA>" of Tue 7 Jun 83 14:18:54-PDT
A new title was suggested, Gio. I believe it was Bob Kahn who didn't like
it.
Betty
-------
∂08-Jun-83 1446 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA Re: ARPA proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 8 Jun 83 14:46:21 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Wed 8 Jun 83 14:38:32-PDT
Date: Wed 8 Jun 83 14:26:57-PDT
From: Gio <Wiederhold%SRI-AI@SU-SCORE.ARPA>
Subject: Re: ARPA proposal
To: BSCOTT@SU-SCORE.ARPA
cc: rpg@SU-AI.ARPA, jmc@SU-AI.ARPA, pickering%SRI-AI@SU-SCORE.ARPA,
wiederhold%SUMEX-AIM@SU-SCORE.ARPA
In-Reply-To: Your message of Wed 8 Jun 83 13:54:53-PDT
who knows what it is, or should we make it up?
Changing it will invalidate references in the task proposals to the umbrella,
so the onle chnge should be adding a post-fix, or perhaps a prefix to the title:
{Advanced/Collected/Miscellanous/...
Research and Development in Computer Science
at Stanford/ Problems / Methods/ ... }
?
-------
∂09-Jun-83 0947 WIEDERHOLD@SUMEX-AIM revision of proposal
Received: from SUMEX-AIM by SU-AI with PUP; 09-Jun-83 09:47 PDT
Date: Tue 7 Jun 83 10:34:26-PDT
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM>
Subject: revision of proposal
To: rpg@SAIL, bscott@SCORE
cc: jmc@SAIL
do you have the on-line version of the enire Arpa umbrella proposal.
I am making an attempt at a revision to satisfy ARPA. Gio
-------
∂10-Jun-83 0000 JMC*
Ask Fran about secretarial work.
∂10-Jun-83 1020 ULLMAN@SU-SCORE.ARPA facilities committee
Received: from SU-SCORE by SU-AI with TCP/SMTP; 10 Jun 83 10:20:10 PDT
Date: Fri 10 Jun 83 10:15:06-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: facilities committee
To: jmc@SU-AI.ARPA
As it is appearing all the time more likely that Ralph is leaving,
I think that we should start a search for his successor.
I have already caused an ad to be placed in CACM, for all the good
it will do. However, I would like you and your facilities committee
to take charge of the selection and hiring process, rather than
appoint an ad-hoc committee.
You should know that even if you wanted to appoint Bosack, he may not
be available, so the sitiuation could become desparate if we are not
careful.
Tom Rindfleisch has offered some assistance in generating names of
candidates, and you may wish to consult him.
-------
I'll take charge of successor problem.
∂10-Jun-83 1026 DUMAS%SUMEX-AIM.ARPA@SU-SCORE.ARPA prolog
Received: from SU-SCORE by SU-AI with TCP/SMTP; 10 Jun 83 10:26:14 PDT
Received: from SUMEX-AIM by Score with Pup; Fri 10 Jun 83 10:29:43-PDT
Date: Fri 10 Jun 83 10:26:15-PDT
From: Jean-Pierre Dumas <DUMAS@SUMEX-AIM>
Subject: prolog
To: mccarthy@SCORE
Dear Dr. Mc Carthy,
Can you tell me what and where about finding a portable Prolog
interpreter...is there any available (free or almost).
Are the prolog interpreters mainly compatible or not ?
I have heard that the performances are not as good as INTERLISP,
as i think that INTERLISP is aquite mediocre system , i wonder.!!!
Any information on prolog will be very much welcomme.
Jean-Pierre Dumas.
-------
∂10-Jun-83 1031 TOB
To: JMC, NAN
Purchasing
John
Please comment on how to make an effect. I plan to send this to Lieberman.
Tom
Dear
We write to arrange a more efficient purchasing method for our research.
We want to purchase directly without going through the delays and
effort involved in going through the Purchasing Department.
We find that they are only a burden on us. While they may be helpful
in procurement of routine supplies, we have yet to be aware of any
benefit they provide to us.
They have no expertise in the sort of equipment which we buy. We
must do all vendor selection and market survey ourselves.
The reality is that they lose track of things and cause substantial delays.
If we do need something in a hurry, my secretary must walk it through
at the expense of several hours of effort.
I do not think that our research can afford such extravagant waste.
∂10-Jun-83 2147 FEIGENBAUM@SUMEX-AIM Guibas appointment
Received: from SUMEX-AIM by SU-AI with PUP; 10-Jun-83 21:47 PDT
Date: Fri 10 Jun 83 21:49:02-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
Subject: Guibas appointment
To: ullman@SCORE
cc: jmc@SAIL
Jeff,
I do want to read the Guibas file carefully before committing
myself on his case, and I want the opportunity to go over it
with John (McCarthy) before deciding.
Bruce Buchanan told me that there were few of the senior faculty
present at the meeting that voted on the Guibas case. I'm sorry
that Iwas not there. But I think that perhaps this important
decision deserves a better attended meeting, in which the issue
can be more broadly and deeply discussed.
Ed
-------
∂12-Jun-83 0222 @SU-SCORE.ARPA:unmvax!hector@Berkeley Cprolog for VAX's
Received: from SU-SCORE by SU-AI with TCP/SMTP; 12 Jun 83 02:22:36 PDT
Received: from UCBVAX.ARPA (UCB-VAX.ARPA) by SU-SCORE.ARPA with TCP; Sun 12 Jun 83 02:17:47-PDT
Date: 12 Jun 83 02:09:34 PDT (Sun)
From: unmvax!hector@Berkeley
Subject: Cprolog for VAX's
Message-Id: <8306120909.AA00266@UCBVAX.ARPA>
Received: by UCBVAX.ARPA (3.341/3.31)
id AA00266; 12 Jun 83 02:09:34 PDT (Sun)
To: PROLOG-HACKER@SCORE, periera@sri-ai.ARPA
Cc: hector@Berkeley, luger@Berkeley
I am sending this to two addresses, because I have had little luck with
the arpa-link at ucbvax. Maybe one will make it.
We have version 1.1 of CProlog here, and I have discovered a strange
occurance. When I have been tracing, and come to a construction like:
X = Y, where X is uninstantiated, and Y is, a strange thing happens.
It is like I turned debug off. Here is a larger example:
CProlog version 1.1
[ restoring file /u1/hector/.plinit ]
yes
| ?- [test].
test consulted 420 bytes 0.066667 sec.
yes
| ?- trace,one(yes).
(1) 1 exit: trace ?
(2) 1 call: one(yes) ?
(3) 2 call: two(yes,←131148) ?
(3) 2 exit: two(yes,no) ?
(4) 2 call: ←131149=no ?
three
Four.
(2) 1 exit: one(yes) ?
yes
| ?- ↑D
[ Prolog execution halted ]
Here is the file test:
one(X) :- two(X,Y), Z = Y, three(X,Y,Z), four(Z).
two(yes,no).
three(X,Y,Z) :- write('three'),nl.
four(←) :- write('Four.'),nl.
What is causing this?
Kenneth Ingham
unmvax!hector@ucbvax or Department of Computer Science
University of New Mexico
Albuquerque, NM 87131
It may be more reliable to use US Mail.
∂12-Jun-83 2146 DEK Leo Guibas
I'm excited about his potential appointment, for four reasons: he is
first rate as (1) researcher in graphics, (2) practical system builder,
(3) teacher, (4) entrepreneur. This is a real opportunity for us,
in my opinion. But he has an offer from Berkeley and will almost
surely get one soon from MIT. If we are to get permission to move
on this, we will want to have a strong case to make to the dean;
hence I hope you will be able to look at his papers soon. As far as
I know, the vote is unanimous so far (but don't let that
prejudice your opinion!).
This case, and the case for a new chairman, should not be left
for September or (shudder) October... I think it's time to move
now on both of them, and I'm hoping you will be able to help me and
Jeff lobby for such good causes! (In spite of the fact that you are
probably even busier than I am, especially after an extensive trip.)
∂13-Jun-83 1638 PKANERVA@SUMEX-AIM My dissertation
Received: from by SU-AI with PUP; 13-Jun-83 16:37 PDT
Date: Mon 13 Jun 83 16:34:39-PDT
From: Pentti Kanerva <PKANERVA@SUMEX-AIM>
Subject: My dissertation
To: McCarthy@SAIL
I have left for you to read a copy of my dissertation. I hope to have
completed my degree by the end of the summer quarter. The dissertation
is the only thing left to do. My advisor is Suppes, and the third
reader is Jaakko Hintikka, who is currenty in the Philosophy Department
at Florida State University. You can send me computer mail at Sumex
or call me at work, 7-3269, or at home, 327-8594. I am interested in
having your comments. - Pentti
-------
∂13-Jun-83 2010 @SU-SCORE.ARPA:HFISCHER@USC-ECLB Anybody Trying to stuff Prolog into an 8088?
Received: from SU-SCORE by SU-AI with TCP/SMTP; 13 Jun 83 20:10:43 PDT
Received: from USC-ECLB by SU-SCORE.ARPA with TCP; Mon 13 Jun 83 20:07:12-PDT
Date: 13 Jun 1983 2005-PDT
From: HFISCHER@USC-ECLB
Subject: Anybody Trying to stuff Prolog into an 8088?
To: prolog-hackers at SCORE
Just wondering. Herm Fischer (HFischer@eclb)
-------
∂14-Jun-83 0017 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #10
Received: from SU-SCORE by SU-AI with TCP/SMTP; 14 Jun 83 00:16:43 PDT
Date: Monday, June 13, 1983 5:35AM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #10
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Tuesday, 14 Jun 1983 Volume 1 : Issue 10
Today's Topics:
Query - Using Prefix Operators,
Applications - Objects,
POPLOG Availability
----------------------------------------------------------------------
Date: Sun 12 Jun 83 19:42:35-PDT
From: SHardy@SRI-KL.ARPA
Subject: Help Needed Using Prefix Operators
I'm using Prolog to build a system to help teach first order predicate
calculus. I need help in using operators to define the grammar I
want.
* Does anyone have any suggestions on how I can solve my immediate
problem (outlined below).
* My problem arises because of weaknesses in the tools PROLOG provides
for implementing embedded languages. Does anyone have ideas for
improving matters?
My specific problem
The following is an example of the type of statement my system must be
able to process.
if (exists x: all y: if person(y) then loves(x, y))
then (all y: if person(y) then (exists x:loves(x, y)))
This syntax was chosen because, with suitable operator definitions, it
can be manipulated directly; I.e. the above expression is a Prolog
term. Here is a first cut at the necessary definitions:
:- op(1, fx, [all, exists]).
:- op(950, xfy, :).
:- op(945, fy, if).
:- op(940, xfy, then).
The problems with these modes are:
* They accept too many things as legal, for example 'if if x'. This
doesn't bother me too much.
* They ''mis-parse'' apparently legal things. For example:
if all x:happy(x) then arrived(utopia)
gets parsed as:
(if all x):happy(x) then arrived(utopia).
Of course, I could get this particular expression parsed right by
lowering the precedence of ':' to below that of 'if' and 'then', but
then expressions like the following would cause problems instead:
all x:if human(x) then mortal(x)
I don't suppose I'm the first person to face this particular problem.
Does anyone have a simple solution for me? Currently, I'm thinking of
writing a conversion program that takes input from the user and then
re-arranges it using transformation rules such as:
(if all X):P then Q ---> if (all X:P) then Q
Has anyone tried this approach?
Why not use DCG's
While Prolog is a fantastic language for writing symbolic manipulation
programs, the two tools for defining new grammars have serious
weaknesses. This is a great pity because domain specific grammars
often make programs much much clearer. In my case, for example, the
rules for conversion to conjunctive normal form are very clear:
not(all X:P) ---> exists X:P.
However, using OP to define new grammars is too complicated. With
over ten types of operator it's hard just to see the implications of a
set of defintions. I have to use DISPLAY to find out how something
has been parsed, then edit the OP defintions and then use DISPLAY
again. Anyhow, operator precedence languages are not that powerful -
for example, you couldn't use them to define the syntax of Prolog
lists.
On the other hand, definite clause grammars are not integrated into
Prolog. You can't define a grammar using DCGs and then use it to help
write your Prolog program. Getting syntax error messages is hard.
You have to duplicate a lot of code - such as the lexical analyser and
operator precedence parser. In fact, DCGs have more of the flavor of
a user-group library routine than of a carefully thought out feature
of the language.
Specifically I'd like to see two things:
(1) The ability to make simple extensions to the syntax of Prolog
itself beyond the capabilities provided by operators. For my
program now, I'd rather not be hacking precedences, or writing
lexical analysers, I'd just like to be able to write things like:
newsyntax all X:P.
newsyntax if P then Q.
(2) I'd like the internals of the Prolog system to be more accessible,
especially the routines for reading and writing. That way, DCGs
could be a one page library procedure and I could write my own
version if I didn't like the one supplied. I might also be able
to write "newsyntax.pl" as a new user group library routine!!
-- Steve Hardy,
Teknowledge
------------------------------
Date: Tuesday, 7 Jun 1983 19:20-PDT
From: Narain at Rand-Unix
Subject: Reply to Abbott
"p" is equivalent to "q" means:
1) When a call to "p" is made, all clauses under "p" as well as
under "q" must be tried.
2) When a call to "q" is made, all clauses under "q" as well as
under "p" must be tried.
The first condition can be expressed by writing:
p(X):-q(X).
If the second condition is expressed as:
q(X):-p(X)
an infinite loop will result.
Instead this last clause is modified, so that whenever a call to "q"
is made, bodies of all clauses under "p" will be tried except the body
of the clause p(X):-q(X).
This is done by writing:
q(X):-clause(p(X),Body),not Body=q(X),Body.
Instead of computing which clauses to use at run-time, one may, at
compile-time assert new clauses under "p" each of which has the body
of a distinct clause under "q". Similarly new clauses are asserted
under "q".
-- Sanjai
------------------------------
Date: Sun 12 Jun 83 19:44:10-PDT
From: SHardy@SRI-KL.ARPA
Subject: Prolog For The Vax
Implementation For VAX/VMS
The Sussex Poplog system is a multi-language programming environment
for AI tasks. It includes:
(a) A native mode Prolog compiler, compatible with the Clocksin and
Mellish book. The system supports floating point arithmetic.
(b) A POP-11 compiler. POP-11 and Prolog programs may share data
structures and may call each other as subroutines; they may also
co-routine with each other. (POP is the British derivative of
LISP; functionally equivalent to Lisp, it has a more conventional
syntax.)
(c) VED, an Emacs like extendible editor, is part of the run time
system. VED is written in POP-11 and so can easily be extended.
It can also be used for input (e.g. simple menus) and for output
(simple cellular graphics). VED and the compilers share memory,
making for a well integrated programming environment.
(d) Subroutines written in other languages, e.g. Fortran, may be
linked in as new built in predicates.
Prolog's complex architecture was designed to help build blackboard
systems working on large amounts of numerical data. The intention is
that Fortran (or a similar language) be used for array processing;
POP-11 will be used for manipulating agendas and other procedurally
oriented tasks and Prolog will be used for logical inference.
However, the components of Prolog can be used individually without
knowledge of the other components. To some users, Poplog is simply a
powerful text editor, to others it just a Prolog system.
Poplog has been adopted, along with Franz LISP and DEC-20 Prolog, as
part of the "common software base" for the IKBS program (Britain's
response to The Fifth Generation).
The system is being transported to the PERQ and Motorola 68000, as
well as being converted for VAX/UNIX.
Although full details haven't yet been announced, the system will be
commercially supported. The license fee will be approx $10,000 with
maintenance approx. $1,000 per annum. For more details, write to:
Dr Aaron Sloman
Cognitive Studies Programme
University of Sussex
Falmer, Brighton, ENGLAND
(273) 606755
-- Steve Hardy,
Teknowledge
------------------------------
End of PROLOG Digest
********************
∂14-Jun-83 1030 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA Informal Session Dedicated to Lerner
Received: from SU-SCORE by SU-AI with TCP/SMTP; 14 Jun 83 10:30:36 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 14 Jun 83 10:31:56-PDT
Date: 14 Jun 1983 1014-PDT
From: Nilsson at SRI-AI
Subject: Informal Session Dedicated to Lerner
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-MC,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL
Jack Minker has requested that AAAI hold a special session at
AAAI-83 to honor Soviet refuse-nik Alexander Lerner's 70th birthday.
We cannot afford a special session at the conference proper, but
Mike Genesereth (the program chairman) suggested that Jack Minker
could organize an informal session on Tuesday evening. This
informal session could be announced in the program and dedicated to
Lerner. Presumably, Jack Minker would make a short announcement of
that fact during the session. The program committee and I
recommend that this be done. I'd be interested to know if there
are strong objections from the excom (no need to respond if you
concur). -Nils
-------
∂14-Jun-83 1300 @SU-SCORE.ARPA,@MIT-MC:Hewitt%MIT-OZ@MIT-MC please change my name on this list
Received: from SU-SCORE by SU-AI with TCP/SMTP; 14 Jun 83 12:59:57 PDT
Received: from MIT-MC by SU-SCORE.ARPA with TCP; Tue 14 Jun 83 12:53:42-PDT
Date: Tuesday, 14 June 1983, 15:51-EDT
From: Carl Hewitt <Hewitt%MIT-OZ@MIT-MC>
Subject: please change my name on this list
To: prolog-hackers@SU-SCORE
Cc: Hewitt%MIT-OZ@MIT-MC
Please change my name on this list to
HEWITT-DIGEST at MIT-MC
Thanks,
Carl
∂14-Jun-83 1315 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA Proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 14 Jun 83 13:15:25 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 14 Jun 83 13:16:16-PDT
Date: Tue 14 Jun 83 11:59:08-PDT
From: Gio <Wiederhold%SRI-AI@SU-SCORE.ARPA>
Subject: Proposal
To: bscott@SU-SCORE.ARPA
cc: pickering%SRI-AI@SU-SCORE.ARPA, jmc@SU-AI.ARPA, dcl@SU-AI.ARPA, dek@SU-AI.ARPA,
zm@SU-AI.ARPA, tob@SU-AI.ARPA
I assume you got notice about ohlanders phone call.
I hope we can MSG the umbrella out today, I dont think any further
changes can be made.
If you want me to merge the budget I can do that, send me a message,
but I believe that the SCRIBE copy can just be attached.
Gio
-------
∂14-Jun-83 1410 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA ExCom Draft Agenda
Received: from SU-SCORE by SU-AI with TCP/SMTP; 14 Jun 83 14:09:45 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 14 Jun 83 14:11:18-PDT
Date: 14 Jun 1983 1319-PDT
From: Nilsson at SRI-AI
Subject: ExCom Draft Agenda
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-MC,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL
Over the next few weeks I'll be collecting agenda items for the AAAI
ExCom Meeting to be held in conjunction with our annual conference
in August in Washington, D.C. Here is a start:
DRAFT
AAAI EXCOM AGENDA TOPICS
Wednesday Evening 7:00 pm, August 24, 1983
Dinner Meeting
(Place to be Announced)
[Any additional agenda items should be sent to Nilsson@sri-ai]
Approval of Minutes
Staff Reports and Discussion
Financial -- Raj Reddy and Don Walker
Conference Local Arrangements
Corporate Affiliates Program
AAAI Office Procedures and Capabilities
. Committee Reports and Discussion
Financial (see above)
Personnel -- Nils Nilsson
Elections -- Marvin Minsky
Program -- Mike Genesereth
Publications -- Lee Erman
Membership -- Bruce Buchanan
Tutorials -- Charles Rich
Relations with IJCAI -- Saul Amarel
Future Conference Sites -- Marty Tenenbaum
Possible Expansion of AAAI Interests and Activities
Industrial and Business -- Rick Hayes-Roth
Educational -- Claudia Mazetti
On-Line Abstracts and Reports -- Michael Genesereth
Other?
Perspectives from the New President -- John McCarthy
Other Business
Installation of New President and Officers
-------
∂14-Jun-83 1749 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA Draft revised umbrella proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 14 Jun 83 17:49:15 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 14 Jun 83 17:48:29-PDT
Date: Tue 14 Jun 83 17:42:22-PDT
From: Gio <Wiederhold%SRI-AI@SU-SCORE.ARPA>
Subject: Draft revised umbrella proposal
To: Ohlander%USC-ISI@SU-SCORE.ARPA, Mach%USC-ECLB@SU-SCORE.ARPA
cc: jmc@SU-AI.ARPA, zm@SU-AI.ARPA, dek@SU-AI.ARPA, tob@SU-AI.ARPA, dcl@SU-AI.ARPA,
wiederhold%SUMEX-AIM@SU-SCORE.ARPA, oliger@SU-SCORE.ARPA, bscott@SU-SCORE.ARPA
Proposal to
The Defense Advanced Research Projects Agency
for
ADVANCED RESEARCH AND DEVELOPMENT
IN ALGORITHMIC AND ARTIFICIAL INTELLIGENCE METHODS
IN COMPUTER SCIENCE
April 1983 (revised June 1983)
Departments of Computer Science and Electrical Engineering
Schools of Humanities and Sciences, and Engineering
Stanford University
!
