perm filename E77.OUT[LET,JMC]1 blob
sn#307078 filedate 1977-09-28 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00447 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00042 00002 ∂09-May-77 0953 MJL ai qual
C00043 00003 ∂09-May-77 1107 FTP:KAY at PARC-MAXC Scandinavian Summer School
C00045 00004 ∂09-May-77 1602 LES YTD
C00047 00005 ∂10-May-77 0008 FTP:MRC at MIT-AI (Mark Crispin )
C00050 00006 ∂10-May-77 0048 JAB
C00052 00007 ∂10-May-77 1057 PAT aiqual
C00053 00008 ∂10-May-77 1258 FTP:Taynai at SUMEX-AIM Borrowed Book
C00054 00009 ∂10-May-77 1455 ZM
C00055 00010 ∂10-May-77 1620 MJL ai qual
C00056 00011 ∂10-May-77 2229 FTP:MRC at MIT-AI (Mark Crispin ) Weizenbaum's hackers
C00058 00012 ∂10-May-77 2245 DCL
C00059 00013 ∂11-May-77 1314 JBR
C00060 00014 ∂11-May-77 1444 DPB News concerning MTC Qual for Fall 1977.
C00061 00015 ∂12-May-77 1322 FTP:Wilber at SRI-KL LISP PROTOHISTORY
C00064 00016 ∂13-May-77 0918 MJL dept. meeting
C00065 00017 ∂13-May-77 1057 FTP:Hart at SRI-KL Job openings
C00066 00018 ∂13-May-77 1227 LES Foreign journalists tour
C00067 00019 ∂14-May-77 0433 JAB meeting
C00068 00020 ∂15-May-77 0613 FTP:PRATT at MIT-AI (Vaughan Pratt )y )
C00070 00021 ∂16-May-77 1515 PAM
C00071 00022 ∂16-May-77 1616 MJL autumn 1977
C00072 00023 ∂16-May-77 1740 DEW Thesis research
C00076 00024 ∂15-May-77 2323 FTP:STEF at SRI-KA
C00079 00025 ∂17-May-77 1115 MJL ai qual
C00080 00026 ∂18-May-77 0310 FTP:Wilber at SRI-KA LISP ARTICLES
C00082 00027 ∂18-May-77 0644 RDR via AMET
C00083 00028 ∂18-May-77 1200 MDD*
C00084 00029 ∂18-May-77 1248 FTP:Wilber at SRI-KL refereeing
C00085 00030 ∂18-May-77 1649 ZM MTC QUALS
C00086 00031 ∂18-May-77 2046 DCO
C00092 00032 ∂18-May-77 2047 DCO
C00093 00033 ∂18-May-77 2110 FTP:Wilber at SRI-KL LISP PAPERS
C00094 00034 ∂19-May-77 0025 REM via AMET Attempts to get DIALNE[W77,JMC] to agree with GEOFF/STEF version
C00097 00035 ∂19-May-77 0118 JP running E from maclsp
C00098 00036 ∂19-May-77 0122 ZM
C00099 00037 ∂19-May-77 0739 RDR via AMET LOTS
C00103 00038 ∂19-May-77 0755 FTP:Lots at SRI-KL the real point
C00105 00039 ∂19-May-77 1000 JMC*
C00106 00040 ∂19-May-77 1142 FTP:Geoff at SRI-KA Dialnet papers.
C00108 00041 ∂20-May-77 1505 FTP:BOBROW at PARC-MAXC IJCAI panel
C00109 00042 ∂20-May-77 2053 MJC
C00110 00043 ∂20-May-77 2056 ARG
C00112 00044 ∂20-May-77 2055 PMF
C00115 00045 ∂20-May-77 2058 BZM
C00116 00046 ∂20-May-77 2058 EJG
C00117 00047 ∂20-May-77 2100 WD
C00118 00048 ∂21-May-77 1526 ME
C00121 00049 ∂22-May-77 0328 FTP:FARBER at USC-ISI The current state of ''ai'' according to Consumers
C00124 00050 ∂22-May-77 1615 JAB Dialnet
C00125 00051
C00126 00052 ∂23-May-77 1358 JAB
C00127 00053 ∂23-May-77 1534 DEA Course Evaluations
C00128 00054 ∂23-May-77 1717 JBR
C00147 00055 ∂24-May-77 1058 RDR via TYMT cache
C00148 00056 ∂24-May-77 1336 RCM borrowed book
C00149 00057 ∂24-May-77 1613 DEA Course Evaluation
C00150 00058 ∂24-May-77 1637 LES Satellite communication meeting at MIT
C00158 00059 ∂25-May-77 1138 MJL
C00162 00060 ∂25-May-77 1939 PMF
C00163 00061 ∂25-May-77 2009 JBR E's buffer size
C00164 00062 ∂26-May-77 0615 JRA stoyan
C00165 00063 ∂26-May-77 0949 ME
C00167 00064 ∂26-May-77 1357 FTP:GLS at MIT-AI (Guy L. Steele, Jr. ) Visit by GLS to SAIL
C00170 00065 ∂26-May-77 1535 FTP:GLS at MIT-AI (Guy L. Steele, Jr. ) Visit
C00171 00066 ∂26-May-77 1926 FTP:Wilber at SRI-KL A REMINDER
C00174 00067 ∂26-May-77 2252 RPG
C00175 00068 ∂27-May-77 0933 JAB Dialnet
C00176 00069 ∂27-May-77 0957 FTP: RAJ REDDY(A610RR29) at CMU-10B IJCAI INVITED PAPER
C00177 00070 ∂27-May-77 1759 JAB via AMET Dialnet meeting
C00178 00071 ∂28-May-77 1208 TOB
C00179 00072 ∂29-May-77 1740 DCL FREELOADERS and HANGERS-ON
C00180 00073 ∂29-May-77 2300 RAK John Borcheck
C00181 00074 ∂29-May-77 2356 JAB
C00182 00075 ∂30-May-77 0022 JAB
C00183 00076 ∂30-May-77 0204 JAB logout and deleting empty ufds
C00184 00077 ∂30-May-77 1224 JBR
C00185 00078 ∂02-Jun-77 0433 DEW CSD Black Tuesday
C00186 00079 ∂30-May-77 1705 DON
C00187 00080 ∂30-May-77 1943 JP
C00188 00081 ∂30-May-77 2159 FTP:EGK at MIT-MC (Edward Kochanowski )
C00191 00082 ∂30-May-77 2347 100 : don adventure game
C00193 00083 ∂31-May-77 0155 JBR
C00194 00084 ∂31-May-77 0755 FTP:JM at MIT-MC (Joel Moses )
C00195 00085 ∂31-May-77 1421 RDR via AMET CS105/106
C00197 00086 ∂31-May-77 1514 FTP:Wiederhold at SUMEX-AIM Re: CS105/106
C00200 00087 ∂31-May-77 1923 FTP:JAB at MIT-ML (John A. Borchek ) egk
C00202 00088 ∂31-May-77 2016 FTP:JAB at MIT-ML (John A. Borchek )
C00203 00089 ∂31-May-77 2021 RAK Another friend of Borchek's
C00205 00090 ∂31-May-77 2120 FTP:JAB at MIT-ML (John A. Borchek )
C00208 00091 ∂31-May-77 2143 JAB via ML
C00210 00092 ∂01-Jun-77 0856 FTP:Wiederhold at SUMEX-AIM (Response to message)
C00212 00093 ∂01-Jun-77 0900 FTP:Wiederhold at SUMEX-AIM (Response to message)
C00215 00094 ∂01-Jun-77 0957 FTP:Boyer at SRI-KL Talk by Woody Bledsoe
C00217 00095 ∂01-Jun-77 1014 FTP:Wiederhold at SUMEX-AIM sail speed
C00218 00096 ∂01-Jun-77 1141 LES
C00219 00097 ∂01-Jun-77 1228 PAM
C00220 00098 ∂02-Jun-77 0336 DON
C00222 00099 ∂02-Jun-77 2143 FTP:GLS at MIT-AI (Guy L. Steele, Jr. ) LISP Historian from Germany
C00224 00100 ∂02-Jun-77 2041 RDR via AMET charging for use of LOTS
C00227 00101 ∂02-Jun-77 2154 RDR via AMET last point
C00230 00102 ∂03-Jun-77 0911 FTP: RAJ REDDY(A610RR29) at CMU-10B Invited Paper
C00233 00103 ∂03-Jun-77 1017 FTP:GLS at MIT-AI (Guy L. Steele, Jr. )
C00234 00104 ∂03-Jun-77 1103 JP New ET.FAS on MAC,LSP
C00236 00105 ∂03-Jun-77 1629 JB My Progress Report (in view of Black Friday on Tuesday!).
C00237 00106 ∂04-Jun-77 1100 JMC*
C00238 00107 ∂04-Jun-77 2050 DCO
C00240 00108 ∂04-Jun-77 2339 RDR via AMET cheap 103 modems
C00242 00109 ∂05-Jun-77 0037 DPB DIStribution for discussion on choice of language in Intro courses
C00243 00110 ∂05-Jun-77 0044 DPB MAINSAIL for Intro Courses?
C00244 00111 ∂06-Jun-77 0610 FTP:Wilber at SRI-KL CHEAP 103 MODEMS
C00246 00112 ∂06-Jun-77 1131 PAM
C00248 00113 ∂06-Jun-77 1254 PAM
C00249 00114 ∂06-Jun-77 1639 FTP:EF at MIT-AI (Edward Fredkin )
C00250 00115 ∂06-Jun-77 1641 FTP:EF at MIT-AI (Edward Fredkin )
C00251 00116 ∂07-Jun-77 1122 JAB via AMET
C00253 00117 ∂07-Jun-77 1128 LES
C00254 00118 ∂07-Jun-77 1443 LES Tuition for PAT
C00255 00119 ∂07-Jun-77 1449 FTP:JAB at MIT-MC (John A. Borchek ) DEC
C00257 00120 ∂07-Jun-77 1521 RPG New Maclisp
C00258 00121 ∂08-Jun-77 0948 DCL MONITOR LIMIT ON JOB SIZE
C00259 00122 ∂08-Jun-77 1210 JBR
C00260 00123 ∂08-Jun-77 1247 LES ARPA approval
C00261 00124 ∂08-Jun-77 2101 REF Core war
C00262 00125 ∂09-Jun-77 1022 TOB mech eng project
C00263 00126 ∂09-Jun-77 1025 RWW NSF PROPOSAL
C00264 00127 ∂09-Jun-77 1415 RPG sewer OLDIO
C00266 00128 ∂09-Jun-77 1431 FTP:EF at MIT-AI (Edward Fredkin )
C00268 00129 ∂09-Jun-77 2316 CLT Assistantship
C00269 00130 ∂10-Jun-77 1145 JP E - MACLSP - E
C00270 00131 ∂10-Jun-77 1617 FTP:Feigenbaum at SUMEX-AIM TIP to Stanford
C00271 00132 ∂11-Jun-77 1745 RAK The appearance of evil
C00272 00133 ∂12-Jun-77 1331 JAB via AMET networks
C00273 00134 ∂13-Jun-77 0031 RAK Re--Aint broke
C00274 00135 ∂13-Jun-77 1013 FTP: RAJ REDDY(A610RR29) at CMU-10B IJCAI-PAPER BY YOU AND HAYASHI et al.
C00276 00136 ∂13-Jun-77 1050 JMC
C00277 00137 ∂14-Jun-77 0003 DPB MTC syllabus
C00278 00138 ∂14-Jun-77 1035 DCL
C00279 00139 ∂14-Jun-77 1121 JAB via AMET
C00281 00140 ∂14-Jun-77 1254 EJG Core War
C00282 00141 ∂15-Jun-77 1104 PAT article
C00283 00142 ∂15-Jun-77 1322 WP
C00284 00143 ∂15-Jun-77 2056 RWW whats up
C00285 00144 ∂16-Jun-77 1001 JMC*
C00286 00145 ∂16-Jun-77 1001 JP another freeloader?
C00287 00146 ∂16-Jun-77 1016 RCM
C00289 00147 ∂16-Jun-77 1221 ROZ
C00290 00148 ∂16-Jun-77 1253 RWW nsf
C00291 00149 ∂16-Jun-77 1643 PAM
C00292 00150 ∂17-Jun-77 1237 FTP:GLS at MIT-AI (Guy L. Steele, Jr. )
C00293 00151 ∂17-Jun-77 1333 PAM
C00294 00152 ∂19-Jun-77 0354 RDR via AMET a dialnet thought
C00295 00153 ∂20-Jun-77 0026 GFF
C00297 00154 ∂20-Jun-77 1026 DEW My next few months work
C00298 00155 ∂21-Jun-77 1312 PAT phone call
C00299 00156 ∂22-Jun-77 0810 FTP:TK at MIT-AI (Tom Knight ) Dialnet and Mark Crispin
C00300 00157 ∂22-Jun-77 1023 FTP:PRATT at MIT-AI (Vaughan Pratt )
C00310 00158 ∂22-Jun-77 2213 MWK A banquet
C00311 00159 ∂23-Jun-77 0153 FTP:GEOFF at SRI-KA [Steve Caine <Farber at sri-kl>: A ''Pocket sized terminal'' from UNILEVER]
C00314 00160 ∂23-Jun-77 0835 FTP:PRATT at MIT-AI (Vaughan Pratt ) LISP article for the Belzer et al Encyclopaedia.
C00316 00161 ∂23-Jun-77 1101 FTP:PRATT at MIT-AI (Vaughan Pratt )
C00318 00162 ∂23-Jun-77 1246 FTP:PRATT at MIT-AI (Vaughan Pratt )
C00319 00163 ∂23-Jun-77 1254 FTP:PRATT at MIT-AI (Vaughan Pratt )
C00321 00164 ∂23-Jun-77 1532 TJW ORALS
C00322 00165 ∂24-Jun-77 1634 LES
C00323 00166 ∂25-Jun-77 0919 FTP: Ijcai-77 at CMU-10B IJCAI invited paper.
C00325 00167 ∂25-Jun-77 1155 FTP:PBARAN at USC-ISI (Response to message)
C00326 00168 ∂25-Jun-77 1239 FTP:REM at MIT-MC (Robert Elton Maas) Compatibility between dialnet and pcnet, discuss at meeting
C00329 00169 ∂25-Jun-77 1515 FTP:GLS at MIT-AI (Guy L. Steele, Jr. )
C00332 00170 ∂25-Jun-77 2342 DON
C00333 00171 ∂26-Jun-77 0013 NS
C00334 00172 ∂26-Jun-77 0926 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00346 00173 Carolyn: Here are some typos and other comments from Vaughan Pratt at
C00353 00174 ∂26-Jun-77 1759 FTP: IJCAI-77(X355IJ00) at CMU-10B IJCAI contributed paper.
C00356 00175 ∂27-Jun-77 0915 BPM OK, everybody...
C00357 00176 ∂27-Jun-77 1501 MDD FUTURE OF PROVER
C00359 00177 ∂27-Jun-77 2247 FTP:DanG at SRI-KL (Dan Gerson) Dialnet opening
C00360 00178 ∂27-Jun-77 2300 RDR via AMET
C00361 00179 ∂27-Jun-77 2324 DON Quick adventure
C00362 00180 ∂28-Jun-77 0401 FTP: Ijcai-77 at CMU-10B Mailing of your invited paper.
C00364 00181 ∂28-Jun-77 1249 RCM your IJCAI paper
C00366 00182 ∂28-Jun-77 1443 MJL
C00369 00183 ∂28-Jun-77 1505 MJL
C00370 00184 ∂28-Jun-77 1511 MJL
C00371 00185 ∂28-Jun-77 1521 MJL
C00372 00186 ∂28-Jun-77 2056 JMC at TTY15 2056
C00373 00187 ∂29-Jun-77 2247 FTP:Geoff at SRI-KL Password.
C00374 00188 ∂30-Jun-77 0646 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00376 00189 ∂30-Jun-77 1125 FTP:PBARAN at USC-ISI PCNET 27 June mtng minutes
C00377 00190 ∂30-Jun-77 1351 DCL first order equivalents
C00379 00191 ∂30-Jun-77 1629 FTP:GLS at MIT-AI (Guy L. Steele, Jr.)
C00380 00192 ∂01-Jul-77 0624 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00382 00193 ∂01-Jul-77 0635 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00384 00194 ∂01-Jul-77 2212 DCO MAIL in MACLSP
C00386 00195 ∂01-Jul-77 2229 DCO Reasoning about List Structure
C00388 00196 ∂01-Jul-77 2338 BPM [SRI-KL]<SU-AI>
C00389 00197 ∂02-Jul-77 0055 100 : DAS RECORDS
C00390 00198 ∂02-Jul-77 1903 FTP:TAA at MIT-DMS ZORK
C00391 00199 ∂03-Jul-77 1101 FTP:PDL at MIT-DMS ZORK Bug
C00392 00200 ∂03-Jul-77 2115 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00394 00201 ∂04-Jul-77 0231 DON
C00395 00202 ∂04-Jul-77 1203 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00414 00203 ∂04-Jul-77 1717 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00415 00204 ∂04-Jul-77 1814 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00417 00205 ∂04-Jul-77 2134 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00421 00206 ∂05-Jul-77 0743 FTP:GLS at MIT-AI (Guy L. Steele, Jr.)
C00424 00207 ∂05-Jul-77 0748 FTP:GLS at MIT-AI (Guy L. Steele, Jr.) LAMBDA considered harmful
C00426 00208 ∂05-Jul-77 0822 FTP:GLS at MIT-AI (Guy L. Steele, Jr.) LISP and efficiency
C00431 00209 ∂05-Jul-77 0824 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00437 00210 ∂05-Jul-77 1030 FTP:GLS at MIT-AI (Guy L. Steele, Jr.)
C00439 00211 ∂05-Jul-77 1706 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00440 00212 ∂05-Jul-77 2129 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00441 00213 ∂06-Jul-77 1057 FTP:Doug Gage Dialnet
C00443 00214 ∂06-Jul-77 1936 RDR via AMET Dialnet et al.
C00444 00215 ∂06-Jul-77 2138 MRC via AI Dialnet
C00448 00216 ∂06-Jul-77 2229 RPG
C00451 00217 ∂07-Jul-77 1317 JBR
C00452 00218 ∂07-Jul-77 1339 RAK System default buffers
C00453 00219 ∂07-Jul-77 2110 MRC via AI Call
C00454 00220 ∂07-Jul-77 2304 LES MRC
C00455 00221 ∂08-Jul-77 0003 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00456 00222 ∂08-Jul-77 1716 WP FOL representation
C00458 00223 ∂08-Jul-77 2117 FTP:Blohm at SUMEX-AIM Account
C00459 00224 ∂09-Jul-77 2247 GFF Hacking NS.
C00462 00225 ∂10-Jul-77 0101 TVR
C00463 00226 ∂10-Jul-77 1620 DAV MLISP
C00464 00227 ∂11-Jul-77 0243 FTP:Geoff at SRI-KA (Geoffrey S. Goodfellow)
C00465 00228 ∂11-Jul-77 1102 FTP:Sherwood at SRI-KL Dial-Net
C00466 00229 ∂13-Jul-77 1100 FTP:PRATT at MIT-AI (Vaughan Pratt) LISP article for Encyclopedia of CS and T (Belzer et al eds)
C00468 00230 ∂13-Jul-77 1324 CCG nsf
C00469 00231 ∂13-Jul-77 2111 MRC via RTGT my arrival
C00470 00232 ∂14-Jul-77 1350 PAT SUMMER AIR FARE INFO
C00473 00233 ∂15-Jul-77 1257 FTP:Geoff at SRI-KA (Geoff Goodfellow) NS
C00474 00234 ∂15-Jul-77 1442 DWW via AMET
C00480 00235 ∂15-Jul-77 1511 FTP:Boyer at SRI-KL A Paper Request
C00481 00236 ∂15-Jul-77 1520 FTP:REM at MIT-AI (Robert Elton Maas) Holofile
C00485 00237 ∂16-Jul-77 1354 RPG New Ncomplr
C00486 00238 ∂17-Jul-77 1353 PAM possible bug in newest ncompl
C00487 00239 ∂17-Jul-77 2102 TVR
C00488 00240 ∂18-Jul-77 1439 FTP:TAA at MIT-DMS Saving state
C00491 00241 ∂18-Jul-77 1635 RWW SAMEFRINGGE
C00492 00242 ∂19-Jul-77 0900 RPG NCOMPL
C00493 00243 ∂19-Jul-77 1039 FTP:Taynai at SUMEX-AIM IBM Fellowship
C00494 00244 ∂19-Jul-77 1436 JP
C00495 00245 ∂20-Jul-77 0808 FTP: Ijcai-77 at CMU-10B Model Theory paper.
C00497 00246 ∂20-Jul-77 1612 FTP:SUZUKI at PARC-MAXC concerning the decidability of list
C00499 00247 ∂21-Jul-77 0137 REM via AMET PCNet
C00502 00248 ∂21-Jul-77 1313 BPM
C00513 00249 ∂21-Jul-77 1319 FTP:SUZUKI at PARC-MAXC (Response to message)
C00514 00250 ∂22-Jul-77 0512 FTP:CERF at USC-ISI dungeon
C00516 00251 ∂22-Jul-77 1145 RWW STATUS
C00517 00252 ∂23-Jul-77 1410 FTP:TAA at MIT-DMS ZORK
C00519 00253 ∂24-Jul-77 1003 RDR via AMET LOTS, Provost & misc.
C00522 00254 ∂24-Jul-77 1519 WP
C00526 00255 ∂24-Jul-77 2210 RAK Time to go home
C00527 00256 ∂25-Jul-77 1149 RWW samefringe
C00528 00257 ∂25-Jul-77 2352 DCO
C00530 00258 ∂27-Jul-77 1628 FTP:CARL at MIT-AI (Carl Hewitt)
C00535 00259 ∂27-Jul-77 2136 FTP:TAA at MIT-DMS Candle bug
C00536 00260 ∂28-Jul-77 1340 CJS summer absence
C00537 00261 ∂28-Jul-77 1544 TOB quid pro quo
C00538 00262 ∂29-Jul-77 0015 DON
C00539 00263 ∂29-Jul-77 0324 DON Auto-zork
C00541 00264 ∂29-Jul-77 0330 DON
C00542 00265 ∂29-Jul-77 1032 FTP:Hart at SRI-KL Lynn Quam
C00543 00266 ∂29-Jul-77 1600 DBG
C00545 00267 ∂30-Jul-77 0237 ME Clobbered files
C00546 00268 ∂30-Jul-77 1542 REM via AMET PCNET protocol committee, progress report.
C00548 00269 ∂30-Jul-77 1925 FTP:TAA at MIT-DMS ZORK
C00549 00270 ∂30-Jul-77 2010 RWW japanese
C00550 00271 ∂30-Jul-77 2011 RWW last message
C00551 00272 ∂30-Jul-77 2016 RWW fol
C00552 00273 ∂30-Jul-77 2145 RWW samefringe
C00554 00274 ∂31-Jul-77 1857 MRC via AMET DIALnet
C00555 00275 ∂31-Jul-77 1903 FTP: Ijcai-77 at CMU-10B Change in session for your ''Model Theory'' paper.
C00556 00276 ∂01-Aug-77 1156 CJS absence from lab
C00557 00277 ∂03-Aug-77 0010 WP Implementation of LISP functions
C00560 00278 ∂04-Aug-77 2311 RAK Mistak
C00561 00279 ∂05-Aug-77 0317 MRC DIALnet
C00562 00280 ∂05-Aug-77 0758 TVR
C00563 00281 ∂05-Aug-77 0858 MRC queue of things for me to do
C00564 00282 ∂05-Aug-77 1132 JBR
C00565 00283 ∂05-Aug-77 1322 MFB MTC QUAL
C00566 00284 ∂05-Aug-77 1556 RDR via AMET CS105/6 TA Consulting
C00568 00285 ∂05-Aug-77 1637 MRC
C00569 00286 ∂05-Aug-77 1817 MRC [DLN,MRC]
C00570 00287 ∂05-Aug-77 2130 DON
C00571 00288 ∂05-Aug-77 2324 DON zork[e77,jmc]/7p contains weights.
C00573 00289 ∂06-Aug-77 2024 GLS MACLSP on SYS
C00574 00290 ∂07-Aug-77 0715 FTP:REM at MIT-MC (Robert Elton Maas) DATA-COMPRESSING
C00577 00291 ∂07-Aug-77 1643 REM via AMET
C00578 00292 ∂07-Aug-77 1825 DON zork
C00579 00293 ∂07-Aug-77 1918 DON
C00580 00294 ∂07-Aug-77 2308 DON
C00583 00295 ∂07-Aug-77 2331 RAK 11pm news--in case you missed
C00584 00296 ∂08-Aug-77 1104 FTP:BOBROW at PARC-MAXC Some proposed questions for panel discussion
C00586 00297 ∂08-Aug-77 2111 BPM
C00587 00298 ∂08-Aug-77 2134 DON green light
C00588 00299 ∂08-Aug-77 2248 DON zork
C00589 00300 ∂09-Aug-77 0231 DON
C00590 00301 ∂09-Aug-77 0232 DON
C00591 00302 ∂09-Aug-77 0236 DON
C00592 00303 ∂09-Aug-77 0308 DON
C00593 00304 ∂09-Aug-77 1414 FTP:CARL at MIT-AI (Carl Hewitt)
C00594 00305 ∂09-Aug-77 1719 REF
C00595 00306 ∂09-Aug-77 1901 FTP:MRC at MIT-AI (Mark Crispin) What's up
C00601 00307 ∂09-Aug-77 2329 TVR ZORKing
C00602 00308 ∂10-Aug-77 0328 DON Blast!
C00603 00309 ∂10-Aug-77 0431 DON input files
C00605 00310 ∂10-Aug-77 0844 REM via AMET PCNet
C00607 00311 ∂10-Aug-77 2050 DON
C00610 00312 ∂11-Aug-77 0201 DON
C00612 00313 ∂11-Aug-77 0403 FTP:RWG at MIT-MC (Bill Gosper)
C00613 00314 ∂11-Aug-77 0444 FTP:PRATT at MIT-AI (Vaughan Pratt)
C00614 00315 ∂11-Aug-77 0447 RPG New MACLSP
C00615 00316 ∂11-Aug-77 1226 REF NSF proposal bibliography -- possible change:
C00617 00317 ∂12-Aug-77 1024 RWW via AI
C00619 00318 ∂12-Aug-77 1058 MRC via AMET
C00620 00319 ∂12-Aug-77 1131 DCO
C00621 00320 ∂12-Aug-77 1858 FTP:POSTEL at USC-ISIB Dialnet
C00622 00321 ∂13-Aug-77 0054 WD
C00623 00322 ∂13-Aug-77 0524 FTP:CERF at USC-ISI dial-net
C00625 00323 ∂13-Aug-77 0954 DON
C00626 00324 ∂13-Aug-77 1215 MS
C00627 00325 ∂13-Aug-77 1503 PAM
C00669 00326 ∂13-Aug-77 2011 RWW via AI mtc
C00670 00327 Could the following have been intended for you?