Contents:
1 Background
2 Scope
3 Statement of Work
4 Equipment Transfer
5 Technical Deliverables
6 Design to Cost
7 Licensing
8 Budget
!
Background.
The faculty of the Computer Science Department with the help of
associated faculty in the Electrical Engineering Department proposes
to continue and initiate research in promising areas of computer science.
We are concerned with making the computer as a tool as powerful and as
easy to use as we can.
We want to increase the simplicity of telling the computer what we
want it to do; we are also concerned with the effective execution of the
actions that are requested.
Our proposal will replenish and enhance a stock of ideas that everyone
who deals with computers must draw from and apply these ideas to
problems common to the community in ways useful to DARPA/NAVALEX.
Computer Science faculty, research associates, and students at
Stanford represent an unusual intellectual resource.
Specifically the areas of Artificial Intelligence, analysis of algorithms,
database, and program analysis are highly developed.
The expertise of the individual investigators is documented by the vitae
attached to this proposal.
Outline.
In the remainder of this section we highlight the
background we have in specific areas and indicate the problems where the
boundaries of computer science can be expanded.
Detailed progress reports for the several areas are appeneded to this
document.
They provide some specific examples of problems being addressed within
the scope of the proposal and background and qualification information
to demonstrate capabilities and research record.
Section 2 outlines the general scope of the proposal.
Section 3 contains the formal Statement of Work.
Sections 4, 5, 6, and 7 deal with equipment transfer, deliverables, design
to cost, and licensing.
A budget to delimit the resources required for this research in presented
in Section 8
We expect this proposal to become further refined through
individual tasking statements to be attached to this proposal.
Formal Reasoning.
Stanford has contributed greatly to the development of former reasoning.
Formal Reasoning is a methodology for studying the basic modes of thought,
decision-making, planning, and reasoning that people use in both their everyday
lives and in their roles as computer programmers. This methodology makes
explicit and simple the objects that fill the world as well as the
properties they have and the relations between them. The representation
used is either first-order predicate calculus or one of the well-studied
variants of it. In this way the issues of belief, knowledge, ignorance,
and epistemology are at the surface of attention rather than buried beneath
{ad hoc} representations. Since these logical languages have a well-known
semantics there is never any doubt about meaning.
These techniques are applied to artifical intelligence problems, to the
study of the mathematical theory of computation, to specialized automatic
programming, and to programming languages.
Lisp.
LISP was invented by faculty at Stanford.
We are collaborating in further development of the language
and its support systems.
LISP has been and continues to be the primary language for
Artificial Intelligence research. Increasingly it is being used as a
system language as well. Operating systems and compilers are being
routinely written. Since the semantics of the language are not
operational semantics neither the tactics nor the strategy of
implementation for most facets are predetermined. Diverse runtime
systems adapt less programs to the requirements covered by the
environment. For instance,
there is
no statement of the requirement for either deep-binding or shallow-binding
for variables. Therefore there is much room for experimentation with the
most effective strategies and tactics for implementation of Lisp given
specific machines and applications.
Being a programming system as well as simply a programming language,
a programming environment is necessary for the success and usefulness
of a Lisp implementation. A portable Lisp implementation requires a
portable Lisp environment.
Because Common Lisp has been widely accepted by the Lisp and
artificial intellegence communities, it is the most attractive Lisp dialect
for this portability.
On-going work at Stanford includes not only portable Lisp implentations
but also a portable runtime environment based on a rich dialect of
Lisp, the common Lisp version being developed at Stanford and other
sites is the most attractive choice. The variety of hardware
architectures available requires that the runtime systems support
these architectures. Work at Stanford is directed having portable
Lisp run on personal computers, work stations, supercomputers, and
multiprocessors.
Software for High Performance Machines.
Principal investigators in this proposal have participated in the research and
development of software for the high performance multiprocessor, the S-1
system
at Lawrence Livermore Laboratories. Such work has included a basic
operating system, a language translation system including as modules
several compilers, optimizers, code-generators, an interpreter, and linker
and loaders. A portable file system, developed initially under joint
sponsorship, is intended to help manage the large data quantities expected
when systems as the S-1 reach full operation. Important research results
of this work include experience in developing software which is more
modular than previous systems, so that improvements in software technology
can be more rapidly moved to evolving computer systems. High-performance,
low quantity systems present important issues of software design of which
we will be conscious in our research plans.
We are considering applications and problems that
require extreme processing power.
High-performance and multiprocessed computers present special problems in
terms of software, especially operating systems. These problems arise
since the ability to execute instructions rapidly often requires a
non-uniform architecture (special registers, vector representations, etc),
a complex instruction set (to take advantage of internal
parallelism---increasing the work per instruction), complex addressing
modes, or non-uniform memory hierarchies.
Beyond this any multiprocessing is an added problem.
Database Management and the Use of Knowledge.
On-going research at Stanford has demonstrated the effectiveness of
applied advanced heuristic techniques to the management of large
quantities of data. Models and application of such techniques have
been published and are finding increased use in other research
organizations and in industry.
The increased size of computers requires that larger amounts of data
are to be manipulated. Traditional algorithms require greater than
linear and often exponential computing resources for data analysis of
large volumes of data. Techniques developed at Stanford in the
design storage and retrieval of information demonstrates the success
of heristic techniques where algorithmic approaches become
unmanageable. The techniques which have been developed address both
centralized and distributed databases. The architecture of computers
that are designed for these applications is not limited to standard
von Neumann machines.
The specific published accomplishments include a methodology for the
design of user models and the integration of user models to design
large computer database management systems and the development of physical
design algorithms for centralized and distributed databases,
Comprehensive techniques for concurrency control
and recovery in distributed databases have been developed, evaluated,
and published.
We have also developed, tested and published methods
for the update of data in situations where the user does not have
access to the full database, querying and cooperative response systems
for natural language interfaces with databases, new methods to apply
database technology to the design of large VLSI databases,
optimization of database queries based on semantics and the results of
previous queries using a new technique called "Query Graphs", new
algorithms for effective processing of complex queries using the
parllel capabilities possible with VLSI technology. In partial support
of this work we have also developed file access systems techniques
which have been widely distributed.
Deductive Programming.
Stanford faculty has contributed to program synthesis since 1968. An
example of earlier work is the DAEDALUS programming synthesis system
which applied a transformational approach under which the program was
constructed by applying transformationals directly to the program
specification.
Subsequently a more powerful deductive approach has been
developed and an automatic version of that system has been implemented.
An interactive version is also under develoment. Work under way gives
a deductive tablau formalism the abilities needed to deal with commonly
used relations, including equality and ordering. These
abilities result in shorter derivations and a dramatically reduced
search space in the synthesis of programs that rely on these special
relations.
The deductive approach has been applied to the synthesis of several
non-trivial programs. Example - a detailed synthesis of a version of
Robinson's unification algoithm was devised. Portions of that
derivation have since been reduced by interactive systems of
Erickson and Paulson.
Advanced Programming Environments.
The program analysis and verification group at Stanford has designed
an implemented PASCAL verifier. This system is one of the most
powerful tools for automated and interactivated forms of verification
of annotated PASCAL programs. The theory and prover component of the
verifier incoporates novel technology for efficient decision
procedures. It has been used extensively as a teaching tool and in
comprehensive case studies including the verification of a PASCAL
compiler itself. It has been widely distributed to research
institutions. Recent work has concentrated on language design and
implementation, the study of distributed and concurrent systems, and
the design of tools for advanced programming environments. Recent
accomplishments in these areas include the following.
o ADAM compiler.
This compiler supports a very large subset of
Ada inclusing task. ADA was demonstrated at the ADA Symposium in
Boston, December 1980 and is currently being updated to support the
Ada 82 redesign and is being distributed to commercial and academic
organizations by the Stanford Software Distribution Center. Reports,
user manuals, and publications are available.
o ADA Task Supervisor Pack.
ADA runtime task supervisor packages
have been designed and implemented in ADA and compiled using ADAM.
The purpose is to develop a standard package designed for ADA support.
As a consequence a standard interface between compiler and runtime
supervisors is defined facilitating portability of compilers. Past
supervisor packages are bing distributed and reports on supervisor
design and implementation are available.
o ANNA Language Design.
ANNA is a specification and annotation language for ADA programs.
It introduces features for formally
specifying the intended behavior of ADA programs or fragments of
programs.
ANNA is a very general formal specification language and is available
to other ALGOL-like languages such as PASCAL, PL1, concurrent PASCAL
and MODULA.
A preliminary ANNA reference manual is available and has been widely
distributed for comment.
o ADA Display-oriented Structure Based Editor.
A display-oriented base editor has been designed and implemented.
A version supporting ADA is available, provides a central facility for
automated teaching courses for programming languages such as ADA.
A user manual is available.
o Runtime Monitor for Deadness Errors in
Parallel Processes.
Deadness errors can occur in parallel programs from communication
failures between the threads of control.
We have defined some basic principles for runtime detection of the class of deadness errors.
We have applied this principle in experimental implementation of a runtime
monitor for ADA tasking errors.
Reports on runtime monitoring are available.
The Analysis of Algorithms.
Stanford faculty has developed and contributed greatly to the field of
algorithms.
Improved algorithms are being developed which will have significant
effect on implementations in hardware.
Previous work has been widely published.
Current work in progress includes simplified methods for diplaying graphics
on raster displays and new methods of cryptanalysis.
Important practical results from this research are the TEX typesetting
system, now in wide use, and--more recently--the WEB system.
The WEB system is a programming tool combines documentations and
programs into one integrated whole.
Presentations on this approach have been given and we expect that
further deomonstrations will convince a substantial number of people
outside Stanford to use this method.
Algorithms for improved computations of large numerical programs have
been developed and their effectiveness has been proven.
This work is already finding transfer into machines with systolic
parallel architectures being developed by local industry.
Image Understanding and Robotics.
Research in the Stanford Robotics Laboratory has centered on the
ACRONYM system, an intelligent knowledge based system for image understanding
and planning.
ACRONYMS's performance depends on domain independent capabilities.
It incoporates geometric modelling with rule-base geometric reasoning.
Recognition of aircraft has been demonstrated in ACRONYM.
ACRONYM has been distributed to about ten research groups.
Recent efforts are going toward successor.
This system is being extended to include negative volumes, multiple
instances of objects, and a general name scoping mechanism.
A large set of aircraft models have been built.
Problems with the rule set for recognition are being addressed, a
constraint resolution system has been tested.
Work is jointly performed with MIT in order to increase productivity
of our resources.
The scope of research and development at Stanford illustrated by the
accomplishments listed above demonstrate the capability of Stanford to
lead the way in further research in the
development of algorithms and artificial intelligence techniques,
large scale scientific computation,
and database processing problems with
an emphasis on the human interaction with computing systems.
As part of the research performed at Stanford many tools are being
developed which are widely distributed over the ARPAnet and to industry.
Dissemination of the research and development at Stanford follows
many paths.
More than xxx technical reports are produced annually by
the Stanford Computer Science Department.
Many papers are published every year in the scientific literature and
presentations are given at conferences of specialists and to
applications-oriented groups.
Our graduates, both Masters and Ph.D. students, are one of the most
important means of research dissemination.
Our students are found throughout the United States in leading
universities and research laboratories.
They contribute greatly to the advanced state we enjoy in this highly
technical field.
!
2 Scope.
Stanford University is proposing a program of research and development in
the following areas of computer science:
formal reasoning, image understanding and robotics, databases,
deductive programming, program analysis and verification, and analysis
of algorithms.
In all these areas the common focus is improved performance of
computer systems.
Improved performance ranges from improved interfaces, improved
capability to implement, improved algorithms to carry out the
computation, and improved execution and control of these algorithms on
new distributed and parallel hardware architectures.
There will be demonstrations of these techniques applied to realistic
and large scale problems.
Specifically proposed tasks will:
o apply formal reasoning techniques to artificial
intelligence problems, to the mathematical theory of computation, to
specialized automatic programming, and to programming languages.
o develop Common LISP and its routine system, to
provide methods and documentation to make Common Lisp available to
computers, work stations, and multiprocessors, to develop software
tools, to aid in the effective use of complex super-computers and
large networks of computers.
o investigate, develop, and demonstrate applications of
advanced techniques from the algorithmic and artificial intelligence
domain to the management of large databases.
o develop, analyze, and demonstrate deductive techniques in
program synthesis.
o develop tasking and verification techniques for
programs in an ADA environment.
o develop algorithms which improve the performance of graphics
interfaces.
o develop algorithms which improve cryptanalysis.
o develop and demonstrate the effectiveness of the WEB system
of integrated programming and documentation.
o develop algorithms for improved performance
numerical algorithms in parallel and VLSI processors.
o perform research and develop systems extending the
concepts of generalized cylinders in the image understanding and robotics area.
This work will be disseminated through reports, scientific papers,
demonstrations of the software techniques, and participation in scientific
symposia.
!
3 Statement of Work.
Stanford University proposes to conduct a broad program of research in the
field of computer science.
The specific research will include: fundamentals of artificial intelligence;
advanced programming techniques and environments; data management; image understanding and robotics; analysis of algorithms; and
relations between algorithms and architectures.
The following specific task areas are proposed:
1} Basic Research in Artificial Intelligence.
Stanford University proposes to conduct research in aspects of basic artificial
intelligence technology with the objectives of making significant advances in
machine reasoning capabilities, automatic speedup of programs, and artificial
intelligence language design and implementation. Examples of the kinds of
tasks that will be undertaken in this area include: development of
formalizations of common sense reasoning and about facts and knowledge used in
the common sense world; development of languages for communications among
computers and between computers and people; development of systems for computer
reasoning and computer-assisted human reasoning and systems that take advice
from users and give advice in return; improvement of artificial intelligence
languages and also performance of research aimed at new languages; and
experimentation with techniques of automatically specializing programs to make
them run faster.
2} Research in Advanced Programming Techniques and Environments.
Stanford University also proposes to carry out research in the development of
advanced programming environments and software production with the goals of
providing much better programming tools than currently exist. Examples of the
kinds of tasks that will be undertaken in this area include: design and
implementation of new high level languages for formulation of systems
requirements, design specifications, formal annotation of implementations, and
documentation; design and implementation of an advanced programming environment
that supports software production in these languages; development of techniques
and tools for analyzing the parallel activity in systems and for verifying and
synthesizing concurrent programs; study of the formalization and codification
of programming knowledge and introduce machine reasoning techniques for
software production applications; design of a high-level logic programming
language using new deductive techniques; development of tools for rigorously
establishing properties of software and hardware systems (using logic of
programming such as dynamic logic and temporal logic); and investigation of
program manipulation techniques in such areas as maintenance, transformation,
and optimization.
3} Data Management.
Stanford University also proposes to carry out a research project in data
management with a goal of improving data management capabilities by the
development of better algorithms and the incorporation of machine intelligence.
Examples of the kinds of tasks that will be undertaken in this area include:
investigation of advanced techniques to improve database access, storage,
update, and management; development of new techniques for the use of databases
in VLSI design; investigation of techniques for maintaining reliability and
integrity in distributed data base systems; and exploration of methods of
inferring new knowledge from data.
4} Image Understanding and Robotics.
Stanford University also proposes to carry out research in image understanding
and robotics with the goals of achieving automatic understanding of images and
sensing and intelligent planning by robotic devices. Examples of the kinds of
tasks that will be undertaken in this area include: analysis, design and
implementation of intelligent systems for interpretation and planning actions;
study of planning, navigation and path-finding, motion control map-making and
world modeling for mobile robots; study, design and development of mechanisms
for interpretation of images; and investigation of architecture of image
algorithms and their implementation in VLSI.
5} Analysis of Algorithms.
Stanford University also proposes to conduct research in the analysis of
algorithms with the objective of achieving dramatic speedups in certain classes
of algorithms. Examples of the kinds of tasks that will be undertaken in this
area include: development of new computer algorithms for a class of practical
problems in order to explore general issues of efficiency of algorithms;
development of the mathematics required to determine efficiency of algorithms;
and extension of programming methodologies so that such algorithms can be
implemented more quickly and reliably than with present techniques.
6} Relations between Algorithms and Architectures.
Stanford University also proposes to carry out research in examining the
relations between algorithms and architectures in order to gain an appreciation
of the processing power required for certain classes of algorithms. Examples
of the kinds of tasks that will be undertaken in this area include;
determination of lower bounds on the optimal ratio of processor power to
communication capability for representative problems; relation of algorithmic
requirements in terms of topology, data rates, and processor capability to
physical and technological restrictions; investigation of inherent tradeoffs in
convergence rates and complexity with concurrency; and investigation of the
possibility of a metalanguage for the description of application problems in
terms of standard computational processes which express both the complexity of
the computation and its topology and which allows accurate simulation of the
process for various architectures.
!
4 Equipment Transfer.
<<< provided by Betty Scott >>>
5 Technical Deliverables.
Technical reports as defined in the individual task descriptions
and quarterly research and development reports will be delivered to DARPA
and Navalex.
6 Design to Cost.
Design to cost does not apply because this is a research effort.
7 Licensing.
Stanford University will provide the required no-cost licensing
agreements to DOD contractors to use resulting software for work to be
performed under DOD contracts. Stanford University will provide the
right to copy all printed material as required by DOD and DOD contractors.
8 Budget.
<<< to be provided by Betty Scott >>>
-------
∂14-Jun-83 1815 @SU-SCORE.ARPA:WIEDERHOLD@SRI-AI.ARPA Draft revised umbrella proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 14 Jun 83 18:15:15 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 14 Jun 83 18:14:55-PDT
Date: Tue 14 Jun 83 17:42:22-PDT
From: Gio <Wiederhold%SRI-AI@SU-SCORE.ARPA>
Subject: Draft revised umbrella proposal
To: Ohlander%USC-ISI@SU-SCORE.ARPA, Mach%USC-ECLB@SU-SCORE.ARPA
cc: jmc@SU-AI.ARPA, zm@SU-AI.ARPA, dek@SU-AI.ARPA, tob@SU-AI.ARPA, dcl@SU-AI.ARPA,
wiederhold%SUMEX-AIM@SU-SCORE.ARPA, oliger@SU-SCORE.ARPA, bscott@SU-SCORE.ARPA
Proposal to
The Defense Advanced Research Projects Agency
for
ADVANCED RESEARCH AND DEVELOPMENT
IN ALGORITHMIC AND ARTIFICIAL INTELLIGENCE METHODS
IN COMPUTER SCIENCE
April 1983 (revised June 1983)
Departments of Computer Science and Electrical Engineering
Schools of Humanities and Sciences, and Engineering
Stanford University
!
Contents:
1 Background
2 Scope
3 Statement of Work
4 Equipment Transfer
5 Technical Deliverables
6 Design to Cost
7 Licensing
8 Budget
!
Background.
The faculty of the Computer Science Department with the help of
associated faculty in the Electrical Engineering Department proposes
to continue and initiate research in promising areas of computer science.
We are concerned with making the computer as a tool as powerful and as
easy to use as we can.
We want to increase the simplicity of telling the computer what we
want it to do; we are also concerned with the effective execution of the
actions that are requested.
Our proposal will replenish and enhance a stock of ideas that everyone
who deals with computers must draw from and apply these ideas to
problems common to the community in ways useful to DARPA/NAVALEX.
Computer Science faculty, research associates, and students at
Stanford represent an unusual intellectual resource.
Specifically the areas of Artificial Intelligence, analysis of algorithms,
database, and program analysis are highly developed.
The expertise of the individual investigators is documented by the vitae
attached to this proposal.
Outline.
In the remainder of this section we highlight the
background we have in specific areas and indicate the problems where the
boundaries of computer science can be expanded.
Detailed progress reports for the several areas are appeneded to this
document.
They provide some specific examples of problems being addressed within
the scope of the proposal and background and qualification information
to demonstrate capabilities and research record.
Section 2 outlines the general scope of the proposal.
Section 3 contains the formal Statement of Work.
Sections 4, 5, 6, and 7 deal with equipment transfer, deliverables, design
to cost, and licensing.
A budget to delimit the resources required for this research in presented
in Section 8
We expect this proposal to become further refined through
individual tasking statements to be attached to this proposal.
Formal Reasoning.
Stanford has contributed greatly to the development of former reasoning.
Formal Reasoning is a methodology for studying the basic modes of thought,
decision-making, planning, and reasoning that people use in both their everyday
lives and in their roles as computer programmers. This methodology makes
explicit and simple the objects that fill the world as well as the
properties they have and the relations between them. The representation
used is either first-order predicate calculus or one of the well-studied
variants of it. In this way the issues of belief, knowledge, ignorance,
and epistemology are at the surface of attention rather than buried beneath
{ad hoc} representations. Since these logical languages have a well-known
semantics there is never any doubt about meaning.
These techniques are applied to artifical intelligence problems, to the
study of the mathematical theory of computation, to specialized automatic
programming, and to programming languages.
Lisp.
LISP was invented by faculty at Stanford.
We are collaborating in further development of the language
and its support systems.
LISP has been and continues to be the primary language for
Artificial Intelligence research. Increasingly it is being used as a
system language as well. Operating systems and compilers are being
routinely written. Since the semantics of the language are not
operational semantics neither the tactics nor the strategy of
implementation for most facets are predetermined. Diverse runtime
systems adapt less programs to the requirements covered by the
environment. For instance,
there is
no statement of the requirement for either deep-binding or shallow-binding
for variables. Therefore there is much room for experimentation with the
most effective strategies and tactics for implementation of Lisp given
specific machines and applications.
Being a programming system as well as simply a programming language,
a programming environment is necessary for the success and usefulness
of a Lisp implementation. A portable Lisp implementation requires a
portable Lisp environment.
Because Common Lisp has been widely accepted by the Lisp and
artificial intellegence communities, it is the most attractive Lisp dialect
for this portability.
On-going work at Stanford includes not only portable Lisp implentations
but also a portable runtime environment based on a rich dialect of
Lisp, the common Lisp version being developed at Stanford and other
sites is the most attractive choice. The variety of hardware
architectures available requires that the runtime systems support
these architectures. Work at Stanford is directed having portable
Lisp run on personal computers, work stations, supercomputers, and
multiprocessors.
Software for High Performance Machines.
Principal investigators in this proposal have participated in the research and
development of software for the high performance multiprocessor, the S-1
system
at Lawrence Livermore Laboratories. Such work has included a basic
operating system, a language translation system including as modules
several compilers, optimizers, code-generators, an interpreter, and linker
and loaders. A portable file system, developed initially under joint
sponsorship, is intended to help manage the large data quantities expected
when systems as the S-1 reach full operation. Important research results
of this work include experience in developing software which is more
modular than previous systems, so that improvements in software technology
can be more rapidly moved to evolving computer systems. High-performance,
low quantity systems present important issues of software design of which
we will be conscious in our research plans.
We are considering applications and problems that
require extreme processing power.
High-performance and multiprocessed computers present special problems in
terms of software, especially operating systems. These problems arise
since the ability to execute instructions rapidly often requires a
non-uniform architecture (special registers, vector representations, etc),
a complex instruction set (to take advantage of internal
parallelism---increasing the work per instruction), complex addressing
modes, or non-uniform memory hierarchies.
Beyond this any multiprocessing is an added problem.
Database Management and the Use of Knowledge.
On-going research at Stanford has demonstrated the effectiveness of
applied advanced heuristic techniques to the management of large
quantities of data. Models and application of such techniques have
been published and are finding increased use in other research
organizations and in industry.
The increased size of computers requires that larger amounts of data
are to be manipulated. Traditional algorithms require greater than
linear and often exponential computing resources for data analysis of
large volumes of data. Techniques developed at Stanford in the
design storage and retrieval of information demonstrates the success
of heristic techniques where algorithmic approaches become
unmanageable. The techniques which have been developed address both
centralized and distributed databases. The architecture of computers
that are designed for these applications is not limited to standard
von Neumann machines.
The specific published accomplishments include a methodology for the
design of user models and the integration of user models to design
large computer database management systems and the development of physical
design algorithms for centralized and distributed databases,
Comprehensive techniques for concurrency control
and recovery in distributed databases have been developed, evaluated,
and published.
We have also developed, tested and published methods
for the update of data in situations where the user does not have
access to the full database, querying and cooperative response systems
for natural language interfaces with databases, new methods to apply
database technology to the design of large VLSI databases,
optimization of database queries based on semantics and the results of
previous queries using a new technique called "Query Graphs", new
algorithms for effective processing of complex queries using the
parllel capabilities possible with VLSI technology. In partial support
of this work we have also developed file access systems techniques
which have been widely distributed.
Deductive Programming.
Stanford faculty has contributed to program synthesis since 1968. An
example of earlier work is the DAEDALUS programming synthesis system
which applied a transformational approach under which the program was
constructed by applying transformationals directly to the program
specification.
Subsequently a more powerful deductive approach has been
developed and an automatic version of that system has been implemented.
An interactive version is also under develoment. Work under way gives
a deductive tablau formalism the abilities needed to deal with commonly
used relations, including equality and ordering. These
abilities result in shorter derivations and a dramatically reduced
search space in the synthesis of programs that rely on these special
relations.
The deductive approach has been applied to the synthesis of several
non-trivial programs. Example - a detailed synthesis of a version of
Robinson's unification algoithm was devised. Portions of that
derivation have since been reduced by interactive systems of
Erickson and Paulson.
Advanced Programming Environments.
The program analysis and verification group at Stanford has designed
an implemented PASCAL verifier. This system is one of the most
powerful tools for automated and interactivated forms of verification
of annotated PASCAL programs. The theory and prover component of the
verifier incoporates novel technology for efficient decision
procedures. It has been used extensively as a teaching tool and in
comprehensive case studies including the verification of a PASCAL
compiler itself. It has been widely distributed to research
institutions. Recent work has concentrated on language design and
implementation, the study of distributed and concurrent systems, and
the design of tools for advanced programming environments. Recent
accomplishments in these areas include the following.
o ADAM compiler.
This compiler supports a very large subset of
Ada inclusing task. ADA was demonstrated at the ADA Symposium in
Boston, December 1980 and is currently being updated to support the
Ada 82 redesign and is being distributed to commercial and academic
organizations by the Stanford Software Distribution Center. Reports,
user manuals, and publications are available.
o ADA Task Supervisor Pack.
ADA runtime task supervisor packages
have been designed and implemented in ADA and compiled using ADAM.
The purpose is to develop a standard package designed for ADA support.
As a consequence a standard interface between compiler and runtime
supervisors is defined facilitating portability of compilers. Past
supervisor packages are bing distributed and reports on supervisor
design and implementation are available.
o ANNA Language Design.
ANNA is a specification and annotation language for ADA programs.
It introduces features for formally
specifying the intended behavior of ADA programs or fragments of
programs.
ANNA is a very general formal specification language and is available
to other ALGOL-like languages such as PASCAL, PL1, concurrent PASCAL
and MODULA.
A preliminary ANNA reference manual is available and has been widely
distributed for comment.
o ADA Display-oriented Structure Based Editor.
A display-oriented base editor has been designed and implemented.
A version supporting ADA is available, provides a central facility for
automated teaching courses for programming languages such as ADA.
A user manual is available.
o Runtime Monitor for Deadness Errors in
Parallel Processes.
Deadness errors can occur in parallel programs from communication
failures between the threads of control.
We have defined some basic principles for runtime detection of the class of deadness errors.
We have applied this principle in experimental implementation of a runtime
monitor for ADA tasking errors.
Reports on runtime monitoring are available.
The Analysis of Algorithms.
Stanford faculty has developed and contributed greatly to the field of
algorithms.
Improved algorithms are being developed which will have significant
effect on implementations in hardware.
Previous work has been widely published.
Current work in progress includes simplified methods for diplaying graphics
on raster displays and new methods of cryptanalysis.
Important practical results from this research are the TEX typesetting
system, now in wide use, and--more recently--the WEB system.
The WEB system is a programming tool combines documentations and
programs into one integrated whole.
Presentations on this approach have been given and we expect that
further deomonstrations will convince a substantial number of people
outside Stanford to use this method.
Algorithms for improved computations of large numerical programs have
been developed and their effectiveness has been proven.
This work is already finding transfer into machines with systolic
parallel architectures being developed by local industry.
Image Understanding and Robotics.
Research in the Stanford Robotics Laboratory has centered on the
ACRONYM system, an intelligent knowledge based system for image understanding
and planning.
ACRONYMS's performance depends on domain independent capabilities.
It incoporates geometric modelling with rule-base geometric reasoning.
Recognition of aircraft has been demonstrated in ACRONYM.
ACRONYM has been distributed to about ten research groups.
Recent efforts are going toward successor.
This system is being extended to include negative volumes, multiple
instances of objects, and a general name scoping mechanism.
A large set of aircraft models have been built.
Problems with the rule set for recognition are being addressed, a
constraint resolution system has been tested.
Work is jointly performed with MIT in order to increase productivity
of our resources.
The scope of research and development at Stanford illustrated by the
accomplishments listed above demonstrate the capability of Stanford to
lead the way in further research in the
development of algorithms and artificial intelligence techniques,
large scale scientific computation,
and database processing problems with
an emphasis on the human interaction with computing systems.
As part of the research performed at Stanford many tools are being
developed which are widely distributed over the ARPAnet and to industry.
Dissemination of the research and development at Stanford follows
many paths.
More than xxx technical reports are produced annually by
the Stanford Computer Science Department.
Many papers are published every year in the scientific literature and
presentations are given at conferences of specialists and to
applications-oriented groups.
Our graduates, both Masters and Ph.D. students, are one of the most
important means of research dissemination.
Our students are found throughout the United States in leading
universities and research laboratories.
They contribute greatly to the advanced state we enjoy in this highly
technical field.
!
2 Scope.
Stanford University is proposing a program of research and development in
the following areas of computer science:
formal reasoning, image understanding and robotics, databases,
deductive programming, program analysis and verification, and analysis
of algorithms.
In all these areas the common focus is improved performance of
computer systems.
Improved performance ranges from improved interfaces, improved
capability to implement, improved algorithms to carry out the
computation, and improved execution and control of these algorithms on
new distributed and parallel hardware architectures.
There will be demonstrations of these techniques applied to realistic
and large scale problems.
Specifically proposed tasks will:
o apply formal reasoning techniques to artificial
intelligence problems, to the mathematical theory of computation, to
specialized automatic programming, and to programming languages.
o develop Common LISP and its routine system, to
provide methods and documentation to make Common Lisp available to
computers, work stations, and multiprocessors, to develop software
tools, to aid in the effective use of complex super-computers and
large networks of computers.
o investigate, develop, and demonstrate applications of
advanced techniques from the algorithmic and artificial intelligence
domain to the management of large databases.
o develop, analyze, and demonstrate deductive techniques in
program synthesis.
o develop tasking and verification techniques for
programs in an ADA environment.
o develop algorithms which improve the performance of graphics
interfaces.
o develop algorithms which improve cryptanalysis.
o develop and demonstrate the effectiveness of the WEB system
of integrated programming and documentation.
o develop algorithms for improved performance
numerical algorithms in parallel and VLSI processors.
o perform research and develop systems extending the
concepts of generalized cylinders in the image understanding and robotics area.
This work will be disseminated through reports, scientific papers,
demonstrations of the software techniques, and participation in scientific
symposia.
!
3 Statement of Work.
Stanford University proposes to conduct a broad program of research in the
field of computer science.
The specific research will include: fundamentals of artificial intelligence;
advanced programming techniques and environments; data management; image understanding and robotics; analysis of algorithms; and
relations between algorithms and architectures.
The following specific task areas are proposed:
1} Basic Research in Artificial Intelligence.
Stanford University proposes to conduct research in aspects of basic artificial
intelligence technology with the objectives of making significant advances in
machine reasoning capabilities, automatic speedup of programs, and artificial
intelligence language design and implementation. Examples of the kinds of
tasks that will be undertaken in this area include: development of
formalizations of common sense reasoning and about facts and knowledge used in
the common sense world; development of languages for communications among
computers and between computers and people; development of systems for computer
reasoning and computer-assisted human reasoning and systems that take advice
from users and give advice in return; improvement of artificial intelligence
languages and also performance of research aimed at new languages; and
experimentation with techniques of automatically specializing programs to make
them run faster.
2} Research in Advanced Programming Techniques and Environments.
Stanford University also proposes to carry out research in the development of
advanced programming environments and software production with the goals of
providing much better programming tools than currently exist. Examples of the
kinds of tasks that will be undertaken in this area include: design and
implementation of new high level languages for formulation of systems
requirements, design specifications, formal annotation of implementations, and
documentation; design and implementation of an advanced programming environment
that supports software production in these languages; development of techniques
and tools for analyzing the parallel activity in systems and for verifying and
synthesizing concurrent programs; study of the formalization and codification
of programming knowledge and introduce machine reasoning techniques for
software production applications; design of a high-level logic programming
language using new deductive techniques; development of tools for rigorously
establishing properties of software and hardware systems (using logic of
programming such as dynamic logic and temporal logic); and investigation of
program manipulation techniques in such areas as maintenance, transformation,
and optimization.
3} Data Management.
Stanford University also proposes to carry out a research project in data
management with a goal of improving data management capabilities by the
development of better algorithms and the incorporation of machine intelligence.
Examples of the kinds of tasks that will be undertaken in this area include:
investigation of advanced techniques to improve database access, storage,
update, and management; development of new techniques for the use of databases
in VLSI design; investigation of techniques for maintaining reliability and
integrity in distributed data base systems; and exploration of methods of
inferring new knowledge from data.
4} Image Understanding and Robotics.
Stanford University also proposes to carry out research in image understanding
and robotics with the goals of achieving automatic understanding of images and
sensing and intelligent planning by robotic devices. Examples of the kinds of
tasks that will be undertaken in this area include: analysis, design and
implementation of intelligent systems for interpretation and planning actions;
study of planning, navigation and path-finding, motion control map-making and
world modeling for mobile robots; study, design and development of mechanisms
for interpretation of images; and investigation of architecture of image
algorithms and their implementation in VLSI.
5} Analysis of Algorithms.
Stanford University also proposes to conduct research in the analysis of
algorithms with the objective of achieving dramatic speedups in certain classes
of algorithms. Examples of the kinds of tasks that will be undertaken in this
area include: development of new computer algorithms for a class of practical
problems in order to explore general issues of efficiency of algorithms;
development of the mathematics required to determine efficiency of algorithms;
and extension of programming methodologies so that such algorithms can be
implemented more quickly and reliably than with present techniques.
6} Relations between Algorithms and Architectures.
Stanford University also proposes to carry out research in examining the
relations between algorithms and architectures in order to gain an appreciation
of the processing power required for certain classes of algorithms. Examples
of the kinds of tasks that will be undertaken in this area include;
determination of lower bounds on the optimal ratio of processor power to
communication capability for representative problems; relation of algorithmic
requirements in terms of topology, data rates, and processor capability to
physical and technological restrictions; investigation of inherent tradeoffs in
convergence rates and complexity with concurrency; and investigation of the
possibility of a metalanguage for the description of application problems in
terms of standard computational processes which express both the complexity of
the computation and its topology and which allows accurate simulation of the
process for various architectures.
!
4 Equipment Transfer.
<<< provided by Betty Scott >>>
5 Technical Deliverables.
Technical reports as defined in the individual task descriptions
and quarterly research and development reports will be delivered to DARPA
and Navalex.
6 Design to Cost.
Design to cost does not apply because this is a research effort.
7 Licensing.
Stanford University will provide the required no-cost licensing
agreements to DOD contractors to use resulting software for work to be
performed under DOD contracts. Stanford University will provide the
right to copy all printed material as required by DOD and DOD contractors.
8 Budget.
<<< to be provided by Betty Scott >>>
-------
∂15-Jun-83 0828 DFH AI Qual
To: "@AIQUAL.[1,DFH]"@SU-AI
From: Diana Hall (DFH@SU-AI)
Subject: AI Qual
This is a reminder to all concerned that the AI Qual is scheduled for Tues.,
June 21 starting at 1 pm in room 352. I have Stuart Russell scheduled at
1 pm, Dave Chelberg at around 3 pm.
∂15-Jun-83 1338 BSCOTT@SU-SCORE.ARPA ARPA Umbrella
Received: from SU-SCORE by SU-AI with TCP/SMTP; 15 Jun 83 13:38:43 PDT
Date: Wed 15 Jun 83 13:37:42-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: ARPA Umbrella
To: JMC@SU-AI.ARPA, Wiederhold@SUMEX-AIM.ARPA, ZM@SU-AI.ARPA, DCL@SU-AI.ARPA,
TOB@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
For your information, it is my intention to express mail the revised
umbrella proposal to ARPA tomorrow. In order to do that I must have
it to SPO by noon tomorrow. Any changes, etc., must be received by
10:00 a.m. tomorrow in order to be incorporated.
You will all receive a copy--I'm sorry that it will be "after the fact."
Betty
-------
∂15-Jun-83 1514 GROSOF@SU-SCORE.ARPA out of town
Received: from SU-SCORE by SU-AI with TCP/SMTP; 15 Jun 83 15:13:41 PDT
Date: Wed 15 Jun 83 15:15:03-PDT
From: Benjamin N. Grosof <GROSOF@SU-SCORE.ARPA>
Subject: out of town
To: genesereth@SUMEX-AIM.ARPA
cc: jmc@SU-AI.ARPA
I will be out of town for the next week, returning Wednesday , 6/23
Benjamin
-------
∂16-Jun-83 1035 DFH
The following are your phone messages
5/31. Tom Alexander, Fortune. Wants and interview...doing book on AI. Will
call you again.
5/31. A Dorothy Martin Bradley called. She thinks she may be a relative of
yours (your father was her mother's brother?) Her mother's name was
Catherine McCarthy. She was going to be in SF at her son's home for part of
June, 921-7667, but has probably left by now. Her address is 32 Lancaster St.,
Quincy, MA 02169.
6/7 Hurd. wants to get together with you.
6/13 Tod Kushner called to thank you for the use of your notes for his class.
He would like to talk with you about them. 408-946-5660 days.
6/14 Penny Park, Canadian Broadcasting, (416) 925-3311 x4341, call collect.
Other items:
1. Marilyn needs to know what books you need ordered for the fall. Do you
want the same books--Lisp (McCarthy) and MAClisp (Moon)?
2. Ben Grosof--although it appears fairly likely that Genesereth will be
able to support him this summer, he would still like to talk with you about
his project. He'll contact you.
3. The student I mentioned in the telex, Chris Tucci, will probably be
contacting you also. Despite the best efforts of several people, he did not
succeed in graduating.
4. Visitors. Corky Cartwright arrived OK. Got letter from Kuo and had
Dina Bolla do a prepay on Pan Am, Shanghai/SF. However, have not heard
anything since (done 6/7). Since his wife was anxious that he get here as
soon as possible, I charged the ticket to your American Express number (so
I could save a day or two getting a check written) --there should be a check
here to reimburse you shortly ($763). In any case, since we haven't heard
from Pan Am, it doesn't seem that he's made reservations yet (I sent an open
ticket).
5. AI Qual is Tues., June 21, 1 - 5 pm room 352. Students Chelberg and Russell.
6. Got a call from an S. Takeo on behalf of Ohm Publishing Co. in Tokyo.
Apparently they sent you an invitation to serve on the advisory board of
a new journal they are publishing in conjunction with Springer-Verlag--I
can't see that it is in your stack of mail, so I don't know that you ever
received it. Anyway, if you can say yes or no without seeing the info.,
I have their address and they would like an answer ASAP. They are to re-send
info. in any case.
∂16-Jun-83 1100 JMC*
Mitch Waldrop, Science, 202 467-4442
∂16-Jun-83 1349 ULLMAN@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 16 Jun 83 13:49:08 PDT
Date: Thu 16 Jun 83 13:50:49-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Thu 16 Jun 83 12:45:00-PDT
Thanks very much.
By the way, do you have any comments on the Guibas case.
It met the approval of a tenured faculty meeting, but we have no slot,
and nothing has really happened, so I am inviting public (i.e.,
mail to tenured-faculty) discussion if you or anyone else likes.
-------
∂16-Jun-83 1415 DIETTERICH@SUMEX-AIM RE: Learning Meaning
Received: from SUMEX-AIM by SU-AI with PUP; 16-Jun-83 14:15 PDT
Date: Thu 16 Jun 83 14:14:28-PDT
From: Tom Dietterich <DIETTERICH@SUMEX-AIM>
Subject: RE: Learning Meaning
To: JMC@SAIL
cc: DIETTERICH@SUMEX-AIM
In-Reply-To: Your message of Thu 16 Jun 83 12:40:00-PDT
Date: 16 Jun 83 1240 PDT
From: John McCarthy <JMC@SU-AI>
It's Minsky, and I'll bring in my copy if you haven't already
found one.
I sent a msg to Minsky, and he is having a copy of the paper sent to
me, so I guess I won't need yours. Thanks anyway.
--Tom
P.S. Could I get an appointment with you sometime to discuss my
dissertation work? I'm doing a dissertation with Bruce Buchanan on
scientific experimentation and theory formation. I'm trying to build
a program that interacts with a UNIX system (over the net) and forms a
theory of its file system by experimenting with the executive-level
file manipulation commands. The theory that is to be formed includes
a description of the data structures and system calls. This domain
provides an opportunity to investigate issues such as
- the deliberate choice of what data to observe (and hence, what to
ignore).
- the role of experiments in theory generation as well as in theory
testing.
- the nature of explanation.
- the formation of theoretical terms.
I'll be out of town next week, but perhaps the week after that (the
27th) would be good.
-------
∂16-Jun-83 1426 CAB Altos
To: REG
CC: LER, JMC, DEK, CAB
Ralph,
As you may know, we are having an international seminar on computer-aided
font design here at Stanford during the first week of August (1-6).
The Altos and their font software (FRED, PrePress) are needed for the
Seminar. I would like to make sure that the Altos will be working and available
during that week (after that, I don't have any strong opinions as to what
should become of them).
Lately, the Altos seem very flakey, and the IFS seems especially troublesome.
For the seminar, we need to do at least these things:
a) Find out what the Alto problems are;
b) Fix the fixable problems;
c) Maintain the fixes through August 6.
I hope you can tell us what the actual problems are.