C00671 00328 ∂14-Aug-77 1742 PAM
C00675 00329 ∂15-Aug-77 1716 JBR
C00693 00330 ∂16-Aug-77 0105 MRC
C00695 00331 ∂16-Aug-77 0919 FTP:BOBROW at PARC-MAXC Panel Time
C00696 00332 ∂16-Aug-77 2217 FTP:PAM at SU-AI (Paul Martin) mail file
C00698 00333 ∂16-Aug-77 2312 PAM
C00699 00334 ∂17-Aug-77 1432 RAK FROM HERE TO THERE....
C00700 00335 ∂18-Aug-77 1047 FTP:BOBROW at PARC-MAXC The russians have cancelled
C00701 00336 ∂19-Aug-77 0953 FTP:RSMITH at RUTGERS-10 copies of your unpublished papers
C00702 00337 ∂20-Aug-77 2040 FTP:Lots at SRI-KL LOTS Resource allocation
C00705 00338 ∂20-Aug-77 2042 FTP:Lots at SRI-KL correction
C00706 00339 ∂22-Aug-77 1149 MJL CS 206 classnotes
C00707 00340 ∂23-Aug-77 1110 FTP:MORRIS at PARC-MAXC A debate on program verification
C00708 00341 ∂23-Aug-77 1438 PAT PHONE CALL
C00709 00342 ∂23-Aug-77 1517 REG via AMET Replacement for Roode
C00712 00343 ∂25-Aug-77 0350 REM via AMET CRUNCHING SOFTWARE, DFTP
C00715 00344 ∂25-Aug-77 1619 MJL
C00716 00345 ∂26-Aug-77 1304 FTP:Moore at SRI-KL Processing at function definition time
C00723 00346 ∂26-Aug-77 2017 CLT chapter 5
C00724 00347 ∂27-Aug-77 1627 TVR ZORK
C00725 00348 ∂27-Aug-77 2208 DON
C00726 00349 ∂27-Aug-77 2214 DON
C00727 00350 ∂27-Aug-77 2215 DON
C00728 00351 ∂28-Aug-77 0106 GFF
C00729 00352 ∂29-Aug-77 0041 MRC via AMET finger problems
C00731 00353 ∂29-Aug-77 0311 DON Balloon
C00733 00354 ∂29-Aug-77 1317 FTP:DCROCKER at USC-ISI Social Statistic of Computer-based Human Communication
C00735 00355 ∂29-Aug-77 1446 FTP:PSZ at MIT-ML (Peter Szolovits) LISP commemorative conference (?)
C00737 00356 ∂29-Aug-77 2329 DON
C00739 00357 ∂29-Aug-77 2359 DON
C00743 00358 ∂30-Aug-77 0031 DON
C00745 00359 ∂30-Aug-77 0221 DON thief
C00746 00360 Can you get me a xerox of the article referred to.
C00747 00361 ∂30-Aug-77 1629 JJK CSD Research Reports
C00748 00362 ∂31-Aug-77 0641 FTP:EF at MIT-AI (Edward Fredkin)
C00749 00363 ∂31-Aug-77 0808 RDR
C00751 00364 ∂31-Aug-77 1049 RDR constructive proposal
C00755 00365 ∂31-Aug-77 1107 FTP:Wiederhold at SUMEX-AIM Re: TA office space
C00756 00366 ∂31-Aug-77 2106 FTP:MF at MIT-AI (Michael H. Fredkin)
C00757 00367 ∂01-Sep-77 0000 JMC*
C00758 00368 ∂01-Sep-77 2052 MRC SUPDUP
C00759 00369 ∂02-Sep-77 1120 FTP:PBARAN at USC-ISI Computers and communications
C00760 00370 ∂02-Sep-77 1128 FTP:Farber at Rand-Unix Re: Computers and communications
C00761 00371 ∂02-Sep-77 1818 FTP:TAA at MIT-DMS Dungeon
C00763 00372 ∂02-Sep-77 2341 WD
C00765 00373 ∂03-Sep-77 0855 FTP:PDL at MIT-DMS The Record
C00766 00374 ∂03-Sep-77 1201 BPM
C00767 00375 ∂03-Sep-77 1657 REM via AMET
C00768 00376 ∂03-Sep-77 1729 FTP:Lynch at SRI-KL Re: Crowding on your machine.
C00771 00377 ∂03-Sep-77 2053 REF
C00772 00378 ∂03-Sep-77 2101 REF birthday
C00773 00379 ∂03-Sep-77 2356 MRC
C00774 00380 ∂04-Sep-77 0148 MRC
C00775 00381 ∂04-Sep-77 0800 JMC*
C00776 00382 ∂04-Sep-77 0800 JMC*
C00777 00383 ∂04-Sep-77 1453 FTP:Raphael at SRI-KL Re: Datamedias
C00779 00384 ∂04-Sep-77 2233 FTP:Boyer at SRI-KL Crowding on the SRI-KL
C00782 00385 ∂06-Sep-77 2033 FTP:Raphael at SRI-KL Review
C00783 00386 ∂07-Sep-77 1232 MRC
C00784 00387 ∂07-Sep-77 2140 DGR via NBST DIALNET
C00786 00388 ∂08-Sep-77 0258 MRC DIALnet paper LPT version
C00788 00389 ∂08-Sep-77 0927 FTP: Lee Erman at CMU-10B Re: Book with invited addresses
C00790 00390 ∂08-Sep-77 1342 LES Dialnet blurb & Bell modem
C00792 00391 ∂08-Sep-77 1801 FTP:Raphael at SRI-KL Ad
C00793 00392 ∂09-Sep-77 0014 DON zork
C00796 00393 ∂09-Sep-77 1733 FTP:Blohm at SUMEX-AIM LOTS Student Coordinator
C00800 00394 ∂09-Sep-77 1753 FTP:Creary at SUMEX-AIM Request for Papers
C00802 00395 ∂09-Sep-77 1808 FTP:Creary at SUMEX-AIM Oops!
C00804 00396 ∂09-Sep-77 1828 FTP:Blohm at SUMEX-AIM
C00810 00397 ∂10-Sep-77 0133 FTP:MRC at MIT-AI (Mark Crispin) We should consider this!
C00812 00398 ∂10-Sep-77 0153 MRC LOTS wheel who abuses his privileges
C00813 00399 ∂10-Sep-77 0510 MRC DIALnet & DECUS
C00814 00400 ∂10-Sep-77 0612 MRC
C00815 00401 ∂10-Sep-77 0907 MRC
C00816 00402 ∂10-Sep-77 0943 FTP:RYLAND at RUTGERS-10 guest account
C00817 00403 ∂10-Sep-77 1513 RDR MRC
C00818 00404 ∂11-Sep-77 0855 MRC Person from Columbia University
C00819 00405 ∂11-Sep-77 0937 MRC System notice REM put up today
C00821 00406 ∂11-Sep-77 1321 RDR via AMET
C00822 00407 ∂11-Sep-77 1323 RDR via AMET idea
C00824 00408 ∂11-Sep-77 1333 RDR via AMET yes there are a few summer PPL alumni around now
C00825 00409 ∂11-Sep-77 1802 MRC via AMET REM mail
C00826 00410 ∂12-Sep-77 1643 FTP:Feigenbaum at SUMEX-AIM weizenbaum review
C00827 00411 ∂14-Sep-77 0433 CLT Chapter 4
C00828 00412 ∂14-Sep-77 2127 LLW via AMET meeting
C00829 00413 ∂15-Sep-77 1135 HVA TELEPHONE PROBLEMS
C00831 00414 ∂15-Sep-77 1348 EJG via AMET Stanford-1 time-sharing system
C00832 00415 ∂15-Sep-77 1409 FTP:Taynai at SUMEX-AIM CSAC Meeting - Oct. 17/18
C00834 00416 ∂15-Sep-77 1424 LES Swedish visitors
C00836 00417 ∂15-Sep-77 1451 LCW meeting
C00837 00418 ∂15-Sep-77 2102 FTP:Geoff at SRI-KA UPI
C00838 00419 ∂15-Sep-77 2305 LLW via AMET Meetings on Friday AM
C00839 00420 ∂16-Sep-77 0203 LES UPI at UTexas
C00840 00421 ∂16-Sep-77 0334 CLT The state of things.
C00842 00422 ∂16-Sep-77 1053 HVA Quarterly Report - NASA
C00843 00423 ∂16-Sep-77 1300 JMC*
C00844 00424 ∂16-Sep-77 1309 ZM LOGIC QUALS
C00845 00425 ∂16-Sep-77 2029 CLT LISP notes
C00846 00426 ∂16-Sep-77 2128 LLW via AMET JMC letter to trimble/onr
C00847 00427 ∂17-Sep-77 1011 JRA BOOK
C00848 00428 ∂18-Sep-77 1630 LES
C00849 00429 ∂19-Sep-77 1059 PAT note from Ringle
C00851 00430 ∂19-Sep-77 1636 PAM Hooray
C00852 00431 ∂19-Sep-77 1725 CLT
C00853 00432 ∂21-Sep-77 0036 FTP:PBARAN at USC-ISI PCNET Advisory Board
C00855 00433 ∂21-Sep-77 0306 CLT
C00857 00434 ∂21-Sep-77 0825 100 : TPS PHONE CALL
C00858 00435 ∂21-Sep-77 1052 FTP:Taynai at SUMEX-AIM Terman Dedication
C00860 00436 ∂22-Sep-77 1039 FTP:Taynai at SUMEX-AIM ( Forwarded Mail )
C00862 00437 ∂22-Sep-77 1353 FTP:Feigenbaum at SUMEX-AIM private guesses from pete hart re interlisp/please dont propagate
C00865 00438 ∂22-Sep-77 1357 FTP:Wiederhold at SUMEX-AIM Program verification course
C00867 00439 ∂23-Sep-77 1739 JBR at TTY43 1739
C00868 00440 ∂26-Sep-77 0130 FTP:Hedberg at SUMEX-AIM Credit for consulting at LOTS
C00869 00441 ∂26-Sep-77 0410 DPB List of your advisees
C00870 00442 ∂26-Sep-77 1049 100 : patte phone call
C00871 00443 ∂26-Sep-77 1421 FTP:Feigenbaum at SUMEX-AIM Congratulations
C00872 00444 ∂27-Sep-77 0911 MJL LOTS meeting
C00873 00445 ∂27-Sep-77 1137 FTP: Bob Sproull at CMU-10A dial-up mail
C00877 00446 ∂27-Sep-77 1438 PAM
C00879 00447 ∂27-Sep-77 2020 DPB Wagner's degree
C00880 ENDMK
C⊗;
∂09-May-77 0953 MJL ai qual
Hi, is there any update on the AI Qual?
∂09-May-77 1107 FTP:KAY at PARC-MAXC Scandinavian Summer School
Mail from host Maxc2 received at 9-MAY-77 1059-PDT
Date: 9 MAY 1977 1059-PDT
From: KAY at PARC-MAXC
Subject: Scandinavian Summer School
To: McCarthy at SAIL
This is to amplify our telephone conversation this morning.
I am playing host to 15 scandinavian students of computational linguistics
during the period 13-24 June during which time I shall be organizing the
fourth and last of a series of summer schools for the Nordic
Cultural Comission. They are coming here for the last one so that they
can have access to some of the harware and software we have. It now appears
that there are regulations ot prevent me letting foreign students spend as
long as would be necessarey on the premises here. It occurred to me
that this might be a low time at the AI-lab and that you might be so
kind as to let us have the use of a room with some terminals giving access
to the network for some time during that period. I will be happy to take
whatever we can get. If there is any time during which we could be
there without getting in anybody's way, we will very gratefully accept.
Martin.
-------
∂09-May-77 1602 LES YTD
To: JAB
CC: JMC, JBR, RP
ytd.doc[act,jab] looks reasonable, but I forsee the need for at least one
additional field, representing disk storage utilization (e.g. K*days).
When it becomes available, this could go into your "noise word" (2).
I don't think that we should attempt to collect this information until
the expanded UFD data becomes available. Since JBR is pushing that off
into the new version of DSKSER, it may take awhile.
Do you have any interest or intention to gather some of the other system
statistics that we discussed (e.g. line utilization, remote user utilization,
system uptime percentage, MTBF, etc.)? Some of that data should be carried
in cumulative files, much like YTD.
∂10-May-77 0008 FTP:MRC at MIT-AI (Mark Crispin )
Date: 10 MAY 1977 0308-EDT
From: MRC at MIT-AI (Mark Crispin )
To: JMC at SU-AI
Hi. I read the recent New York Times article about all the controversy
about Weizenbaum's book and AI and hackers. While AI has gotten a lot
of defense from the AI people (as indeed it should!), not much has been
said for the hackers, who are portrayed as "compulsive programmers"
who are incapable of writing readable documentation, have no social life
to speak of (except for perhaps being in love with the computer), and
are unproductive to the point of being parasitic.
Well, as an ITS hacker, and one who has often been praised for good
documentation, who has a social life, and is reasonably productive (I
think), and yet who also spends long hours on a computer and enjoys
himself doing so, I thought that I should try to answer some of the
arguments that Weizenbaum presented.
The results are on the MIT-AI computer as
MRC;WEIZEN LETTER
which I sent off to Weizenbaum a couple of days ago. I thought that since
you published a rebuttal to his book in the SIGART Newsletter that you
might be interested in something generated by a hacker, especially one
who, if you look at him with Weizenbaum's exaggerated point of view, would
fit Weizenbaum's description of the "compulsive programmer". I would
appreciate any comments you have to make on the letter, and if you are
interested in what I get as a reply from Weizenbaum, I can forward you
a copy.
Regards,
Mark Crispin (MRC@MIT-AI)
[PS. I intend to visit California as sort of a graduation present to
myself sometime in the next couple of weeks. I have never been
further west than Alabama. In particular I intend to visit SAIL
during that time. So maybe you'll meet me then.]
∂10-May-77 0048 JAB
Could we talk sometime about my plans for the upcoming year? I would like
to find a university related position which would also allow me to finish
the course or two I need. I have unoffical approval to finish without
being in residence at JHU, however I won't be able to get final approval
until I present a definite plan.
Could I look to the AILab for some funding for next year? Something on the
order of half-time so I would have time for completing my degree work. I
have looked into working for DEC but that would take me out of an acedemic
environment. There might also be some funding for me at JHU but probably
not enough for half-time support. A position at Columbia University is
also a possiblity.
I would also appreciate if you would act as a reference for me with
regards to the experience I have gained here at AILab. Thanx.
∂10-May-77 1057 PAT aiqual
Feigenbaum called and has two questions regarding the qual:
Is Nils Nilsson on the committee...he was unaware of when asked?
Will the committee divide into to parallel groups, each with six
students or will the committee stay as a group and ask each student questions
individually?
∂10-May-77 1258 FTP:Taynai at SUMEX-AIM Borrowed Book
Date: 10 MAY 1977 1259-PDT
From: Taynai at SUMEX-AIM
Subject: Borrowed Book
To: McCarthy at SAIL
cc: Taynai
Dr. McCarthy,
On March 1 you borrowed from Prof. Feigenbaum a book entitled
"Language and Perception" by George A. Miller and Philip N.
Johnson-Laird. Prof. Feigenbaum would appreciate receiving it
back.
Thank you,
Carolyn Taynai
-------
∂10-May-77 1455 ZM
∂09-May-77 1347 MJL ai qual
hi patte, the following people are taking the ai qual:
Filman, Bob
Appelt, Doug
Goldman, Ron
Farmwald, Paul
Pattis, Rich
Stefik, Mark
the examiners are:
McCarthy, Feigenbaum, Winograd, Green, Nilsson, Buchanan (i think that is all)
∂10-May-77 1620 MJL ai qual
hi, as of today, five people are signed up for the qual
(two had their names taken off the list, Davidson & Stefik).
those taking the exam are: Appelt, Farmwald, Filman, Goldman,
and Pattis. Students seems anxious to know what the format
is going to be, do they have to sit around waiting while others
are being examined? do they examine with one examine, all of
them at once? so, if you can give me any info along these lines
to ease their little minds, i would be appreciative. Thanks -- moira
∂10-May-77 2229 FTP:MRC at MIT-AI (Mark Crispin ) Weizenbaum's hackers
Date: 11 MAY 1977 0128-EDT
From: MRC at MIT-AI (Mark Crispin )
Subject: Weizenbaum's hackers
To: JMC at SU-AI
Yes, actually I can see where he gets his ideas. He merges every type
of person that you might find around a computer, from the hacker to the
non-productive random high school kid to the poor programmer to the
grad student... and somehow generalizes it to his mythical "computer
bums". Perhaps he once observed a hacker happily on the trail of a
truly wonderful bug, and with his preconceived opinions the conclusions
he reached in CP&HR might be understandable (not that they are in any
way rational, but still understandable!).
∂10-May-77 2245 DCL
To: LES, JMC, JC
What has been decided about all these priviledged 220 students
∂11-May-77 1314 JBR
∂10-May-77 2106 JMC
Would you kill off TOG again? His directory is protected.
done. --jbr
∂11-May-77 1444 DPB News concerning MTC Qual for Fall 1977.
To: CG, MFB, JMC, ZM
Due to JMC's recent trip, and Zohar Manna's imminent one, the MTC Qual will
postponed. JMC and ZM are willing to give it in July when ZM returns, but
early fall quarter will probably be better for students. The three-year
deadline for area qual completion will be extended appropriately. Comments?
People interested in taking MTC Qual.
Goad, Chris
Lengauer, Tom
Brooks,Martin
∂12-May-77 1322 FTP:Wilber at SRI-KL LISP PROTOHISTORY
Date: 12 May 1977 1320-PDT
From: Wilber at SRI-KL
Subject: LISP PROTOHISTORY
To: JMC at SAIL
HELLO JOHN -- I'M INVOLVED IN EDITING A COUPLE OF ARTICLES INTRODUCING
LISP TO NEW PEOPLE, AND I HAVE A HISTORICAL QUESTION THAT YOU'RE BEST
QUALIFIED TO ANSWER: IS MY HAZY RECOLLECTION CORRECT THAT YOU ARE THE
PERSON PRIMARILY RESPONSIBLE FOR THE INITIAL DEVELOPMENT OF LISP INTO
A REAL PROGRAMMING LANGUAGE AND SYSTEM AND THAT YOU DID SO ON A 709 AT
MIT IN THE EARLY 60'S?...IT'S A SMALL POINT, BUT I'D LIKE TO GET THE
RECORD STRAIGHT AT THE OUTSET...ALSO, ASSUMING I CAN GET PERMISSION TO
ASK YOU, WOULD YOU BE WILLING TO READ AND PERHAPS COMMENT ON THESE
ARTICLES PRIOR TO THEIR PUBLICATION?...THANKS -- MIKE
-------
I am responsible for both the theoretical and practical development of
LISP. car and cdr were devised in late 1956 in connection with what
later became FLPL (Fortran list processing language) designed and
implemented by Gelernter and Gerberich at IBM. The first LISP programs
were written in the fall of 1958, and an interpreter was working sometime
in 1959. The first paper "Recursive functions of symbolic expressions
and their computation by machine" appeared in CACM April 1960, but appeared
earlier as report of the M.I.T. Research Laboratory of Electronics.
LISP was first implemented on the IBM 704, the immediate predecessor of
the 709. There is a transcript of a speech giving some of this history
and if you read German, I have a copy of a fuller history by Herbert
Stoyan of Dresden.
∂13-May-77 0918 MJL dept. meeting
hi, just a reminder about the department meeting
on monday, may 16 @ noon, law school, room 290.
also, EAF is hosting a lunch at the Faculty Club
afterwards for the awardee of the Forsythe Award
and Sandra Forsythe. If you are interested in
attending, please let me know today. thanks
∂13-May-77 1057 FTP:Hart at SRI-KL Job openings
Date: 13 May 1977 1059-PDT
From: Hart at SRI-KL
Subject: Job openings
To: McCarthy at SU-AI
cc: Hart
John,
We're actively looking for some AI people-- primarily new PhD. types,
but possibly also a good support person who implements well. Any
suggestions?
Peter
-------
∂13-May-77 1227 LES Foreign journalists tour
To: JMC, JC
CC: PAT, RWW
The Office of Foreign Visitors now says that the tour group will arrive
at 9:00 AM next Thursday and they must leave at about 10:45 in order to
get to campus for a presentation by Ed Parker at 11.
I suggest that JMC take them from 9 to 10:15 and JC from there on, but
feel free to negotiate (with each other).
∂14-May-77 0433 JAB meeting
when will you have some time to talk to me about next year?
whenever you can catch me
∂15-May-77 0613 FTP:PRATT at MIT-AI (Vaughan Pratt )y )
Date: 15 MAY 1977 0853-EDT
From: PRATT at MIT-AI (Vaughan Pratt )
To: JMC at SU-AI
In the introduction to my Encylopaedia article on LISP, I am torn
between introducing LISP the way it was done originally, with heavy
emphasis on lists, and introducing it as a language that does justice
to the word-object dichotomy. The former is historically more
accurate, the latter seems to me to be the direction the LISP
philosophy ought to be going. I can indeed combine both in the
introduction, but I am not sure whether I can explicitly deal with
both aspects without lessening the impact of the introduction. Well,
actually, just writing this note was helpful to me in sorting out my
problem, so whether or not you reply to this you will have
(indirectly) been of help in this issue.
What encyclopedia is it? It seems to me that an encyclopedia article
ought to be based on its subject as it is, not as the author thinks
it should become. Remarks on the latter are more appropriate for
a concluding section than for an introduction. I would be interested
in commenting on drafts of the article.
∂16-May-77 1515 PAM
John the dentist moved my appointment earlier, so I'll take care of
the homework and such.. I may be a little late, but I'll be there before
its's over for sure.... Thanx anyway. Paul
∂16-May-77 1616 MJL autumn 1977
hi, are you planning to use a text that need ordering
for CS 206 next fall? Please let me know ASAP, thanks.
∂16-May-77 1740 DEW Thesis research
To: BERLINER at CMU-10A
CC: JMC
Hi Hans,
I have finished debugging my completely rewritten productions. The knowledge
has been completely reorganized, the production language has more power,
and the knowledge source structure which the productions are accessed
through has been made more powerful (and more complex).
I am happy with the way the system reasons now and with the way the productions
are organized. It is all debugged and running fine on my example positions
(nos. 1, 4 and 5 from WIN AT CHESS).
The tree search has been polished to the point where it now searches all
these positions for many different lines of play and returns an accurate
answer with a small tree (like tens of nodes). There are no arbitrary or
depth cutoffs--it always searches until no likely threats are found.
If am working on two problems currently: 1) how to fix offensive plans
to thwart defensive threats at the same time. It seems that each threat
production will have to check for defensive threats and fix them in its plan
if possible. It is not desirable check this so many times but each threat
is complex and allows particular fixes that other threats don't.
2) I'm enlarging the production knowledge base to make the program handle
more problems. I've gone thru the first 12 positions in WIN AT CHESS
(skipping 2 and 6 -end game) and the results are encouraging. Some of the
positions should be solved by the current program and all the others
can be handled by writing a few more productions. Thus the system has
the necessary ability to express what is going on in these positions.
(Actually, problem 1) above is the one problem these postions presented.)
(Also, these positions have not been given to the program, this is my
analysis.)
That is what I'm doing. I feel like I've been so embroiled in making
the program perform that I may be out of tune with the overall worth
of what I'm doing. Perhaps I should stop and sit back and think about
where it is going.
The problem of embodying knowledge in such a way that it is useable and
easily added to or changed seems to be getting solved.
However, none of the fancy tree communication stuff has been done and
if I need this for a thesis perhaps I should concentrate on switching
my effort to that as soon as possible. Any advice?
Thanks,
Dave
∂15-May-77 2323 FTP:STEF at SRI-KA
Date: 15 May 1977 2319-PDT
Sender: STEF at SRI-KA
From: STEF at SRI-KA
To: REM at SAIL
Cc: STEF, Geoff
Message-ID: <[SRI-KA]15-May-77 23:19:17.STEF>
In-Reply-To: Your message of May 15, 1977
Ah Yes! It is as bad as I feared.
I cannot distribute anything to MsgGroup that is not properly
sanctioned, so can i ask that you guys somehow convert that file
to clear text in a standard TENEX sequential file, and send it to
MsgGroup@ISI for distribution to the members?
It is curious to me that this should be so apparently difficult.
How do you guys ship stuff to sites that are not blessed with
XGP's?
Just curious and hopeful - I feel that the DIALNET paper is quite
important for MsgGroup to see.
Cheeers, Stef
-------
[REM - I tried SRCCOM between DIANET.[W77,JMC] and the TENEX plaintext
format version, but total discord, not enough in common for any sync.
Perhaps you know how to make PUB generate a formatted but plaintext
file, or know enough about your own file to be able to create and
authenticate a plaintext version for export?]
It isn't difficult to edit out all the PUB control information.
Get rid of all lines beginning with .. Then get rid of all %<character>
in the text. There may be a little garbage left, but the text will
be quite readable.
∂17-May-77 1115 MJL ai qual
hi, please send me a memo with the results of the
AI qual. Send it over the system if easier. Thanks, Moira
∂18-May-77 0310 FTP:Wilber at SRI-KA LISP ARTICLES
Date: 18 May 1977 0302-PDT
From: Wilber at SRI-KA
Subject: LISP ARTICLES
To: JMC at SAIL
cc: WILBER at SRI-KL
HELLO JOHN -- PERHAPS YOU RECALL MY GETTING IN TOUCH WITH YOU LAST WEEK
ABOUT SOME INTRODUCTORY LISP ARTICLES?...SINCE THEN I'VE TALKED TO MY
SOURCE, CARL HELMERS, EDITOR OF BYTE MAGAZINE, WHO WOULD BE DELIGHTED
TO HAVE YOU LOOK OVER THE ARTICLES AND COMMENT ON THEM...WOULD
YOU BE WILLING?...IF SO, I CAN GET COPIES OF THEM TO SAIL AFTER WORK
TODAY; CARL WOULD LIKE YOU TO FUNNEL YOUR COMMENTS THROUGH ME (I'M ALSO
ASKING DAVE MOON AND WARREN TEITELMAN TO COMMENT) SO THAT I CAN GET A
RESPONSE TO HIM BY JUNE FIRST...ALSO, CARL WOULD BE JUST DELIGHTED IF
YOU COULD WRITE A HISTORICAL (OR OTHER)ARTICLE (I THINK IN THE SAME TIME
FRAME(!))...ALSO, I'D LIKE TO TAKE YOU UP ON YOUR OFFER OF THE TWO
HISTORICAL ARTICLES...DOES THAT SOUND REASONABLE?...THANKS -- MIKE
-------
I will be willing to look over the articles. What time will you come?
∂18-May-77 0644 RDR via AMET
The guest lecturer from Xerox PARC teaching CS246/EE386 this quarter
has told the class he has a surprisingly effective 10-20
line patch to TENEX to circumvent its eagerness to page things out.
Maurice and Ralph refuse to investigate. I oppose in principle
rejecting any easy solution merely because it seems too good to
be true --that is rejecting it out of hand. Still,
any request on LOTS' behalf would have to come from Mo
or Ralph or at least they must not be hostile to the guy. He isn't going
to come and beg to do LOTS a favor.
His name is Lynch.
∂18-May-77 1200 MDD*
To: JMC
CC: RWW
I'll be here tomorrow.
∂18-May-77 1248 FTP:Wilber at SRI-KL refereeing
Date: 18 May 1977 1248-PDT
From: Wilber at SRI-KL
Subject: refereeing
To: jmc at SAIL
thanks -- will be at sail early evening today and leave them on your
desk if you're not there then -- mike
-------
∂18-May-77 1649 ZM MTC QUALS
I mailed you a tentative draft of the MTC Syllabus.