As for the fixing and maintaining, we could ...
ask Xerox's El Segundo division for some help, since they still use
and maintain Altos down there. I can't guarantee what they will say, but
they do have some interest in having the Altos working during the Seminar.
make the repairs here, and pay for them from the Seminar budget, rather
than from normal maintence funds.
ask whatever part of Xerox actually donated them in the first place for
further assistance.
Please do let me know how we should proceed, and send a copy to Lynn Ruggles
(ler) also.
∂16-Jun-83 1706 CLT talk
I'm going to give my talk again nex tuesday at 3
for Michael,David and Chris, in case you want to come.
I'll be there. Forget not Teller reception and dinner that evening.
∂16-Jun-83 1657 BSCOTT@SU-SCORE.ARPA Umbrella proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 16 Jun 83 16:57:01 PDT
Date: Thu 16 Jun 83 16:58:31-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Umbrella proposal
To: Ohlander@USC-ISI.ARPA, Mach@USC-ECLB.ARPA
cc: JMC@SU-AI.ARPA, DCL@SU-AI.ARPA, TOB@SU-AI.ARPA, ZM@SU-AI.ARPA,
Wiederhold@SUMEX-AIM.ARPA
The revised umbrella proposal is in Express Mail. You should have it to-
morrow.
Please let me know if the task proposals should be revised.
Thank you,
Betty Scott
-------
∂16-Jun-83 1632 JJW EKL proofs
To: JK
CC: JMC
∂16-Jun-83 1548 JK
To: JMC, JJW
I have constructed a small library of current EKL proofs on [PRF,JK]
LISPAX.LSP is the set of lisp axioms
REVERS.LSP contains facts about reverse
APPEND.LSP associativity of append
RAC.LSP contains facts about rac,rdc,snoc
LENGTH.LSP contains facts about the length function on lists
ARITH.LSP basic facts about integers and sets of integers
RAMSEY.LSP redone version of ramseys theorem - only 30 lines!
Note that all function definitions outside lispax.lsp come from DEFINE
using high order unification.
Ramsey also uses heavily high order unification.
Another added feature: Unification can now use equalities from simpinfo.
The proofs are quite elegant.
JJW - Which of these do you think should be in the manual? I plan to
spend some time on that this weekend.
∂16-Jun-83 1615 JK
Bellin, a student of Kreisel, would like to work with EKL during
the summer. He needs some financial support for this. I think this would
be a good idea.
I also found another potential EKL user: A graduate student of
Arrow, who does a lot of formal theorem proving. It might be useful
to give him an account for playing with EKL.
∂16-Jun-83 1548 JK
To: JMC, JJW
I have constructed a small library of current EKL proofs on [PRF,JK]
LISPAX.LSP is the set of lisp axioms
REVERS.LSP contains facts about reverse
APPEND.LSP associativity of append
RAC.LSP contains facts about rac,rdc,snoc
LENGTH.LSP contains facts about the length function on lists
ARITH.LSP basic facts about integers and sets of integers
RAMSEY.LSP[prf,jk] redone version of ramseys theorem - only 30 lines!
Note that all function definitions outside lispax.lsp come from DEFINE
using high order unification.
Ramsey also uses heavily high order unification.
Another added feature: Unification can now use equalities from simpinfo.
The proofs are quite elegant.
∂16-Jun-83 2017 JK trip
To: JJW, JMC
I am leaving for Finland Sunday - will be back July 3.
∂17-Jun-83 0545 OHLANDER@USC-ISI Re: Umbrella Proposal
Received: from USC-ISI by SU-AI with TCP/SMTP; 17 Jun 83 05:44:50 PDT
Date: 17 Jun 1983 0540-PDT
Sender: OHLANDER at USC-ISI
Subject: Re: Umbrella Proposal
From: OHLANDER at USC-ISI
To: BSCOTT at SU-SCORE
Cc: Machado at USC-ECLB, Wiederhold at SUMEX-AIM
Cc: TOB at SU-AI, JMC at SU-AI, ZM at SU-AI
Message-ID: <[USC-ISI]17-Jun-83 05:40:42.OHLANDER>
In-Reply-To: Your message of Tue 14 Jun 83 11:44:52-PDT
Betty,
The task proposals submitted by McCarthy, Wiederhold, Binford, and
Manna are not OK as submitted. They are all asking for more money and the
additional money is simply not in the budget. I am currently trying to find
out what additional funding we might be able to come up with. No one will
be cut from previous years' levels but it is not clear that additional
funding will be forthcoming either. I hope to get back to the parties
concerned by next week.
Ron
∂17-Jun-83 0809 BSCOTT@SU-SCORE.ARPA Re: Umbrella Proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 17 Jun 83 08:09:10 PDT
Date: Fri 17 Jun 83 08:10:10-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Re: Umbrella Proposal
To: OHLANDER@USC-ISI.ARPA
cc: Machado@USC-ECLB.ARPA, Wiederhold@SUMEX-AIM.ARPA, TOB@SU-AI.ARPA,
JMC@SU-AI.ARPA, ZM@SU-AI.ARPA, BSCOTT@SU-SCORE.ARPA
In-Reply-To: Message from "OHLANDER at USC-ISI" of Fri 17 Jun 83 05:40:00-PDT
Thank you for your message. I will wait until we hear from you next week
before starting any revision effort.
Betty
-------
∂17-Jun-83 0940 WIEDERHOLD@SUMEX-AIM [OHLANDER at USC-ISI: Re: Umbrella proposal]
Received: from SUMEX-AIM by SU-AI with PUP; 17-Jun-83 09:40 PDT
Date: Fri 17 Jun 83 09:32:53-PDT
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM>
Subject: [OHLANDER at USC-ISI: Re: Umbrella proposal]
To: BSCOTT@Score, jmc@SAIL, tob@SAIL, zm@SAIL, dek@SAIL, dcl@SAIL, oliger@Navajo
For your information, message from Ohlander:
---------------
Return-Path: OHLANDER@USC-ISI
Received: from USC-ISI by SUMEX-AIM.ARPA with TCP; Fri 17 Jun 83 06:37:01-PDT
Date: 17 Jun 1983 0642-PDT
Sender: OHLANDER at USC-ISI
Subject: Re: Umbrella proposal
From: OHLANDER at USC-ISI
To: WIEDERHOLD at SUMEX-AIM
Message-ID: <[USC-ISI]17-Jun-83 06:42:52.OHLANDER>
In-Reply-To: Your message of Tue 14 Jun 83 17:12:34-PDT
Gio,
I am very sorry to hear about your broken vertebrae. I hope
it hasn't been too serious.
The most important thing to resolve right now is the umbrella
proposal. We will then have to address the actual task funding
levels. I am scratching around for some additional funds but it
is evident that I am not going to come up with anything
significantly above the levels for the previous effort. I'll get
back to you soon on this issue.
In regards to Luckham, any budgetary changes that he has should
not affect the umbrella. Duane had worked out an agreement with
him before he left and that won't be changed.
As I remarked to Betty, the umbrella proposal should go to the
NAVELEX contracting officer with copies to me and Machado.
NAVELEX will then ask if we want to fund it. Of course, I will
be working with the copy that you send me to get the paperwork
set so that we can make an instant reply. We will then indicate
to them what portions of the tasks we want to immediately fund.
This will correspond to our usual research effort. If the tasks
are all worked out this can happen simultaneously with the action
on the umbrella proposal. If not, it will come a little later
but the delay won't be significant because there are audits,
etc., that have to be accomplished for the umbrella. The actual
tasking can come as a simple modification.
Ron
-------
∂17-Jun-83 1113 DFH This morning's phone messages
1. Jack Minker, Univ. of Maryland called, regarding panel session on
non-monotonic logic honoring Dr. Lerner at IJCAI. The following
have agreed to participate: Robert Moore, Drew McDermott,
Ray Reiter, and himself. He also tentatively listed your name,
and is assuming this is OK with you and will consider things
firm unless he hears from you today (by 5 pm his time), as the
program is going to the printer. His phone is 301-454-4251.
2. Hiroshi Takeo, 503-244-4097. Called again today. This is about the
new journal Ohm Publishing is putting out called New Generation
Computing--I believe I mentioned something about this to you in
my messages yesterday. Anyway, they want an answer right away.
As I noted before, I don't see that we ever received any
material from them. If you just want to say no, I will phone him
back. If you want more info, please phone him, though I'm not
sure just how much he actually knows about it all.
∂17-Jun-83 1324 YM PhD reading committee
May I ask you to be on my reading committee?
The proposed title is "Non Clausal Logic Programming".
I'm trying to build a logic programming language based on nonclausal
resolution that allow the use of negation, iff and equality.
thanks,
Yoni Malachi.
∂17-Jun-83 1337 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA Award Suggestion
Received: from SU-SCORE by SU-AI with TCP/SMTP; 17 Jun 83 13:37:41 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Fri 17 Jun 83 13:39:03-PDT
Date: 17 Jun 1983 1320-PDT
From: Nilsson at SRI-AI
Subject: Award Suggestion
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-MC,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL
Mike Genesereth and the program committee have made what I think is
an excellent suggestion: that the AAAI award to Marvin Denicoff
"The AAAI Commendation for Excellence" in the form of a certificate
with appropriate verbiage to be given to Marvin at the AAAI
conference in WDC in August. (The title "The AAAI Commendation for
Excellence" is provisional only--we may think of something even
better.) Such an award could be given by AAAI from time to time
as excellence might happen to appear.
As we all know, Marvin has labored long and hard, with sensitivity
and taste, in his support of AI. He will be retiring from ONR later
this year. His sort of dedication and performance certainly
ought to be spotlighted. I'd be interested in your comments on this
matter BUT we have to act reasonably fast because we think it
appropriate that the award be noted in the soon-to-be-published
program.
--Nils
-------
∂18-Jun-83 0108 minker.umcp-cs@UDel-Relay Panel on Non-Monotonic Logic
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 18 Jun 83 01:08:11 PDT
Date: 17 Jun 83 17:54:25 EDT (Fri)
From: JACK MINKER <minker.umcp-cs@UDel-Relay>
Return-Path: <minker.umcp-cs@UDel-Relay>
Subject: Panel on Non-Monotonic Logic
To: Bundy@Rutgers
Cc: "uw-beaver!ubc-vision!Reiter"@LBL-UNIX, MOORE@SRI-AI, MCDERMOTT@YALE,
JMC@SU-AI
Via: UMCP-CS; 17 Jun 83 18:31-EDT
Dear Alan:
In response to your message,
1. Thursday at 1:30 is very good. I'm delighted that it does
not conflict with anyone's schedule and that it is after
Moore's talk.
2. John McCarthy has been out of the country for the past several
weeks. I phoned him today, but he was not available. I spoke with
his secretary. We agreed that if I did not hear from him by
5:00 PM my time, I could assume, by circumscription, that he
agrees to be on the panel in accordance with my earlier discussion
with him. He is being sent a copy of this message. If he disagrees,
he will send you a message over the net.
3. The panelists have all been informed that the session is to honor
Dr. Aleksandr Lerner's 70th birthday and that we will discuss
non-monotonic logic. The panelists are:
John McCarthy
Drew McDermott
Jack Minker (Session Chairman)
Robert Moore
Raymond Reiter
4. My phone numbers are:
Office-(301)454-4251
Home -(301)229-6480
5. I wrote to the Russian refusenik about the possibility of reading
his paper for him. I sent you a copy of the letter. However, I have
not received a response from him. I doubt that I will since he may
not receive his mail. I have only recently received his paper and
have not yet had an opportunity to read it. When I do I will inform
you if I feel qualified to do a good job in presenting his material.
Unless I feel comfortable with the paper, it will neither serve him
nor the scientific community to have me read the paper.
I will be at the Logic Programming Workshop in Portugal starting
June 25. I will return home on July 11. I believe that I have
responded to all of your questions. Thanks again for all of your help and support. I will be in touch with the panelists over the net to
discuss with them the issues we will cover in the panel. I will
keep you posted with respect to the panel's discussions.
Sincerely yours,
Jack
∂18-Jun-83 0603 BUNDY@RUTGERS.ARPA
Received: from RUTGERS by SU-AI with TCP/SMTP; 18 Jun 83 06:03:11 PDT
Date: 18 Jun 83 09:03:00 EDT
From: BUNDY@RUTGERS.ARPA
To: JMC@SU-AI.ARPA
cc: bundy@RUTGERS.ARPA, minker.umcp-cs@UDEL-RELAY.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of 18 Jun 83 04:17:00 EDT
And yet you felt it necessary to send a confirmatory message.
I see you do not trust default logic, or do not believe I do.
Thanks anyway.
Alan
-------
∂18-Jun-83 1152 GENESERETH@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 18-Jun-83 11:52 PDT
Date: Sat 18 Jun 83 11:44:47-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
To: JMC@SAIL
In-Reply-To: Your message of Fri 17 Jun 83 23:26:00-PDT
I wrote down 11:00 but 11:30 on Wednesday is fine. I'll see you then.
mrg
-------
∂19-Jun-83 1417 FEIGENBAUM@SUMEX-AIM profile of you
Received: from SUMEX-AIM by SU-AI with PUP; 19-Jun-83 14:16 PDT
Date: Sun 19 Jun 83 14:09:30-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
Subject: profile of you
To: jmc@SAIL
John,
I just finished the profile of you in that book. I will return the book to
you tomorrow.
I found it not 'bad", but not good either. He uses you as a platform to
discurse on a variety of things that interest him about AI, and his
discussions are not silly or frivolous, and probably are good for the
purpose--popular science writing and informing the people (especially young
people) about some exciting fields of science. But in the process, you
come across as a kind of "pasteboard" figure, i.e. the portrait is not
fully realized. Perhaps that is inevitable given the relatively few pages
he has to do the job. But you deserve better.
Ed
-------
∂19-Jun-83 1437 FEIGENBAUM@SUMEX-AIM Xerox and AI
Received: from SUMEX-AIM by SU-AI with PUP; 19-Jun-83 14:37 PDT
Date: Sun 19 Jun 83 14:02:14-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
Subject: Xerox and AI
To: hpp@SUMEX-AIM, su-bboards@Score, jmc@SAIL, tw@SAIL
It was startling to see an ad from the Xerox Corporation (the copier
part) on television last night (during one of the major-network
11PM news broadcasts). The ad touted Xerox's commitment to
Artificial Intelligence, in connection with the improvement of
the interface between people and copiers.
The ad began by showing infants in an nursery, and indicated that these
babies will grow up to be fully communicating, fully literate adults.
Then it switched to a mention of Artificial Intelligence, in analogy,
that would "grow up" to be fully communicative and helpful to people
in the use of increasingly more complicated copiers. Finally a cut back
to the nursery for more.
Xerox, of course, is not a monolithic place, any more than Stanford is.
It's interesting, though, in view of the disparate attitudes about AI
prevalent around PARC, to see where the advertising and copier people
are (in particular, the public image they want to present).
-------
∂19-Jun-83 1456 CL.BOYER@UTEXAS-20 correctness of program for computing pi or e
Received: from UTEXAS-20 by SU-AI with TCP/SMTP; 19 Jun 83 14:56:22 PDT
Date: Sunday, 19 June 1983 16:51-CDT
From: CL.BOYER at UTEXAS-20
To: John McCarthy <JMC at SU-AI>
Subject: correctness of program for computing pi or e
Thanks for the idea.
∂19-Jun-83 1557 FEIGENBAUM@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 19-Jun-83 15:57 PDT
Date: Sun 19 Jun 83 15:50:07-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
To: JMC@SAIL
In-Reply-To: Your message of Sun 19 Jun 83 14:18:00-PDT
Yes, Hilts.
-------
∂20-Jun-83 0014 FAHLMAN@CMU-CS-C lisp on perq
Received: from CMU-CS-C by SU-AI with TCP/SMTP; 20 Jun 83 00:14:40 PDT
Received: ID <FAHLMAN@CMU-CS-C>; 20 Jun 83 03:13:26 EDT
Date: Mon, 20 Jun 1983 03:13 EDT
From: Scott E. Fahlman <Fahlman@CMU-CS-C>
To: John McCarthy <JMC@SU-AI>
Subject: lisp on perq
In-reply-to: Msg of 19 Jun 83 2334 PDT from John McCarthy <JMC at SU-AI>
John,
Our version of Common Lisp is now running on the Perq under CMU's Accent
operating system (which Three Rivers plans to market). A few
"sympathetic users" around CMU and at Three Rivers are using this system
for real work, and our Hemlock editor (EMACS written in Common Lisp) is
now turning over, but doesn't yet have Lisp mode and other hair. The
Lisp runs only on the version of the Perq with 16K microstore and 1
Mbyte main memory.
The whole system needs a lot of polishing before it will be ready for
random users. As you know if you've been watching the Common Lisp
mailing list, we will have a real manual in a few days, but it will take
us a while to incorporate all the last-minute changes and shake out the
remaining bugs and polish the rough edges.
The roughest edge right now is performance. On the TAK benchmark the
Perq is currently about 10 times slower than Franz Lisp or Common Lisp
on the Vax 780, and the Perq gets even slower when it starts paging. We
think that there is an easy factor of 4 to be gained by tuning, but I
just lost my two Perq microcoders, so it will take a while to break in
someone new and get this tuning done. The faster disk on the Perq 2
will help a lot, as will the hardware page map and cache that 3RCC is
working on. With this new hardware and a fair amount of work, we ought
to be able to get the Perq running Lisp at least as fast as a Vax 750,
at which point it should be a fairly nice little program development
station. Our hard-core AI work will probably live on the Symbolics
machines.
It is likely that our Lisp and the rest of Accent will be solid enough
for real users by September. 3RCC plans to distribute a version as soon as
it's stable, but I don't think that either 3RCC or ICL have enough
in-house Lisp talent to maintain this system for a large customer base.
That would worry me if I lived in Singapore, unless there are more Lisp
hackers there than I suspect.
For us, the main point of the Perq implementation was not to get Common
Lisp up on the Perq, which we have always viewed as a transitional
pre-Spice machine, but to develop a Lisp that we could easily port to
better machines. I think that we've been successful in that --
already our Lisp has been ported to Vax/VMS and Vax/Unix by DEC and the
people at Rutgers will very soon have it running on the Dec-20 with
extended memory. Several other manufacturers are beginning similar
efforts.
-- Scott
∂20-Jun-83 0800 JMC*
passport photo and passport
∂20-Jun-83 1052 RJB βxMAIL modification
To: ME, JMC
RJB - To ME: Impossible? Yes, it is impossible as stated, but a nearly
equivalent task is possible: E can refuse to mail an entire page *of*
bboard (unless the user confirms). To do so, just check what file is
being edited to see if it is bboard and check to see if there is an
argument to the βxMAIL command or if the user is in attach mode. I
believe that this feature is more desirable than the one JMC suggested,
since one rarely wants to mail a whole page of bboard to anyone, and one
sometimes wants to mail a whole page to bboard.
∂20-Jun-83 1107 Mailer failed mail returned
In processing the following command:
MAIL "minsky%oz"%mc,feigenbaum%sumex,nilsson%sri-ai,newell%cmu-a
The following message was unsent because of a command error:
------- Begin undelivered message: -------
∂20-Jun-83 1107 JMC
To: minsky%oz@MIT-MC, feigenbaum@SUMEX-AIM, nilsson@SRI-AI
∂20-Jun-83 1052 RJB βxMAIL modification
To: ME, JMC
RJB - To ME: Impossible? Yes, it is impossible as stated, but a nearly
equivalent task is possible: E can refuse to mail an entire page *of*
bboard (unless the user confirms). To do so, just check what file is
being edited to see if it is bboard and check to see if there is an
argument to the βxMAIL command or if the user is in attach mode. I
believe that this feature is more desirable than the one JMC suggested,
since one rarely wants to mail a whole page of bboard to anyone, and one
sometimes wants to mail a whole page to bboard.
------- End undelivered message -------
∂20-Jun-83 1400 JMC*
Waltuch,Creary,CSDCF
∂20-Jun-83 1533 DFH Facilities Committee Meeting
To: "@FACCOM.[1,DFH]"@SU-AI
From: Diana Hall (DFH@SU-AI)
Subject: Facilities Committee Meeting
JMC would like to have a Facilities Committee meeting sometime this week or
next. The times open are: afternoon of 6/22 (Wed), 6/23 or 6/24 (Thurs - Fri)
or Thurs or Fri, June 30 or July 1. Please respond, letting me know the best
times for you and the times you couldn't come. Thanks.
∂20-Jun-83 1554 RPG Interesting Orals
Next monday at 2pm, Paul Cohen (no relation) on reasoning about uncertainty.
-rpg-
∂20-Jun-83 2109 HURD@SU-SCORE.ARPA Cuthbert C. Hurd is back on SCORE
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 Jun 83 21:09:36 PDT
Date: Mon 20 Jun 83 21:10:57-PDT
From: Elizabeth Hurd <HURD@SU-SCORE.ARPA>
Subject: Cuthbert C. Hurd is back on SCORE
To: mccarthy@SU-SCORE.ARPA, gorin@SU-SCORE.ARPA, golub@SU-SCORE.ARPA,
feigenbaum@SU-SCORE.ARPA, ullman@SU-SCORE.ARPA, tvr@SU-AI.ARPA
Dear friends of Cuthbert Hurd,
A little over a year ago, some of my father's friends in the CS
department at Stanford encouraged him to get an account on one of the
machines so that they could communicate with him electronically. My
father liked this idea, and acquired a terminal, modem, an account on
SCORE, and was taught how to use it. Having enthusiastically joined
his friends on the network, he then received either no, or very few,
messages, lost interest, and stopped logging on.
Last night, after having not logged on in over a year, we got him
back on the system. But he's going to lose interest again unless we
start sending him mail. I admit that I'm at fault, too: I'm used
to talking to him, and always call him up when I want to talk to him!
Here my father, a computer pioneer, is interested in using electronic
mail, and everybody else insists on using old-fashioned telephone
systems to communicate with him!
One thing I would like you to discuss with him is this idea:
he has been encouraged to give a class here at Stanford, and I have
suggested that it be a seminar on the history of computing. Maybe
a once-a-week seminar course. What do you think of this idea? Or
do you have other suggestions for a possible course? Please send
your comments to my father.
Since I got my account on SCORE before he did, I have HURD;
my father is CHURD.
Let's all send him mail less he gets disinterested again!
Elizabeth Hurd (daughter #3)
-------
∂20-Jun-83 2242 FEIGENBAUM@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 20-Jun-83 22:42 PDT
Date: Mon 20 Jun 83 22:34:57-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
To: JMC@SAIL, minsky%oz@MIT-MC.ARPA
In-Reply-To: Your message of Mon 20 Jun 83 11:10:00-PDT
I've already told Gunkel I'm not interested in participating in his
conference and white-paper enterprise, but he keeps pestering me,
and keeps telling me to call marvin, which I have not done because I
can't believe Marvin is interested.
Ed
-------
∂20-Jun-83 2247 FEIGENBAUM@SUMEX-AIM Re: Facilities Committee Meeting
Received: from SUMEX-AIM by SU-AI with PUP; 20-Jun-83 22:47 PDT
Date: Mon 20 Jun 83 22:39:43-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
Subject: Re: Facilities Committee Meeting
To: DFH@SAIL
cc: jmc@SAIL, rindfleisch@SUMEX-AIM
In-Reply-To: Your message of Mon 20 Jun 83 15:33:00-PDT
Diana,
because of an unusually heavy crunch before leaving on my long trip on Sunday
I can not come to a facilities committee meeting this week, and I'll be
gone next week, but I will ask Tom Rindfleisch to go in my stead, if he
is not already on the committee.
Ed Feigenbaum
-------
∂21-Jun-83 0102 BUNDY@RUTGERS.ARPA Re: circumscription
Received: from RUTGERS by SU-AI with TCP/SMTP; 21 Jun 83 01:02:37 PDT
Date: 21 Jun 83 04:01:59 EDT
From: BUNDY@RUTGERS.ARPA
Subject: Re: circumscription
To: JMC@SU-AI.ARPA
cc: bundy@RUTGERS.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of 18 Jun 83 14:03:00 EDT
John
Normally yes, but in this case I was explicitly told to use
circumscription as a rule of inference.
Alan
-------
∂21-Jun-83 1024 JRP lunch
interested in lunch today? There are a couple of things
Id like to discuss with you.
∂21-Jun-83 1135 JRP lunch
To: JMC@SU-AI
CC: grosz@SRI-AI
Lets say friday then, at the faculty club. Barabara Grosz will
be there too. What I want to talk to you about:
1. Report on the current status of the proposal to SDF.
2. Bratman and I are editing an introductory philosphy anthology,
and want to include a section on computers and artificial
intelligence, and I want to pick your brain about this.
If you have a reprint of your 1979 article you can spare,
I would appreciate it.
Duplicate to JRP, because I forgot Barbara the first time.
I have made a reservation for three at noon on Friday. Which 1979
paper do you mean?
∂21-Jun-83 1231 WIEDERHOLD@SUMEX-AIM Chris Tucci's programming project for the comp.
Received: from SUMEX-AIM by SU-AI with PUP; 21-Jun-83 12:30 PDT
Date: Tue 21 Jun 83 12:22:55-PDT
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM>
Subject: Chris Tucci's programming project for the comp.
To: Jmc@SAIL, zm@SAIL, pw@SAIL, ars@SAIL, ark@SAIL, RJB@SAIL, JK@SAIL
cc: walker@Score
I am trying to resolve the status and the future of the comprehensive
programmin project of Chris Tucci, and it appears that all of you have been
involved with it at some point or other.
I solicit input on the issue.
The question is then: Is the project basically flawed so that Chris has to
start over, or are there recommendations on how it should be fixed, handled,
evaluated and passed.
Thanks Gio
-------
∂21-Jun-83 1331 ARK Re: Chris Tucci's programming project for the comp.
To: WIEDERHOLD@SUMEX-AIM
CC: JMC@SU-AI, ZM@SU-AI, PW@SU-AI, ARS@SU-AI, ARK@SU-AI, RJB@SU-AI,
JK@SU-AI, mwalker@SU-SCORE, tucci@SU-SCORE
I believe that the project can be salvaged, but requires a fair amount of
cleanup of the algorithm. The major "feature" of the program is that it
uses recursion, however, the way in which the recursion is done could be
better. Chris concentrated mostly on the algorithm (which, by the way,
has some flaws according to Paul Wieneke). I spoke to him about improving
the way it was implemented. He had made some significant improvements
subsequent to that, but apparently it was not sufficient.
Arthur
∂21-Jun-83 1526 BSCOTT@SU-SCORE.ARPA lay-off letter
Received: from SU-SCORE by SU-AI with TCP/SMTP; 21 Jun 83 15:26:03 PDT
Date: Tue 21 Jun 83 15:27:52-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: lay-off letter
To: JMC@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
John, how about the following? It follows your paragraph, and adds a bit
more. You may prefer different wording, but we do need to mention at least the lay-off benefits.
"I have reluctantly concluded that your project is not making sufficient
progress in the direction of achieving my research goals to justify my
continuing it. For this programmatic reason I propose to terminate your
employment effective November 30, l983.
"I have appreciated your work over the years, and wish you success in your
future endeavors. As we discussed, I offer my assistance to you in your
efforts to secure another position to your liking.
"The lay-off benefits to which you will be entitled will be addressed in
a separate letter to you from Betty Scott, the Department administrator."
--------
Betty
-------
∂21-Jun-83 1529 @MIT-MC:perlis.umcp-cs@UDel-Relay logic
Received: from MIT-MC by SU-AI with TCP/SMTP; 21 Jun 83 15:29:22 PDT
Date: 21 Jun 83 17:47:28 EDT (Tue)
From: Don Perlis <perlis.umcp-cs@UDel-Relay>
Return-Path: <perlis.umcp-cs@UDel-Relay>
Subject: logic
To: phil-sci%mit-oz@mit-mc
Cc: perlis.umcp-cs@UDel-Relay, JMC%SU-AI@MIT-MC
Via: UMCP-CS; 21 Jun 83 18:12-EDT
From: John McCarthy <JMC@SU-AI>
I have been in the Far East and have just looked at the
discussion of expressibility in first order logic. The point
I thought I was making in my original message seems to have
been mainly missed.
Not missed. DAM and I were agreeing on what you say here, but
disagreeing on its significance for what 'meaning' is. I think
Tarskian meaning is just fine and NO real limitation on AI or
on FOL. Apparently you also think this.
DAM's example of the transitive closure of a relation provides
a good example of what I meant...
Suppose we don't want to use second order logic, i.e.
we don't want to quantify over relations. We can introduce a
domain of objects that we will call Relations, i.e. we reify
(make things out of) relations... Relations are first order
objects...(but there) is nothing that requires that our domain
Relations reifies all relations.
However, second order logic isn't such a big win
either, because second order logic itself is incomplete. Not
every formula true in all models of second order logic is
provable in second order logic. This contrasts with the
completeness of first order logic. This tradeoff between
incompleteness of theories and incompleteness of the logic is
unavoidable.
Exactly, and for DAM's sake let me add that requiring Relation to reify
all relations depends on a model itself, namely our concept of the
world. How do we know what 'all relations' refers to? We don't, we
just go ahead anyway, content in our ruminations. We can prove what
we think we ought, and that's enough for most of us. This involves
no surrender to those (who are they, DAM?) that may tell us we aren't
'really' thinking about 'all relations'.
A rather powerful choice is to use Zermelo-Frankel set
theory...The computationally difficult part of ZF for theorem
provers is to make them use the axiom schema of comprehension
and the somewhat more difficult axiom schema of
replacement...Making the computer invent properties Foo
relevant to solving problems or deriving theorems is difficult
- even more difficult than making the computer invent suitable
instances of the axiom schema of induction in arithmetic or
Lisp theory. Only Boyer and Moore do much with the latter,
and, to my knowledge, no-one has written a program that invents
formulas for doing comprehension. We won't have powerful AI
until someone does.
Doesn't Lenat's AM do something along these lines? It invents
useful concepts (sets) as it goes. Granted it's not a problem-solver,
but it seems to have an approach to the concept generation issue.
Don Perlis
∂21-Jun-83 1646 PW Chris Tucci's final project
To: Wiederhold@SUMEX-AIM
CC: JMC@SU-AI, ZM@SU-AI, ARS@SU-AI, ARK@SU-AI, RJB@SU-AI, JK@SU-AI,
mwalker@SU-SCORE, CT@SU-AI
Chris has has resolved most of the problems in the one-against-one counterpoint,
and is currently working on two-against-one counterpoint. We decided to limit
the task to achieving the latter. I am only checking the quality of the output
and appropriateness of the code to the task. I will have to defer the decision
on whether this is ambitious enough in scope to the CS faculty.
I expect the program will be ready for final evaluation by this Friday.
∂21-Jun-83 1752 @MIT-MC:DAM%MIT-OZ%MIT-MC@SU-DSN First Order Logic and Epistemological Adequacy
Received: from MIT-MC by SU-AI with TCP/SMTP; 21 Jun 83 17:51:44 PDT
Date: Tue, 21 Jun 1983 20:41 EDT
From: DAM%MIT-OZ%MIT-MC@SU-DSN
To: JMC@SU-AI
cc: phil-sci%MIT-OZ%MIT-MC@SU-DSN
Subject: First Order Logic and Epistemological Adequacy
Date: Monday, 20 June 1983 05:07-EDT
From: John McCarthy <JMC at SU-AI>
The point I thought I was making in my original message seems
to have been mainly missed.
You are claiming that first order logic is in some sense
universal. I will elaborate what I see as the possible
interpretations of this statement and you can tell me which (if any)
of these interpretations you had in mind.
The weakest claim is that first order logic is universal in
some computational sense (i.e. it is Turing complete). This I will
grant completely, there is a technical sense in which a complete
inference engine for first order logic can be programed via axioms to
perform any computation. But of course FORTRAN is also Turing
complete so it is not clear why this property of first order logic
makes first order logic interesting. I assume you have more in mind
for first order logic than just the Turing completeness of first order
logic viewed as a programming language.
The strongest possible is that first order logic is
"epistemologically universal", i.e. that all knowledge we might be
interested in capturing can be expressed in first order logic. If one
is interested in epistemological adequacy rather than computational
adequacy then one should (it seems to me) be interested in the
expressive power of a language independent of computational
properties. I have argued against the notion that first order logic
is epistemologically adequate (I don't think that one can claim that
the frist order axioms you presented in your last message captured the
notion of a trasitive closure in an epistemological sense, what if you
had used the symbol "g0032" instead of the symbol "applies"). I
believe that using formalisms which have a richer expressive power may
be important for acheiving computational adequacy (the notion of an
essential as opposed to contextual property is my favorite example of
an epistemological notion which may be important computationally).
An intermediate claim about first order logic is that
everything can be represented by a first order structure and that
first order logic is useful for reasoning about first order
structures. When you talk about "reifying" certain things it seems
that you mean representing those things as points in the domain of a
first order structure (reification is not "making something a thing",
a relation is already a "thing" in true second order logic). Any
mathematical object can indeed be thought of as a point in the domain
of a first order structure (the structure of "sets" whose only
relation is the binary "member-of" relation). One can rightly claim
that first order logic is universal in the following sense: Let L be
any formal language and let M be any model for the sentences of L (a
given sentence of L is either true or false of M). Any system for
COMPUTING the sentences true of M can be encoded as a set of axioms
about a first order structure M' which represents M (remember anything
can be represented by a first order structure).
This last claim seems to be the most consistent with your last
message. It is really the combination of two claims: that first order
STRUCTURES are EPISTEMOLOGICALLY universal (can represent anything)
and that first order LOGIC is COMPUTATIONALY universal for reasoning
about any given first order structure. Thus it is the models rather
than the sentences which are epistemologically universal.
The universality of first order structures should not
lead one to believe that all of the epistemelogical problems of
representation have been solved. Consider a mathematical object we
might be interested in representing, say a context free grammar.
There is a natural notion of two context free languages being
"isomorphic", of non-terminal symbols "representing" a set of
expressions, and of a langauge (set of strings) as being "more
abstract" or having "less information" than a particular grammar for
that language. The notions of "isomorphism", "representation", and
"more abstract" seem to be universal (they can be used in a precise
and natural way when talking about any mathematical objects).
However these notions are hard to define even when objects (such
as languages and grammars) are "reified" as points in a first order
structure.
David Mc
∂21-Jun-83 1753 REM@MIT-MC Belated voice notes (January, TCP not yet up at SU),
Received: from MIT-MC by SU-AI with TCP/SMTP; 21 Jun 83 17:53:00 PDT
Date: 21 June 1983 20:46 EDT
From: Robert Elton Maas <REM @ MIT-MC>
Subject: Belated voice notes (January, TCP not yet up at SU),
To: rem @ SU-AI
cc: REM @ MIT-MC, HPM @ MIT-MC, jmc @ SU-AI
1983 Jan 24, 1421 PST
A minute ago I just came up with a brilliant idea for how to handle
the toplevel histogram when doing data compression, how to handle
escape from the context nodes. The answer is to store the toplevel
histogram as a complete histogram with every conceivable character in
it, but to store that as a logarithmetically accessible data structure
like Moravec and I were going to do with the multiple-LISP-processor.
Subsets of this can be extracted changing only a logarithmetic amount
of the data rather than having to copy the whole thing, and the actual
counters in the histogram can be shared, so that when the toplevel
histogram is updated the subset histograms automatically get updated
too. Of course when you want to delete an entry from one of the
sub-histograms you simply deduct the current count for that token from
the overall count and delete the structure in the sub-histogram that
points to that token-node in the original histogram. So, when you have
a context histogram, and the escape occurs, some character not in the
actual context-histogram, you encode the escape prefix and then to
encode the actual character instead of using a 7-bit or 8-bit code or
reverting to the null-context histogram, you encode it according to
this sub-histogram of the null-context histogram (or when more than
one level deep in context, you encode the escaped character/token by
the immediate parent (shorter context) which in turn either encodes
directly or escapes yet another level). Of course you update the
counts in both the case of it being escaped from a deeper context and
in the case of it being directly in that context, thus having the
count always reflect all the uses to which it's being put.
An example of where this method is almost essential: Suppose you have
some context that individually recognizes space T O A N and I but does
not recognize E because it just happens to have dropped out of the
context histogram because it just doesn't occur enough in that
context, so E is now included in the escape node for that context
histogram. Now suppose an "E" acctually occurs in data in that
context. Using the old scheme you would encode the escape prefix
followed by a complete 7-bit or 8-bit code for the letter "E", whereas
using this new method you would have the escape prefix followed by how
often E is in that context and with the a priori knowledge that none of
those other common characters occurred, that is a histogram with
normal characteristics but with all those other common characters
except E deleted, in which cave E would have very high probability,
which is in fact what the case would be. So you get a much more
efficient coding of the common characters that just happen not to be
in the particular context histogram you're working with.
Of course the context histogram can still be stored in the traditional
linear way that makes linear search efficient for the most common
tokens. Of course you have to encode context histograms in this
logarithic way too if you want them to share elements with
sub-histograms that are escaped from deeper-context histograms. Also
if you use this method, you can share data structure after histograms
are forked until such time as the counts start to diverge, i.e.
copy-upon-write a la demand paging of pure pages. I.e. between
histogram and escape from deeper context, sharing is permanent (until
the node is deleted from the sub-histogram because it's back in the
main deeper-context histogram), whereas between histogram and main
deeper-context histogram, sharing of a node is temporary until such
time as the node is changed in either histogram. Former is shared
side-effects, latter is copy-on-write.
P.s. the actual data structure for the logarithmetic access is a
balanced binary tree.
One problem with the data structure sharing: When a count bubbles up
towards the front of the histogram because it has been incremented
recently more than other counts previously ahead of it, this bubbling
occurs only in the one data structure, not in other histograms that
share some nodes and structure. Thus the search sequence where you
start at the front of the histogram and work down may actually not be
in the correct order, and thus there'll be a very long search
sometimes. Of course each time you search in a given histogram and
find it not in descending-count order you can rearrange it, so if a
particular histogram is causing a problem it will eventually arrange
itself into the right order like the histogram it was derived from.
Now 1716 PST - Another problem with the data structure sharing: When
you increment the count for a node, you also have to increment the
total not just for the histogram you're working with but also for all
other histograms that include this node as a logical element. So I
guess that means each element of a histogram must have a list of all
histograms that reference it. That's a little bit of extra overhead I
guess.
∂21-Jun-83 1813 REM@MIT-MC Belated voice notes (January, TCP not yet working then), crunch
Received: from MIT-MC by SU-AI with TCP/SMTP; 21 Jun 83 18:13:35 PDT
Date: 21 June 1983 21:12 EDT
From: Robert Elton Maas <REM @ MIT-MC>
Subject: Belated voice notes (January, TCP not yet working then), crunch
To: rem @ SU-AI
cc: REM @ MIT-MC, jmc @ SU-AI
Wednesday Ferbruary 2, afternoon
A couple more ideas on the data compression thing, for copying context
from parent down to sub-node. The method I described last week has
problems in that things are done in sort of the wrong order. When
updating a histogram, you start off with a full histogram (forked from
parent), but it rapidly gets purged. There's so little paralelism that
the method might not work, but I have two other ideas.
One way is to simply make a pointer to the entire parent histogram,
and in the rare event that an escape node happens and you want to
refer to the parent histogram you recompile a virtual new histogram
that has deleted all the entries that are present in the
greater-context histogram. This requires a lot of extra compute time
but not extra permanent memory. If it only happens once in a while, it
may be optimal.
Another method is to have the whole thing be in binary, as suggested
in one of those IBM articles. Simply decode the bits of a character
one at a time instead of en masse, so you always have a two-way
decision. Furthermore the 2-way decisions that you're making at one
level are the same as the 2-way decisions that you're making at the
parent node, so that if one or the other of them is escaped you can
simply point directly to that subnode in the parent without having to
copy anything or having to make a virtual histogram with entries
deleted from the parent histogram; just point to that half-histogram as-is.
∂21-Jun-83 1854 REM@MIT-MC Belated voice notes (Feb., TCP not yet working then), causality debugging
Received: from MIT-MC by SU-AI with TCP/SMTP; 21 Jun 83 18:54:18 PDT
Date: 21 June 1983 21:52 EDT
From: Robert Elton Maas <REM @ MIT-MC>
Subject: Belated voice notes (Feb., TCP not yet working then), causality debugging
To: REM @ SU-AI
cc: REM @ MIT-MC, MINSKY @ MIT-MC, RWG @ MIT-MC, PRATT @ MIT-MC,
JMC @ SU-AI
This is a reminder: I should at some point actually write up an
article on causality-backtracing as a way of debugging; where you
identify the major cause and all the minor causes for any particular
event and you trace these back from the observed bad output through
the program and/or data states until you reach a point at which all
the inputs (causes) are correct and the output (effect) is wrong, or
at least you think all the inputs are correct and the output's wrong,
actually that usually means the major premise, which is the piece of
program code executed at that point, has a bug and you've now isolated
it.
It may very well be that my idea for the causality backtracing is a
major innovation that will make me famous someday as being maybe ahead
of my time like McCarthy and Minsky and other people in artificial
intelligence, Shannon and Babbage...
P.s. these entries Feb 18th in the morning.
Regarding how to actually implement the causality backtracing, two
ways: With conventional technology you use the "all side effects are
contained in modified list structure using balanced binary trees", so
that nothing really ever gets changed, things only get added to, which
means that references back into the way things were don't get modified
out from under them by RPLACA RPLACD SETQ things like that, so you can
have a back pointer of frozen states of the universe at various times,
and roll back to the last state before the part you're trying to
trace, and then step forward adding new states from there up to the
current state where you know the error's already happened. Now with a
video disk, you could store the complete transcript of every
state-change in the entire machine, on the video disk, because it has
such high storage capacity, and simply have back-pointers to all the
causes for each new thing that happens, so you write this transcript
using reverse Polish notation basically, that is the order it's stored
on the disk goes from the leaves up to the nodes that point to them,
and then when you get to the place that the error occurs you simply
read the disk backwards, skipping over I mean, but you start at the
very end where the occur was observed, and you follow a pointer back
to a cause another pointer back to an earlier cause and so on, reading
backwards on the video disk until you find the point where the error
occurred. But you don't have to re-execute anything because the
complete transcript of all state changes is on the disk rather than
just a scattered subset of them.