Please add the revised version to the QUALS.ALL[INF,CSD] file and notify
D. Brown, M. Brooks, C. Goad, T. Lengauer and W. Polak. Thanks Zohar
∂18-May-77 2046 DCO
I will be writing this up in the next few days, but here is an advance
preview of results on decidability and complexity of the first order theory of
list structure and its generalization. The theorem on the complexity of the
quantifier-free conjunction fragment of the theories is jointly due to myself
and Greg Nelson. We have implemented our decision algorithm for this fragment
and it seems to be as fast in practice as it is in theory.
In light of Rabin's theorem and some other results that exist in logic
I do not claim all these results as new. However, apparently no one has seen
the possible applications of the results in logic to problems of interest to us.
First order theory of list structure: (is this a good name?)
Function symbols: CAR, CDR, CONS
Predicate symbols: = (LISP EQUAL), ATOM
In the constructive version of the proof of decidability, it doesnt
make any difference to the results whether or not one takes car of an atom
to be the atom itself, to be some distinguished constant (NIL, say) or to
be undefined (i.e. omega). It also doesnt make any difference whether
or not one allows circular list structure - i.e. whether one models the
results of RPLACA etc. or not. So let's assume that some choice has been made.
Theorem 1:
The first order theory of list structure is decidable.
I have two proofs, one a proof by reduction and the other a constructive
proof exhibiting a decision procedure.
The easy proof follows as a corollary of a theorem by Michael Rabin
that the weak second order theory of two successors is decidable. Idea:
Delete all occurrences of CONS in favour of CAR and CDR. Think of CAR and CDR
as two predecessors, and encode list structure by finite sets. The result
follows! (I had the reduction in mind for a long time but couldn't complete
the proof until last January when I asked Albert Meyer whether or not this
second order theory was decidable; he told me of Rabin's result).
The constructive proof is much more interesting (assuming you agree
that the weak second order theory of two successors is not as intuitive a theory
as the theory of list structure); it exhibits a quantifier-elimination method.
The theory is naturally generalized to the theory of any data structures
with the notion of constructors (e.g. CONS) and selectors (e.g. CAR and CDR).
In fact, generalize the theory further to allow unions of such theories -
i.e. allow various flavours of (indepedent sets of) constructors and selectors.
The essential axioms are the obvious generalizations of the well-known axioms
for LISP [McCarthy 1958].
Theorem 2:
The first order theory of generalized data structures ( name ?) is decidable.
The complexity of the theory is of interest (at least to me). Greg Nelson
and I have proved the following theorem.
Theorem 3:
There is a decision procedure which will decide the satisfiability of
any quantifier-free conjunction of length n of literals in time O(n-squared).
Theorem 4:
The quantifier-free theory of generalized data structures is NP-complete.
This follows immediately from Theorem 3, since Theorem 3 shows that the
theory is in NP. You cannot win with the fully-quantified theory however.
Theorem 5:
The first-order theory of generalized data structures is non-elementary
recursive.
∂18-May-77 2047 DCO
I just realized I may have messed up the "directory updated .." pointer
in your mail file; I noticed an error in my message to you and corrected it.
∂18-May-77 2110 FTP:Wilber at SRI-KL LISP PAPERS
Date: 18 May 1977 2116-PDT
From: Wilber at SRI-KL
Subject: LISP PAPERS
To: JMC at SAIL
HI -- I WAS UP AT SAIL ABOUT AN HOUR AGO AND LEFT THEM ON THE CHAIR
NEAR THE KEYBOARD IN YOUR OFFICE...LOOKING FORWARD TO YOUR COMMENTS...
THANK -- MIKE
-------
∂19-May-77 0025 REM via AMET Attempts to get DIALNE[W77,JMC] to agree with GEOFF/STEF version
To: STEF at SRI-KA
CC: JMC
Well, JMC suggested deleting all PUB command lines (those which begin
with a "." character) and all %<any character> pairs. After doing that
SRCCOM was still unable to find any match whatsoever between the two files.
I looked at them and noticed that the GEOFF/STEF had more even lines, perhaps
it had been left-justified or manually edited, so next I used SOS to re-justify
copies of both files in an attempt to eliminate that source of non-agreement.
After that, SRCCOM showed enough occasional matches to break up the output
into chunks of size varying from 8 to over 20 lines each. At this point I
noticed that the JMC version has single spaces after periods whereas the
GEOFF/STEF version has double spaces, and SOS is to dumb to massage the
two versions into the same, so the GEOFF/STEF version has one extra space
in critical places that force the last word to the next line, displacing
text all the way to end-of-paragraph. I 'uess my next step is to delete
all double-spaces from both files, re-justify, and try SRCCOM again.
If that works, I'll declare the two versions equivalent. If not I'll
give up and suggest GEOFF/STEF discard their version and get JMC's and
start from scratch again.
∂19-May-77 0118 JP running E from maclsp
To: MACLSP.DIS[AID,RPG]:;
File ET.FAS[MAC,LSP] contains LSUBR ET that swaps to E from MACLSP.
Permissible incantations are
(et '(fil)) ;uses E default extensions in disk dir
(et '(fil ext)) ;disk dir
(et '(fil ext (prj prg)))
(et '(fil (prj prg))) ;default ext
(et) ;runs E on last file edited
Notice that the device is always assumed to be DSK.
Bugs, comments → JP
∂19-May-77 0122 ZM
I am sorry! It is now 1:30 a.m. I was busy all evening finishing a paper and
absolutely forgot about our " first-order formalization meeting". Zohar
∂19-May-77 0739 RDR via AMET LOTS
I don7t know how much good this will do but all else has been tried.
Currently at peak times, LOTS is delivering 30-40% to the users. These
peak times last over 12 hours out of each day. With 10 less terminals,
we were delivering 60-75% to the users at peak times. Since
everyone had to wait at the terminal anyway, it is no
gain to be delivering to more users --ie even less (much less, like 50%
as much) is being split many more ways.
People are swearing at the system when they use it. They are not happy at
all. The lines are LONGER than with fewer terminals. True, there is
less variance in the length of the line, but what good is that?
Professors Ferziger, Franklin and Adams have all repeatedly told Ralph that
everyone would rather have to wait more for a terminal and
less AT the terminal. Ralph ignores them politely.
IE 144 was asked to evaluate their use of LOTS. In the papers
that Adams brought over, a recurrent suggestion was that the
system canot handle so many terminals and that the people
wished some would be taken off. Ralph marked in green one lone
comment along the "more terminals" line and ignored all the
other comments.
The problem alleviates itself at times when terminals are left
free, but in 90%+ off other times, the other terminals are doing
tremendous harm. That is the only time it would be OK to
have 40 terminals instead of 30 is when only 30 are in use.
Ralph just likes to have a lot of users on the system and ignores
the fact that they are getting so much less done.
Finally, it is obviously no long term solution to do with fewer
terminals--it is just that for the time being it is clearly
the way to go. they system has generally too many terminals
when 32 are in use so there will be no problem knowing
when it is time to add more terminals as system improvements are
made -- hopefully only a few terminals at a time.
∂19-May-77 0755 FTP:Lots at SRI-KL the real point
Date: 19 May 1977 0803-PDT
From: Lots at SRI-KL
Subject: the real point
To: jmc at SU-AI
I forgot to explain that the consequence of the additional terminals
is that with people spending 3 to 4 time as long at a terminal
as they did with fewer terminals, their time is wasted.
Furthermore, other people are forced to come more and more
at 2am, 4am, 6am resulting in round-the-clock
excessive load averages. They do not enjoy these times
and I don't see why they should be made to come then.
It is not high load averages at issue here as much as LOW efficiency as
measured by cycles delivered to the users out of the potential.
-------
∂19-May-77 1000 JMC*
Call Theberge, Dowd, J.J. Wurtner 202 296-1683.
∂19-May-77 1142 FTP:Geoff at SRI-KA Dialnet papers.
Date: 19 May 1977 1142-PDT
From: Geoff at SRI-KA
Subject: Dialnet papers.
To: Rem at SAIL, Stef, JMC at SAIL
Folks,
The files <GEOFF>dialnet.paper on SRI-KA and DIALNE.[W77,JMC] are THE SAME
file. I spent an hour or so deleting all of the PUB directives from
JMC's file, putting in a special invisiable formatting characters
so that XED could make it look neeter. It then ran it though through
the XED ← command to justify without filling it. I'm sure if you went on
you would have found that the 2 files are exactly the same in fact, it
is that mine has all the the PUB nonsense pulled out and a little
prettying up done.
Hope this clears up all of the confusion that seems to have been flttflying
back and forth!
-------
∂20-May-77 1505 FTP:BOBROW at PARC-MAXC IJCAI panel
Mail from host Maxc2 received at 20-MAY-77 1458-PDT
Date: 20 MAY 1977 1457-PDT
From: BOBROW at PARC-MAXC
Subject: IJCAI panel
To: jmc at SAIL
So Nu Where's the answers? Can I have them by Tuesday at the latest
please (Net mail is fine)
Thanks
danny
-------
∂20-May-77 2053 MJC
LOADING
LOADER 4 PAGES
EXECUTION
302
304
307
311
313
314
316
317
319
323
326
328
331
332
334
337
338
341
344
346
347
349
352
353
356
358
359
361
362
367
368
373
374
376
377
379
382
383
386
388
389
391
394
397
398
401
403
404
407
409
412
416
418
419
421
422
424
428
431
433
436
437
439
442
443
446
449
451
452
454
457
458
461
463
464
466
467
472
473
478
479
481
482
484
487
488
491
493
494
496
499
End of SAIL execution
↑C
.
∂20-May-77 2056 ARG
Exit
↑C
.EX PROB.SAI
Swapping to SYS:SAIL.DMP
SAIL: PROB.SAI 1 2
LOADING
LOADER 4 PAGES
EXECUTION
302
304
307
311
313
314
316
317
319
323
326
328
331
332
334
337
338
341
344
346
347
349
352
353
356
358
359
361
362
367
368
373
374
376
377
379
382
383
386
388
389
391
394
397
398
401
403
404
407
409
412
416
418
419
421
422
424
428
431
433
436
437
439
442
443
446
449
451
452
454
457
458
461
463
464
466
467
472
473
478
479
481
482
484
487
488
491
493
494
496
499
End of SAIL execution
↑C
.
∂20-May-77 2055 PMF
301302304308311313316317319322323326329331332334337338341343344346347352353358359361362364367368371373374376379382383386388389392394397401403404406407409413416418421422424427428431434436437439442443446448449451452457458463464466467469472473476478479481484487488491493494497499
End of SAIL execution
↑C
.s
301302304308311313316317319322323326329331332334337338341343344346347352353358359361362364367368371373374376379382383386388389392394397401403404406407409413416418421422424427428431434436437439442443446448449451452457458463464466467469472473476478479481484487488491493494497499
End of SAIL execution
↑C
.s
301302304308311313316317319322323326329331332334337338341343344346347352353358359361362364367368371373374376379382383386388389392
394397401403404406407409413416418421422424427428431
434436437439442443446448449451452457458463464466467469472473476478479481484487488491493494497499
End of SAIL execution
↑C
.....s
301302304308311313316317319322323326329331332334337338341343344346347352353358359361362364367368371373374376379382383386388389392394397401403404406407409413416418421422424427428
431434436437439442443446448449451452
457458463464466467469472473476478479481484
487488491493494497499
End of SAIL execution
↑C
.......
.
.
.
.s
301302304308311313316317319322323326329331332334337338341343344346347352353358359361362364367368371373374376379382383386388389392394397401403404406407409413416418421422424427428431434436437439442443446448449451452457458463464466467469472473476478479481484487488491493494497499
End of SAIL execution
ss
↑C
.Too few arguments
?
.s
301302304308311313316317319322323326329331332334337338341343344346347352353358359361362364367368371373374376379382383386388389392394397401403404
∂20-May-77 2058 BZM
302
304
307
308
310
311
316
317
319
322
323
325
329
331
332
334
337
338
341
344
346
347
349
352
355
356
359
361
362
364
368
370
371
374
376
377
382
383
385
386
389
391
394
397
398
400
401
404
407
409
412
413
415
416
421
422
424
427
428
430
434
436
437
439
442
443
446
449
451
452
454
457
460
461
464
466
467
469
473
475
476
479
481
482
487
488
490
491
494
496
499
∂20-May-77 2058 EJG
302
304
307
311
313
314
316
317
319
323
326
328
331
332
334
337
338
341
344
346
347
349
352
353
356
358
359
361
362
367
368
373
374
376
377
379
382
383
386
388
389
391
394
397
398
401
403
404
407
409
412
416
418
419
421
422
424
428
431
433
436
437
439
442
443
446
449
451
452
454
457
458
461
463
464
466
467
472
473
478
479
481
482
484
487
488
491
493
494
496
499
∂20-May-77 2100 WD
302
304
311
313
316
317
319
322
323
325
329
332
334
337
338
340
343
344
346
350
353
355
358
359
361
364
365
367
371
374
376
379
380
382
385
386
388
392
395
397
400
401
403
406
407
409
413
416
418
421
422
424
427
428
430
434
437
439
442
443
445
448
449
451
455
458
460
463
464
466
469
470
472
376
479
481
484
485
487
490
491
493
497
∂21-May-77 1526 ME
∂21-May-77 0914 JMC
To: ME, ALS
E seems to have the following bug - at least on an Imlac. If a line overflows
what can be displayed on one screen line and then is shortened to fit by
deleting characters from the early part, there appear two identical lines
when the user does <ctrl><cr>. It's not a serious bug, because the
extra can be deleted, but it makes a user nervous, because one can
wonder whether E is fully aware that it has two lines.
[ME - I was unable to duplicate this problem myself (using the DM
simulator on a DD), and I'd like to know if you think it always happens
when you do what you describe. Secondly, I'm pretty sure that the second
"copy" of the line is just a display error, not an internal data error in
E. That is, there probably are not two real text lines with the same
text, but the second displayed version is accidentally drawn over the
wrong line, thus obscuring the real text of some line. Anyway, let me
know if you can make this doubling happen at will, and if so, what
commands you give to do it.
By the way, when something strange like this happens, you can immediately
give the command ⊗XTELLME which will make a report to ALS and me of what
the state of E was at the given time, including a list of the recent
commands you gave to E; this helps us track down bugs. Also, you can
verify what text is really on the page by giving the ⊗V command to force
the screen to be redrawn with the correct text.]
∂22-May-77 0328 FTP:FARBER at USC-ISI The current state of ''ai'' according to Consumers
Mail from SRI-KL rcvd at 22-May-77 0242-PDT
Mail from RAND-UNIX rcvd at 22-May-77 0236-PDT
Date: 21 MAY 1977 1550-PDT
Sender: FARBER at USC-ISI
Subject: The current state of "ai" according to Consumers
From: FARBER at USC-ISI
To: list:
Message-ID: <[USC-ISI]21-MAY-77 15:50:40.FARBER>
Either I didn't know how successful AI research has been or the
following is not ai based or what ever.
Consumers June 1977 P 334 "Selling it".
Everybodies home companion
"The trade publication Electronics Retailing tells of an absolute
must - a robot that will be programmed to do 12 'basic household
tasks such as mopping the floor, and will be equipped with a full
personality permitting him to speak and interact in any human
situation. Feed him a tape and he`ll teach the kids French; flip
a switch and he`ll continue teaching them while they sleep'.
The gizmo also reportedly can serve as a dishwasher, lawn mower
and cocktail server. Called the DA (Domestic Android) II, its
the product of Quasar Industries, stands 5'4", weighs 240 # (its
prototype at least) and supposedly will go on sale in about two
years. Yours for about $4000. , no marriage license required"
And that is NOT the April issue of consumers
Dave
-------
Redistributed-To: JMC at SAIL, Hart, Raphael, Sacerdoti
Redistributed-By: GEOFF
Redistributed-Date: 22 May 1977
∂22-May-77 1615 JAB Dialnet
Ralph and Mo told me that there was a meeting at 9:30 am. Would you
mind if I came along. If I sta in Calf. I will probably want to
take an interest in the project.
You can come. Meeting may be at 11 though. I'll mail you a note if so.
∂23-May-77 1358 JAB
I'd like to work on Dialnet on a full-time basis. By this I mean I would
make the project be my first and only responsibility if in return it would
by itself provide a decent income. I have no desire to split myself as
that would make difficult the developement of any loyalty to one project.
I would like over 7.5k per year for which I will provide a good deal over
30 hours a week.
∂23-May-77 1534 DEA Course Evaluations
John,
I am on the course evaluation committee, and I have the responsability
of evaluating CS 206. Please let me know what day would be most convenient
for devoting 10 minutes or so of class time to filling out evaluation
questionaires. Thank you very much.
- Doug Appelt
The last day of classes June 2 would suit me best.
∂23-May-77 1717 JBR
You have exceeded your disk quota.
The files listed below have been purged to reduce your disk
area to your quota of 2000
Before purging, your files occupied 2551
BACKUP.TMP[F76,JMC]
PPSAV.TMP[F76,JMC]
BACKUP.TMP[W77,JMC]
QQPUB.RPG[W76,JMC]
QQPUB.RPG[F76,JMC]
QQDO.RPG[206,JMC]
QQPUB.RPG[S77,JMC]
QQPUB.RPG[F75,JMC]
QQPUB.RPG[W77,JMC]
SIGART.LST[LET,JMC]
ENERGY.LST[LET,JMC]
GTREE.LST[206,JMC]
LEADER.LST[LET,JMC]
AIQUAL.LST[W77,JMC]
NEWELL.LST[LET,JMC]
DIALNE.LST[W77,JMC]
CRYPT.DMP[ 2,JMC]
CODE.DMP[ 2,JMC]
SOURC2.LAP[ 1,JMC]
2P.LAP[W77,JMC]
REPRES.PRO[ 1,JMC]
EPIS[ 1,JMC]
FUNS[ 1,JMC]
P1[ 1,JMC]
PART[ 1,JMC]
PERM[ 1,JMC]
PATH[ 1,JMC]
PATH2[ 1,JMC]
ANTIN[ 1,JMC]
SYLL[ 1,JMC]
MEET[ 1,JMC]
COMPIL[ 1,JMC]
TIMES[ 1,JMC]
TESTA.SAI[ 1,JMC]
TESTB.SAI[ 1,JMC]
TESTC.SAI[ 1,JMC]
TESTD.SAI[ 1,JMC]
DADDA[ 1,JMC]
PROB1[ 1,JMC]
ORDIN[ 1,JMC]
ROTA.SAI[ 1,JMC]
ROTAT.SAI[ 1,JMC]
ROTB.SAI[ 1,JMC]
ROTC.SAI[ 1,JMC]
SEMAA[ 1,JMC]
INTER[ 1,JMC]
SEMAB[ 1,JMC]
CONVER.SAI[ 1,JMC]
RELREP[ 1,JMC]
MC[245,JMC]
FWGC[245,JMC]
PATH[245,JMC]
PATH2[245,JMC]
DEMO[ 1,JMC]
LISPAD[ 1,JMC]
TESTE.SAI[ 1,JMC]
ICOM[ 1,JMC]
TRANS1[ 1,JMC]
SIMUL[ 1,JMC]
COMPU2[ 1,JMC]
SOURC2[ 1,JMC]
SOURCE[ 1,JMC]
PUZZ.SAI[ 1,JMC]
TCLUBA[ 1,JMC]
TCLUB[ 1,JMC]
BOARDS.SAI[ 1,JMC]
NEWDOC[ 1,JMC]
REFER.ENC[225,JMC]
LIB.RLS[206,JMC]
ECHO.FAI[206,JMC]
OUTLIN[206,JMC]
MCCRAC.LET[ 1,JMC]
PATHS.RLS[225,JMC]
SYMFUN.RLS[206,JMC]
PARKER[ 1,JMC]
GRPDAT.RLS[225,JMC]
GRPALG.RLS[225,JMC]
GROPER.RLS[225,JMC]
DIMEN.RLS[225,JMC]
CHAR.RLS[225,JMC]
PERMU2.RLS[206,JMC]
LARRY2[ 1,JMC]
COMPU2.LET[ 1,JMC]
DEG5.IN[225,JMC]
REPRES.RLS[225,JMC]
S4.REP[225,JMC]
DEG6.IN[225,JMC]
R42.IN[225,JMC]
PUZZ.SAI[225,JMC]
PUZZA.SAI[225,JMC]
PANIC.SOS[225,JMC]
ANNOUN[225,JMC]
PUZZB.SAI[225,JMC]
ENCYC1.PRO[ESS,JMC]
PUZZE.SAI[225,JMC]
MINE.DIR[ AI,JMC]
PTCH2.DIR[ AI,JMC]
DRIVE.DIR[ AI,JMC]
KNOW2.AI[ESS,JMC]
ARCH.PRO[ESS,JMC]
ARTNA1.ART[ESS,JMC]
STUD[206,JMC]
RECOG.LET[ 1,JMC]
MEMMTC.QUA[ESS,JMC]
LANG1.AI[ESS,JMC]
S5.QUA[ESS,JMC]
MTC71.QUA[ESS,JMC]
R1303.ART[ESS,JMC]
R1301.ART[ESS,JMC]
KNOW3.AI[ESS,JMC]
N30[ 1,JMC]
QA.PRO[ESS,JMC]
ARTNAT.ART[ESS,JMC]
IMAGE.PSY[ESS,JMC]
SEMAN.AI[ESS,JMC]
FRANK.STA[ESS,JMC]
MTCSYL.QUA[ESS,JMC]
COMPUT.STA[ESS,JMC]
AISY69.QUA[ESS,JMC]
FILEC.FAI[ESS,JMC]
LETTVI.REV[ESS,JMC]
CH4A[206,JMC]
R1300.ART[ESS,JMC]
LCFMEM.MTC[ESS,JMC]
PEARL.NOT[ESS,JMC]
PUZZ.RLS[226,JMC]
PUZB.SAI[226,JMC]
PUZZ.SAI[226,JMC]
TRANS.DOC[226,JMC]
BLISET.PRF[226,JMC]
BLISET.COM[226,JMC]
IMPROV.DOC[226,JMC]
SETRUL.DOC[226,JMC]
PCHECK.M2[226,JMC]
COND.PRF[226,JMC]
BACKSL[ 1,JMC]
SYSTEM.HAC[226,JMC]
SORTS.DOC[226,JMC]
PCHECK.M1[226,JMC]
ADD.COM[226,JMC]
ADD.PRF[226,JMC]
TH1.PRF[226,JMC]
TH1.COM[226,JMC]
NETJO2.PRO[ESS,JMC]
FALSE.PRF[226,JMC]
DI[226,JMC]
SCWORL.AX2[226,JMC]
ONE.PRF[226,JMC]
IPT73.REP[ESS,JMC]
NSFX.2[ESS,JMC]
NSFX.6[ESS,JMC]
NSFX.5[ESS,JMC]
NSFX.7[ESS,JMC]
NSFX.9[ESS,JMC]
NSFX.8[ESS,JMC]
NSFX.4[ESS,JMC]
NSFX.3[ESS,JMC]
HAYES.FRM[ESS,JMC]
ADTEL2.NOT[ESS,JMC]
PLATT.NOT[ESS,JMC]
REPRES.BIB[226,JMC]
ONTO.NOT[226,JMC]
TASK2.MAR[ESS,JMC]
TASK1.MAR[ESS,JMC]
LETTER.HE2[ESS,JMC]
XLET[ESS,JMC]
ZF.AX[ESS,JMC]
FOO73A[ESS,JMC]
FOON.PR1[ESS,JMC]
ROTE.SAI[ 1,JMC]
ROTD.SAI[ 1,JMC]
ROTF.SAI[ 1,JMC]
SWOPSI.RE[ESS,JMC]
HART.ME1[ESS,JMC]
GGGG[ESS,JMC]
PRO[ESS,JMC]
OUTPUT[206,JMC]
OUT2[206,JMC]
SORT.RLS[206,JMC]
SHUFF2.RLS[206,JMC]
SORTIT.RLS[206,JMC]
M29AUG.MES[ESS,JMC]
SYMPOS[ESS,JMC]
CH1.BH2[ESS,JMC]
CLOCK.FAI[ESS,JMC]
FONTS[ESS,JMC]
TASKS.206[206,JMC]
SLIDES[ESS,JMC]
FACMIN.MEM[ 1,JMC]
TODO[ESS,JMC]
W74.PLN[226,JMC]
CLASS.REG[226,JMC]
GAME.CHK[226,JMC]
FOO1[ESS,JMC]
FOL2.MEM[226,JMC]
FOL1.MEM[226,JMC]
SUPPES.QUE[226,JMC]
WUTHER.QUE[ESS,JMC]
ROY[LET,JMC]
TOLLES.LET[LET,JMC]
FORDH.LET[LET,JMC]
AD.NET[ESS,JMC]
COMPUT[ 1,JMC]
KNUTH.SAI[225,JMC]
NEWRUL.DOC[226,JMC]
RELIG.SIM[ 1,JMC]
TUCKER.LET[LET,JMC]
TUCKER.LE2[LET,JMC]
TOLLES.LE1[LET,JMC]
TCLUB.MEM[ESS,JMC]
PUZZLE[ 1,JMC]
NOTES2[ AI,JMC]
PUTCH3[ AI,JMC]
RUSS[ AI,JMC]
CHYLIN.LE1[ AI,JMC]
GEN1[ AI,JMC]
DRIVE3[ AI,JMC]
MPRO[ AI,JMC]
PUTCH[ AI,JMC]
PUTCH.DIR[ AI,JMC]
DRIVE[ AI,JMC]
AIBIB[ AI,JMC]
MINE[ AI,JMC]
SLAMEC.LE2[LET,JMC]
BOWDEN.LE1[LET,JMC]
CHERN.LE2[LET,JMC]
MATCH.RLS[206,JMC]
CARTES.RLS[206,JMC]
III[206,JMC]
SIMP.PRO[206,JMC]
FUNS.MLS[206,JMC]
BASIC.OLD[206,JMC]
MAY.ME[LET,JMC]
NSFX.10[ESS,JMC]
HENRY[ESS,JMC]
MC.AI[ESS,JMC]
BIRKHO.LE1[LET,JMC]
GAME.OLD[206,JMC]
FLANAG.LE1[LET,JMC]
ERIK.LE6[LET,JMC]
TAMBOV[ 1,JMC]
BARAB[ 1,JMC]
HARPER.LE1[LET,JMC]
PUZZE.F4[225,JMC]
BELIEF.AI[ESS,JMC]
ERIK.LE7[LET,JMC]
BAJCSY.LE1[LET,JMC]
STARKM.LE1[LET,JMC]
MT[ 1,JMC]
BRIT.LET[LET,JMC]
IRANI.LE1[LET,JMC]
BOWEN.LE1[LET,JMC]
LCOM5.RLS[206,JMC]
JAF.STA[ESS,JMC]
TERMIN.BLA[ESS,JMC]
EARLID.DOC[ESS,JMC]
RECUR[ESS,JMC]
ANDERS.RE1[ 1,JMC]
PEOPLE[226,JMC]
LCOM4.RLS[206,JMC]
OLSEN.LE1[LET,JMC]
ERIK.LE0[LET,JMC]
SUBCOM.LE1[LET,JMC]
MICHIE.LE2[LET,JMC]
NEWSCI.LE1[LET,JMC]
AHLHAU.LE1[LET,JMC]
MEMOS.CUR[LET,JMC]
PUBINT.LE1[LET,JMC]
HORNIN.LE1[LET,JMC]
FELDST.LE1[LET,JMC]
CHECK.DEP[LET,JMC]
TUCKER.LE1[LET,JMC]
SCANDU.LE1[LET,JMC]
MELTZ.LE1[LET,JMC]
PIASET.LE1[LET,JMC]
MILNER.REC[LET,JMC]
FLESS.LE1[LET,JMC]
TRESSL.LE1[LET,JMC]
PIASET.LE2[LET,JMC]
LICKL.LE1[LET,JMC]
COOP.LE1[LET,JMC]
LICKL.LE2[LET,JMC]
POE.LE1[LET,JMC]
JMILLE.LE2[LET,JMC]
IGARAS.LE1[LET,JMC]
ITO.LE1[LET,JMC]
CLARK.LE1[LET,JMC]
FLESS.LE2[LET,JMC]
MICHIE.LE3[LET,JMC]
TAVARE.LE1[LET,JMC]
SCIENC.LE1[LET,JMC]
SCRIB.LE1[LET,JMC]
ATKIN.LE1[LET,JMC]
KOCHEN.LE1[LET,JMC]
GREER.FAC[LET,JMC]
CU.LE1[LET,JMC]
RALSTO.LE2[LET,JMC]
MESSAG.SCI[LET,JMC]
MCCRAC.LE2[LET,JMC]
RALSTO.LE3[LET,JMC]
CHRON.LE1[LET,JMC]
LOGIC.QUA[LET,JMC]
SHIGER.LE1[LET,JMC]
PAMELA.LE1[LET,JMC]
TAUGNE.LE1[LET,JMC]
MORGAN.LE2[LET,JMC]
FROOK.LE1[LET,JMC]
LOMONA.LE1[LET,JMC]
APR.ME[LET,JMC]
LOGIC.PUB[LET,JMC]
CONSUL.LE1[LET,JMC]
VERE.LE1[LET,JMC]
PAWLAK.LE1[LET,JMC]
TSUDA.LE1[LET,JMC]
NOLL.LE1[LET,JMC]
ROBERT.LE1[LET,JMC]
RILI.LE1[LET,JMC]
RALSTO.LE1[LET,JMC]
MORGAN.LE1[LET,JMC]
MIYA.LE1[LET,JMC]
HEALTH.LE1[LET,JMC]
WILKS.LE1[LET,JMC]
LAST.LE1[LET,JMC]
SEDEL.LE1[LET,JMC]
CAMPB.LE1[LET,JMC]
ARDEN.LE1[LET,JMC]
BURNS.LE1[LET,JMC]
SHOCK.LE1[LET,JMC]
CU.LE2[LET,JMC]
DAVID.LE1[LET,JMC]
SCHOEL.LE1[LET,JMC]
CALHOU.LE1[LET,JMC]
FINCH.LE1[LET,JMC]
LYKOS.LE1[LET,JMC]
JMILLE.LE1[LET,JMC]
MITCH.LE1[LET,JMC]
JERRY.LE1[LET,JMC]
MCQUIL.LE1[LET,JMC]
PAM.LE1[LET,JMC]
DRLUK.LE1[LET,JMC]
OCT.ME[LET,JMC]
LCOM7.RLS[206,JMC]
MARK.LE2[LET,JMC]
FORSHA.LE1[LET,JMC]
COMPT.LE1[LET,JMC]
STORK.LE1[LET,JMC]
R1302.ART[ESS,JMC]
ASTRO2.ART[ESS,JMC]
FRIEDM.LE1[LET,JMC]
FF[ESS,JMC]
AIRLIN.MTC[ESS,JMC]
JFOL.USE[ESS,JMC]
MEARS.LET[LET,JMC]
GRIMM.BKP[ESS,JMC]
SCIAM.LE1[LET,JMC]
NYT.LE1[LET,JMC]
NAVROZ.LE1[LET,JMC]
POOPA[206,JMC]
TECH[ 1,JMC]
LISPN[206,JMC]
NOTES1[ AI,JMC]
LISPN[ 1,JMC]
TXP1.LSP[ 1,JMC]
STARK.LE1[LET,JMC]
PAT.LE1[LET,JMC]
REPRES[LET,JMC]
MDAVIS.LE1[LET,JMC]
IGARAS.LE2[LET,JMC]
KNOW.ART[ AI,JMC]
GLUGG[LIT,JMC]
GLUGG2[LIT,JMC]
FLESS.LE0[LET,JMC]
KNOW.NOT[ AI,JMC]
LARRY[ 1,JMC]
NAS.NOT[ESS,JMC]
TELLER.LE1[LET,JMC]
PERSON[ 1,JMC]
YOUNG.LE1[LET,JMC]
JAF.STA[LET,JMC]
AAAS.LE1[LET,JMC]
MEIR.LE1[LET,JMC]
PIERCE[ 1,JMC]
MCCRAC.LE1[LET,JMC]
JAF.LE2[LET,JMC]
KILPAT.LE1[LET,JMC]
STULET.AIL[ESS,JMC]
CRYPT.FAI[ 2,JMC]
MICHIE.LE1[LET,JMC]
WEISSK.LE1[LET,JMC]
AAAS.PRO[ESS,JMC]
ONEILL.NOT[ESS,JMC]
3D.NSF[ESS,JMC]
ISRAEL.ART[LET,JMC]
ISRAEL.LE1[LET,JMC]
CHENEY.LE1[LET,JMC]
OCTOBE.OUT[LET,JMC]
ONEILL.LE1[LET,JMC]
CORBAT.LE1[LET,JMC]
TES1.RLS[206,JMC]
TES1.MLS[206,JMC]
COMTOP.LIS[ESS,JMC]
COMNEE[ 1,JMC]
EFOO[ESS,JMC]
CALHOU.LE2[LET,JMC]
MICHIE.LE4[LET,JMC]
ALUMNI[ 1,JMC]
INFO.NEE[ESS,JMC]
NAMES[ESS,JMC]
INFO[ 1,JMC]
LIB.PRO[ 1,JMC]
AIMEM2.JMC[226,JMC]
NET1[ 1,JMC]
NETDOC.PLN[ESS,JMC]
EVANS.LE1[LET,JMC]
EVANS.LE2[LET,JMC]
SUFFIC.FRE[LET,JMC]
HOBBIT.2[LIT,JMC]
HOBBIT.3[LIT,JMC]
HOBBIT.4[LIT,JMC]
HOBBIT.TRU[LIT,JMC]
HOBBIT.BIO[LIT,JMC]
OCT.USE[ESS,JMC]
OLIVER.LE1[LET,JMC]
SOLIT.WRU[206,JMC]
OBVI.RLS[206,JMC]
SOLIT.RLS[206,JMC]
INTRO.AI[ 2,JMC]
COMMON.NS[LET,JMC]
DRELL.LE1[LET,JMC]
ADMIS.LET[LET,JMC]
SIGART.LE1[LET,JMC]
SIGART.BUL[LET,JMC]
ELKIND.LE1[LET,JMC]
MASSY.LE1[LET,JMC]
SIGART.LE2[LET,JMC]
BOWEN.LE2[LET,JMC]
SAFE1[206,JMC]
NOTE.MJH[206,JMC]
RUSSOF.NOT[ESS,JMC]
SOLF73[206,JMC]
INVOIC[LET,JMC]
ACM.LE2[LET,JMC]
ACM.LE1[LET,JMC]
MINSKY.ME1[LET,JMC]
FERRIS.LE1[LET,JMC]
JOSHI.LE1[LET,JMC]
CARGO.LIT[LIT,JMC]
NYT.LE2[LET,JMC]
SOCIAL[ 1,JMC]
BDAY.LIT[LIT,JMC]
BLOWIN.LIT[LIT,JMC]
1421.LIT[LIT,JMC]
AILAB.AIL[ESS,JMC]
ALGOL.SNG[LIT,JMC]
CORREC.ABS[258,JMC]
BACKUP.MCP[258,JMC]
HOMOMO.NOT[258,JMC]
MCP.PR1[258,JMC]
COND.PRF[258,JMC]
PQR.PRF[258,JMC]
SCHEM.WRU[258,JMC]
GREEK.JMC[258,JMC]
MCCUS.PRO[258,JMC]
FOLSCO.MEM[258,JMC]
BRACEW.LE1[LET,JMC]
EXTFOR.MEM[258,JMC]
ACTOR0[258,JMC]
NONTER.PRB[258,JMC]
FEMANO.LE1[LET,JMC]
NAUR.LE1[LET,JMC]
MTC.GRP[ESS,JMC]
PETERS.LE1[LET,JMC]
JAN.ME[LET,JMC]
ENERGY.ST1[ 1,JMC]
LEGUIN.CRI[LIT,JMC]
SCIFI.CRI[LIT,JMC]
HAYS.LE1[LET,JMC]
JAN14.DBX[258,JMC]
SUM.CAR[ 1,JMC]
GELLMA.LE1[LET,JMC]
ESSEX.LE1[LET,JMC]
SHELDO.NS[LET,JMC]
APPEND.CON[258,JMC]
LOOKUP.SPE[ 1,JMC]
FOLMAN.MOD[258,JMC]
HOBBS.LE1[LET,JMC]
BREMER.LE1[LET,JMC]
AILET[LET,JMC]
CONSUL.POX[ESS,JMC]
REYNOL.LE1[LET,JMC]
ACM.LE3[LET,JMC]
LEAST.CM1[258,JMC]
NORBYE.LE1[LET,JMC]
JAN.OUT[LET,JMC]
DEC74.IN[LET,JMC]
OUTLIN.256[258,JMC]
HAND.2[258,JMC]
JO.NOT[258,JMC]
MTC.BLA[258,JMC]
TAKASU.LE1[LET,JMC]
LOGIC.QUA[258,JMC]
HAND.W75[258,JMC]
OKAMOT.LE1[LET,JMC]
MANNA.PR2[258,JMC]
MANNA.CM2[258,JMC]
HARDPR.OOF[258,JMC]
MANNA.WRU[258,JMC]
DEMO.PR1[258,JMC]
HOBBIT.NOT[LIT,JMC]
HAND3.W75[258,JMC]
TAKASU.LE2[LET,JMC]
POOP[206,JMC]
VIS1[ 2,JMC]
PERM.EX[206,JMC]
VIS.CAP[ 2,JMC]
UNSOLV[206,JMC]
HAND.73[206,JMC]
EXX.PUB[206,JMC]
HAND1[206,JMC]
CH10[206,JMC]
MKMAP.RLS[206,JMC]
PROJEC.WRU[206,JMC]
NNN[ 2,JMC]
SYNTAX.PUB[206,JMC]
TTY[206,JMC]
VITASK[ 2,JMC]
CHRI.LE1[ 2,JMC]
VIS[ 2,JMC]
FRAGA.LE1[LET,JMC]
FIELDS.ME2[LET,JMC]
GORDON.LE1[LET,JMC]
GORDON.LE2[LET,JMC]
ENERGY.LE1[LET,JMC]
TASK[LIT,JMC]
FEB.ME[LET,JMC]
DEC.ME[LET,JMC]
ADHOC.THE[258,JMC]
SAKAI.LE1[LET,JMC]
COMPLR.PRO[ WD,JMC]
LICK.LE1[LET,JMC]
LICK.ME1[LET,JMC]
LICK[ESS,JMC]
PARTIT.LSP[ 1,JMC]
S4.PAR[206,JMC]
HAND.1[258,JMC]
PERMUT.RLS[206,JMC]
GAME.WRU[206,JMC]
CANNOT.AI[ESS,JMC]
BIBLE[245,JMC]
USER[245,JMC]
IDEAS[225,JMC]
SUZUKI.REC[LET,JMC]
SUSIE[ WD,JMC]
MUFTIC.LE1[LET,JMC]
SAKAI.LE2[LET,JMC]
VOR1[ 1,JMC]
SET.AX[226,JMC]
INT.AX[226,JMC]
SCWORL.AX[226,JMC]
W0.AX[226,JMC]
INTEG2.AX[258,JMC]
COND.AX[258,JMC]
EQUA.AX[226,JMC]
ORD2.AX[226,JMC]
SEQ.AX[226,JMC]
LISP.AX[226,JMC]
ADMIND.AX[226,JMC]
INDUC.AX[258,JMC]
INTEGE.AX[258,JMC]
COMPIL.AX[258,JMC]
SCOTT.AX[226,JMC]
SET3.AX[226,JMC]
SET2.AX[226,JMC]
BLISET.AX[226,JMC]
COND.AX[226,JMC]
ZF.AX[258,JMC]
EQUAL.AX[258,JMC]
EXTFOR.AX[258,JMC]
IGR.AX[226,JMC]
SCINT.AX[226,JMC]
ORD1.AX[226,JMC]
LIST.AX[226,JMC]
REST.AX[226,JMC]
MCP.AX[258,JMC]
MCP2.AX[258,JMC]
MCP3.AX[258,JMC]
TERMS.AX[226,JMC]
MRHUG.QUE[226,JMC]
WEINER.ME1[LET,JMC]
ASHLEY.LE1[LET,JMC]
WEINER.LE1[LET,JMC]
WEINER.GRP[LET,JMC]
74DEC1.AJT[LET,JMC]
WEINER.ME2[LET,JMC]
SMITH.LE1[LET,JMC]
GAREY.LE1[LET,JMC]
BERG.LE2[LET,JMC]
KOREA.POL[S75,JMC]
KOREA.2[S75,JMC]
RUSSIA.LE1[LET,JMC]
ENGINE.LE1[LET,JMC]
GELFAN.LE1[LET,JMC]
MT.PRO[ 1,JMC]
MAY75.OUT[MSG,JMC]
JUN75.OUT[MSG,JMC]
JUL75.OUT[MSG,JMC]
AUG75.OUT[MSG,JMC]
SEP75.OUT[MSG,JMC]
JAN75.OUT[MSG,JMC]
DEC74.OUT[MSG,JMC]
∂24-May-77 1058 RDR via TYMT cache
turns out it wasn't on. now with that knowledge it is
too soon to tell with these figures from watch, but
idle: 1.0% used 29.5% swpw 58.1% sked 9.8%
I'd say it says we should have taken off terminals before,
although not now as long as there is idle time.
∂24-May-77 1336 RCM borrowed book
I loaned you a library book (essays in honor of Carl Hempel)
which I believe is due in a few days. Could you please return it
either to me or to the library.
∂24-May-77 1613 DEA Course Evaluation
John,
I will be in your CS 206 class on June 2, at 11:00 to pass out
course evaluation forms, unless you would prefer to do the evaluation
at the end of the class period instead of the beginning.
Also, this is a friendly reminder of the panel discussion on
problems of funding computer science research for the CS department
colloquium on May 31 in which you agreed to participate.
The other panel members will be Nils Nilsson, Peter Hart,
and Bruce Buchanan.
I'll see you then. Thanks for your time.
- Doug Appelt
∂24-May-77 1637 LES Satellite communication meeting at MIT
To: Feigenbaum at SUMEX-AIM, Weiderhold at SUMEX-AIM, JMC
Ed,
This note records my impression of the meeting on May 19 at MIT's Endicott House
among representatives of CMU, IBM, MIT, Satellite Business Systems, Stanford,
and RPI. My general conclusion is that there doesn't seem to be a much payoff
for us (CSD) in acting now, but there might be for the University as a whole.
If the price is right, instructional networks might permit major universities
not only to assist each other but to provide much-needed continuing education
programs throughout the country.
The meeting was chaired by Michael Dertouzos and was heavily attended by the
MIT community (15 people). The other groups sent two to four persons each.
The SBS people (who are actually still working for IBM currently) discussed
organizational issues first (e.g. SBS is a partnership equally owned by IBM,
Aetna, and Comsat). They then described the communication services they
plan to offer.
TECHNICAL CHARACTERISTICS: their synchronous satellite will provide 8
channels, each with a capacity of 43 million bits/second.
All stations that are to talk to each other must share the same channel.
Each channel will be time division multiplexed into subchannels of from
2.4 Kb/s to 6.312 Mb/s. The low end is intended for teletype
communication and the high end for data communication, high speed
facsimile, and slow-scan television (it can't support commercial grade
TV). Digitized voice will use about 32 Kb/s.
The multiplexing frame time will be 10 to 20 ms and the one-way
communication delay will be about 1/4 second. Bandwidth allocations will
be controlled from a central station and can be altered (on request) in one
to two seconds.
Each ground station will include a sizable (around 10 foot) antenna, and
a room full of transmitters, receivers, multiplexers, controllers, etc.
The investment cost per ground station will be about $250,000 (not paid
directly by the user).
MARKETING CONSIDERATIONS: they expect to sell mainly to large commercial
users, each of which must contract for a certain data rate on a continuous
(24 hour) basis. Users may be permitted to temporarily exceed their
allocation but will have to pay substantially higher rates for this
supplemental capacity. They expect the system to be used mainly for voice
communication during the daytime, with data communications and facsimile
traffic moving at night. All scheduling and subchannel multiplexing must
be handled by the user.
Looked at just as a telephone communication system, they expect their
rates will be comparable to or slightly higher that WATS service. From
this viewpoint, the data transmission services ride "free", which should
make it attractive to large banking and commercial customers.
DISCUSSION BY UNIVERSITY REPRESENTATIVES
Attendees were invited to brainstorm on potential applications and produced
a somewhat random list. This was subsequently divided into lists of things that
can and cannot reasonably be done via Arpanet. Here (without comment)
are the lists.
CAN BE DONE (AT LEAST PARTLY) BY ARPANET, BUT MIGHT BE DONE BETTER
Library services
On-line journalism
Access to lab reports
Shared instructional materials
Mail and message systems
Load sharing
Networking studies
Satellite research
POSSIBLE NEW CAPABILITIES VIA SATELLITE
Classroom/seminar sharing
Continuing education
Teleconferencing
Shared mass computer memories
Multiprocessing
Educational graphics
I found the subsequent discussions and amplifications by participants to be
less than inspiring, though some were clearly excited by their ideas.
I remarked that we would be interested as users (not developers) in seminar
sharing and teleconferencing and that shared mass memories were of interest
to us, though we should probably do it first in a local network.
There was some discussion of current experiments, such as a freeze-frame
TV system (using a 50 Kb/s link) being used for teleconferencing by IBM;
as I recall it links San Jose with some place in Colorado. They mentioned
the CTS satellite, which has a 500 MHz bandwidth and is available for
experimental use of a month or two duration any time between now and some
time in 1978 when it exhausts its fuel. They have portable ground stations
(on trailers) to facilitate experimental hookups.
[In a subsequent conversation with Ken Down, I learned that the CTS satellite
was used to send a Knuth seminar to Carlton University in Canada a while back,
using a microwave link to Ames, where there is a ground station.
I also learned that there will be an experimental teleconference on satellite
communication between Stanford and the University of Montreal on June 9 and 10.]
In conclusion, I am not currently motivated to grab the ball and run.
∂25-May-77 1138 MJL
To: TOB, JMC, ZM, CCG, DEK, DCL, TW
May 20, 1977
TO: CSD Faculty and Ph.D. Students
FROM: Denny Brown
SUBJECT: Black Friday
The next faculty meeting is scheduled for Tuesday, June 7 in Polya 204.
The schedule and agenda for the meeting is below.
The Black Friday (sic) meeting to evaluate the progress of PhD students
will begin at 3:00 pm. Faculty members are expected to attend this
meeting to report on the progress of their advisees.
Students who were asked to provide further evidence about their continuing
progress will be reevaluated. Such students should make sure that they
have assembled the requested evidence, and that their advisors are well
informed as to their current activities. Advisors should be sure that
advisees are making reasonable progress according to Department guidelines
and that they are personally satisfied with their students' progress.
A list of students who will be discussed, including evidence requested at
the previous meeting, is available in the file BLACK.TUE[INF,CSD] or from
Moira. A list of the advisees of each faculty member is available in
ADVSRS.PTY[INF,CSD] or from Moira.
If you need any additional information regarding the Black Friday
meeting, please contact me. (DPB at SAIL)
Agenda for Faculty Meeting on June 7, Polya 204.
1. 1:30 - 2:15 Discussion of new admittees. Assignment to Advisors
and research projects. Folders are available in Moira's office
for those wishing to see them before the meeting. 2. 2:30 - 3:00 Regular
2. 2:30 - 3:00 Regular meeting.
a. Conferral of Spring Degrees.
b. Out-of-area Quals.
c. Report of Admissions Committee on Masters candidates.
3. 3:00 - 4:30 Black Tuesday.
DPB/ml
∂25-May-77 1939 PMF
I passed the AA qual. I still plan to get some sort of thesis proposal together.
I guess there is no rush however and so I plan to go home (Indiana) for a couple of
weeks.
∂25-May-77 2009 JBR E's buffer size
To: ME, PMF
CC: JMC
Is it true that E uses fewer than 19 buffers for reading files. It seems
to me that it would be a great improvement if the single most important
system program were as disk efficient as possible. E should especially use
19 buffers when reading in or reformatting a file, searching over page boundaries,
rippling, page switching, etc. I consider its implementation to have much higher
priority than fancy macros.
∂26-May-77 0615 JRA stoyan
humm, i'll start xeroxing one (at HP). sent him the version i gave
you but made the mistake of telling him that i had revised it. he's
something of a time and energy sink. we wants me to call everyone at
mit,bbn, and xerox who has been part of the lisp folklore, and prod them
to write to him.
tony hearn gave stoyan my name; i've been trying to think of an
appropriate way to "express my appreciation" to tony.
∂26-May-77 0949 ME
To: JBR
CC: PMF, JMC
∂25-May-77 2009 JBR E's buffer size
To: ME, PMF
CC: JMC
Is it true that E uses fewer than 19 buffers for reading files. It seems
to me that it would be a great improvement if the single most important
system program were as disk efficient as possible. E should especially use
19 buffers when reading in or reformatting a file, searching over page boundaries,
rippling, page switching, etc. I consider its implementation to have much higher
priority than fancy macros.
[ME - E does not use buffered mode. As a result, it will be non-trivial
to change the (dump mode) buffer size in E. I have been considering this
problem and haven't given up on it--I'm well aware of its significance to
efficient system operation as well as to the individual user of E.
However, macros are mostly done and I'm not going to stop what I'm doing,
especially since I would have to undo what I've done in order to work on
the buffering problem. (By the way, the one thing E does efficiently is
rippling.)]
The answer you received from ME confirms the opinion that we made the
right decision about him.
∂26-May-77 1357 FTP:GLS at MIT-AI (Guy L. Steele, Jr. ) Visit by GLS to SAIL
Date: 26 MAY 1977 1654-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr. )
Subject: Visit by GLS to SAIL
To: JMC at SU-AI, RPG at SU-AI
CC: GLS at MIT-AI, GJS at MIT-AI, DIANNE at MIT-ML, JM at MIT-MC
CC: ELLEN at MIT-MC, JONL at MIT-MC
It's probably about time we got all these details of my
visit squared away. I was thinking of going out on
Tuesday 26 July, for the MACSYMA Users Conference which is
Wednesday-Friday 27-29 July, and staying for about two
weeks following 26 July (that is, for about 1.6 weeks
after the conference is done) leaving sometime like
Wednesday fortnight. Jon L White (JONL) planes, I think,
to stay in California for about three days after MUC;
if so, we will probably rent a car and cruise around
to PARC and IBM San Jose and SRI and other bizarre places.
I figure on spending a solid week at least at SAIL proper
(not counting dinnertime, which should obviously be spent
at all the ∨best Chinese restaurants!). By that time I may
have some significant changes to LISP to test at that time
on the SAIL machine.
-- Guy
You will be welcome at SAIL at time mentioned. Anything you can do
or advise to help implement your changes at the LOTS Decsystem 20
site will also be much appreciated.
∂26-May-77 1535 FTP:GLS at MIT-AI (Guy L. Steele, Jr. ) Visit
Date: 26 MAY 1977 1832-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr. )
Subject: Visit
To: jmc at SU-AI
CC: GLS at MIT-AI
Thanks much. Assuming the number stuff I'm working on
gets done within a couple of weeks, I hope to attack
SAIL code and TOPS-20 code, and perhaps TENEX code,
all in one sweel foop, and have something usable by time
of visit.
∂26-May-77 1926 FTP:Wilber at SRI-KL A REMINDER
Date: 26 May 1977 1918-PDT
From: Wilber at SRI-KL
Subject: A REMINDER
To: JMC at SU-AI, MOON at MIT-MC
HELLO -- THE MONTH IS WEARING ON, WE'VE GOT THAT LONG WEEKEND
AHEAD OF US, AND YOU'VE BOTH BEEN SILENT SO FAR...OF COURSE THAT COULD
MEAN THAT THE LISP ARTICLES ARE ENTIRELY SATISFACTORY TO YOU AND THAT
I'LL NOT NEED TO COLLATE YOUR COMMENTS...IF THAT ISN'T THE CASE, THOUGH,
DON'T FORGET THAT I PROMISED TO GET THE COLLATED RESULTS BACK BY JUNE
FIRST, WHICH MEANS THAT I SHOULD HAVE IT TOGETHER AND IN THE (U.S.) MAIL
ON MONDAY, THE THIRTIETH...PLEASE TRY TO GET BACK TO ME BY MONDAY NOON
SO I CAN GET IT ALL TOGETHER AND INTO THE MAIL SOMETIME THAT AFTERNOON
OR EVENING...IF YOU HAVE IMPORTANT COMMENTS THAT YOU CAN'T GET TOGETHER
BY THEN, PLEASE LET ME KNOW BY THEN SO I CAN AT LEAST ASK THE BYTE FOLKS
TO ACCOMMODATE US...AND THANKS AGAIN FOR YOUR WILLINGNESS TO LOOK AT THE
ARTICLES...CHEERIO -- MIKE
-------
I am sorry to have to say this, but I am afraid that the LISP articles
are unsalvageably bad - at least with the time I am willing to give it.
The main problem is that the examples are written as though they were
Fortran programs, i.e. almost all PROGs. Someone with more time might
be of more assistance.
∂26-May-77 2252 RPG
∂26-May-77 1531 JMC
To: gls at MIT-AI
CC: RPG
You will be welcome at SAIL at time mentioned. Anything you can do
or advise to help implement your changes at the LOTS Decsystem 20
site will also be much appreciated.
Almost all improvements to MacLisp for SAIL will be directly
transferrable to LOTS (in fact I cannot imagine any that Steele
is now working on or contemplating that would not be), and I will
be glad to install same. GLS and JONL will be staying with me during
their visit.
-rpg-
∂27-May-77 0933 JAB Dialnet
Is there anything you would like to discuss with me about Dialnet?
I'd like to get an idea of what kind of work you were thinking of
having me doing so I could see how well it fits in with what I think
I can do.
John
PS: I leave for Boston Monday return Friday.
∂27-May-77 0957 FTP: RAJ REDDY(A610RR29) at CMU-10B IJCAI INVITED PAPER
Date: 27 May 1977 1257-EDT
Sender: RAJ.REDDY at CMU-10B
Subject: IJCAI INVITED PAPER
From: RAJ REDDY(A610RR29) at CMU-10B
To: MCCARTHY at SU-AI
- - - -
DEAR JOHN,
WE ARE FRANTICALLY WAITING FOR YOUR INVITED PAPER FOR IJCAI-77!