∂21-Jun-83 2350 RPG
∂21-Jun-83 2341 JMC rms
While I suspect your characterization is correct, messages such as yours
should not be part of the Common Lisp discussion file. In the heat of
a controversy such remarks may be reasonable, but remember that they'll
still be in the file a year hence. I would suggest that you edit it
out.
Well, you may not have noticed that that message was cc'ed to the file,
not sent to the list. I've re-mailed many private messages to the file
as a record. Perhaps, you're suggesting, I should do a global fork?
-rpg-
∂22-Jun-83 0140 ARK Chris Tucci's CS293 Project
To: ZM@SU-AI, JMC@SU-AI
CC: ARK@SU-AI, MWalker@SU-SCORE
I'd suggest that Chris be given an Incomplete for Spring. (If he already
has an "L" grade, that is a bad idea, because it would allow him to
graduate without ever changing it to some letter grade.) If Chris finishes
the project to people's satisfaction, that it can be changed to the
appropriate letter grade. Furthermore, by giving him an incomplete, he
won't have to pay extra tuition. Marilynn, please advise us on whether he
can get away with paying tuition for, say, 1 unit over the summer to be
registered to complete the CS293 and do the programming project. It really
doesn't make sense for us to force Chris to pay extra tuition. What does
everyone think of this approach? Thanks.
Arthur
∂22-Jun-83 0700 JMC*
Phone Bradley
∂22-Jun-83 0805 JRP 1979 paper
The one on ascribing mental properties to machines.
∂22-Jun-83 0912 DFH my schedule today
To: JMC, ZM
If it does not conflict with your needs, I will take
an early lunch today from 11:30 am - 12:30 pm. (I have
to go sign a rental agreement.) -- Diana
∂22-Jun-83 0933 ATP.BLEDSOE@UTEXAS-20 Re: strange mail
Received: from UTEXAS-20 by SU-AI with TCP/SMTP; 22 Jun 83 09:33:11 PDT
Date: 22 Jun 1983 1130-CDT
From: Woody Bledsoe <ATP.Bledsoe@UTEXAS-20>
Subject: Re: strange mail
To: JMC@SU-AI
cc: ATP.Bledsoe@UTEXAS-20
In-Reply-To: Your message of 22-Jun-83 0917-CDT
Sorry John, My new secretary goofed. The correct stuff is coming now.
Woody
-------
∂22-Jun-83 0935 DFH Facilities Committee Meeting
To: "@FACCOM.[1,DFH]"@SU-AI
From: Diana Hall (DFH@SU-AI)
Subject: Facilities Committee Meeting
Unless there are strenous objections, I have scheduled the Facilities Committee
meeting from 1:30 - 3:00 pm on Thursday, June 30, room 252. I realize that this
may not be a particularly good time for some of you, but it seemed the best
compromise. People who did not respond to my first message have not been taken
into account.
∂22-Jun-83 1107 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract
Received: from SUMEX-AIM by SU-AI with PUP; 22-Jun-83 11:07 PDT
Date: Wed 22 Jun 83 10:57:56-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Re: Equipment Contract
To: Feigenbaum@SUMEX-AIM, JMC@SAIL, CSL.JLH@Score, Ullman@Diablo, RPG@SAIL,
REG@SAIL, Bosack@Score
cc: CSD.BScott@Score, Rindfleisch@SUMEX-AIM
In-Reply-To: Your message of Mon 23 May 83 17:25:49-PDT
DSSW is supposed to have mailed the equipment contract last Friday and
it is expected at Stanford any day now. As I said in an earlier
message, we will be funded with an initial increment of $1.6M through
December and since some of this is "old" money, it is important that
we spend it soon. I got no replies to my earlier message asking for
priorities so below is my tentative plan for subdividing the first
increment. It is based on what I think are the most pressing needs.
If I have gored anyone's ox in the process, we should probably have a
meeting of the principals.
Lisp machines $700K LM-2, 5 3600's, 5 1108's, 1 1132, glue
File server 230 Full configuration
11/750 workstations 320 I think DEC's deal self-destructs soon
SUN workstations 190 13 of 27
Printers 60 2 of 4
Ether TIP's, gateways,
test equip, SUN clust 70 Gross estim -- SUN devel after Sept
Other personnel, etc. 30
-----
$1600
We also need to agree on how the budget will be managed. I propose
setting up 4 accounts by major group so everyone has a common model of
how funds are divided. These would be for HPP Lisp Machines, Formal
Reasoning Lisp Machines, CSL Equipment, and CSD-CF. Clearly we can
reallocate between these by joint agreement at any time. Comments?
Tom R.
-------
∂22-Jun-83 1331 ullman@Diablo Re: Equipment Contract
Received: from SU-HNV by SU-AI with PUP; 22-Jun-83 13:31 PDT
Date: Wed, 22 Jun 83 13:32 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Re: Equipment Contract
To: Bosack@SU-Score, CSL.JLH@SU-Score, Feigenbaum@Sumex-Aim, JMC@Sail,
REG@Sail, RPG@Sail, Rindfleisch@Sumex-Aim
Cc: CSD.BScott@SU-Score
Sounds good. However, we did leave in limbo the question of
how many of the first SUN's will go to CSD, how many to ERL.
Ditto the matter of the VAX750's;, e.g., Gio asked for one
and as an ARPA contractor he has some rights in this regard.
Perhaps CSD-bound equipment should go in the CSD-CF pile.
∂22-Jun-83 1204 RJB
To: JMC@SU-AI, ZM@SU-AI, PW@SU-AI, ARS@SU-AI, ARK@SU-AI, JK@SU-AI,
wiederhold@SUMEX-AIM
RJB - I decided not to pass Chris's project for 7 reasons. Three of them
were stylistic and one was a trivial bug; I expect that he will have no
trouble fixing these four problems. Another problem is that I don't think
that his heuristic works well for some inputs; he *should* be able to fix
this.
The remaining problems are more disturbing. According to the Charge to
the Comprehensive Committee, "The project must exhibit the use of
sophisticated algorithms and data structures." Tucci's most complicated
algorithm is a depth-first search. His most complicated data structures
are a list and a look-up table (both implemented by arrays). In light of
this, I believe that it would be worth discussing whether any amount of
improvement to the project will enable it to meet the project
requirements.
Finally, I am concerned about the amount of aid Chris received before turning
in his project and the amount of aid he is receiving in his revision. I
would like feedback on this.
∂22-Jun-83 1451 DFH Arpa pre-award audit
Please contact Betty Scott about this right away.
She has tentatively set the meeting with them for
2 pm Monday, June 27, and feels it is extremely
important that you be there. (I note that this
conflicts with IBM Seminar). -- Diana
∂22-Jun-83 1500 JMC*
cleaning
∂22-Jun-83 1521 BSCOTT@SU-SCORE.ARPA Auditors
Received: from SU-SCORE by SU-AI with TCP/SMTP; 22 Jun 83 15:20:55 PDT
Date: Wed 22 Jun 83 15:22:47-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Auditors
To: JMC@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
John, how about 10:00 a.m. on Monday? Gio is complete tied up with
class and seminar on Friday.
Betty
-------
Monday is impossible. I'll be in San Jose all day. Also Tuesday.
∂22-Jun-83 1532 DFH summer support for Luigi
He has decided to take our RA appointment. He has filed
for candidacy in the Philosophy dept., so we can actually
pay him at a the higher of two rates (1476 vs 1368 per month)
if you want. There is apparently no set rule on this when the
student is from another department, so you'll have to let me know.
--Diana
In view of his inexperience in the area he will be working, I have
decided on the lower rate.
∂22-Jun-83 1547 BSCOTT@SU-SCORE.ARPA Auditors
Received: from SU-SCORE by SU-AI with TCP/SMTP; 22 Jun 83 15:46:51 PDT
Date: Wed 22 Jun 83 15:48:35-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Auditors
To: JMC@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
John, Thursday and Friday are out for Gio. So we'll just have the audit
meeting on Monday without you. Since you are the senior P.I., I thought
it would be good to have you there. But I think Gio and I can take care
of it.
Thanks,
Betty
-------
∂22-Jun-83 2029 JMC*
common[s83,jmc]
∂22-Jun-83 2041 perlis.umcp-cs@UDel-Relay Re: First Order Logic and Epistemological Adequacy
Received: from SU-DSN by SU-AI with PUP; 22-Jun-83 20:41 PDT
Received: From UDEL-RELAY by SU-DSN.ARPA; Wed Jun 22 19:27:09 1983
Date: 22 Jun 83 22:14:03 EDT (Wed)
From: Don Perlis <perlis.umcp-cs@UDel-Relay>
Return-Path: <perlis.umcp-cs@UDel-Relay>
Subject: Re: First Order Logic and Epistemological Adequacy
To: perlis.umcp-cs@UDel-Relay, DAM%MIT-OZ@MIT-MC, JMC%SU-AI@SU-DSN
Cc: phil-sci%MIT-OZ@MIT-MC
Via: UMCP-CS; 22 Jun 83 22:23-EDT
I am butting in here in an exchange between JMC and DAM. Of course I
am not speaking for JMC; I am simply carrying on the previous series
DAM and I were tied up in, as it impinged directly on DAM's present
points.
From: DAM%MIT-OZ@MIT-MC
[JMC is] claiming that first order logic is in some sense
universal. I will elaborate what I see as the possible
interpretations of this statement...
The weakest claim is that first order logic is
universal in some computational sense (i.e. it is Turing
complete)... But of course FORTRAN is also Turing complete so it
is not clear why this property of first order logic makes first
order logic interesting.
First-order logic has a semantics, in that true sentences (in any
interpretation) lead via rules of inference to other true sentences.
The strongest possible is that first order logic is
"epistemologically universal", i.e. that all knowledge we
might be interested in capturing can be expressed in first
order logic. If one is interested in epistemological adequacy
rather than computational adequacy then one should (it seems to
me) be interested in the expressive power of a language
independent of computational properties. I have argued against
the notion that first order logic is epistemologically
adequate...
But your objection as given in past messages has a peculiar aspect.
After all, if I write A(x) in the standard syntax of first-order logic,
and if I tell you I MEAN by that formula that x is an apple (and if I
similarly tell you what I mean by the other symbols) then I have told
you an intended structure for this first-order language, and hopefully
one that is a model of the axioms as well (let us suppose I have also
stated the axioms, viz, A(x) & B(x) .--> C(x) etc).
Now I may really intend such a meaning, and I may further state that
the ONLY structures that are of interest to me are the mentioned one
(apples etc) and a few other alternates (but say always ones with fruit
for the predicates). Now by your definition earlier, this is a logic:
a set of 'models' and a language with truth functions over sentences on
those models. I may even admit 'apples etc' as the ONLY model.
Yet I can hardly deny that there do exist functions f and mathematical
structures M such that the pairs <f,M> do in fact form other
interpretations of the language and even ones in which the axioms will
come out true. I may not be interested in them but they do exist; I
may not call them models, but they have the indicated relationship to
the language-plus-axioms, like it or not. This is just the upward
Lowenheim Skolem theorem, whether we use the word model or just
'related structure'. So I don't see that using a word differently in
any way changes the actual situation.
I see no problem with saying that there are only certain models (call
them what you will) that we want to talk about. The others still exist
and cause no trouble; why can't we have friendly models and unfriendly
ones? Indeed logicians speak this way: standard models and
non-standard ones. The former are intended, the latter are still there
although sometimes (not always) as a burden. Introducing a new name
for them (calling them non-models as DAM seems to urge) accomplishes
nothing of either technical or philosophical/semantic significance that
I can see.
I'm perfectly willing to believe DAM when he says he means
such-and-such things by the symbols, even while recognizing that other
meanings are possible. What's the fuss? And why in the world no
longer call it first-order logic? The other models DO exist, nothing
can stop that. Where in the world is the greater POWER in denying
what exists, or in urging a preference between two different kinds
of structure? If your telling me what you mean by certain symbols
is a kind of increase in expressive power, fine, I guess it is true
that you have given me more information about your linguistic usage.
But why change the generic name of the logic for that? No one else
seems to.
When we use first-order logic, we write formulas, choose
axioms, discuss particular models. Never is there any fuss about
what to call the models we aren't interested in: we just say we aren't
interested in them, and we have no need to say we're doing a different
kind of logic. It's when we chamge the SYNTAX that we call it a
different kind of logic, eg, new inference rules or new kinds of
formulas, such as more than one kind of variable, or
non-truth-functional connectives.
I realize DAM has a different (non-standard) definition of first-order
logic, in which a specification of the things to call models figures
prominently. But not only is this non-standard, I don't see the
point of it.
...When [JMC] talks about
"reifying" certain things it seems that [he] means representing
those things as points in the domain of a first order structure
(reification is not "making something a thing", a relation is
already a "thing" in true second order logic).
Only relations over first-order individuals in the second-order logic
are already things in the second-order logic. We may want to have
things for relations much more ambitious than this, and it can be done
all in first-order logic with reifications. Indeed that's what the
comprehension principle does in Zermelo-Fraenkel (first-order) set theory.
So what's the point of your comment?
Don Perlis
∂22-Jun-83 2205 LLW@S1-A Festschrift, Etc.
Received: from S1-A by SU-AI with TCP/SMTP; 22 Jun 83 22:04:47 PDT
Date: 22 Jun 83 2159 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Festschrift, Etc.
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
∂22-Jun-83 0725 JMC@SU-AI Shackleton for Festschrift
Received: from SU-AI by S1-A with TCP/SMTP; 22 Jun 83 07:25:21 PDT
Date: 22 Jun 83 0720 PDT
From: John McCarthy <JMC@SU-AI>
Subject: Shackleton for Festschrift
To: llw@S1-A
I didn't know about the Festschrift or this idea might have occurred
to me sooner, but maybe it's not too late. When Edward spoke of going
to the moon again, it occurred to me that you and Rod and I might write
up Shackleton in its present state for the Festschrift. It would provide
an obscure but respectable place from which to push it, and from which
Edward could push too if his initial favorable reaction holds up.
I could have a draft in two weeks if that were reasonable or a week if
it made the critical difference.
How about the 29th or the 5th or 6th for my next visit to the Lab?
[John: Excellent idea re Shackleton; I'll let you know if Rod has any
problems with it ASAP; I think two weeks is sufficiently soon. The 29th
looks likes a Washington day from the present vantage point, but I may
succeed in getting back from there on the 28th; since it's uncertain from
my standpoint, would you care to come on over that day and take potluck
(working with Rod on the paper and/or laying the lash on JJW, if I don't
show), or should we make it the 6th for certain? Lowell]
I'll come the 29th - and the 6th also if it seems worthwhile.
∂22-Jun-83 2218 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract
Received: from SUMEX-AIM by SU-AI with PUP; 22-Jun-83 22:18 PDT
Date: Wed 22 Jun 83 22:06:07-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Re: Equipment Contract
To: ullman@Diablo, Bosack@Score, CSL.JLH@Score, Feigenbaum@SUMEX-AIM, JMC@SAIL,
REG@SAIL, RPG@SAIL
cc: CSD.BScott@Score, Rindfleisch@SUMEX-AIM
In-Reply-To: Your message of Wed 22 Jun 83 13:24:22-PDT
I guess I don't recall the agreement in the original submission about
the SUN/VAX subdivision between CSD and CSL, Jeff -- do you? If the
overall subdivision is clear, an easy approach would be to allocate the
incremental purchases in the same proportions. It seems reasonable to
lump CSD-bound stuff under CSD-CF -- that is where I thought the file
server should go for example.
Tom R.
-------
∂22-Jun-83 2221 LLW@S1-A Treasures Of The East
Received: from S1-A by SU-AI with TCP/SMTP; 22 Jun 83 22:21:15 PDT
Date: 22 Jun 83 2215 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Treasures Of The East
To: rpg@SU-AI
CC: LLW@S1-A, TM@S1-A, JBR@S1-A, jmc@SU-AI
Dick, JMC mentioned to JBR, TM and myself yesterday that he had brought
back all sorts of goodies from the exotic Orient regarding
high-performance computers. He was even reckless enough to suggest that
some of them might compare favorably with the Mark IIA, and then adroitly
forestalled demands to inspect said documents by remarking that he had
given them all to you. If you're not likely to be over here in the quite
near future, would you please either mail me xeroxes of these, or else
scrawl your name all over them, list their titles somewhere, and mail them
to me for xeroxing and return to you when you next come by? Thanks,
Lowell
∂23-Jun-83 0131 TUCCI@SU-SCORE.ARPA CS 293 etc
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 Jun 83 01:30:59 PDT
Date: Thu 23 Jun 83 01:32:11-PDT
From: Christopher Tucci <TUCCI@SU-SCORE.ARPA>
Subject: CS 293 etc
To: jmc@SU-AI.ARPA
cc: tucci@SU-SCORE.ARPA
Dear Prof McCarthy,
Hello, this is your sometimes advisee, troublemaker, and all-around
black sheep of the CS Dept, Chris Tucci. I don't know if you've heard
about all of the things that have been going wrong with CS 293 and the
Comprehensive programming project due mostly to a lack of communication
between me and several others. Oh, well, that's all behind me now.
What I need from you is a decision as my advisor. There are two
possibilities:
1) Meet with me very soon (i.e. Thursday or Friday) and discuss
my project, which I have all but completed.
This probably would not take very long, but we would
have to go over the project. Then you could give me a grade in
CS 293 and I could submit the Comp Programming Project whenever
you were happy with it.
2) Because of my lack of communication with you in the past, or
other various reasons, you may not want anything more to do
with this ill-fated, hapless project.
In which case I would ask that you give me names of others whom I
could possibly seek out and ask them to be my new advisor. Of
course, this takes only a few seconds, but is not as rewarding.
Well, that's about it -- I hope you will choose Door #1, but I won't
blame you otherwise. I will also try to get in touch with you personally
today (Thurs) -- I think Marilynn will also. Goodbye, and thank you.
Chris Tucci
-------
I'll be in both Thursday and Friday. Catch me if you can.
∂23-Jun-83 0844 MWALKER@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 Jun 83 08:44:44 PDT
Date: Thu 23 Jun 83 08:46:35-PDT
From: Marilynn Walker <MWALKER@SU-SCORE.ARPA>
To: jmc@SU-AI.ARPA
cc: mwalker@SU-SCORE.ARPA
John, I need to talk to you about one of our students who needs some
guidance and is not having too much luck getting any.
Marilynn
-------
∂23-Jun-83 1059 DFH lunch today
Hurd called and wants to have lunch with you today
if you have no other plans.
∂23-Jun-83 1159 RV Let's talk
As you suggested, I think we should get together for a little while
and talk about my qual situation. I'll try to find you tomorrow (Friday);
can you suggest a good time?
Rick
∂23-Jun-83 1306 RSC Afternoon meeting
I've gone to get a sandwich -- will be back by 1:45 P.M., probably earlier.
Corky
∂23-Jun-83 1422 ULLMAN@SU-SCORE.ARPA meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 Jun 83 14:22:35 PDT
Date: Thu 23 Jun 83 14:19:56-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: meeting
To: super@SU-SCORE.ARPA
This probably isn't the time to mention it, but since there were
only positive votes for holding the meeting at 3PM, let's start doing that
today. We'll assume ERL401 is available; if not, we'll hope there is
a small group meeting there.
-------
∂23-Jun-83 2240 FEIGENBAUM@SUMEX-AIM Re: making life complicated for ourselves
Received: from SUMEX-AIM by SU-AI with PUP; 23-Jun-83 22:40 PDT
Date: Thu 23 Jun 83 22:40:58-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
Subject: Re: making life complicated for ourselves
To: JMC@SAIL
In-Reply-To: Your message of Thu 23 Jun 83 18:24:00-PDT
John,
you're right. I wish things could be made simpler in our department.
Things are getting so time-consuiming and complicated. (Sigh...)
Ed
-------
∂23-Jun-83 2322 ULLMAN@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 Jun 83 23:22:05 PDT
Date: Thu 23 Jun 83 23:23:58-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Thu 23 Jun 83 18:40:00-PDT
Please do come if you can make it.
It looks, by the way, as if the meetings have moved to
the CIS conference room.
P.S. I think you may have something about the programming project.
It is handy as it is to encourage students to do real projects,
but I too have felt myself in an awkward situation like the Tucci
business.
-------
∂24-Jun-83 0024 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #11
Received: from SU-SCORE by SU-AI with TCP/SMTP; 24 Jun 83 00:24:40 PDT
Date: Thursday, June 23, 1983 2:55PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #11
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Friday, 24 Jun 1983 Volume 1 : Issue 11
Today's Topics:
Implementations - VAX Prolog & Paranoia,
Applications - Using Prefix Operators & Definitions
----------------------------------------------------------------------
Date: Wednesday, 15-Jun-83 19:24:56-BST
From: RAE (on ERCC DEC-10) <Rae@EDXA>
Subject: Prolog For The VAX
Steve,
You correctly state that POPLOG and Franz have been identified
by the UK IKBS initiative as systems for getting people off the ground
in IKBS. DEC-20 Prolog is not classified with them, unfortunately, as
the other vital ingredient for the software infra-structure is the
operating system, and UNIX has been adopted. So DEC-20 Prolog will
not be relevant.