RAJ
-------
∂27-May-77 1759 JAB via AMET Dialnet meeting
To: LES
CC: JMC
I leave for Boston on Monday morning. I expect to return
Friday night. I'd like to talk either this weekend or next,
I would prefer as soon as possible.
∂28-May-77 1208 TOB
To: LES, JMC
Ram Nevatia, who is now a prof at USC,
wants to spend some time here in the summer,
about 2 months. He will be traveling back
and forth. I want him. He is asking that
we pay expenses. I think that is a good
deal.
Tom
∂29-May-77 1740 DCL FREELOADERS and HANGERS-ON
To: JMC, LES, JC
Our system is currently unusable for much although the CPU may sit
around doing nothing. This is usually due to disk accesses which lock
programs in core and thus exclude others. Something trivial may thus
be soaking up cycles indirectly. I notice certain users who do not
seem to be doing anything relevant to music 220 although they appear
under that guise:
PAS and DAS.
If these people are freeloaders, and any others, they should be stopped
at all times.
∂29-May-77 2300 RAK John Borcheck
To: LES, JMC
JAB is running one hidden channel+1 detached channgel. But tonight
I found a pirate using the initials EGK. I found him--he was someone
who said he was here because he is a "friend of JAB." Whatever
that guy is supposedly doing for free for the lab, it isn't worth it,
I bet.
Dick
An UNKNOWN called EGK was challenged by RAK and said he was a friend
of JAB. Using this machine is not a transitive relation.
∂29-May-77 2356 JAB
I HAVE PREVIOUSLY EXPLAINED THE RULES TO HIM BUT HE OPTED TO IGNORE MY
WORDS. SINCE THERE IS NO PRACTICAL WAY FOR ME TO ENFORCE SUCH RULES ALL I
CAN REALLY DO IS VOICE MY OBJECTIONS AND HOPE THAT THEY ARE HEEDED. IT IS
NOT A VERY PLEASANT SITUATION WHEN A GUEST EXCEEDS HIS WELCOME AND ONE I
DO NOT LIKE BEING IN.
∂30-May-77 0022 JAB
I WILL GET RID OF HIS STUFF AND MAKE SURE HE CANT LOGIN FROM THE NET WHEN
HE RETURNS HOME. I WOULD LIKE TO KNOW WHAT ONE SHOULD DO WITH VISITORS?
LETTING THEM TRY OUT THE SYSTEM CERTAINLY SEEMS A REASONABLE THING, BUT
WHERE DOES IT STOP? ONCE THEY LEARN A FEW TRICKS THEY CAN BE VERY HARD TO
KEEP IN HAND.
JOHN
I haven't had that problem, so I don't suppose it will happen often.
∂30-May-77 0204 JAB logout and deleting empty ufds
To: JMC, JBR
logout does not delete an empty ufd!
∂30-May-77 1224 JBR
To: JAB
CC: JMC, LES
the reason logoutt does not delete an empty ufd is
because that's what les asked for.
∂02-Jun-77 0433 DEW CSD Black Tuesday
1) Do you have enough information to evaluate my progress for Black Tuesday,
and if not, what would you like?
2) I'm supposed to get a reading committee. I'm not sure what that entails
but I would like you to be on it. I'll talk to you as soon as I can.
I plan to take a vacation starting June 1 so I'd like to do this immediately.
(The system clock is screwed up: now is 1pm on 30 May. Thanks, Dave
∂30-May-77 1705 DON
I'm implementing a serious side to a light issue: There is a new
game that's been making the rounds am I'm building in a hack
to keep it from being used during "peak" hours.
∂30-May-77 1943 JP
yes. thanks. I found that out a while ago when I digged into the ARPANET resource
handbook (hadn't thought it would be there). -J
∂30-May-77 2159 FTP:EGK at MIT-MC (Edward Kochanowski )
Date: 31 MAY 1977 0058-EDT
From: EGK at MIT-MC (Edward Kochanowski )
To: jmc at SU-AI
CC: reg at SU-AI
Dear Dr. McCarthy
I would like to extend my deepest apologies to you and to all concerned for
my presence at SAIL and LOTS, and for any trouble my presence has caused.
Please be assured that I had no intentions of being malicious or mischievous
in any way at all. My sole purpose in using these systems was to further my
knowledge and education in the field of Computer Science. I sincereley regret
that I had neglected to ask the permission of those in authority to use the
systems. I realize now that it was a serious act of misconduct on my part not
to do so. For this also I apologize and ask your forgiveness. From now on
before I even think of attempting to do anything of this sort I shall consult
with the people in charge and abide by their wishes.
Again, please except my sincerest apologies.
Respectfully yours,
Ed Kochanowski
Your apology is accepted. I was somewhat short-tempered, because I am
being pressed by users of both AI and LOTS complaining aobut lack of
compute capacity, and you were on both machines at crowded times. At
uncrowded times, requests for limited use for something of mutual
interest can sometimes be honored.
∂30-May-77 2347 100 : don adventure game
To: RAK
CC: JMC
The "prime-time" restrictions are implemented and working fine.
I set it to disallow anything longer than a quick demo between
noon and 5pm and between 9pm and midnight (RAK's suggested times).
(Incidentally, lest you think I've been spending all my time on
this since we spoke...no I haven't. It all worked quite well
quite soon.) Let me know if people seem to be figuring out ways
of evading the security and running during the day, or if you
think the hours should be further restricted. Also, how long
would you guess it will be before the end-of-quarter crunch dies
down? I'm leaving in a few days but would like to set a "holiday"
before I do--the program is willing to remember one upcoming
holiday which may span any number of days. I figure we can set
up "holiday hours" and set up a holiday called "summer".
∂31-May-77 0155 JBR
klh has been deleted. yes a purge deletes unknown users.
∂31-May-77 0755 FTP:JM at MIT-MC (Joel Moses )
Date: 31 MAY 1977 1055-EDT
From: JM at MIT-MC (Joel Moses )
To: JMC at SU-AI
I TOLD DERTOUZOS ABOUT YOUR REQUEST FOR PAYMENT.
∂31-May-77 1421 RDR via AMET CS105/106
To: DPB, CGN, WTL, MRB, MJC, ARS, PLW, wiederhold at SUMEX-AIM, JMC
Course descriptions say 105 and 106 are essentially the same course
except that 106 sometimes uses examples that are "mathematical
in nature". I have observed a desire for something significantly
different from 106 on the part of undergaduates at least. There are
people who want to see what programming is without encountering
the infamous hours at the computer of 105/106, even so much so
that they do not care about credit.
Anyway, I just wanted to mention this possibly unobserved point.
The math department has 5 or 6 different flavors of Calculus, only two
of which are even possibly as close as CS105 is to CS106.
How about a CS 10? A possible first step would be to prohibit
enrollment by graduate students, although the number alone would probably
accomplish this.
∂31-May-77 1514 FTP:Wiederhold at SUMEX-AIM Re: CS105/106
Date: 31 MAY 1977 1515-PDT
From: Wiederhold at SUMEX-AIM
Subject: Re: CS105/106
To: RDR at SAIL, DPB at SAIL, CGN at SAIL, WTL at SAIL,
To: MRB at SAIL, MJC at SAIL, ARS at SAIL, PLW at SAIL,
To: JMC at SAIL
In response to the message sent 31 May 1977 1421-PDT from RDR at SU-AI (David Roode)
I agree with the concept of a lower level introductory
course. I believe we can safely assume that all admitees to
CS PhD or Masters students as well as EE/CS
Computer Eng. students have this competence, or can aquire it
trivially in the summer. There used to be in fact a CS50, with
the same course description as CS 105. I dont know why the
level was changed.
I believe however that a fast , highly interactive terminal
system is a better teacher than nearly anything else. An
interpreter or incremental compiler of an Algol W like
language would be adequate. Pascal might be a candidate, the
language constrains users to relatively clean programs, and
allows semantically meaningfull diagnostics. Chiron
is yet another candidate although I understand that Floyd
feels that interactiveness interferes with good programming
practice.
I have a proposal drafted on a change of the systems curriculum,
Dot Dale (74365) can send you a copy if you are interested. Gio
-------
Students who take 105 and 106 will want to use the language
they learned in that course afterwards. If this language is too
inefficient, it may swamp LOTS. Apparently ALGOL W was OK, and I
hope CSD will choose a language that is no worse.
∂31-May-77 1923 FTP:JAB at MIT-ML (John A. Borchek ) egk
Date: 31 MAY 1977 2221-EDT
From: JAB at MIT-ML (John A. Borchek )
Subject: egk
To: jmc at SU-AI
unfortunately he also made himself a directory on lots where there
is little that can be done to stop it especially when i am in boston.
he is supposed to leave anyday now and i expect for him to be gone
when i return. unwanted guests can be hard to get rid of. i think
he is gone from sail and will be from lots shortly.
EGK was driven from LOTS, and I received a polite note of apology
promising not to use our machines without permission.
However, Dick Karp (RAK) just encountered someone else (AVB)
who claimed you told him it would be OK to use the machine at night.
∂31-May-77 2016 FTP:JAB at MIT-ML (John A. Borchek )
Date: 31 MAY 1977 2313-EDT
From: JAB at MIT-ML (John A. Borchek )
To: jmc at SU-AI, les at SU-AI
Would Monday be a good time to discuss Dialnet?
Suggest Monday at 2pm.
∂31-May-77 2021 RAK Another friend of Borchek's
To: LES, JMC
A pirate with the initials AVB was logged in tonight. He said that
John Borchek told him it was ok to use the system at night when it
wasn't busy. This is the second time in two days this has happened.
The last fellow, I understand, Borchek denied giving permission to.
Well, statistically, this thing isn't looking so believable to me
anymore.
I'm not as polite as I used to be--if you can imagine me being even
less polite than usual--so I told him that LES had to authorize all
users at all times.
Note that this instance was despite a direct rebuke from JMC to JAB
the last time it happened.
Dick
Borchek is in Boston, so it must have been before. I noticed AVB
but wasn't in a mood to give chase. I trust you repelled the
invasion.
∂31-May-77 2120 FTP:JAB at MIT-ML (John A. Borchek )
Date: 1 JUN 1977 0017-EDT
From: JAB at MIT-ML (John A. Borchek )
To: jmc at SU-AI, les at SU-AI
I am quickly coming to resent Mr. Karps behavior. He seems to blame
me fully for anything with which I am remotely connected. I make it
a habit of explaining the rules to everyperson. I tell them that while
it is possible to use sail that it should be first OK`d by Les or Jmc.
It is very difficult to make this point when the system makes no effort
to prevent illegal usage. I would like to know what are my responsiblities
with regards to this beyond reminding people of the rules. I see no
way I can prevent the actions of independent human beings, especially
when i am 2500 miles away! I also see little point in Mr. Karps tactics
there is no reason to hostily confront unknown users, often they need
only have the situation explained politely by an "authority" figure which
I am not. There is certainly no reason to treat these people as criminals.
I addition I resent this going over my head. If somebody objects to
something I do I think I deserve to have them discuss it with me first.
I will probably be very happy to fix anything that bothers somebody
but I cant do that until I am told there is a problem. Why must it
be assumed I will be unreasonable necessitating going over my head?
I am upset about this because it seems so silly to waste all this time
talking about such a thing.
∂31-May-77 2143 JAB via ML
∂31-May-77 2124 FTP: ANDY BECHTOLSHEIM(X335AB20) at CMU-10B car
Date: 1 Jun 1977 0023-EDT
Sender: ANDY.BECHTOLSHEIM at CMU-10B
Subject: car
From: ANDY BECHTOLSHEIM(X335AB20) at CMU-10B
To: jab at SAIL
- - - -
i called the police and they said they would rather like
to wait for you to get complete information.
the radio was not stolen. i really don't know what these
burglars were looking for maybe cash or checks and i hope
you hadn't any of that left in the car.
i remember that you used to have magnetic tapes in the car
but couldn't see any tapes now.
sorry for sail. i didn't know these people.
one person which i didn't know scared me off aand said
thaat it would be the second time in two daays that a friend
of borchek would use the system illigimatedly.
i wonder then why they allow local unknown people to log in.
if this causes only aggressions it would be better if
this would be disabled.
i will talk to les tomorrow about that.
sorry for causing you this trouble.
-------
∂01-Jun-77 0856 FTP:Wiederhold at SUMEX-AIM (Response to message)
Date: 1 JUN 1977 0856-PDT
From: Wiederhold at SUMEX-AIM
Subject: (Response to message)
To: JMC at SAIL
In response to your message sent 31 May 1977 1554-PDT
I believe a choice of language will be a concern for many, and
adequate performance at LOTS is an important, and underrated concern. After
using BLISS, and having to give up assigning homework in CS140B
because of intolerable compiletimes I have been super-sensitized.
I do believe that a compiler of the scope of ALGOL-W or Pascal
can both compile fast, avoid excessive recompiles due to the
straightforwardness of the languagr, and generate efficient code
for programs where listproceessing is not the dominant
aspect.
How does SAIL rate in your assessment?
The current state of CHIRON precludes judgement, the compiling
and execution times may be initially large, the error recovery
aspects may reduce the frequency of recompilation.
Gio
-------
∂01-Jun-77 0900 FTP:Wiederhold at SUMEX-AIM (Response to message)
Date: 1 JUN 1977 0900-PDT
From: Wiederhold at SUMEX-AIM
Subject: (Response to message)
To: JMC at SAIL, RDR at SAIL, DPB at SAIL, CGN at SAIL,
To: WTL at SAIL, MRB at SAIL, MJC at SAIL, ARS at SAIL,
To: PLW at SAIL
In response to the message sent 31 May 1977 1554-PDT from JMC at SU-AI (John McCarthy)
In response to your message sent 31 May 1977 1554-PDT
I believe a choice of language will be a concern for many, and
adequate performance at LOTS is an important, and underrated concern. After
using BLISS, and having to give up assigning homework in CS140B
because of intolerable compiletimes I have been super-sensitized.
I do believe that a compiler of the scope of ALGOL-W or Pascal
can both compile fast, avoid excessive recompiles due to the
straightforwardness of the languagr, and generate efficient code
for programs where listproceessing is not the dominant
aspect.
How does SAIL rate in your assessment?
The current state of CHIRON precludes judgement, the compiling
and execution times may be initially large, the error recovery
aspects may reduce the frequency of recompilation.
Gio
I think SAIL is OK as far as speed goes though not great. When SAIL
was written, I believe they neglected speed for features though not
to the extent of some other languages. Ralph Gorin (REG at SAIL)
has made some comparisons and should be included in this discussion.
∂01-Jun-77 0957 FTP:Boyer at SRI-KL Talk by Woody Bledsoe
Date: 1 Jun 1977 1150-PDT
From: Boyer at SRI-KL
Subject: Talk by Woody Bledsoe
To: ccg at SU-AI, jmc at SU-AI
Title: A Maximal Method for Set Variables in Automatic Theorem Proving
W.W. Bledsoe
Abstract: A procedure is described which gives values to set variables
in a theorem. This techniques has been used to prove theorems in
intermediate analysis (the intermediate value theorem), topology,
logic, and proram verification. The method is "maximal" in that a
largest (or maximal) set is usually produced if there is one.
Time: 2 p.m. Tuesday June 7
Place: 237 Digital Systems Lab
-------
∂01-Jun-77 1014 FTP:Wiederhold at SUMEX-AIM sail speed
Date: 1 JUN 1977 1013-PDT
From: Wiederhold at SUMEX-AIM
Subject: sail speed
To: reg at SAIL, jmc at SAIL
I would be interested in data on SAIL speeds(c&e) that you have,
and on size - since that may affect paging load critically.
I believe BLISS was greatly affected by its size in compiling.
Gio
-------
∂01-Jun-77 1141 LES
To: JMC, JAB
∂31-May-77 2116 JMC
To: JAB, LES
Suggest Monday at 2pm.
[Fine with me - LES]
∂01-Jun-77 1228 PAM
John I need to see you about this summer and about cs206. I'll be in
my office except for class tomorrow..Paul
∂02-Jun-77 0336 DON
To: wiederhold at SUMEX-AIM, RDR, DPB, CGN, WTL, MRB, MJC, ARS, PLW
To: JMC
From some correspondence I've seen floating around, I gather you might
be interested in some statistics I gathered last Fall comparing
execution (not compilation, I'm afraid) speeds for SAIL, Algol W,
Fortran H, PL/1-X, and PL/C on 4 programs with various "principal"
features (e.g. recursion, subroutine calls, etc.). I've got a sheet
summarising the statistics and could xerox it and give it to Gio for
reference. (If you have any questions about the test programs or
anything else related to this, catch me soon, because I'm leaving Friday
evening to fly east for two weeks in Boston.)
I would like a copy.
∂02-Jun-77 2143 FTP:GLS at MIT-AI (Guy L. Steele, Jr. ) LISP Historian from Germany
Date: 3 JUN 1977 0042-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr. )
Subject: LISP Historian from Germany
To: jmc at SU-AI
CC: GLS at MIT-AI
Fredkin recently mentioned to me this person from
E. Germany (I forget his name) who is interested in
LISP history and has been collecting material. I have
never interacted with him, but I know he and Jon L White
here have exchanged a letter or two. I am also
interested in LISP history, but haven't had much time to
devote to it, and I think this fellow ought to be
encouraged. Fredkin said you know something about him,
and maybe there is something I can do in the way of
sending him material, getting other people to help
him or invite him to U.S., etc.?
Dr. Herbert Stoyan
DDR 806 Dresden
Togliattistr. 40
East Germany∞
I think it is entirely worthwhile helping him. The best thing
you could do would be to find reports from Research Laboratory of
Electronics mentioning LISP and send him copies. RLE has a document
room and Fredkin could probably finance the copying. If you read
German, I could sent you his drafts.
∂02-Jun-77 2041 RDR via AMET charging for use of LOTS
When libraries weren't free, didn7't students in Universities
have access to libraries included in their attendance at no charge?
Now libraries are commonly public. It doesn't seem that computing
is to do anything other than get cheaper.
Also, a lot more value can be gotten out of students in the form of
labor to support LOTS than by charging for use of LOTS. Enthusiasm
is not yet so high that monetary discouragement will fail to
have an adverse effect. If even a quarter of the student body were
regular users, the increased interest would generate
enthusiasm on the part of the administration for funding LOTS.
Also, Stanford has a policy against lab fees. Chemistry classes
charge DEPOSITS, REFUNDABLE at the end of the quarter. I think
it's part of Mrs. Stanford's will or the founding grant or some
such, thank goodness. i think you do not realize the
effect of usage related fees on those students who are only
here with financial assistance. I also believe that such fees
would be held in viloation of the principle of making
facilities of the University equally available to all students.
Protests would soon reach the University committees like CUAFA
were charges ever implemented for LOTS, assuming
implementation in a quick and sneaky manner precluded such
protests in advance.
I meat to be negative but not quite this wordy. It isjust that
A LOT of LOTS positive nature is the lack of accounting, not
to mention the increased expense in processing any sort of
bills for 2000 students.
∂02-Jun-77 2154 RDR via AMET last point
Someone just said you didn't think charges would discourage use or
make a sunstantial contribution to the operaing budget.
He sad you thought there would be objections if usage
expanded much and was paid for out of general funds.
No one objects to the wasted million in administrative
computing. For a million a year, lots could have 8 times as much
hardware. Objections to funding of LOTS because the objector
cannot get sufficient service can only result in a net increase in
funding. Peopl who don7t want LOTS' service won't complain if
the funding is under a million a year serving at least 25% of
the university. It all comes under the tradeoffs of libraries versus laboratories
and football stadiums. Stanford, even with LOTS, doesn7t stack up
very well against other Universities in such services.
Next years expansion just starts approaching reasonability.
The external advisory committee was basicaly appalled at
the state of computing at Stanford -- computing FOR THE STUDENTS,
in view of the facilities for research.
The points you make are good but not necessarily decisive. As long
as LOTS remains as inexpensive as it is, there is no need for fees.
The range of usage between 15 percent and 50 percent is the range
of maximum potential strain. You realize that LOTS could not itself
decide for fees. The University would indeed have to decide to
reverse all the traditions you mention.
∂03-Jun-77 0911 FTP: RAJ REDDY(A610RR29) at CMU-10B Invited Paper
Date: 3 Jun 1977 1213-EDT
Sender: RAJ.REDDY at CMU-10B
Subject: Invited Paper
From: RAJ REDDY(A610RR29) at CMU-10B
To: jmc at SU-AI
- - - -
Dear John,
Because of the delays by various invited speakers, we have decided not to
include it a part of the proceedings but rather publish it separately.
The present plans are to publish the invited papers as a
separate book by MIT-Press. It
would be highly desirable to have a comprehensive paper representing
your views in that volume because I expect it will be refered
to for many years to come. It would be a shame to include either
an incomplete, unreviewed, or unpolished paper in the book. Could
you please try to work on the complete version of the paper
and send it to me as soon as possible.
I am also concerned about your oral presentation since there will
be over a thousand people present. Many of them will not be familiar
with all the relevant issues of your data. It would be worth trying out your
talk on groups that are not overly familiar with your work to get
their reactions. Otherwise you will be missing a golden opportunity
at the IJCAI-77 to make your views understood and appreciated.
Sincerely,
Raj
-------
∂03-Jun-77 1017 FTP:GLS at MIT-AI (Guy L. Steele, Jr. )
Date: 3 JUN 1977 1317-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr. )
To: JMC at SU-AI
CC: GLS at MIT-AI
I do read German, and would like very much to
see his work. I will see what I can dig up around
here for him.
∂03-Jun-77 1103 JP New ET.FAS on MAC,LSP
To: MACLSP.DIS[AID,RPG]:;
There is a new version of ET.FAS (the MACLSP → E → MACLSP swapper) that
permits specifying page and line numbers, as well as swapping back to a
clean MACLSP after using E. The LSUBR ET expects from 0 to 3 arguments,
where
arg1 : Filename in format shown below
(fil) ;uses E default extensions in disk dir
(fil ext) ;disk dir
(fil ext (prj prg))
(fil (prj prg)) ;default ext
arg2 : Page number
arg3 : Line number
In case of missing page or page/line numbers default values are used.
Notice that page/line #s are interpreted according to your current fixnum
base. ET with no args swaps to E on your tmpcor file, if any:
(et) ;runs E on last file edited
If you swap to E with a filename specified, no tmpcor file gets
written. That is, getting back into MACLSP and doing (et) will NOT do
what you expect. This is a feature of E. If you want to write a tmpcor
file use the Xtend command TMPCOR within E.
Finally, ⊗XRUN will swap you back to a fresh MACLSP (not your old core image).
Bugs, comments → JP
∂03-Jun-77 1629 JB My Progress Report (in view of Black Friday on Tuesday!).
I put a progress report on your terminal. Should it have disappeared,
it is in REPORT.1[GOL,JB].
Of course I will be glad to supply any additional information, for
I did not mean to do this report unnecessary detailed.
∂04-Jun-77 1100 JMC*
don't forget resche.xgp
∂04-Jun-77 2050 DCO
To: MACLSP.DIS[AID,RPG]:;
I have added one function to CGN's line editor load package.
(inited) initializes a buffer to receive input destined for the line editor.
(added <whatever> <switch>) adds <whatever> to the existing contents of this
buffer.
(loaded) loads the temporary buffer into the line editor buffer.
The new function is (loadcr) which is the same as (loaded) except
that after the line editor buffer is loaded, it activates the contents as
though a type-ahead carriage return had been typed on the terminal. If
anyone wants to, loadcr could easily be modified so that any string of
activation characters (including control and/or meta bits) would be activated
after loading the line editor buffer.
∂04-Jun-77 2339 RDR via AMET cheap 103 modems
To: cbf at MIT-ML, mlk at MIT-ML, archy at MIT-ML
To: lsp at MIT-ML, geoff at SRI-KL, lynch at SRI-KL, REM, JMC
There is an Ad in April BYTE magazine for $105 originate AND answer
modems, "fully assembled and tested". 10 day money back guarantee.
RS232 and EIA. 90 day full warranty. $25/year service contract
available. Vendor is US Robotics, Box 5502 Chicago 60680
(312)528-9045.
Does anyone have any experience with these modems or the company?
They appear to be 1/2 or 1/3 of the cost of 103 modems in the past
and as such could prove useful for dialnet, home computers,and LOTS.
∂05-Jun-77 0037 DPB DIStribution for discussion on choice of language in Intro courses
To: INTRO.DIS[1,DPB]:;
I will maintain a distribution list of people who are interested in
the discussion of languages at LOTS. INTRO.DIS[1,DPB] will contain
the current list of interested parties. Suggestions for additions and
deletions are welcome. -Denny
∂05-Jun-77 0044 DPB MAINSAIL for Intro Courses?
To: INTRO.DIS[1,DPB]:;
Clark Wilcox recently sent me a message suggesting MAINSAIL for Intro
courses. He (and it) should be included in our discussions. WILCOX%SUMEX
has been added to INTRO.DIS[1,DPB]. -Denny
∂06-Jun-77 0610 FTP:Wilber at SRI-KL CHEAP 103 MODEMS
Date: 6 Jun 1977 0604-PDT
From: Wilber at SRI-KL
Subject: CHEAP 103 MODEMS
cc: CBF at MIT-ML, MLK at MIT-ML, ARCHY at MIT-ML, LSP at MIT-ML,
cc: GEOFF at SRI-KL, LYNCH at SRI-KL, REM at SU-AI,
cc: JMC at SU-AI
Date: 6 Jun 1977 0556-PDT
From: Wilber
Subject: (Response to message)
To: RDR at SU-AI
cc: WILBER, PBARAN at ISI
In response to your message sent 4 Jun 1977 2339-PDT
HI DAVID -- I DON'T HAVE ANY EXPERIENCE WITH THE PRODUCT OR THE COMPANY,
BUT I NOTICE THEY'VE ALSO ADVERTISED THAT MODEM IN THE FEBRUARY AND
MARCH ISSUES BUT NEITHER BEFORE NOR SINCE, AND THAT THE ADS ARE IN THE
LOW-RENT DISTRICT...DAVE CAULKINS (PBARAN@ISI) POINTS OUT THAT A DIALOUT
LINE CONTROLLER (E.G., A FAST AUTO-ANSWER MODEM) CONTROLLED BY A BESERK
PROGRAM CAN PHONE AN INNOCENT BYSTANDER EVERY FIVE MINUTES THROUGH THE
NIGHT AND BRING THE PHONE COMPANY DOWN ON YOU HARD, AND THAT OBSERVATION
SORT OF DAMPENED THE PCNET GLOBAL ISSUES COMMITTEE'S OUTLOOK ON THE
FLEXIBILITY A PCNET MIGHT ATTAIN...LET ME KNOW IF YOU HEAR ANYTHING
EITHER WAY ABOUT USROBOTICS OR THEIR CHEAP MODEM / LINE CONTROLLER...
AND GOOD LUCK -- MIKE
-------
-------
∂06-Jun-77 1131 PAM
John .. Do you want me to give the exam or will you? I have tickets for a
concert that night. Also, do you need any suggestions for the final? I looked
over your old exams to give a midterm makeup, and I liked the fibonacci number
problem (where you give the explosive definition, ask why it's bad, and ask
for a better one). Sadly, the one guy who took the makeup midterm has already
seen it, but I thought I'd suggest that kind of problem anyway. He didn't seem
to understand the problem very well, but that was consistent with his other work.
I will make up the exam and administer it, but I must go to Washington
on Thursday, so I want to have a big grading party Wednesday night
starting 8pm here. We will grade the exams and the projects and make
out grades.
∂06-Jun-77 1254 PAM
Sounds good to me... I'll see you at the lab Wednesday night..Paul
∂06-Jun-77 1639 FTP:EF at MIT-AI (Edward Fredkin )
Date: 6 JUN 1977 1937-EDT
From: EF at MIT-AI (Edward Fredkin )
To: jmc at SU-AI
hi, let me know if you
get this message, I assume that the telegram to the
the soviets exxx will not mention the list
of recipients.
∂06-Jun-77 1641 FTP:EF at MIT-AI (Edward Fredkin )
Date: 6 JUN 1977 1939-EDT
From: EF at MIT-AI (Edward Fredkin )
To: jmc at SU-AI
Why not send Sharansky a copy, it
will be good for his morale!
I got your message, and doubt that prisoners are allowed to receive
telegrams, but we can send it anyway.
∂07-Jun-77 1122 JAB via AMET
To: LES, JMC
I HAVE TO TELL DEC ON WEDS IF I WILL ACCEPT. MY PREFERENCE IS TO REMAIN
HERE AND WORK ON DIALNET. IF THERE IS ANY REASON I SHOULD NOT DO THIS
ASIDE FROM WHAT HAS ALREADY BEEN MENTIIONED LET ME KNOW.
Already sent to JAB
I know no such reason. However, we still have to go through the
procedure of listing the job. We have never failed to get the person
we want so the risk to you seems negligible. My experience suggests
that if you tell D.E.C. that you prefer the Dialnet job, but would take theirs
in case of a hitch and need three weeks, the probabilities are that
they'll keep their offer open, and it would have a good long range
effect if you ever want to work for them in the future.
∂07-Jun-77 1128 LES
To: JAB
CC: JMC
Dialnet position
I believe that we will be able to generate a reasonable offer, as we discussed.
∂07-Jun-77 1443 LES Tuition for PAT
Patte wants to learn SAIL this summer by taking CS105. I think that she
should take it for credit, rather than auditing. The University will kick
in about $75 of the $325 (approx.) tuition but expects the Department to
cover the rest. I tried to get it out of Ed, but he thinks that you should
cover it out of unrestricted funds. What say you?
I say my unrestricted funds are broke. Until I get an NSF grant for
theory, I am running into the hole and will eventually be forced
to lay off RWW.
∂07-Jun-77 1449 FTP:JAB at MIT-MC (John A. Borchek ) DEC
Date: 7 JUN 1977 1614-EDT
From: JAB at MIT-MC (John A. Borchek )
Subject: DEC
To: jmc at SU-AI
Thanx for the suggestion.
I can write up a blurb for CACM in a few days. I would put an invitation
to join our mailing list in it. Setting up a conference seems
premature for the first announcement.
I also thought about the modem. It might be easier to store a digital image
of each character rather than try to generate it each time. When we were
playing with the 11/20 we found complex programs were at times too slow.
Of course it was an old 11/20.
There is something called the X-25 protocol for information transmission,
and EDUCOM is supposed to be using it. Professor Gene Franklin knows who
knows about it. Please find out if it is something we should use.
∂07-Jun-77 1521 RPG New Maclisp
To: MACLSP.DIS[AID,RPG]:;
There is a new Maclisp up with 6 input/output buffers;
any MACLISP systems should be regenerated. Report bugs to RPG.
-rpg-
∂08-Jun-77 0948 DCL MONITOR LIMIT ON JOB SIZE
To: JMC, LES, JBR
I have been told that the Monitor is supposed to limit jobsize to
256k pages, and that it does not yet do so because of a bug.
If this is true, and the bug is fixed, we will be unable to
run any of our current experiments on verifying complicated programs
or the absence of runtime errors. This often requires up to
350 pages.
Please let me know what the proposed jobsize limit is.
∂08-Jun-77 1210 JBR
To: DCL
CC: LES, JMC
[∂08-Jun-77 0948 DCL MONITOR LIMIT ON JOB SIZE
To: JMC, LES, JBR
I have been told that the Monitor is supposed to limit jobsize to
256k pages, and that it does not yet do so because of a bug.
If this is true, and the bug is fixed, we will be unable to
run any of our current experiments on verifying complicated programs
or the absence of runtime errors. This often requires up to
350 pages.
Please let me know what the proposed jobsize limit is.]
There is no bug, and the limit is 196K words = 392 pages.
∂08-Jun-77 1247 LES ARPA approval
To: JMC, TOB, CCG, DCL, RWW
Good news: ARPA Headquarters approved our proposal this morning, apparently
without a struggle.
∂08-Jun-77 2101 REF Core war
Implementation suggestions in COREWA.REF[G,REF]. Looks (relatively)
easy to implement, but damn slow to run.
∂09-Jun-77 1022 TOB mech eng project
Dick Liu asked about possible projects for
student design for next year, like the piano
mover. What do you think about a powered
mobility aid, with computer control, for
disabled children and adults?
Tom
I have no more money for any kind of project.
∂09-Jun-77 1025 RWW NSF PROPOSAL
Would you like me to do anything, we probably should start soon.
Yes, we should start next week.
∂09-Jun-77 1415 RPG sewer OLDIO
To: MACLSP.DIS[AID,RPG]:;
At SAIL Maclisp has been modified as follows: ordinary disk file
IO now uses 6 buffers in the input (output) buffer rings and fasload reads
1400(8) words at a time from disk in dump mode while fasloading. Since the
bottleneck at SAIL is the disk right now these changes will greatly
improve the performance of Maclisp AND the rest of the lusers on the
SAIL machine. However, there are a few things to watch out for:
1. the correctness of Require in this new scheme has not been proven.
2. BPS will need to be increased for fasload. Essentially fasload
uses array space for its dump mode buffering. So, the increased size
(200 to 1400 (8)) means that BPS should be about 1200(8) [640.] words larger
than before. You ought to go through your *.ini files and make this
update or face "no core -- array" errors.
-rpg-
∂09-Jun-77 1431 FTP:EF at MIT-AI (Edward Fredkin )
Date: 9 JUN 1977 1729-EDT
From: EF at MIT-AI (Edward Fredkin )
To: jmc at SU-AI
John, I haven't been able to get through to Michie yet, but I
hope to do so Friday AM. Please mail (on the net) me
another copie of the current telegram. I lost my old copy.
After quite a bit of thought I have decided to switch from
a somewhat reluctant telegram signer to an aroused activist.
I would appreciate more suggestions on what to do.
How about a group of us requesting a meeting with the
Ambassador, giving him a chance to explain his view, but
with us forcefully making the point that they may lose
tangible benefits by alienating the people who might
have helped them with technological problems. My guess
is that he would refuse to see us but some good might
come of the attempt.
Ed F
I think you might be good at seeing an ambassador. I don't think I am.
I just got back from a trip. I'll think of some more ideas.
∂09-Jun-77 2316 CLT Assistantship
I accept. I plan to spend next week at home (in Santa Cruz). I will be
back sometime Monday June 20th. If you wish to contact me before then, the
phone number in Santa Cruz is: 408-426-3238.
∂10-Jun-77 1145 JP E - MACLSP - E
To: MACLSP.DIS[AID,RPG]:;
ET.FAS[MAC,LSP] has been modified so ⊗XRUNning from E returns you
to the calling MACLSP image. LSUBR ET now returns nil. Comments → JP
∂10-Jun-77 1617 FTP:Feigenbaum at SUMEX-AIM TIP to Stanford
Date: 10 JUN 1977 1608-PDT
From: Feigenbaum at SUMEX-AIM
Subject: TIP to Stanford
To: lederberg, jmc at SAIL, les at SAIL
cc: rindfleisch, feigenbaum
I consulted with John and Les about their OK for putting the
TIP at SAIL rather than at SUMEX. Having received their OK,
I asked Walker if that was acceptable to ARPA. He said
that was OK with him, but he did want to see it at Stanford
(i.e. where to put it was up to us).
ed
-------
∂11-Jun-77 1745 RAK The appearance of evil
To: JAB, JMC
Alas, when one continually hides a channel, others assume that there
is a reason to hide that channel--i.e. doing LOTS work from the AI
lab instead of AI work. Presumably that is not happening, so it might
seem desirable to avoid the appearance of evil as well as evil itself....
Dick
If it ain't broke, don't fix it. - Bert Lance
∂12-Jun-77 1331 JAB via AMET networks
The May issue of Datamation has an article by Vint. Cerf you may wish
to glaance through. There is also an annoucement that x-25 protocol
is thought likely to become some sort of standard. It seems alot
of the manufactureers plan to use it for their network products.
∂13-Jun-77 0031 RAK Re--Aint broke
You're right, of course, I regreted that message just after
I sent it but it is ungentlemanly to munge others mail.
However, he was doing non-lab stuff at the time I sent the
message. There are lots of ways to read hidden channels,
and I know at least two......
Dick
∂13-Jun-77 1013 FTP: RAJ REDDY(A610RR29) at CMU-10B IJCAI-PAPER BY YOU AND HAYASHI et al.
Date: 13 Jun 1977 1313-EDT
Sender: RAJ.REDDY at CMU-10B
Subject: IJCAI-PAPER BY YOU AND HAYASHI et al.
From: RAJ REDDY(A610RR29) at CMU-10B
To: JMC at SU-AI
- - - -
Dear Dr. McCarthy:
I had a letter from Dr. Hayashi today (dated June 6, 1977) saying that he had
not received any information of the acceptability of the paper "On the Model
Theory of Knowledge" ID#275. It was accepted as a short paper and the model
paper for the final version was sent. The problem is did we mistakenly send
the notice etc. to you instead of Dr. Hayashi?? If so, can you take care of
it as we are nearing the deadline. Please let us know something soon.
Beverly Howell, Secretary
Raj Reddy.
-------
To Beverly Howell:
You didn't send the notification to me on the paper by Hayashi et. al.
∂13-Jun-77 1050 JMC
To: RAJ, JMC, reddy at CMU-10B
To Beverly Howell:
You didn't send the notification to me on the paper by Hayashi et. al.
This is a repeat, since there may have been a transmission error.
∂14-Jun-77 0003 DPB MTC syllabus
To: ZM, JMC
Is the official word that the syllabus is unchanged from the old (couple
of years old) one in Moira's files? I thought some changes were planned
but I haven't heard. -Denny
There is a new syllabus that Zohar made. I'll have to find it.
∂14-Jun-77 1035 DCL
To: JC, LES, JMC
Is MMM, McNabe a privileged 220 student too?
∂14-Jun-77 1121 JAB via AMET
I have given some more thought to DEC and decided to accept their offer.
The main reason for this is my basic unhappiness with living in Calf. I
fear that this will impact on my work. It perhaps is more being half-time
that upsets me but I am not sure if going to full out here is the answer.
It certainly is not something to risk. I would still like to maintain an
interest in Dialnet and the AILab in general. I think I would be able to
help out now and then especially with my being at DEC.
Thanx for the offer but I think the responibility on top of being so far
away from my roots maybe too much. Please keep me informed on events and I
will keep my eyes open for any good things from DEC.
John
PS: I hope to talk to you before I leave.
∂14-Jun-77 1254 EJG Core War
John, I just wanted to say that if there exists such a thing as a
Core War "distribution list" or "interest list," I would like to
be included - to get reports of its development status, other
people's remarks about it, etc. Thanks.
∂15-Jun-77 1104 PAT article
Cuthbert Hurd called - there is an article in THE KEY REPORTER (Phi Betta Kappa)
publication by Dyson entitled "The Next Industrial Revolution", vol. 42, #3,
Spring 1977, that he thinks you would be interested in reading.
∂15-Jun-77 1322 WP
∂14-Jun-77 2325 JMC
I have forgotten where I told you the MTC syllabus was located.
**********************************************************************
Answer:
The MTC syllabus I know about is on page 5 of QUALS.ALL[INF,CSD].
Wolf
Thanks. I had forgotten we had moved it there.
∂15-Jun-77 2056 RWW whats up
in case your wondering what i've been up to
1. writing (4 notes)
2. helping todd wagner with hardware verifier
3. working with dsb on reflection
4. working withjb on goal structure (slow going)
5. working with scott kim on axiomatizing facts about polyhedra
he wants to prove euler's theorem.
6. working with ray elschlager may be a reader on his thesis
7. some minor debugging
reflection almost here
problem with tauteq i can't find
etc
rww
nsf[s77,jmc] is a start on an nsf proposal.
∂16-Jun-77 1001 JMC*
Call fredkin, loebner, lerman, fernbach
∂16-Jun-77 1001 JP another freeloader?
To: LES
CC: JMC
Les,
I just noticed this guy logging in thru the net, with a directory
but completely unknown to the system (its 9:59 jun 16 right now)
.fing grm
Person Job Jobnam Idle Line Room Location
GRM UNKNOWN 23 NS PTY121 job 22 Arpanet site MTRT
↑C
.find grm
0 HITS ON KEY = grm
↑C
.r ppsav
∂16-Jun-77 1016 RCM
Re SRI - I had a job interview with Peter Hart and some other people last
Thursday. I am waiting for the results. We have been discussing my going
to work there after the AI conference in September, but it is by no means
certain that they will offer me a job.
Re proving "¬ provable" - If we axiomatize provability by a set-of-support
based algorithm, then only "relevant" axioms will be examined. In many cases
(including, I believe, all the puzzles we have been examining) proofs of
non-provability will be obtained by running the algorithm to exhaustion.
Since this ammounts to running a fixed program on fixed data, no axiom
schema would be required. It is true that there is a wider class of proofs
that could be obtained by using induction, however, these proofs, by the few
examples we have, seem to be so incredibly difficult that I don't think there
is any possiblity of having a machine discover one in the forseeable future.
∂16-Jun-77 1221 ROZ
How is the "Terminal Project" doing? Any meetings coming up?
∂16-Jun-77 1253 RWW nsf
On nsf[pro,rww] is a copy of your start with some remarks by me
mostly addressed to you rather than for inclusion. I will say
more but wouldn't mind some feedback.
rww
∂16-Jun-77 1643 PAM
John do you want the problems and answers and midterm from cs206?
They are on my directory [206,pam]. The files themselves are still protected, but
I have removed protection from the directory, so you can log in as 206,pam and
slurp them; alternatively, I'll be glad to send them wherever you like....Paul
∂17-Jun-77 1237 FTP:GLS at MIT-AI (Guy L. Steele, Jr. )
Date: 17 JUN 1977 1536-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr. )
To: rpg at SU-AI
CC: GLS at MIT-AI, jmc at SU-AI
So I will be flying in on July 26, and will stay at hotel
for MACSYMA Users Conference the nights of 26, 27, 28.
I fly back out on the morning of August 10.
Is this okay with all concerned?
∂17-Jun-77 1333 PAM
All there; index.s77 describes them..Paul
∂19-Jun-77 0354 RDR via AMET a dialnet thought
To: JMC
CC: JAB
It might help with the practicality of the protocols to do parallel
implementations on two different machines -- like a 10 and the
SCIP 370 for example. Then again it might confuse matters, but
maybe SCIP would finance their end, and the secondary implementation
could be done as an undergraduate research project or master's
project. Usefulness would certainly be locally enhanced if SCIP
participated and they have their own decnet-by-telephone plans
that might take forever, be unusable, and preclude other paths
of action.
∂20-Jun-77 0026 GFF
I am thinking about leasing a line from SRI to my house, where I have
a NOVA 1220 computer with attached Plasma display, and I was curious if you
could tell me (1) What the current baud rate of your connection frm SAIL to
your house is and (2) How much the line charges are per month [I live closer
to SRI than you do to SAIL so I susprct it would be cheaper, but I am just
trying to get a rough idea of total cost/hassles] from SAIL to your house,
and what type of modems you have on each end of the line and how much they
lease (or were bought) for.
I would appreciate any info you could give me.
[Geoff]
1. I believe the baud rate is 1800, but it may be 1200. It is not as
fast as one would like, but the main delays come from slowness of the
time-sharing system and not from an inadequate modem.
2. The line charges are something like $22.00 per month.
3. The modems are Sangamo and cost about $800 per end about 5 to 7
years ago. They have been entirely reliable.
∂20-Jun-77 1026 DEW My next few months work
John, I've had a few conversations with Hans Berliner and we've outlined
what I'll be doing from now til I finish (hopefully I'll start writing a
thesis in 6 months or so). If you'd like to discuss it and comment on it,
I'd like to arrange a time to come and talk. It shouldn't take long.
Thanks, Dave
I am sorry not to have had more time to talk, because your topic is one
that interests me. I'll need a real lesson some time soon. This week
I'm working on my IJCAI paper so we should get together next week.
∂21-Jun-77 1312 PAT phone call
Lucia Vaena of Berkeley would like you to call her between 8:30 and 10:00am
concerning a visitor from japan - 642-7383 - on Wednesday.
∂22-Jun-77 0810 FTP:TK at MIT-AI (Tom Knight ) Dialnet and Mark Crispin
Date: 22 JUN 1977 1107-EDT
From: TK at MIT-AI (Tom Knight )
Subject: Dialnet and Mark Crispin
To: JMC at SU-AI
I will be here all day till about 7:30 at either 253-7807 or 253-6765.
∂22-Jun-77 1023 FTP:PRATT at MIT-AI (Vaughan Pratt )
Date: 22 JUN 1977 1321-EDT
From: PRATT at MIT-AI (Vaughan Pratt )
To: jmc at SU-AI
CC: (FILE [PRATT;LISP MAIL]) at MIT-AI, GLS at MIT-AI
I finally made the time to read a little of your "Recursive Programming
in LISP." It really was a delight to read, delicately balancing the
status quo of LISP against a denotational semanticist's (or for that
matter, a common-sense mathematician's) view of mathematical objects.
My own preference is to match the language to the prevailing mathematical
viewpoint, but your paper impressed on me how close one could come
without doing any violence to LISP as implemented. While I will
continue to try to bring mathematics into programming without feeling
any obligations to established languages, I cannot do so in the LISP
article I am writing (as you pointed out to me recently). I will
therefore take your paper as a guide to how to do mathematical justice
to LISP in my article. (It is AI:PRATT;LISP > if you care to XGP it.
If you find it inconvenient to remove backspace-underscore from it at
your end, let me know. It is trivial for me to do it, and in fact it
is also trivial for me to put in your local XGP commands so you can
XGP it at your end without delay).
There is an aspect of LISP that I have trouble with, namely that
functions are treated differently from other objects. This agrees
with the older mathematical viewpoint that insisted on distinguishing
functionals from functions, but it then becomes awkward to talk about
functions in general, which is necessary in the metadiscussion.
My approach to this problem has been simply to present this as a
problem, and to cope with it as best I could. Some things come out
clumsily this way, but I don't see any way out beyond lying about
LISP. What is your position on this?
The business about numbers and T and NIL not needing quoting is also
awkward, but something that I guess LISP purists just learn to live
with. It is hard to keep word and object separate in the reader's
mind when one does not do so in one's writing. This comes from
treating abstract-syntax programs as S-expressions, which though
elegant leads to just such difficulties. Do you have any idea as to
how much of a pedagogical obstacle this presents in practice?
I appreciate your use of commas in place of semicolons in
M-expressions. I found semicolons a strain to read. While I don't
particularly mind your retention of brackets where conventional
notation would call for parentheses, I think I shall experiment in my
article with the latter for this purpose. I believe a reasonable
degree of unambiguity can be attained without using brackets.
Do you have any preference between (QUOTE (A B)) and (LIST (QUOTE A)
(QUOTE B))? I prefer the latter, since it means that the use of QUOTE
in ordinary programming can be confined to atoms, where it can be
thought of as string quote or symbol quote. I like your convention of
using upper case as implicit QUOTE, which in M-notation then lets you
write list[A,B], or list(A,B) as I would say.
It is of course convenient to be able to quote when doing symbolic
computation, for example if one wishes to differentiate 'x+1', so that
one can talk about the operator and operand (list of arguments) of the
expression, using CAR and CDR. The drawback of M-notation it seems to
me is that CAR and CDR are unlikely names for these concepts, whereas
they are much more reasonable when you write expressions as
S-expressions, whether as (PLUS X 1) or [PLUS,X,1]. Perhaps 'x+1'
should not be considered to have a CAR and a CDR but rather an
OPERATOR and an OPERAND (list). Thus car('[A,B]') would be undefined
rather than LIST, and instead operator('[A,B]') would be LIST, and
operator('A.(B.NIL)') would be CONS even though [A,B] = A.(B.NIL).
Let me try out a half-baked idea on you, some version of which
may have occurred to you already. With no use for brackets in my
version, we may co-opt them as an abbreviation for "list," permitting
us to further contract list(A,B) to [A,B]. (Smith's MLISP uses <A,B>,
which unfortunately rules out the use of > as an infix relational
operator.) This not only makes the standard notation redundant (by
removing the argument that (LIST ...) is less readable than (QUOTE
...) in complex expressions having many LISTs), but gives the
additional flexibility over quoted S-expressions that expressions to
be evaluated may appear inside the expression; thus [a,B] would be a
list of two items whose first was the value of a and whose second was
the atom B. One might object that [A,B] causes needless consing every
time it is evaluated, but in fact that is a drawback of the
implementation rather than the language; the implementation could be
arranged to recognize that [A,B] was a constant and represent it
suitably internally (the usual optimizer trick). In fact, now that I
think of it, why can't one represent [a,B] internally as (#A B)
together with an environment specifying the value of #A, the pair
forming a closure? Maybe this is really just approaching lazy
evaluation from a different direction.
One possible objection to the above is that one needs a name for
objects; just as 3+2 has 5 as its name, so should the value of (LIST
(QUOTE A) (QUOTE B)) get the name (A B). However, the only property a
name really needs to satisfy is that it be canonical, and [A,B] seems
no less reasonable a candidate than (A B). The objection to 3+2 as a
name for 5 is that + does not preserve information. This is not a
problem for the function LIST. Of course, [a,B] is not an object, so
[a,B] is not an instance of a name, rather it acts like a variable, or
a function of the environment, in the same sense as does x+1.
For the article, I will continue to use the standard notation for
S-expressions. However, I will give some more thought to the question
of what makes sense in the interaction between the word-object
dichotomy and the implementation of LISP-like languages. Any thoughts
you have on either the article or on these explorations would be
welcome.
Vaughan
∂22-Jun-77 2213 MWK A banquet
Carrying on in fine tradition, I am organizing a chinese banquet before I
depart. It will be held next Wednesday at 7pm at Hsi Nan. It will feature
a cold plate, a crab, a fish, a Peking Duck and other stars. The cost per
person will be $10. This includes tax and tip. You are invited to bring
friends! Make your reservations as soon as convienient. You should give a
check to me BEFORE the banquet. If you can't make the banquet, you are
invited to a small party afterward at 458 Channing, where we will have
some drink (alcoholic and not) and merriment.
Hope to hear from you,
Mark.
∂23-Jun-77 0153 FTP:GEOFF at SRI-KA [Steve Caine <Farber at sri-kl>: A ''Pocket sized terminal'' from UNILEVER]
Date: 23 Jun 1977 0149-PDT
Sender: GEOFF at SRI-KA
Subject: [Steve Caine <Farber at sri-kl>: A "Pocket sized terminal" from UNILEVER]
From: GEOFF at SRI-KA
To: JMC at SAIL, Les at SAIL, REM at SAIL, PBaran at ISI, Mathis at SRI-KL, Retz at SRI-KL, Kunzelman at USC-ISIC, Wilber at SRI-KL
Message-ID: <[SRI-KA]23-Jun-77 01:49:20.GEOFF>
Begin forwarded message
--------------------
Mail from SRI-KL rcvd at 23-Jun-77 0135-PDT
Mail from RAND-UNIX rcvd at 22-Jun-77 2250-PDT
The following appeared on page 90 of the 18 June issue of the
ECONOMIST:
"The world's first pocket computer terminal, measuring five inches by
two, is being launched by Unilever. It costs POUNDS 900 (POUNDS 800
in bulk), has video display, 800 characters of memory, and can be
hooked up to any telephone. A salesman visiting supermarkets could
put an order straight through to the warehouse. Unilever reckons on
selling POUNDS 4m worth, or 5000 units, to its own subsidiaries in the
next three years. The terminal can be linked to any manufacturer's
computer. Unilever has bought sole production and marketing rights
for Britain, Europe (except Sweeden), and North America from a
Sweedish company that developed the terminal."
-------
-------
Date: 22 Jun 1977 2248-PDT
From: Steve Caine <Farber at sri-kl>
Subject: A "Pocket sized terminal" from UNILEVER
To: pc at RAND-UNIX
--------------------
End forwarded message
-------
∂23-Jun-77 0835 FTP:PRATT at MIT-AI (Vaughan Pratt ) LISP article for the Belzer et al Encyclopaedia.
Date: 23 JUN 1977 1131-EDT
From: PRATT at MIT-AI (Vaughan Pratt )
Subject: LISP article for the Belzer et al Encyclopaedia.
To: LISP-DISCUSSION at MIT-AI
Finally I have found some time to write some more. I would welcome
feedback from all and sundry, which is why I am sending this message
to everyone on the LISP-DISCUSSION mailing list here rather than just
to those that requested interrupts. The file is AI:PRATT;LISP > .
Those of you that cannot cope with backspace-underscore, please let me
know and I can easily produce another file with the underscores done
however you want. (Just tell me what your XGP or other device wants.
I have things set up here so that our XGP treats underscoring as
simply specifying a different font, which I set to be italics, and the
associated mechanism makes it painless for me to tailor a file to your
requirements. I'll let you supply the appropriate XGP header info.)
∂23-Jun-77 1101 FTP:PRATT at MIT-AI (Vaughan Pratt )
Date: 23 JUN 1977 1401-EDT
From: PRATT at MIT-AI (Vaughan Pratt )
To: jmc at SU-AI
Do you happen to know what character sequence gives you font change and/or
underline-start-and-end? Alternatively, I can just put in PUB commands
and you can pub it with everything turned off except recognition of
those commands. Since this problem has come up with a number of things I have
written that Stanfordites have wanted to xgp at their end, it would be
nice to know your XGP format for either font-change or underscore, once
and for all so I can just write a TECO macro that allows me to
SAILify anything without hassle.
I don't know about underline, but %2 preceding and %1 following text will
put it into italics provided the first line of the file is
.require "suitable" source;
where suitable (which must be surrounded by quotes) is a file that
defines a page size and fonts. MEMO.PUB[LET,JMC] is a suitable file.
MEMO.PUB has much more in it than necessary, so maybe I can get someone
to make up a suitable file and put it in your directory here if you
can give the full set of facilities you want.
∂23-Jun-77 1246 FTP:PRATT at MIT-AI (Vaughan Pratt )
Date: 23 JUN 1977 1545-EDT
From: PRATT at MIT-AI (Vaughan Pratt )
To: JMC at MIT-AI
How do you specify Greek lambda? ASCII 8 or what?
According to the table on my terminal, that seems to be correct.