You should also, to be fair, point out that C-Prolog has also been
identified for providing Prolog capability.
-- Robert
------------------------------
Date: Sat 18 Jun 83 12:49:21-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Re: Prolog For The Vax
As a result of the paranoia induced by the Japanese 5th Generation
proposals, there was a lot of discussion about what the UK should do
to keep up with the foreign competition in AI and computing in
general. Eventually several government initiatives where started,
amounting to several 100 million dollars spread over five years or so.
In particular, the Science and Engineering Research Council (SERC),
whose closest US analogue is the NSF, started the Intelligent
Knowledge Based Systems initiative (IKBS), which is applied AI under a
different name (it seems the name "AI" is not very popular in UK
government and academic circles). Discussions sponsored by the IKBS
initiative have decided on a common software base, built around Unix
{a trademark of Bell Labs.}, Prolog (POPLOG and C-Prolog) and Lisp
(Franz). The machines to be used are VAXes and PERQs (the UK computer
company ICL builds PERQs under license, have implemented a derivative
of Unix on it, so this is a case of "support your local computer
manufacturer").
The fact that none of the systems mentioned above is nearly the ideal
for AI research is recognized by many of the UK researchers, but less
so by the administrators. Efforts to build a really efficient
portable compiler-based Prolog that would be for the new machines what
DEC-10/20 Prolog is for the machines it runs on have been hampered by
the sluggish response of The Bureaucrats, and by uncertainty about how
that huge amount of money was going to be allocated.
However, implementation of a portable compiler - based Prolog is now
going on at Edinburgh. Robert Rae is certainly in a better position
than I to describe how the project is progressing.
-- Fernando Pereira
------------------------------
Date: 21 June 1983 1023-EDT (Tuesday)
From: Vijay.Saraswat@CMU-CS-A (C410VS90)
Subject: Defining 'Equal' in Prolog
Here is a much shorter brain←teaser than the usual ones:
Define a predicate equal(X, Y) which is true iff X and Y are the
'same' terms up to renaming of variables (X and Y should not be
instantiated as a result of the call to equal). Note however that
\+ equal(foo(X,X), foo(Y,Z))
but of course:
equal(foo(X,Y),foo(D,O)).
Here is a 'quick and dirty' implementation:
eq(X,Y):-
assert(temp1(X)),
assert(temp2(Y)),
retract(temp1(A)),
retract(temp2(B)),
numbervars(A, 0,M),
!, numbervars(B, 0, M),
!, A == B.
------------------------------
Date: Monday, 20-Jun-83 22:29:19-BST
From: RICHARD HPS (on ERCC DEC-10) <okeefe.r.a.@edxa>
Reply-to: okeefe.r.a.%edxa%ucl-cs@isid
Subject: Operators and Prolog, a reply to Steve Hardy
Using Operators in Prolog.
As Steve Hardy says, using :-op to tell the Prolog operator
precedence parser about his operators gives him two problems (by
the way, please don't call syntax properties "modes"; confusion
between :-op declarations and :-mode declarations is something the
Prolog community can do without)
1) The parser accepts more things than he expected.
This is a general property of operator precedence parsing,
and is one of the reasons people bother about LL and LR
parsing. Possibly the only way around it would be to have
an extra "legality" check in the routine that bolts terms
together.
2) They do not ascribe the structures he expected to some terms.
He cites "if all x:happy(x) then arrived(utopia)".
It is important to realise that the Prolog reader is NOT
''mis-parsing'' [sic] these terms, it is doing exactly what
he told it to. The problem is that he couldn't tell it what
he really wanted. All he wants is to give ':' narrow left
scope and wide right scope, and almost all operator precedence
parsers would be let him have exactly that. This is NOT a
problem with operator precedence parsing.
In a fragment "A + B", where "+" indicates any binary operator,
there are five priorities to be considered:
OL -- the priority the Operator wants on its Left
AL -- the priority the form A actually has
OR -- the priority the Operator wants in its Right
AR -- the priority the form B actually has
OC -- the priority the Operator ascribes to the Combination
To be well formed, the fragment must satisfy AL <= OL & AR <= OR.
The priority of the whole will then be OC. As you see, there are
THREE priorities associated with the operator. Dec-10 Prolog and the
systems based on it impose the restriction that
OC-1 <= OL <= OC & OC-1 <= OR <= OC
This was largely in the interests of simplicity, so that people would
only have to bother about one number.
All the Prolog parsers known to me interface to the syntax
properties through a table
isop(+Operator, +Class, -OL, -OC, -OR)
where Class is prefix|infix|postfix. The representation depends
on whether the implementation language is Prolog, Pseudo-Prolog,
IMP, C, or PDP-11 assembler, but the basic idea is the same. The
parsers, therefore, are *already* able to cope with Steve Hardy's
example, and it is only a matter of generalising :-op suitably.
This need not involve anyone in system hackery. The Dec-10
Prolog tokeniser has been available to user programs for nearly
two years. If you look in the VCHECK program in the toolkit part
of the <PROLOG> directory at SRI-AI (see Fernando Pereira's
message in the Prolog digest about [V1 #8 -ed] that directory
and what is in it) you will find an example of its use. Beware:
it is only available to COMPILED code. Alan Mycroft of the Edinburgh
CS department keep asking for changes to the syntax of Prolog, and I
finally ripped out the 'read'/1 routine, made it work as a
user-level routine, tidied it up, added a few comments, and said
"here you are, you want it, you write it, and tell us about it
when you've done". He did in fact find one or two minor bugs in
the 'read'/1 predicate, and fixed them in this version, but
decided to live with the syntax as it was. This utility should
also be in <PROLOG>@SRI-AI; at Edinburgh the files are
RDTOK.PL -- interface to the tokeniser
NREAD.PL -- editable 'read'
NREAD is "semi-portable"; if you provide a tokeniser which
delivers the right sort of tokens, it doesn't require anything
that a good Prolog shouldn't already have. The one thing that
might be lacking in some (and is lacking in older versions of C
Prolog) is current←op, but that is precisely what Steve Hardy
needs to change.
Steve Hardy also has some critical comments about DCGs. He
says "DCGs have more of the flavour of a user-group library
routine than of a carefully thought out feature of the language."
In fact DCGs (or rather "grammaires de metamorphose") are the
reason why Prolog exists. DCGs are just a notational variant of
Horn clauses, nothing more. They were intended as a tool for
writing NATURAL language grammars, e.g. French, Spanish,
Portuguese, English &c, and have proved to be very good at it.
Writing a grammar that is driven off keywords isn't too hard.
Writing a grammar based on a large set of operators at several
different priority levels, especially in Prolog where the syntax
properties of an atom may change from one call to the next, is
very much harder. DCGs and operator precedence are different
solutions to the same parsing problem, and you can't expect
combining them to be any easier than combining LL and LR parsing
in the same parsing machine.
There is no great problem in getting error messages from a
DCG. The problem is in relating it to the original text, which
may never have existed. NREAD gets around the problem by saving
the length of the unparsed end, failing out, and then read's 2nd
alternative is to print the erroneous text. Given that DCGs use
lists, and do not work one token at a time, it is difficult to
see what better could be done. Can anyone tell me how you get
good error messages out of an ATN? Don't forget backtracking!
It is also agreed that more extensibility in the syntax of
Prolog would be a Good Thing. Distfix operators (e.g. all ← : ←
or if ← then ← else ←) are the most promising candidate, if only we
can find someone to adapt the HOPE parser. If anyone gets distfix
working, please do it by adapting NREAD.PL, please maintain the
property that an atom can be infix, prefix, postfix, any two, or
all three with possibly differing priorities, and please tell the
Prolog Digest about it when it works.
Steve Hardy's complaint "I'd like the internals of the Prolog
system to be more accessible, especially the routines for reading
and writing" is understandable, but years late (unless he is
referring to PopLog?). The tokeniser IS accessible in Dec-10
Prolog. {It isn't in C-Prolog, but it would be about a week's
work to make it accessible. I have a thesis to write. Volunteers?}
The parser IS available as a library file you can edit. In C Prolog,
though the parser and tokeniser aren't modifiable, you CAN apply
your own macro-processing to everything read, by adding your own
clauses to expand←term/2. {expand←term/2 in Dec-10 Prolog isn't
user-extensible, but could be made so if Dec-10 Prolog were still
being worked on.} The write routine, under the name 'print'/1, IS
extensible, just add clauses to 'portray'/1. (Again, some versions
of C Prolog have been released where print doesn't call portray at
each level. You can tell by looking at pl/init. If you have an
old version, ask for a new one.)
DCGs, by the way, are implemented by macro-expansion in
expand←term. They ARE a "one[2] page library procedure", and exactly
what DCGs are and how they are implemented is fully described in
the Clocksin and Mellish book. (Chris Mellish is at Sussex, and
wrote the Prolog component of PopLog.) The preprocessor in that
book actually contains a bug, and I recently submitted a corrected
and improved version of the whole thing to net.sources.
Pop-2, by the way, provides the analogue of yfx operators with
priorities 1..9 . Dec-10 WonderPop also provides 'form' declarations
form keyword {argument keyword}... ;
template
endform
where argument --> reading-function-name : argname
which lets different arguments be parsed by different routines. This
gives you a sort of distfix, and is quite painless to use. I do not
know what Pop-11 (the Pop part of PopLog) lets you do.
------------------------------
End of PROLOG Digest
********************
∂24-Jun-83 0954 RSC Prolog
I presume that the Prolog literature includes an exploration of the
connection between Prolog and non-monotonic logic. In fact, to the
extent that I understand circumscription, it seems directly applicable.
Comments?
Corky
∂24-Jun-83 1002 WIEDERHOLD@SUMEX-AIM Re: making life complicated for ourselves
Received: from SUMEX-AIM by SU-AI with PUP; 24-Jun-83 10:02 PDT
Date: Fri 24 Jun 83 10:02:14-PDT
From: Gio Wiederhold <WIEDERHOLD@SUMEX-AIM>
Subject: Re: making life complicated for ourselves
To: JMC@SAIL
In-Reply-To: Your message of Thu 23 Jun 83 18:24:00-PDT
You're right. The change has made ane comparison impossible and absolute
standards for programming projects are equally impossible to define. Gio
-------
∂24-Jun-83 1021 RPG Programming Ability Evaluation
To: JMC
CC: rpg-q
John, Lew asked me to write you a short evaluation of his programming
ability so that if any prospective employers were to ask you about it
you would be able to respond with an informed opinion.
Over the last 2 years Lew has produced about 10,000 lines of code.
His previous programming experience was limited, and was restricted
to the Interlisp language. Since the styles of programming natural to
these languages are quite different, I expected that he would have
trouble with the switchover. On the contrary, he managed to switch to
MacLisp very well, and fairly rapidly.
I would characterize his programming style as totally typical of
good, experienced MacLisp programmers. In fact, his ability to abstract
data structures, to integrate macros for control and data structuring, and
to use established techniques for complex data structures is above average
for Lisp programmers in the department.
His control structures are well thought out and are broken into small enough
pieces, through functions, to be understandable by people not familiar with
his code.
The most impressive part of his programming ability is his attention to
detail. In particular, his demonstration program for NFLT and his reasoning
system is extraordinarily well put together. Several excellent Lisp programmmers
to whom I demonstrated this program remarked favorably on this point.
Simply speaking, his code works with no lacunae.
A point that I cannot address well is that of programming speed. To contrast,
when I am programing well, I can produce 10,000 lines in 2 or 3 months. The key
issue is how much auxilliary work was he doing while coding. I know that he was
engaged in much philosophical work while programming. This work seemed necessary
since the code he was producing brought to light issues with his theoretical
underpinnings. I suspect that these activities intertwined in a way that
is difficult to separate for this sort of evaluation.
In short, I would not hesitate to recommend his as an accomplished Lisp
programmer on a par with the majority of Lisp programmers in the
department.
-rpg-
∂24-Jun-83 1349 ARK Comp Programming Project
To: Tucci@SU-SCORE, CT@SU-AI
CC: JMC@SU-AI, ARK@SU-AI, PW@SU-AI
As soon as you have a "final" version you want to hand in, please give it
to me. I will get Paul's approval, check it for "Computer Science"-y type
things that will make the Comp. Committee barf, and then pass it on for
John McCarthy's signature. Hopefully, then, the Comp. Committee will pass
it. Send all of us a message indicating when you will have it ready.
Thanks.
Arthur
∂24-Jun-83 1545 ATP.BLEDSOE@UTEXAS-20 ATP Volume - New information
Received: from UTEXAS-20 by SU-AI with TCP/SMTP; 24 Jun 83 15:44:43 PDT
Date: 24 Jun 1983 1743-CDT
From: Woody Bledsoe <ATP.Bledsoe@UTEXAS-20>
Subject: ATP Volume - New information
To: Peter.andrews@CMU-CS-A, cl.boyer@UTEXAS-20, dwl.duke@UDEL-RELAY,
lenat@SU-AI, jmc@SU-AI, gnelson@PARC-MAXC, wos@ANL
cc: ATP.Bledsoe@UTEXAS-20
Dear Authors:
(I will mail a copy of this to all the authors, including those
not on the ARPA net)
Here is a new development that I did not expect. I got the following
letter from Professor R.j.Milgram (Stanford), Chairman of the AMS
Publication Committee.
"Dear Professor Bledsoe:
This is in regard to your letter of December 14, 1982, in which you
asked if Comtemporary Mathematics would be interested in publishing the
Proceedings of your conference on Automatic Theorem Proving held at the
Denver meeting of the AMS. The referees have looked over the abstracts
and were very favorable. So, if the preceedings are indeed going to be
organized we would be happy to publish them.
I am sending a copy of this letter to R. O. Wells who is the new editor
of Contemporary Mathematics, and I am sure that he will be glad to handle any
further details and questions you may have.
Yours,
R. J. Milgram"
I still think we should go with AMS if we can. Frank Satlow at
MIT press has given us an almost sure OK, but I can still call that
off if I do it promptly. What do you think?
I have talked with Tony Palermino at AMS Headquarters, and he sees
no problem with going ahead with it. Ours would be a volume (Book) in
the Series. He is sending the pertinent specifications, and some of
their special paper that they require -- they require camera ready copy.
The may be able to take laser printer output that has been prepared (or
will be). I am to send him an example of such a laser-printed paper
for his opinion and decision. (I have asked Peter Andrews to do this).
I would verty much like your prompt reaction to this. What is
your preference?
If your reaction is positive, I will get to Saltow promply if we
are not going with MIT press.
Below are some remarks from Don Loveland which are pertinent.
Woody
-------
......
I am pleased that the AMS has endorsed our interest in a
volume to present the papers of the Denver session in hardcopy.
Publishing through the AMS instead of the MIT Press may well
better reach the audience that we aimed for, i.e. the
mathematicians. Let me raise one question that might favor
the MIT Press. I feel that most of the best current research
was summarized at that conference by leading researchers.
This overall summary may not occur again even if the same
researchers agree to speak at another award meeting in several years.
Thus this may be something of a "milestone" document in the
ATP field. Would the MIT Press give better promotion to the
volume, particularly to the computer science audience? Upon
reflection I think I would be happy with either and feel that
there will be little objection to the AMS route since we all
first believed that was to be the route anyway. I might simply
note that since resigning from the AMS I have not been aware
of their publications;they are not broadly advertised.
-------
∂25-Jun-83 0019 FEIGENBAUM@SUMEX-AIM copy of msg to Gabriel
Received: from SUMEX-AIM by SU-AI with PUP; 25-Jun-83 00:19 PDT
Date: Sat 25 Jun 83 00:19:40-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM>
Subject: copy of msg to Gabriel
To: jmc@SAIL
Dick,
If you have a proposal in to Ron, don't create turbulence in the flow. Just
let it go through as submitted.
In any event, the theme of our project is parallelism. It was on that theme
that John and I made contact--he said he had a paper on parallel LISP, and
we were about to propose such research, so it fit perfectly.
I'm off on a very long trip (5 weeks) so nothing is going to happen on our
proposal until August, probably September, anyway. let's talk when I
get back.
Best regards,
Ed
-------
∂26-Jun-83 0220 LLW@S1-A Wednesday
Received: from S1-A by SU-AI with TCP/SMTP; 26 Jun 83 02:20:42 PDT
Date: 26 Jun 83 0218 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Wednesday
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
John, Rod will be glad to see you on Wednesday, at least in part for lunar
colony paper work (and I will be also, to the extent that I'm around).
Lowell
∂26-Jun-83 1600 JMC*
paulso.pre
∂26-Jun-83 1734 RWW your part of the proposal
I have been told by penny to integrate your stuff (as opposed to
as appendices as we discussed) I am doing that now. If
this is objectionable to you, you had better send talk to Ed
and get it straight. I'm not sure what I'm supposed to do.
Richard
∂26-Jun-83 2354 HST ICAI in Karlsruhe
Do you come to Karlsruhe for IJCAI?Have you some time (and interest) to
wvisit Erlangen?I would like to invite you ...
herbert Stoyan
I will be at Karlsruhe, and it would be interesting to visit Erlangen.
Just before and just after the meeting are both feasible at present.
∂27-Jun-83 0003 HST karlsruhe
Both will be possible.When do you think will you decide the termin?
Let me tentatively make it the 15th, i.e. the Monday after IJCAI,
but coming on the weekend. I may want to change it later so let
me know when it should be definite.
∂27-Jun-83 1015 BSCOTT@SU-SCORE.ARPA NSF
Received: from SU-SCORE by SU-AI with TCP/SMTP; 27 Jun 83 10:14:50 PDT
Date: Mon 27 Jun 83 10:16:38-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: NSF
To: JMC@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
John, were you a P.I. on an NSF grant through the University of Oklahoma?
SPO has heard from NSF that you owe them a final report on grant ATM78-1240l,
with end date of 12/81.
Can you shed any light on this?
Betty
-------
I was never a P.I. on an NSF grant through the University of Oklahoma.
I think they have asked about this before with the same reply.
∂27-Jun-83 1223 PAPA@SU-SCORE.ARPA Going Away
Received: from SU-SCORE by SU-AI with TCP/SMTP; 27 Jun 83 12:23:20 PDT
Date: Mon 27 Jun 83 12:25:07-PDT
From: C. Papadimitriou <PAPA@SU-SCORE.ARPA>
Subject: Going Away
To: super@SU-SCORE.ARPA
I'll be away from June 28 to July 20.
I'll try to come back with a proposal draft.
Meanwhile, please don't say anything important on Thursdays.
-------
∂27-Jun-83 1658 RINDFLEISCH@SUMEX-AIM Architecture Proj Hardware
Received: from SUMEX-AIM by SU-AI with PUP; 27-Jun-83 16:58 PDT
Date: Mon 27 Jun 83 16:52:52-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Architecture Proj Hardware
To: JMC@SAIL, RPG@SAIL
cc: Rindfleisch@SUMEX-AIM, Nii@SUMEX-AIM, Delagi@SUMEX-AIM
John/Dick,
In planning the Architecture Project budget, I wanted to get your inputs
about hardware. As you know, DEC is proposing to make available a
32-element "parallel processing array" (PPA) of VLSI VAX's as a
prototyping laboratory 1-2 years into the project. There are 2 separate
questions: what do we need in the meantime and what happens if (when?)
the PPA is late. I've tentatively planned some workstations (2
Dandelions, 1 3600, and 1 Dorado) for year 1. Bruce Delagi (DEC) is
pushing for a VAX 11/784 that may be convertible for the PPA later or
could serve as a hedge against late delivery of the PPA. Could you make
use of a big VAX running VMS and whatever Lisp implementations exist
(DEC CommonLisp?)?
Tom R.
-------
∂27-Jun-83 2111 CLT susie
would like to converse with you.
∂27-Jun-83 2234 Bonnie%UPenn.UPenn@UDel-Relay AAAI executive board meeing
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 27 Jun 83 22:33:59 PDT
Date: Fri, 24 Jun 83 13:14 EDT
From: Bonnie Webber <Bonnie.UPenn@UDel-Relay>
Return-Path: <Bonnie%UPenn.UPenn@UDel-Relay>
Subject: AAAI executive board meeing
To: nilsson@sri-ai, mccarthy@su-ai, walker@sri-ai,
Bonnie Webber <bonnie.UPenn@UDel-Relay>
Via: UPenn; 28 Jun 83 1:31-EDT
I will not be able to attend the executive board meeting in August, as I
will be remaining in Europe for several weeks following IJCAI. If it's
acceptable to you, I would like to have Aravind Joshi take my place at
the meeting. One reason for this is that I would like to propose Philadelphia
as host for a future meeting. Before then I will have material prepared
and sent to you (as well as to the AAAI office) concerning hotel accommodations
and conference facilities that Center City Philadelphia and the university
have to offer. I am confident that we could have a very successful meeting
here.