∂23-Jun-77 1254 FTP:PRATT at MIT-AI (Vaughan Pratt )
Date: 23 JUN 1977 1551-EDT
From: PRATT at MIT-AI (Vaughan Pratt )
To: JMC at MIT-AI
Actually, I think it will be an unnecessary pain for you to go
through PUB at your end. If you can just tell me what XGP characters
correspond to things like font changes, I can produce a file at this
end that you can simply XGP directly. This is no harder for me than
producing a PUBable file, and easier for you. Actually it should
suffice for me to have a Rosetta stone consisting of a PUB source,
its XGP translation, and the corresponding XGP hardcopy, in which
various fonts are used. (I take it that % simply signals a font change.)
% signals a font change to pub. I don't know how it works directly.
Jeff Rubin would know and is here at the moment. I would do more
myself, but I'm worried about finishing my paper.
∂23-Jun-77 1532 TJW ORALS
I am in the process of pubbing my thesis and will be able to give you
a copy as soon as Panofsky fixes the XGP. I would like to schedule my
orals during the week of July 18-22. Are there any specific days that
would be inconvenient for you ?
No present problem with July 18-22.
∂24-Jun-77 1634 LES
∂22-Jun-77 1332 FTP:PBARAN at USC-ISI PCNET General Meeting
Date: 22 JUN 1977 1330-PDT
From: PBARAN at USC-ISI
Subject: PCNET General Meeting
To: CALPCNET:
Dear Cal. PCNET people:
The next general meeting will be at 7:30PM 27 June in ERL401 at Stanford;
a number of new and interesting things will be discussed. Hope to
see you there.
Dave Caulkins
(mailbox PBARAN@ISI)
-------
To Dave Caulkins:
I would like to come to your PCNET meeting on Monday, and, if there
is time and it is convenient, discuss compatibility between Dialnet
and PCNET.
John McCarthy
∂25-Jun-77 0919 FTP: Ijcai-77 at CMU-10B IJCAI invited paper.
Date: 25 Jun 1977 1217-EDT
From: Ijcai-77 at CMU-10B
Subject: IJCAI invited paper.
To: jmc@su-ai
Sender: IJCAI-77 at CMU-10B
Message-ID: [CMU-10B] 25 Jun 1977 12:17:48 Ijcai-77
John,
We are trying to get the "final" program settled. Can you let me know
what the title is for your invited talk?
Also, if you know how many pages the paper will be, that would also be helpful
(along with an estimate on when we might receive it).
thanks,
Lee Erman
-------
Title: Epistemological Problems of Artificial Intelligence
I've got to finish this weekend.
John McCarthy
∂25-Jun-77 1155 FTP:PBARAN at USC-ISI (Response to message)
Date: 25 JUN 1977 1148-PDT
From: PBARAN at USC-ISI
Subject: (Response to message)
To: JMC at SU-AI
cc: PBARAN
In response to your message sent 24 Jun 1977 1749-PDT
By all means. We would love to talk about PCNET and DIALNET, and
will put you on the agenda.
Dave Caulkins
(mailbox PBARAN@ISI)
-------
∂25-Jun-77 1239 FTP:REM at MIT-MC (Robert Elton Maas) Compatibility between dialnet and pcnet, discuss at meeting
Date: 25 JUN 1977 1538-EDT
From: REM at MIT-MC (Robert Elton Maas)
Subject: Compatibility between dialnet and pcnet, discuss at meeting
To: JMC at SU-AI
If possible you should read PROTO.PRI[1,REM] and PROTO.WR1[1,REM]
(both POX source) to get an idea of our specific first-cut design,
as well as Ron Crane and Peter Deutsch's handouts. The former should
be studied beforehand if possible, the latter will be difficult to
obtain other than at (just before) meeting. (Assuming you
haven't already read our protocol specs.)
Considering that our net is based primarily on virtual simplex
connections end-to-end composed of momentary node-to-node duplex
connections, whereas I presume dialnet will be based on end-to-end
duplex connections like the Arpanet, I doubt much compatibility can
be obtained at the lower levels. Perhaps our addressing scheme and
message-header scheme could be compatible (Peter Deutsch). Ron
Crane appears to be having trouble designing message-level protocol,
and nobody in our group has volunteered to do major part of it, so
perhaps you could discuss your ideas along that line if you have any
you like. (I like the idea of "PLEASE DO / I WILL / I WON'T / PLEASE
DON'T" flavors of related node to node messages, but there are so
many details to think of to get the design solidified. Mostly we
are not proceeding well in the design of node to node messages.)
∂25-Jun-77 1515 FTP:GLS at MIT-AI (Guy L. Steele, Jr. )
Date: 25 JUN 1977 1814-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr. )
To: PRATT at MIT-AI
CC: (FILE [PRATT;LISP MAIL]) at MIT-AI, GLS at MIT-AI, jmc at SU-AI
I am pretty sure that it is not meaningful to "close a form",
but only to close functions (or, rather, to close a representation
of a function, thereby producing a function). Thus to close
"(#A B)" in an environment is not meaningful.
However, to write such a form in a place where A has a value is meaningful.
SCHEME provides a notation for this, based on reader macros
going back to CONNIVER, if not MICRO-PLANNER. Namely, one can
write
"( ... ,A ... @B ...)
where ,A means to substitute in the value of A, and @B to splice
in the value of B, which must be a list.
Anyway, this is clearly a notation rather than a notion;
(LIST ... A (APPEND ... B ...))
would do just as well, modulo readability.
[A,B] and '(A B) are equivalent notations iff they do
equivalent things in the language. In a pure LISP it doesn't matter, but if RPLACA
and RPLACD are involved it is impossible in general for an optimizer
to decide whether the CONSing is necessary (though it might be
able to decide in many practical cases). For better or worse,
the contemporary style of LISP programming
relies very heavily on EQ-ness as opposed to EQUAL-ness.
Wether this is a point of fundamental semantics or underlying
implementation I leave to you to decide. (I might point out that
MacLISP explicitly defines the language to allow copying of numbers;
that is, ((LAMBDA (X) (EQ X Y)) Y) may yield NIL if the value of Y
is numeric; but it must yield T if Y is non-numeric,
all this being modulo an interrupt coming along and changing the
value of the variable, of course.)
∂25-Jun-77 2342 DON
∂25-Jun-77 0250 JMC
I didn't know game could end before I got all treasures home.
[Heh heh! Yes, it ends a fixed number of turns after you FIND the
last treasure!]
∂26-Jun-77 0013 NS
To: JMC
Your following News Service notification
request(s) will expire within a week:
(electronic*mail)/AP/NYT
∂26-Jun-77 0926 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 26 JUN 1977 1227-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: RPG at SU-AI, JMC at SU-AI
CC: (FILE [PRATT;LISP MAIL]) at MIT-AI, GLS at MIT-AI
How is CGOL faring at SAIL these days? Anyone using it?
I've been thinking about implementing BB in CGOL, it has some nice features.
I like writing S-expressions just as they are, parenthesized upper case
thingies, so using brackets for syntactic grouping outside of
S-expressions maybe isn't that big a price to pay after all. I
appreciate John's change from semicolons to commas.
Two changes I think I would make to BB to make it more convenient as
a programming language are:
1. boldface treatable as italic - thus can write italic "if" instead of
boldface "if" etc;
2. boldface "a" written as italic "car" (otherwise the users will go
crazy),
and similarly boldface "at" as italic "atom".
(Of course, italics means simply lower case - you could have a special font
that has its lower case letters in italics, for printing BB programs.
Pity you can't get the boldface ones that way, which is one reason for
considering the above two changes to be substantive changes and typing
lower case roman for lower case italic not substantive).
One thing that BB (a cause de LISP) is a little confusing about is the distinction
between parentheses and angle brackets. The two differences are:
1. You can't write variables inside parentheses; in fact you can't
write anything that is intended to be evaluated inside angle brackets.
2. Users of EQ, RPLACA and RPLACD are the only ones who can detect a
difference between <A,B,C> and (A B C) (as GLS recently reminded me).
Here is a suggestion for coping with 1 and 2. 1 is easily disposed of
simply by regarding the parentheses of any S-expression containing
forms to be evaluated to be angle brackets instead. 2 is a bit
trickier. Here is my (flaky) suggestion. Eliminate angle brackets
entirely, using my suggestion for 1 for those cases where an
S-expression does contain something to be evaluated. Constant
S-expressions that the user wants processed as in (LIST 'A) should be
written (in this case) A.NIL to indicate explicitly that the LISP cell
so constructed is going to be hacked by EQ, RPLACA or RPLACD.
This is unclean because it still leaves CONS
representation-dependent. However, as long as EQ, RPLACA and RPLACD
remain in the language you aren't going to do much better than this.
Actually, maybe you could distinguish pure and impure cons.
"Good" users will in fact only have one cons in their language anyway,
while "bad" users (maybe they should be called Type I and Type II
users to avoid pejorative terminology) aren't going to complain.
On the other hand I suppose if you do it that way then you can
go back to one cons and instead require users to answer the
implementation's question "Are your intentions towards my
implementation of cons honorable?" meaning "If I give you a cell, do
you promise not to look at its address, or tamper with its insides?"
On the other hand (the first hand we were on), there's a little bit of
bad in the best children, so maybe good children should be given both
good and bad cons anyway because they will every now and then want to
use bad cons.
I guess my use of bad cons is going the way my use of goto's
went some years ago. I don't believe there is a single bad cons in
the entire Presburger-theorem-prover I wrote recently. Ah yes, just
one, to build a circular list of period one so that I could mapcar
along a constant in parallel with a list. Usage: plus[x,circ [1]] has
the effect of add1[x], where circ x does "cdr last x := x; x". You
can think of the implementation of circ as being outside the
theorem-prover if you like, but no, that doesn't help, now we have a
primitive circ that is just as bad as rplacd. Ah, I know, what is
needed for this application is a slightly less useful function,
constant x, which does circ [x], guaranteeing that no one else could
own the cell and so keeping the user honest. Circ of course remains
bad, but constant is good even though it is implemented with circ.
Actually a good circ could be defined by making it copy its argument first.
You might say, well there goes Pratt, typical purist, doesn't
give a damn about speed when it comes to language design, better avoid
those techniques for our applications where speed really counts. Well
then, we need a benchmark, and Stanford has one in the form of Derek
Oppen's Presburger package, Stanford's opposite number to the
above-mentioned program. Derek tells me his program represents an
improvement over the algorithms of Cooper, Bundy, Bledsoe, etc. I
took a particularly big theorem that our system had to deal with in
the course of proving the total correctness of the Knuth-Morris-Pratt
pattern matcher and gave it to the two systems. (Its THM31 on both
my MITAI and SAIL directories.) On your KL10 Oppen's program takes 48
seconds (EBOX 26 seconds) to see that Thm31 held, whereas on our KL10
(MC) my program took 0.5 seconds. (I suspect Derek's program may go
faster than mine on Presburger problems representing really odd-shaped
polytopes in hyper-space - my program is tuned for relatively
rectangular polytopes, which is what our dynamic-logic proof-checker
encounters in practice.) When faced with variations of this magnitude
between programs solving the same problem, it is pathetic to hear
people whimpering about little improvements that are possible by using
EQ, RPLACA and RPLACD. In fact, there is a way in which they can slow
things down in the end even when they have been effectively used
initially to gain a small improvement in speed. This happens when one
wants to modify one's code to incorporate some change to the algorithm
that may speed things up a bit. If everything has been coded really
cleanly, the modification should be routine, but if the coding is
obscure, which is what it tends to be once you get into the habit of
using RPLACA etc, the modification may take more work than it is worth
to you, so you end up not bothering. Of course this brings us back
once again to the eternal question of, when does it pay to do
low-level coding? But there is another argument that goes beyond
that. If the compiler knows for sure that nothing "clever" (like
RPLACA) is going to be done, it may have a much better chance of
producing really good code for a particular list. For example, it may
be able to see that the list should really be implemented as a vector,
safe in the knowledge that the programmer isn't going to come along
later and try to RPLACD the vector in the middle.
Actually what I would like best of all is to be able to
program using such an abstract notion of list that I don't need the
datatype vector as well. To get the implementation to treat my list
as a vector, I should just be able to say, treat that list as a
vector, as a remark separate from the program. All pieces of code
that the optimizer wants to optimize that can be seen not to try to
RPLACD into that vector can then be optimized properly. Any other
code will be treated with kid gloves by the interpreter when
approaching that vector to make sure that it doesn't try anything
funny on it. If something funny happens, the vector reverts to a
list. This of course generalizes to other data types.
Cheers
Vaughan
Carolyn: Here are some typos and other comments from Vaughan Pratt at
M.I.T.
∂26-Jun-77 1002 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 26 JUN 1977 1305-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: jmc at SU-AI
CC: (FILE [PRATT;LISP MAIL]) at MIT-AI
Are you interested in typos and other remarks for "Recursive
Programming"? Here's a bunch. (6,-3 means 3rd line up from bottom of
page 6. I'll just give the corrected form. Things in parens are
remarks rather than corrections. If you don't have a copy with this
numbering and care about the typos, I'll just mail you this copy and
you can mail me a fresh one. Come to think of it, other order since I
wouldn't like to be without it for more than a day.)
6,8 PLUS.
13,14 expressions
13,19 programs
15,-2 (You ought to disown such an implementation.)
16,11 (then why is difference[x,y] denoted by x-y? Certainly not
because it is associative.))
16,-2 Here
18,8 (You ought to mention that these examples are over the natural
numbers rather than the integers. Otherwise your definition of mod
doesn't make sense.)
21,-12to-5 (Did you intend the prettyprinting to turn out that
way, or is there a difficulty with your prettyprinter?)
26,middle+2 ...compilers that know that multiplication is
associative should...
27,-9 (Thought it was Cadiou, but I'm not entirely sure)
36,-1 and
38,13 (I don't know, maybe one day machines will be so cheap and
people so fussy about their user interface that code generation will
be considered the side issue, if in fact any code generation is done
at all.)
39,first line of code c(ac)
40,7 <subr>
40,11 your
44,4-7 (To avoid this, which has nothing to do with compiling, how
about making the output a list? That should approximately halve
"file", which seems to be doing two different things anyway, unless
you consider DE to be something associated with files.)
48,-13 append
70,middle non-terminating
77,15 (You probably should give the expression, in the absence of
exercises bringing the user up to the necessary expertise to do it
himself.)
What are you planning to do with "Recursive Programming?" I would
think it would make an excellent monograph on pure LISP. I am getting
to the point where I would like to do all my programming in pure LISP.
Unfortunately our LISP compiler, by not doing a good job on eliminating
recursion, inter alia, doesn't encourage this, and so to continue to
cope with large problems I have to continue to program with loops and
setq's, which makes it hard to form good programming habits.
Thanks for your comments: Since Carolyn Talcott, a graduate student
and already a PhD in Chemistry, is helping me rewrite the notes
for my Fall class which will be given on Stanford's instructional
TV, they are particularly helpful, and I'll send you a copy of the
next draft she gives me. I have been seriously thinking about replacing
the bold face a and d by italic car and cdr, just to reduce the number
of notations. However, the primary objective of the BB notation is
still something that will look good in a book, so it doesn't deter
me that it isn't exactly convenient as an input notation. Look at it
this way: Even in the long run when terminals are super, it will still
be justified to make the machine do a lot of work to produce good-looking
and easily readable output that it won't be justified to make a human
do for input. On the other hand, there is no need to have unnecessary
differences.
I will study your ideas about how best to go beyond pure LISP
this week - as soon as I have sent of my IJCAI paper. My tentative
idea was to have an Algolic structure for that, which would use statements
like "cadar x ← y" to make replacements. I have a whole song that I'll
sing for you sometime about how one should have a class of entity formed
by Cartesian product that is used non-algebraically. Thus while an
employee record is the Cartesian product of a name and a salary (say),
it is not like a complex number, because we never form new employees
by algebraic operations on old ones. Most networks used in AI programs
have that same usage characteristic though a more complex structure.
These ideas haven't resulted in notational proposals yet, though.
∂26-Jun-77 1759 FTP: IJCAI-77(X355IJ00) at CMU-10B IJCAI contributed paper.
Date: 26 Jun 1977 1919-EDT
Sender: IJCAI-77 at CMU-10B
Subject: IJCAI contributed paper.
From: IJCAI-77(X355IJ00) at CMU-10B
To: jmc at SU-AI
cc: IJCAI-77
- - - -
John,
The McCarthy, Sato, Hayashi, & Igarashi paper for IJCAI, On the Model
Theory of Knowledge, had been accepted as a short paper; i.e.,
conditionally that it be cut down to one page (or two if a $100 page
charge is paid). It is thus also allocated 15 minutes for
presentation.
We received a letter from Japan several weeks ago saying they had not
received any notice from us about the disposition of the paper. We
have not heard anything from then since.
It is getting fairly late to get it included in the proceedings. Do
you want to produce a one or two page version for the proceedings
Please let us know how you would like to proceed.
thanks,
Lee Erman
-------
Lee:
I can't do anything, because I have no time and not even a final
version of the paper. I mentioned what I heard from you previously
in a letter to Sato who has just moved to Tokyo University. Igarashi
has also moved, so something may have gotten lost. The paper owed its
inspiration to some lectures I gave at Kyoto, but the model theory results,
i.e. almost all of it, are theirs. So, if they respond, they respond.
Otherwise, I fear you'll have to give up on it. I hope the papers you
accepted for full presentation are better than it is. Sato is very
good. I fear that theoretical papers may have gotten short shrift;
perhaps there should be a section for relevant theory at the next
conference.
∂27-Jun-77 0915 BPM OK, everybody...
To: DON, JMC, LES
% tn kl
Open
@
SRI-KL, TOPS-20 Monitor 101B(104)
System shutdown scheduled for Sun 26-Jun-77 18:00:00,
Up again at Sun 26-Jun-77 19:00:00
@sys russell
2 230 ADVENT Russell (NBS-TIP)
@logout
System shutdown scheduled for Sun 26-Jun-77 18:00:00,
Up again at Sun 26-Jun-77 19:00:00
Logout Job 30, TTY 233, at 26-Jun-77 17:22:59
Used 0:0:0 in 0:0:27
∂27-Jun-77 1501 MDD FUTURE OF PROVER
To: RWW
CC: JMC, DSB
A working linked-conjunct theorem prover is now available on an experimental
basis as part of FOL. It is invoked by a PROVE command in which premises
and conclusion are stated as for TAUT. Preliminary trials are quite encour-
aging, and I hope it won't die when I leave. (For a couple of examples see
INTER on my file using a SHOW PROOF command.) Two gaping holes are:
1. There is no way provided for dealing with equality;
2. There is no way of handling sorts.
Equality can be incorporated easily by writing a list of equality axioms,
including e.g. ∀x y z.((x=y∧Q(x,z))⊃Q(y,z)) that can be accessed by a
user.
Sorts can be dealt with by generating a list of SORT axioms and including
a procedure for processing VL's to make sorts explicit before feeding them
to PROVE. A better method would be to revise PROVER to take explicit account
of sorts, e.g. in the unification algorithm.
∂27-Jun-77 2247 FTP:DanG at SRI-KL (Dan Gerson) Dialnet opening
Date: 27 Jun 1977 2248-PDT
From: DanG at SRI-KL (Dan Gerson)
Subject: Dialnet opening
To: JMC at SAIL
cc: DanG
After hearing your talk on Dialnet at the Hobby computer conference up
in SF a while back, I have been interested in following what is happening.
Then at tonight's PCNET meeting you mentioned that a position is open for
a programmer to implement the protocols. I was wondering what the
job responsibilities, qualifications, time demands, and salary would be
for such a job.
Thanks,
Dan
-------
∂27-Jun-77 2300 RDR via AMET
re dialnet on 370 --computer research group at slac may be best bet
Good thought.
∂27-Jun-77 2324 DON Quick adventure
I've got it down to 237 turns. This included a couple of mistakes
(such as trying to go West from the west pit of the Twopit Room
instead of first getting out of the pit), but all in all I doubt
that it can be done in much less than 230.
∂28-Jun-77 0401 FTP: Ijcai-77 at CMU-10B Mailing of your invited paper.
Date: 28 Jun 1977 0700-EDT
From: Ijcai-77 at CMU-10B
Subject: Mailing of your invited paper.
To: JMC at SU-AI (John McCarthy)
Sender: IJCAI-77 at CMU-10B
Message-ID: [CMU-10B] 28 Jun 1977 07:00:59 Ijcai-77
In-Reply-To: Your message of June 27, 1977
John,
As long as the number of pages doesn't change from what you said (7),
and as long as the title doesn't change (Epistemological Problems of
Artificial Intelligence), if you get it in the mail by Friday, there
should be no problem on getting it printed properly.
Lee Erman
-------
lee,
Thanks, it will be done as you say.
John McCarthy
∂28-Jun-77 1249 RCM your IJCAI paper
In your IJCAI paper, some of your comments about blocks world programs'
understanding of space are not quite correct. Winograd's original blocks
world program had a Euclidian model of space and understood quantitatively
the size of its blocks. It did not have any knowledge of gravity (I think),
and therefore required every block to be completely over whatever supported
it. Fahlman's BUILD program, however, had a very sophisticated model that
included not only the geometry of space, but also gravity and friction. This
enabled it to build a wide range of structures using bridges, cantelevers,
ramps, balanced substuctures, etc. It also was able to test for stability
under lateral acceleration, so it could realistically model moving complex
subassemblies around. It was actually the later programs such as Sussman's
and Sacerdoti's that had the simpler models of space. I really think that
you should modify the statements in your paper about this subject.
Thanks for the comments. I'll modify the paper accordingly.
∂28-Jun-77 1443 MJL
To: JB, JMC, RWW
To: CSD Faculty and PhD students
From: Denny Brown
Subject: Results of the June 7 Black Friday meeting
On June 7, the faculty met to re-evaluate the progress of some of the PhD
students. There was strong sentiment at that meeting that:
(1) Each member of a student's Reading Committee should be informed
about the nature and status of the student's dissertation
research throughout the research. Post-facto, rubber stamp
Reading Committees should be eliminated.
(2) Committees are urged to require a thesis proposal early in the
student's research, so that each reader can rationally evaluate
the potential of the research. The form and formality of the
proposal requirement will vary from field to field, and from
committee to committee. Committees should report to the Graduate
Study Committee (Denny Brown) when a thesis proposal has been
approved. A short memo to the student's file stating the nature
of the research and the containing the committee's approval is
sufficient.
(3) Nth year students will be required to form committees immediately,
even though early committee formation has not been the practice
in the past.
During the meeting, advisors often gave assurances that students were
making progress, while reading committee members were uninformed and
unable to corroborate the advisor's statements. This resulted in many
requests to have research programs approved by Reading Committees.
Individual memos containing comments and suggestions from the meeting are
being prepared.
∂28-Jun-77 1505 MJL
To: JB
CC: JMC, RWW
Results of Black Tuesday
John McCarthy reported that you are making progress toward completion. He
stated that you should encounter no problem in finishing in one year.
Your continued good standing in the PhD program, and student assistantship
support, will be terminated if you do not finish by June 1978.
∂28-Jun-77 1511 MJL
To: DBG
CC: TOB, JMC, levinthal at SUMEX-AIM
Results of Black Tuesday
Tom Binford reported that you are progressing well toward completion. By
October 15, you should have a thesis proposal approved by your Reading
Committee.
∂28-Jun-77 1521 MJL
To: HPM
CC: JMC, TOB, levinthal at SUMEX-AIM
Results of Black Tuesday
John McCarthy reported that you are making progress toward completion of
your dissertation. He said that you should finish during the next
academic year.
Your continued good standing in the PhD program, and student assistantship
support, will be terminated if you do not finish by June 1978.
∂28-Jun-77 2056 JMC at TTY15 2056
subject :mazda
since you have several mazdas, you can always swap the alternaters around
among them.
whit
∂29-Jun-77 2247 FTP:Geoff at SRI-KL Password.
Date: 29 Jun 1977 2248-PDT
From: Geoff at SRI-KL
Subject: Password.
To: BPM at SAIL
cc: LES at SAIL, JMC at SAIL
What would be a good password for the <SU-AI> directory I have created on
our KL for file transfer excursions or what have you?
I will take the first suggestion sent to me, and then rely on that
person to tell all the others.
-------
∂30-Jun-77 0646 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 30 JUN 1977 0943-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: GLS at MIT-AI
CC: (FILE [PRATT;LISP;MAIL]) at MIT-AI, GJS at MIT-AI, jmc at SU-AI
The main aspect of the lambda-calculus that endears it to me above,
say, a TRAC- or GPM-like language, or general recursive functions (domain
the natural numbers) is that I can use a subset of it, the typed
lambda calculus, to name mathematical objects about which I can reason
easily. The universality of the language is, to me, a curiosity on
a par with the universality of a string-processing or number-processing
language. Thus my reasons for liking lambda-calculus are more semantic
than syntactic.
(This a propos of your quite detailed syntactic perspective on connectivity
in lambda-calculus terms.)
∂30-Jun-77 1125 FTP:PBARAN at USC-ISI PCNET 27 June mtng minutes
Date: 30 JUN 1977 1124-PDT
From: PBARAN at USC-ISI
Subject: PCNET 27 June mtng minutes
To: jmc at SU-AI
A draft of the minutes is available as page 18 of PCNET.COL at SU-AI
or as HOBBYNET.MINUTES at SRI-KL. Comments/corrections/additions
welcome. I expect to send a pointer msg to the minutes to the PCNET
list in a week.
Dave Caulkins
(mailbox PBARAN@ISI)
-------
∂30-Jun-77 1351 DCL first order equivalents
At one point in our conversation yesterday, you mentioned
that you thought it a good idea to get rid of Hoare's logic.
1. Hoares logic is merely a convenient formalism for expressing
statements about algol-like programs.
2. VCG is an automated procedure for reducing such statements to
first-order form. It does in fact get rid of Hoare's logic.
you can regard our current experiments in program verification
as testing whether this reduction-or elimination of Hoares logic-
makes automating proofs of statements about programs any easier.
QUESTION: Why do you stress first order formalism so much?
It isnt clear to anyone-to my knowledge- that proofs get easier
or statements clearer. Even in first order logic we see such
devices as Padoa's definitions; Hoares logic is little more
than that.
Do you subscribe to elimination definitions too?
-David
∂30-Jun-77 1629 FTP:GLS at MIT-AI (Guy L. Steele, Jr.)
Date: 30 JUN 1977 1929-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr.)
To: PRATT at MIT-AI
CC: GLS at MIT-AI, GJS at MIT-AI, jmc at SU-AI
I am quite confused by your last note. It would seem to me that
universality is a very deep semantic property, while you
seem to use it primarily for notational purposes, which are syntactic.
I was attempting to explain certain aspects of the "usual" semantics
in terms of syntactic features so as to free myself to apply
other interpretations, or to invent new notations.
∂01-Jul-77 0624 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 1 JUL 1977 0922-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: GLS at MIT-AI, PRATT at MIT-AI
CC: GJS at MIT-AI, jmc at SU-AI
One can use the lambda-calculus to compute or to denote. For computing
purposes, the lambda-calculus permits construction of engines that
grind or purr along. For denotational purposes the lambda calculus
permits naming of functions of arbitrarily high type. The extent to
which the lambda-calculus as a computational device is any more
convenient than any of the other mechanisms with universal computing
power is, I believe, the extent to which one can combine the computing
and denoting capabilities of the lambda calculus.