Best regards, Bonnie
∂28-Jun-83 0015 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #12
Received: from SU-SCORE by SU-AI with TCP/SMTP; 28 Jun 83 00:14:59 PDT
Date: Monday, June 27, 1983 6:40AM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #12
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Tuesday, 28 Jun 1983 Volume 1 : Issue 12
Today's Topics:
Implementations - An Explanation,
LP Library Additions Available,
Job Opportunity at Edinburgh
----------------------------------------------------------------------
Date: Sun 26 Jun 83 23:22:07-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: Missing Explanation & Header
The gentle reader (yes, I believe you exist!) may be wondering why I
sent to this Digest a note on the UK IKBS & Prolog scene. Its purpose,
described in a header that somehow got lost (or that I forgot to
include) was to explain to an US audience some of the terms used by
Robert Rae in his note about various Prolog systems in use at
Edinburgh.
Sorry for any confusion I may have caused.
--Fernando Pereira
------------------------------
Date: 17 Jun 83 11:10:15-PDT (Fri)
From: harpo!floyd!vax135!ukc!edcaad!peter @ Ucb-Vax
Subject: Job: Prolog Graphics at Edinburgh
[Reprinted from the AIList]
UNIVERSITY OF EDINBURGH
COMPUTER AIDED DESIGN
RESEARCH WORKER
EdCAAD, the Edinburgh Computer Aided Architectural Design Research
Unit, is actively forging links between knowledge engineering and CAD,
focusing on the Prolog logic programming language. Recent advances at
EdCAAD include C-Prolog for 32-bit machines with C compilers and
Seelog, a graphics front end to Prolog. The Unit offers an excellent
computing environment as a leading UK UNIX site, with its own VAX
11/750, a PDP 11/24 and a large range of text and graphics terminals,
serving a small user community.
Current SERC supported research is aimed at building description tech-
niques, including drawing input with associated meaning attached to
drawings. This project has a vacancy for a research worker preferably
with AI experience. The research post is for an initial period of 18
months, on the research salary scale 1A, with placement according to
qualifications and experience.
Enquiries and applications should be addressed to Aart Bijl, EdCAAD,
Department of Architecture, University of Edinburgh, 20 Chambers
Street, Edinburgh EH1 1JZ, tel. 031 667 1011 ext. 4598.
------------------------------
Date: Sat 25 Jun 83 12:41:43-PDT
From: PEREIRA@SRI-AI.ARPA
Subject: More utilities
More Utilities from Richard O'Keefe
----------------------------------------------------------------------
From: RICHARD HPS (on ERCC DEC-10) <OKeefe.r.a.@EDXA>
Subject: Utilities
You should have
Multil.Pl (18-May)
Occur .Pl (22-May)
ListUt.Pl (20-May) -- replaces ListRo
MetUtl.Pl (20 May) -- from my working paper
Not .Hlp (21-May) -- debugging not/1 checks unbound vars
Not .Pl (21-May) -- Prolog code for not/1
OrdSet.Pl (22-May) -- ordered set utilities
Applic.Pl (28-May)
SetUtl.Pl (28-May)
...
Struct.Pl (1-June)
Some of these you have versions of, but they may be out of date. It
looks as though you only took things that were mentioned in UTIL, not
all the things that were in [140,143,Util].
..
You might find the "Advice Package" Advice.Pl useful; it implements an
InterLisp like "advise" command, e.g.
advise(fred(A,B,C), call, break).
-----------------------------------------------------------------------
I have placed all those files, plus the reader NREAD.PL and the
tokeniser interface RDTOK.PL mentioned in the Digest, in the usual
place, PS:<PROLOG> at SRI-AI.
--Fernando
------------------------------
End of PROLOG Digest
********************
∂28-Jun-83 1038 NILSSON@SRI-AI.ARPA Re: AAAI executive board meeing
Received: from SRI-AI by SU-AI with TCP/SMTP; 28 Jun 83 10:31:42 PDT
Date: 28 Jun 1983 0845-PDT
From: Nilsson at SRI-AI
Subject: Re: AAAI executive board meeing
To: Bonnie.UPenn at UDEL-RELAY, nilsson at SRI-AI,
To: mccarthy at SU-AI, walker at SRI-AI
cc: NILSSON
Ok, Bonnie, we'd be glad to have Aravind attend the meeting in your
place. -Nils
-------
∂28-Jun-83 1042 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 28 Jun 83 10:42:15 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 28 Jun 83 08:55:55-PDT
Date: 28 Jun 1983 0851-PDT
From: Nilsson at SRI-AI
Subject: meeting
To: jmc at SU-SCORE
cc: nilsson
Rick H-R, Bob Engelmore and I are having lunch this Friday,
July 1, at noon at Alouette (401 Lytton Ave., P.A.) to discuss
their concern about AAAI being "too academic." Can you join us?
If so, we'll meet you there. -Nils
-------
∂28-Jun-83 1322 DFH phone messages
1. Your daughter Sarah needs to get ahold of you.
2. Mort Labrecque of Mosaic Magazine (magazine of NSF)
wants to talk with you about article he is doing on
Fifth Generation Computer Architecture. Phone is
(212) 675-8164, day or evening.
∂28-Jun-83 1517 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA AAAI Executive Director
Received: from SU-SCORE by SU-AI with TCP/SMTP; 28 Jun 83 15:17:19 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 28 Jun 83 15:19:00-PDT
Date: 28 Jun 1983 1446-PDT
From: Nilsson at SRI-AI
Subject: AAAI Executive Director
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: stan at SRI-AI, gjs at MIT-MC, tenenbaum at SRI-KL,
To: walker at SRI-AI, dwaltz at BBNG, bonnie.upenn at UDEL
Since Lou Robinson's announcement of his intention to resign as AAAI
Executive Director effective July 1, 1983, Ms. Claudia Mazzetti has
been doing an extremely effective job of substituting for Lou. She
is carrying through with all arrangements for the annual meeting in
August, she is coordinating the AAAI office role in the publication
of the AI magazine, and she is handling all the other details that
occur in the running of a national scientific society. In my opinion
we are extremely fortunate to have Claudia fill this role. I
think it quite appropriate therefore, that we appoint Claudia to
succeed Lou as AAAI Executive Director, effective July 1, and I
hereby recommend that we do so. I will presume that, hearing no
strong arguments to the contrary, the AAAI Council concurs. I
recommend that the financial committee (Don Walker, Raj Reddy, and
myself) establish an appropriate salary.
--Nils
-------
∂28-Jun-83 1550 @SU-SCORE.ARPA:BOBROW.PA@PARC-MAXC.ARPA Re: AAAI Executive Director
Received: from SU-SCORE by SU-AI with TCP/SMTP; 28 Jun 83 15:49:44 PDT
Received: from PARC-MAXC.ARPA by SU-SCORE.ARPA with TCP; Tue 28 Jun 83 15:51:22-PDT
Date: Tue, 28 Jun 83 15:48 PDT
From: BOBROW.PA@PARC-MAXC.ARPA
Subject: Re: AAAI Executive Director
In-reply-to: "Nilsson@SRI-AI.ARPA's message of 28 Jun 83 14:46 PDT"
To: Nilsson@SRI-AI.ARPA
cc: Amarel@RUTGERS.ARPA, bobrow.PA@PARC-MAXC.ARPA, buchanan@SUMEX.ARPA,
BEngelmore@SRI-KL.ARPA, LErman@SRI-KL.ARPA, csd.genesereth@SCORE.ARPA,
grosz@SRI-AI.ARPA, hart@SRI-KL.ARPA, csd.mccarthy@SCORE.ARPA,
mcdermot@YALE.ARPA, minsky@MIT-MC.ARPA, reddy@CMU-10A.ARPA,
rich@MIT-MC.ARPA, stan@SRI-AI.ARPA, gjs@MIT-MC.ARPA,
tenenbaum@SRI-KL.ARPA, walker@SRI-AI.ARPA, dwaltz@BBNG.ARPA,
bonnie.upenn@UDEL.ARPA
I will not allow silence to stand for affirmation, so I want you to know
that I strongly approve of your recommendation. I have had a number of
dealings with Claudia Mazzetti and have found her quite effective in her
role. I think we are quite lucky to have someone so competent already
familiar with AAAI affairs.
danny
∂28-Jun-83 1639 DFH phone message
Dan Robinson, of Science Applications called regarding
possible consulting work. (408) 734-4162.
∂28-Jun-83 2242 JBR@S1-A
Received: from S1-A by SU-AI with TCP/SMTP; 28 Jun 83 22:42:26 PDT
Date: 28 Jun 83 2242 PDT
From: Jeff Rubin <JBR@S1-A>
To: jmc@SU-AI
I saw that in BBOARD. I doubt if my technical Cantonese is good enough, though.
∂29-Jun-83 1618 DFH phone message
Brian Randall from University of Newcastle
called this morning. He is very anxious to get
the title and abstract of your lecture you are
giving at their seminar in Sept. In fact, he
wants us to telex it.
∂29-Jun-83 2016 CLT
if you get home before 11ish call sarah
ow she will call you in the morning
∂30-Jun-83 0700 JMC*
$1766 TO DAVIDSON AUTO SALES
∂30-Jun-83 0750 CLT matadero shell
unless they've changed it the number is 493-1250
∂30-Jun-83 1100 DFH ARPA budget
To: JMC, ZM
When budget was finally sent in, an amount was
included for foreign travel with no detail as
to what specific conferences, etc., it was to be
used for. Betty now needs to have this detail.
Could you let me know name, location, and duration
of the conferences and/or research collaboration
you want listed. --Diana
Please find out from Betty how much foreign travel we asked for. Make
a budget dividing it evenly between importing Keith Clark from
Imperial College in London
Masahiko Sato from Tokyo University, my attending a meeting of IFIP
working group 2.2 in Europe (location unknown at present) and somone
attending AISB (artificial intelligence and simulation of behavior
conference in Europe). This budget is pro forma only as I understand
it.
∂30-Jun-83 1400 JMC*
walsh.xgp
∂30-Jun-83 1604 @SU-SCORE.ARPA,@MIT-MC:MINSKY%MIT-OZ@MIT-MC AAAI Executive Director
Received: from SU-SCORE by SU-AI with TCP/SMTP; 30 Jun 83 16:04:07 PDT
Received: from MIT-MC by SU-SCORE.ARPA with TCP; Thu 30 Jun 83 16:09:42-PDT
Date: Thu, 30 Jun 1983 16:29 EDT
From: MINSKY%MIT-OZ@MIT-MC
To: Nilsson@SRI-AI
Cc: Amarel@RUTGERS, BEngelmore@SRI-KL, bobrow@PARC-MAXC,
bonnie.upenn@UDEL-RELAY, buchanan@SUMEX-AIM, csd.genesereth@SU-SCORE,
csd.mccarthy@SU-SCORE, dwaltz@BBNG, gjs@MIT-MC, grosz@SRI-AI, hart@SRI-KL,
LErman@SRI-KL, mcdermot@YALE, reddy@CMU-CS-A, rich@MIT-MC, stan@SRI-AI,
tenenbaum@SRI-KL, walker@SRI-AI
Subject: AAAI Executive Director
In-reply-to: Msg of 28 Jun 1983 17:46-EDT from Nilsson at SRI-AI
Claudia sounds good to me.
∂30-Jun-83 1618 AAAI-OFFICE@SUMEX-AIM COnference accomodations
Received: from SUMEX-AIM by SU-AI with PUP; 30-Jun-83 16:18 PDT
Date: Thu 30 Jun 83 16:18:21-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM>
Subject: COnference accomodations
To: Nilsson@SRI-AI.ARPA
cc: Genesereth@SUMEX-AIM, Reddy@CMU-CS-A.ARPA, JMC@SAIL
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
The Washington Hilton offers VIP treatment for officers of
organizations attending conferences in their hotel. VIP
treatment from what I could gather means - special check-in and
out, larger rooms, comp continential breakfasts. If you desire
this special treatment, I will need your arrival/departure dates
and times, accodation requirements (single,double), and food
and beverage preferences. If I don't hear from you soon, I can
assume you do not want this special treatment.
Cheers,
Claudia
-------
I would like a single from Sunday night through Friday night. I have
no special food or beverage preferences.
∂30-Jun-83 2000 JMC*
No food after 10.
∂30-Jun-83 2324 HST AFTER IJCAI
THE TIME IS FINE.THE ONLY PROBLEM IS,THE UNI IS EMPTY AT THIS TIME
AND A TALK TO MAKE MONEY WILL NOT BE POSSIBLE.THAT MENAS(MEANS) YOU
WILL BE MY PRIVATE GUEST.IF YOU CAN PAY A HOTEL FORYOURSELF THIS
COULD BE POSSIBLE OR YOU COULD SLEEP IN ONE ROOM IN OUR FLAT.ARE
YOU INTERESTED IN TRAVELING OR WALKING,SEEING OLD CASTLES AND TOWNS?
ERLANGEN IS VERY NEAR TO NUERMBERG.
IS THAT CONVENIENT?HERBERT STOYAN
That would be fine, but in that case it would be best to make it
the weekend after IJCAI. Walking and seeing old castles would
be very interesting, and I thank you much for the invitation.
If it makes a difference, just previous to IJCAI is also possible.
∂01-Jul-83 0129 ARK Facilities Committee
Agenda Items
1) Replacement of Ralph
2) Subcommittee to discuss details of Dover gateway
3) CF actions like incompatible SCORE modem replacements
Thanks.
Arthur
∂01-Jul-83 1204 ullman%SU-HNV.ARPA@SU-SCORE.ARPA conferences
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Jul 83 12:04:07 PDT
Received: from Diablo by Score with Pup; Fri 1 Jul 83 12:05:57-PDT
Date: Fri, 1 Jul 83 12:05 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: conferences
To: super@score
I recently got notices of two conferences that may be of interest.
1. "Frontiers of Supercomputing" Aug 15-19, Los Alamos.
2. "Intl. Conf. on Parallel Processing" Aug. 23-26, Bellaire Mich.
I have prospecti if anyone is interested.
∂01-Jul-83 1207 ullman@Diablo parallel LISP
Received: from SU-HNV by SU-AI with PUP; 01-Jul-83 12:07 PDT
Date: Fri, 1 Jul 83 12:09 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: parallel LISP
To: jmc@sail, rpg@sail
Would you guys be interested in starting a discussion of parallelization
of LISP next Thursday at the 3PM meeting?
∂01-Jul-83 1210 ullman@Diablo Weyrauch
Received: from SU-HNV by SU-AI with PUP; 01-Jul-83 12:09 PDT
Date: Fri, 1 Jul 83 12:11 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Weyrauch
To: jmc@sail
I didn;t want to give the impression that I wanted to exclude Weyrauch
from the supercomputer group. For me to invite him would probably
be construded by Ed as meddling, and I am not anxious to go down that
road again. However, if you are convinced that it would be appropriate
to have him in the group, I'm not trying to exclude him or anybody else.
I be glad to talk about parallelization of LISP next Thursday, and
I'll invite Weyhrauch. Weyrauch I don't know.
It won't take long to tell the little I know, so there will be
plenty of time for discussion.
Abstract: Consider parallel processors (e.g. S-1) all looking at
a single large memory. There are at least two approaches to using
such a system for a single problem. In one a compiler finds opportunities
for parallelism, either because the parallelism is indicated or by
cleverness, and produces a program that explicitly uses a certain
number of processors. In our approach, however, the compiler produces
code that executes forks by starting on one task while putting a
list of the others on a queue. When processors become free they
take tasks from the queue. Not even the object program knows what
processors will execute what code or even how many processors are
available. We consider language extensions to LISP for allowing
and controlling such queue based parallel processing. Not much
has been done, so the presentation will be very short. The files
MULTI[S83,JMC]@SAIL and MULTIP[S83,JMC]@SAIL contain some preliminary
studies emphasizing the S-1.
∂01-Jul-83 1347 jlh@Shasta Re: Equipment Contract
Received: from SU-SHASTA by SU-AI with PUP; 01-Jul-83 13:47 PDT
Date: Friday, 1 Jul 1983 13:49-PDT
To: T. C. Rindfleisch <Rindfleisch at SUMEX-AIM>
Cc: ullman at Diablo, Bosack at Score, CSL.JLH at Score,
Feigenbaum at SUMEX-AIM, JMC at SAIL, REG at SAIL, RPG at SAIL,
CSD.BScott at Score
Subject: Re: Equipment Contract
In-reply-to: Your message of Wed 22 Jun 83 22:06:07-PDT.
From: John Hennessy <jlh@Shasta>
We did promise Gio's group the equivalent of 1 750, although we had
planned to store it with the rest. As far as Sun's go, I believe that
the bulk of the CSD general money is allocated for McSuns. I p[resume
that the systems group (me, Reid, Cheriton, and Lantz) will come up
with an appropriate distribution of system Sun's between ERL and MJH.
∂01-Jul-83 1415 ullman@Diablo Re: Parallelizing LISP
Received: from SU-HNV by SU-AI with PUP; 01-Jul-83 14:15 PDT
Date: Fri, 1 Jul 83 14:17 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Re: Parallelizing LISP
To: JMC@Sail
Cc: RPG@Sail
Thanks. By the way, this group is much more informal than you
might imagine. I'll forward your abstract to super@score, but
no one has ever thought of providing one before. What I'm hoping
will happen is more of a "round table" discussion, with you
providing focus when you feel it necessary.
∂01-Jul-83 1427 ullman@Diablo Re: Equipment Contract
Received: from SU-HNV by SU-AI with PUP; 01-Jul-83 14:27 PDT
Date: Fri, 1 Jul 83 14:28 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Re: Equipment Contract
To: Rindfleisch@Sumex-Aim, jlh@Shasta
Cc: Bosack@SU-Score, CSD.BScott@SU-Score, CSL.JLH@SU-Score,
Feigenbaum@Sumex-Aim, JMC@Sail, REG@Sail, RPG@Sail
I see no reason not to store GIo's 750 with the rest; I recollect he
asked for the console in his office.
I tried to trace the history of the 50 SUN's. The earliest record
I have shows 25 for ERL and 10 "general PI use".
Then there was a discussion with Kahn in which he suggested the number
be raised to 50. My recollection was that the terminals were to
be spread around, including about 5 for Newmath. I also recollect that
there was a distinction between the ERL SUN's and the others:
ERL SUNs have 1meg of memory, the others 0.25meg.
The actual proposal, due to a formatting bug, does not indicate the
disposition of the SUN's. Since the 50 have now become 30, can we
reach some equitable arrangement on their locations?
If we assume 5 were for Newmath (which we shall conveniently forget about),
then the original order was for ERL:MJH in the ratio 5:4.
On the other hand, the ERL machines were more expensive, so on a dollar
basis it was perhaps 60% ERL.
Can we divide the things 18 for ERL, 12 for MJH?
∂01-Jul-83 1430 ullman%SU-HNV.ARPA@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Jul 83 14:30:04 PDT
Received: from Diablo by Score with Pup; Fri 1 Jul 83 14:31:28-PDT
Date: Fri, 1 Jul 83 14:31 PDT
From: Jeff Ullman <ullman@Diablo>
To: super@score
John McCarthy has agreed to lead a discussion on parallelization of
LISP next Thursday. We'll meet 3PM in the CIS conference room, as
usual.
Abstract: Consider parallel processors (e.g. S-1) all looking at
a single large memory. There are at least two approaches to using
such a system for a single problem. In one a compiler finds opportunities
for parallelism, either because the parallelism is indicated or by
cleverness, and produces a program that explicitly uses a certain
number of processors. In our approach, however, the compiler produces
code that executes forks by starting on one task while putting a
list of the others on a queue. When processors become free they
take tasks from the queue. Not even the object program knows what
processors will execute what code or even how many processors are
available. We consider language extensions to LISP for allowing
and controlling such queue based parallel processing. Not much
has been done, so the presentation will be very short. The files
MULTI[S83,JMC]@SAIL and MULTIP[S83,JMC]@SAIL contain some preliminary
studies emphasizing the S-1.
∂01-Jul-83 1507 lantz%SU-HNV.ARPA@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Jul 83 15:07:43 PDT
Received: from Diablo by Score with Pup; Fri 1 Jul 83 15:09:25-PDT
Date: Friday, 1 Jul 1983 15:09-PDT
To: ullman at Diablo
Cc: super at Score
In-reply-to: Your message of Fri, 1 Jul 83 14:31 PDT.
From: Keith Lantz <lantz@Diablo>
I though Werner was speaking next Thursday?
∂01-Jul-83 1528 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Werner finked out. Wants to wait a few weeks.
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Jul 83 15:28:40 PDT
Received: from Diablo by Score with Pup; Fri 1 Jul 83 15:29:39-PDT
Date: Fri, 1 Jul 83 15:29 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Werner finked out. Wants to wait a few weeks.
To: lantz
Cc: super@SU-Score
∂02-Jul-83 1133 BSCOTT@SU-SCORE.ARPA Final Report
Received: from SU-SCORE by SU-AI with TCP/SMTP; 2 Jul 83 11:33:50 PDT
Date: Sat 2 Jul 83 11:35:50-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Final Report
To: JMC@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
This is a reminder that the final report on ARPA Contract MDA903-80-C-0102
is past due. It was extended because of Keith's work, but the contract
did end on 3/31/83.
Betty
-------
∂02-Jul-83 1956 JMC