∂01-Jul-77 0635 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 1 JUL 1977 0934-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: JMC at SU-AI
maintained chiefly at XEROX Palo Alto Research Center
That was a bit cryptic, but I cottoned on eventually. Thanks.
AI:PRATT;LISP > is now developing rather rapidly as I go into
the final sprint to meet a Tuesday night deadline. Thus whenever
you sit down to read a bit of the paper, always grab the latest
edition from AI:PRATT;LISP > itself. Right now I'm cleaning up
the "On denoting" section. I'm afraid I'm being too pedantic, so
any suggestions on toning it down would be welcome.
∂01-Jul-77 2212 DCO MAIL in MACLSP
To: MACLSP.DIS[AID,RPG]:;
MACLSP now accepts mail. If you have the (HELP) option, type
(MAIL) to enter the mail program. You will be prompted to give a destination
list and then the message, both of which must be surrounded by double quotes.
Here is an example of its use:
(MAIL)
Destination? (Any valid MAIL destination list surrounded by "s)
"DCO, RPG%SAIL"
Message? (surrounded by "s)
"Notice that a message may run over any number of lines and may contain
any characters. If you want a double quote in your message, type it twice
as in "". The message now ends."
It is generally faster to send mail using (MAIL) then to use the MAIL
monitor command because (MAIL) does not start up a separate job. The tradeoff
is that the mail is not delivered immediately, but waits until the remind
phantom checks if there is any incoming arpanet mail. This happens roughly
every ten minutes.
∂01-Jul-77 2229 DCO Reasoning about List Structure
To: JMC, RWW, DCO
Greg Nelson and I will be presenting a paper "Fast Decision Algorithms
based on Union and Find" at the fall FOCS (Foundations of Computer Science)
conference. In it I will include a description of an O(n-squared) decision
algorithm for the quantifier-free DNF theory of list structure.
I am also starting to write a more comprehensive paper, perhaps titled
"Reasoning about List Structure" (better suggestions?) in which I will show that
a linear-time decision algorithm for the quantifier-free DNF theory exists if
one excludes circularities from the model (i.e. bars x = car(x)). I will also
show that the quantifier-free theory (without the restriction that the formula
be in DNF) is NP-complete (i.e. of the same complexity as the propositional
calculus) and that the full theory (with quantifiers) is non elementary-recursive
but decidable (the latter will be described as corollaries of theorems of
Rackoff and Tenney respectively.)
I intend submitting the latter to next spring's POPL conference.
∂01-Jul-77 2338 BPM [SRI-KL]<SU-AI>
To: JMC, LES, JBR
CC: geoff at SRI-KA
I have set a password on the <SU-AI> directory at SRI-KL. Let me know if
you or anyone else wants to use it.
∂02-Jul-77 0055 100 : DAS RECORDS
If you have finished with the Tangerine Dream albums, would you return them to
my desk or leave them with TVR or MLB? Thanks. I hope you enjoyed them. Debra
Your message about albums misdirected. Maybe you meant John Chowning?
∂02-Jul-77 1903 FTP:TAA at MIT-DMS ZORK
DATE: 2 JUL 1977 2200-EDT
FROM: TAA at MIT-DMS
SENDER: TAA at MIT-DMS
SUBJECT: ZORK
ACTION-TO: jmc at SU-AI
MESSAGE-ID: <[MIT-DMS].53728>
The bug you reported with regard to exorcising the evil spirits
twice has been fixed, but not in the installed version. Thanks.
/Tim Anderson
∂03-Jul-77 1101 FTP:PDL at MIT-DMS ZORK Bug
DATE: 3 JUL 1977 1400-EDT
FROM: PDL at MIT-DMS
SENDER: PDL at MIT-DMS
SUBJECT: ZORK Bug
ACTION-TO: JMC at SU-AI
MESSAGE-ID: <[MIT-DMS].53755>
Thanks for reporting the bug. It will be fixed when the newest
version of ZORK is installed.
Dave Lebling
(PDL @ MIT-DM)
∂03-Jul-77 2115 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 4 JUL 1977 0015-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: (@FILE [PRATT;ARTICL LIST]) at MIT-AI
I am now on the last leg (and my last legs, too) of my LISP article
for the Belzer et al encyclopedia. The bulk of the thing is now in,
though Monday and Tuesday should see a fair bit more stuff added. I
plan to stop work Tuesday midnight. Any feedback before then would be
welcome. Feedback after then will probably be useful too, up to
Thursday morning I imagine.
I don't have very much historical data on the development of the various
LISPs. Any anecdotes or sagas to tell, anyone?
As always, the file is AI:PRATT;LISP >
∂04-Jul-77 0231 DON
You lose. Odd though; I could have sworn I'd tried that before.
That was to have been my hint.
∂04-Jul-77 1203 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 4 JUL 1977 1430-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: jmc at SU-AI
CC: (FILE [PRATT;LISP MAIL]) at MIT-AI, GLS at MIT-AI
I just love micro manuals, and yours is no exception. It is
the exact opposite of John Allen's book, which I can't possibly hope
to read all of given my schedule. (My apologies to John.) The extent
to which one can combine truth with brevity in describing a useful
subset of a language is a valuable index both of the initial
"learnability" of the language, and of its esthetics. The former has
to do with the cost of initiating users into the language, as opposed
to the longer-term costs associated with their eventual immersion in
the whole pot-pourri. The latter is the sort of thing which is hard
to explain to some people, but which the true esthete is willing to
fight for.
A general rule for forming a micromanual should be to take an
initial segment of a larger pedagogically sound primer. (Maybe I've
just defined "pedagogically sound primer.") Thus one choice to make
is where to make the (single) cut in the primer. However, there is a
much harder choice which affects both micromanual and primer: who is
the manual addressed to? If raw beginners, you can't simply expect to
say that a list may contain either atoms or lists and have them cotton
on right away to the recursion. You need to talk about nested lists
explicitly, THEN get to the recursive definition. Also, "zero or
more" for the number of elements a list may have should be emphasized
by including () as an example. On the other hand, if it is addressed
to computer scientists that haven't yet learned the basics of LISP,
then you can capitalize on their background and use formality to
attain brevity.
In the micromanual numerical expressions appear only as
expressions without meaning. This may be confusing to someone looking
for the number-crunching part of the language, unless you explain that
it is useful to be able to write (PLUS X (TIMES A B)) because you
might want to implement an algebraic simplifier. Maybe the best bet
is either to omit all reference to numeric concepts, or to admit that
numbers get fair treatment in LISP, both as atoms like -2 and 3.14 and
as expressions like (PLUS 1 2) which evaluates to 3. Algebraic
simplification actually isn't a bad example, but I guess implementing
a small simplifier is inconsistent with getting the micromanual really
short.
NIL should be among the examples of atoms. I think it ought
to get a special mention, just so that the reader doesn't get too much
culture shock when he encounters it for the first time under ATOM,
then in again under COND, then in a different role under LIST. While
what you say about NIL is true, you haven't prepared the beginner for
its ubiquity. (I used to think of NIL as just another name for (),
but now that I have taken to thinking of (a ... z) as list[a,...,z], I
see it as the basis case for the definition of list, namely as
list[]=()=NIL and list[a,b,...,z]=(a b ... z)=a.(b ... z). Was
something like that the original rationale for making () an atom? I
found that very puzzling when I learned LISP.)
Your treatment of LABEL is fine, but my thinking these days is
that both LAMBDA and LABEL are inappropriate objects for beginners.
First, observe the equivalence-preserving transformation E → "αf.E"
where "f" does not occur free in E. (Milner's α, LISP's LABEL,
everyone else's mu.) (This transformation is so that remarks about
LAMBDA can be subsumed under remarks about LABEL.) Second, observe
that any reference to "αf.λx.E" can be replaced by "f" provided one
says somewhere (with the necessary scope for f) "where f(x)=E."
Unlike LAMBDA and LABEL, the concepts of "where" and of solutions to
equations are in everyone's bag of tricks, even if the finer points of
solving for function variables in general recursive equations have not
been mastered. No abuse of notation is involved here (unlike the use
of = in FORTRAN for assignment) provided one takes "where f(x) = ..."
to mean "where for all x f(x) = ...". Thus, from a pedagogical point
of view, maybe LISP should have offered a "where"-like way of
expressing LABEL and LAMBDA without having to force the user to choose
between the Scylla of the monstrous DEFUN and the Charybdis of the
vertigo-inducing LAMBDA and LABEL.
Of course not every usage of LABEL (when LABEL is treated
properly as the minimization operator) can be directly translated in
the above way, but such usages are not what your beginning programmer
wants to involve himself with anyway as exemplified by the fact that
LABEL is not treated properly in LISP anyway, due to LISP implementers
thinking about LAMBDA and LABEL only in applications typified by the
applicability of "where".
Actually there is a difference bewteen LABEL (presumably an
operational concept) and "where," if we don't insist that "where f"
means "where f is least such that f". This might be all to the good,
as it might free us from the tyranny of leastness that is inflicted on
us by operational definitions. I see no good reason why imaginative
compilers shouldn't, as a part of their optimization, come up with
better solutions to equations, as suggested by Shamir's notion of
optimal fixpoints. One might consider the duty of compilers to be
finding better solutions to equations, if one considers that in
practice a real-time limit often applies, so that the object actually
defined is not even a solution but just a practical approximation. It
is an easy step from there to being willing to go past the least
fixpoint, especially when the least fixpoint happens to fall well
under the time limit due to its unwanted triviality.
Maybe I'll submit the above three paragraphs to SIGPLAN news
under "LAMBDA considered harmful." I haven't followed my advice in
my LISP article, but to do so might be to represent LISPer sentiment
inaccurately, which counts LAMBDA and to some extent LABEL as virtues,
not vices. I already have enough problems capturing LISPer sentiment
when LISP afficionados can't agree on the right proportion of S- to
M-notation in discussion of LISP concepts. The locals are about ready
to wring my neck for betraying the "self-sufficiency" cause of using
S-notation for everything.
I notice that in your manual everything in italics is lower
case while everything in your S-expression font is in upper case. You
would simplify font changing considerably by having a single font for
both of these. I have done this with my LISP paper; the font is
AI:PRATT;BB KST. (I have no idea whether our AST resp. KST formats
are compatible with your font-editor resp. XGP, but if so you're
welcome to use it. However it is very easy to roll your own.)
I found myself saying "the value of x," where "x" was a
variable, so often that I introduced the abbreviation x-hat (i.e. x
with superimposed caret) for it. If you can stomach the thought of
another abbreviation, it would simplify your micromanual some. (I
counted about 17 places where you could omit saying "value" by using
some such abbreviation.) In fact, it lends itself to tabularization
of LISP forms and their informal equivalents, making such equivalents
easier to grok. I had to raise my hat (or rather BB KST's caret) so
that it would fit over b; I guess Stanford's caret is uparrow so you
may have to grow your own carets.
I am still thinking about abbreviations for car and cdr. One
problem with boldface a and d is that they take getting used to, to
avoid confusing them with variables. α and β for car and cdr don't
seem unreasonable (unless one wants to use α for LABEL as Milner
does, though mu is better known - pity the SAIL character set
designer(s) didn't think of that). However it occurs to me that x sub
1 (maybe typed in programs as x↓1 and in publication using real
subscripts) is an excellent name for car x. (I doubt whether this is
an original suggestion, it is so obvious.) I had been resisting this
because I was already using the notation for metavariables in, e.g.,
prototypical CONDs, but then I realized that in fact there was no
conflict, they were the same thing! Thus (COND (a↓1 b↓1) ... (a↓n
b↓n)) mentions the n elements of the lists a and b consisting of n
S-expressions each. This can be generalized to cadr, caddr etc in the
obvious way. One can also type x↓2↓3 for caddadr. ↓2↓3 should appear
as the subscript 2,3 in publication.
I am unaware of any similarly satisfactory notation for cdr.
David McQueen once showed me a JSL (?) article in which some logician
seemed to be rediscovering the many uses of cons, car and cdr in
recursive function theory, and he might have had a notation for cdr.
Do you know of the article? I have no reference for it. One feeble
possibility is x↓+. (Automata theorists at least can make the
connection by thinking of x↓* as being the list of all elements of x,
i.e. x itself. Others can just say to themselves that x consists of
x↓1 + the rest. Pretty feeble, as I said.)
An alternative possibility is to draw the (vague) analogy with
APL take and drop, and use something like x↓-1, (x take away one
element) or maybe what might be better, x 1. This has the advantage
that cdddddr is easier to read as x 5 than as x↓+↓+↓+↓+↓+, or for that
matter as cdddddr. (In publication you might be able to rely on
context, e.g. the fact that x is clearly a list resp. number, to
denote x 1 resp. x**1, both of which you may wish to print using
superscripts.)
However, maybe the best thing to do is not to try too hard to
capture symmetry between car and cdr that really isn't there when
talking about lists. Something like my treatment of eval might be ok,
with perhaps tilde or bar in place of caret, or something similar to
denote that cdr[x] is a defective x. As with eval, one only bothers
with this abbreviation when a single-letter variable and a single
application of cdr is involved. If cddr, cdddr, are so popular that
they deserve their own abbreviations, then the x n suggestion is the
best I have.
Whether "x↓n" (.e. variable n, as opposed to "x↓1", "x↓2", ...)
should be permitted in BB programs (as opposed only to the discussion
language) depends on whether you think BB ought to offer abbreviations
for such "non-official" concepts such as indexing.
Now you that you have a micromanual for LISP as she is spoke,
you need a micromanual for BB notation. When I get the time I will
implement a variant of CGOL that better approximates BB notation than
does CGOL by itself. Don't be too surprised or offended if I
incorporate some of the above innovations.
Cheers
Vaughan
The objection to using "x↓1" for (CAR X) is that you then need other names
for the functions themselves. Writing them as a subscripts obscures the
fact that they may be treated as any other functions. My present opinion
is that writing (1 X) for (CAR X) and (2 X) for (CADR X) is a good
solution if one doesn't mind the asymmetry. This is in line with an old
memo of Dana Scott in which he defined the numbers as λ-expressions that
take successive elements of a list. One still needs a name for CDR. (2 2
X) would be (CADADR X).
Another possibility is to reserve the numbers as names for the iterates of
CDR, i.e. (0 X) = X, (1 X) = (CDR X), etc. Then we need another name for
CAR, suppose it were A. We would then write (A 2 X) for (CADDR X).
∂04-Jul-77 1717 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 4 JUL 1977 2016-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: GLS at MIT-AI, jmc at SU-AI
I wrote a very detailed reply to David Moon's comments on my LISP
article. I would be interested in your opinions on the issues
raised and positions taken.
∂04-Jul-77 1814 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 4 JUL 1977 2113-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: jmc at SU-AI
CC: (FILE [PRATT;LISP MAIL]) at MIT-AI
Hmm, guess I'm confused then about various notations you use and why
you use them. Why do you write x*y for append[x,y]? Don't you need two
names here? The same goes for all the arithmetic functions, which in
your nomenclature can be written e.g. x+y, but still need a name for
the function itself, plus. I don't see x sub 1 being any different than
x+y in this regard - it is a widely used notation that still needs a
name for the corresponding function. (Maybe you consider subscripting
as application, in which case I guess I see your point, but my remark
about being able to treat x sub 1 as car[x] still seems to stand.)
Well, that can't be the full reason why I don't like the subscript
notation. My initial negative reaction came from looking at Pascal
where no other function names are provided. How about n*x for
(cdr↑n)x and n.x for car cdr↑n x? Do you see objections to using
numbers as functions?
∂04-Jul-77 2134 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 5 JUL 1977 0033-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: PRATT at MIT-AI, JMC at SU-AI
Date: 4 Jul 1977 1939-PDT
From: JMC at SU-AI (John McCarthy)
To: pratt at MIT-AI
Well, that can't be the full reason why I don't like the subscript
notation. My initial negative reaction came from looking at Pascal
where no other function names are provided. How about n*x for
(cdr↑n)x and n.x for car cdr↑n x? Do you see objections to using
numbers as functions?
-------
The tension when choosing notation is between succinctness and
unobtrusiveness. The APL notation is for the most part quite
obtrusive in that most of it is unfamiliar and unsuggestive.
I prefer to stick to widely used notations wherever possible,
to avoid arbitrary proliferation of notations. My preference for
x sub 1 derives from the widespread use of that notation for the
first element of a vector - one can hardly be accused of being
arbitrary there. The only arbitrary choices I feel badly about
are my notations for mapcar and apply, and the overstruck-tilde
notation for cdr that I suggested this morning.
n*x and n.x seem reasonable if you want to give up the ability to
cons 1 onto the front of a list.
I have no objections to using numbers as functions, so long as the
usage isn't too arbitrary. Church numerals struck me as pretty
useful in general, applying their first arg n times to their
second. In that light, Y (the fixpoint operator) can be seen to be
just the Church numeral for infinity, which therefore doesn't need
and therefore doesn't take a second arg. Thus (20 TERPRI NIL) would
generate 20 linefeeds, while (Y TERPRI) would do linefeeds forever.
(3 CDR X) would be (CDDDR X), etc. (3 (2 CDR) X) would be equivalent
to (6 CDR X), if you parse (A B C) as ((A B) C) as you ought. Hence
(M (N SUCC) 0) multiplies M and N. One can do pretty crazy programming
with just Church numerals, but then that's true of the lambda calculus
and combinatory logic in general.
If you run across any prior shorthand for cdr I would be most happy
to know about it. I hate to innovate when it comes to such a touchy
thing as notation.
∂05-Jul-77 0743 FTP:GLS at MIT-AI (Guy L. Steele, Jr.)
Date: 5 JUL 1977 1041-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr.)
To: PRATT at MIT-AI, jmc at SU-AI
CC: (FILE [PRATT;LISP MAIL]) at MIT-AI, GLS at MIT-AI
To JMC: Micromanual looks nice. There are some bugs:
The description of LABEL fails to note that the
replacement of f by (LABEL f ...) is to occur only
within e and not e1, e2, ..., en.
The fact that OR will return the first non-nil value
is an important property and is not obvious from the manual.
I note that the entire issue of scoping is carefully avoided;
was this intentional? In particular, the description of
LABEL is marvellously ambiguous in this respect.
To PRATT: why the great search for abbreviations for
CAR and CDR? While I agree that CAADAR is not very mnemonic
and deserves an alternative, the mad struggle to replace
them with single characters seems a good example of
false economy.
In situations where conses are to be treated symmetrically,
there is actually nothing wrong with car/cdr, so I assume
the case of interest is lists. In this case,
why not the semi-standard (cf. InterLISP)
(nth x n)
for ca<d↑n>r, and perhaps (nthcdr x n) for c<d↑n>r.
These are fine as LISP-level functions, and as special
cases people often use (first x), (fourth x), etc.
Now if you are looking for meta-notation rather than
function names so that you can write, as
the logic texts do:
∀x.f(x) iff for all x, f(x)
which sort of thing has confused thousands of students who
have, perhaps sloppily, been using ∀ already as shorthand
in their class notes for years, then I sympathize little.
∂05-Jul-77 0748 FTP:GLS at MIT-AI (Guy L. Steele, Jr.) LAMBDA considered harmful
Date: 5 JUL 1977 1047-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr.)
Subject: LAMBDA considered harmful
To: PRATT at MIT-AI
CC: GLS at MIT-AI, jmc at SU-AI
There is no point is submitting such a note
to SIGPLAN Notices, since most people already
think it's harmful. The problem is to convince
them that is is less so than they think!
Sussman and I have found, with SCHEME, that the problem
with discovering these fourth-order phenomena about LISP
is that people aren't even aware of the second-order
phenomena. It is very hard to explain the aesthetic
and phiosophical differences between static and dynamic
scoping to someone who has never even heard of funargs.
∂05-Jul-77 0822 FTP:GLS at MIT-AI (Guy L. Steele, Jr.) LISP and efficiency
Date: 5 JUL 1977 1119-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr.)
Subject: LISP and efficiency
To: RMS at MIT-AI, PRATT at MIT-AI, JMC at MIT-AI
CC: GLS at MIT-AI, BAKER at MIT-AI, MACRAK at MIT-AI, MOON at MIT-AI
CC: RG at MIT-AI
RMS has some very cogent remarks about the historical growth
of LISP at MIT and elsewhere as constrained by engineering
tradeoffs. He correctly points out that such decisions
as the "non-standard" denotation in function position
and the dynamic stack-binding implementation were clearly
motivated by desires for speed, and that while they may have
represented a compromise of theoretical aesthetics, they were
a necessary step for providing a production language in
which systems such a SHRDLU and MACSYMA could be written.
(I am going out on a limb here, since JMC was around then
and I wasn't, but this seems as accurate picture from discussions
I have had with RG.) The engineers are to be applauded for
going ahead in whatever way necessary, whether pretty or not,
to get things done.
I would like now to push another side of this coin, however.
I think it would be fair to say that there is some theoretical
ideal (though we may not agree on what it is!), to which
real implementations are all approximations and compromises.
I think we should beware of becoming so used to a given
compromise that, out of habit, we call it "the right thing"
without question. Each engineering compromise has had a reason
behind it, and when the reason disappears we should re-evaluate
the need for the compromise, rather than blindly continue an
established tradition.
A case in point is the storing of functions on property lists.
The need for this disappeared in the LISP Machine, and LISPM in fact
opted for a different representation. On the other hand, I don't
know if they returned to the theoretical notion of uniform
evaluation and then consciously decided, for other reasons,
to continue the non-uniform function denotation, or whther that
was taken for granted and they merely looked for another way to
do it.
Another case in point is scoping, but I won't beat that
horse here.
There are often very good reasons to continue a tradition,
the foremost begin compatibility. Quite a few MacLISP
programs (and programmers) depend on being able to use
the same name for both a function and a variable;
quite a gew also depend heavily on dynamic scoping.
These are reason enough to continue these traditions on
the LISP Machine.
But I think that surprise is unwarranted when these traditions
are questioned. When the original motivations have disappeared,
it may be time for a change. We should beware of enshrining
engineering compromises as theoretical beauty unblemished
(but on the other hnd, of enshrining useless theory for the
sake of theory).
∂05-Jul-77 0824 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 5 JUL 1977 1121-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: GLS at MIT-AI
CC: (FILE [PRATT;LISP MAIL]) at MIT-AI, jmc at SU-AI
Did you mean that ∀x.f(x) shouldn't exist as shorthand, or just that
it oughtn't to be necessary to explain what it means? If the former
then we should agree to disagree, if the latter then I wholeheartedly
agree, which is why I feel nervous about any notation that doesn't
stand out as "obviously right." I don't know of anything for cdr
that is "obviously right", and the only thing for car[x] that I
see as being widely used is x sub 1 (provided you don't start arguing
that x sub y is shorthand for x[y], meaning that the list is actually
a function, which gets into conflict with the LISP usage that car is
the real function).
x.y for cons[x,y] is "right" only insofar as that's the notation in use
anyway for dotted pairs, so it may as well be part of the metanotation.
<x,y> for list[x,y] (the official M-notation) I actually disagree with
on the ground that (x y) is already fulfilling that function. I don't
know whether people will find that confusing or not, but I claim it is
a perfectly workable simplification of the M-notation, and one that
should only confuse people already used to the old ways. I think
having <x,y> around as a separate thing from (x y) takes the M-notation
into the process-oriented part of LISP. While John is willing to go
that far (as must any programming version of the full M-notation), I
think there is some merit to reserving the M-notation primarily for
the denotational part of LISP (and the denoted objects are lists rather
than cells). This is obviously not a closed issue however.
All the arithmetic notation is unarguably satisfactory.
As to the rest, I am pretty content to just stick with f[x,y], especially
because the distance from that to the list (F X Y) is samll and well-defined.
BUT: I kinda dig John's x*y for append[x,y], and since John uses, I dont
feel too bad about using it.
Using aλ↑ for eval[a] is definitely a matter of convenience, since it allows
you to say that the value of →(PLUS a b)_ is aλ↑+bλ↑, which is considerably
easier on the eyes than eval[a]+eval[b].
The only other two thingies I am somewhat attached to are my versions of
append and mapcar, which I believe help to emphasize that their first argument
is being used in each case as a function. For example I find f<g<x>> much
easier to take in at a glance than mapcar[f,mapcar[g,x]], though this may
be a personal thing, I don't know.
So there you have it: some close-to-unarguable abbreviations, plus four
semi-questionable ones for which one can make cases of sorts. I don't
regard that as unreasonable. In fact, what I consider unreasonable is
enforcing your preferred notation on someone else that feels constrained
by it. I find straight LISP a pain to read. If you think that's because
I don't use it, look at LINGOL, which is all straight LISP (CGOL wasn't usable
when I wrote LINGOL). What I would like to be able to use, and to make available
to anyone that shares my preference for the brevity used in mathematical
notation, is a reasonable notation. Given that, the only argument I
want to indulge in is whether any given abbreviation is reasonable.
∂05-Jul-77 1030 FTP:GLS at MIT-AI (Guy L. Steele, Jr.)
Date: 5 JUL 1977 1329-EDT
From: GLS at MIT-AI (Guy L. Steele, Jr.)
To: rpg at SU-AI
CC: GLS at MIT-AI, jmc at SU-AI, jonl at MIT-MC
I am coming in for the MACSYMA Users Conference,
as you know, arriving the evening of the 26th.
The conference will be over the evening of the
29th. I leave at 1:30 PM Wednesday, August 10,
from San Francisco. I am causing a car to be rented
which JONL and I will share while he is out there
(probably about until August 1; ask him to be sure),
and which I will use thereafter (there will probably
be some sightseeing I want to do on my own).
Looking forward to seeing you out there.
-- Guy
∂05-Jul-77 1706 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 5 JUL 1977 2005-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: rpg at SU-AI, jmc at SU-AI, dav at SU-AI
Is MLISP available under UCI-LISP?
∂05-Jul-77 2129 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 6 JUL 1977 0028-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: jmc at SU-AI, Teitelman at PARC-MAXC
I don't have very much that ties together how the various lisps grew
up. Can you both give me any info you have that you think would be of
interest to someone trying to get a mini-picture of how the extant LISPs
got to where they are now?
∂06-Jul-77 1057 FTP:Doug Gage Dialnet
Date: 6 JUL 1977 1051-PDT
Sender: QUERY2 at USC-ISIC
Subject: Dialnet
From: Doug Gage
To: McCarthy at SU-AI
Cc: QUERY2
Message-ID: <[USC-ISIC]6-JUL-77 10:51:09-PDT.QUERY2>
I heard your talk on Dialnet at the west Coast Computer Faire. I am
interested in how the project is coming, and, after I get my
home system (3-9 months, hopefully), in participating, if that would be
appropriate.
Doug Gage
Code 8123
Naval Ocean Systems Center
San Diego, CA 92152
QUERY2 at ISIC
(714) 225-6515
-------
To Doug Cage:
The Dialnet project is getting started, and there may be
experimental protocols available in nine months though not
in three. You may also be interested in PCNET, oriented
specifically to hobbyists. You can inquire about it from
Dave Caulkins who can be reached c/o PBARAN@ISI.
John McCarthy
∂06-Jul-77 1936 RDR via AMET Dialnet et al.
To: MRC, JMC
CC: RDR
First, for what little it matters, I am happy that Mark is going to be
working on Dialnet. I cannot help having some reservations about
LOTS use for the Dialnet project -- basically involving whether
Dialnet DEPENDS on LOTS or just happily coexists.
Anyway, I think Mark makes the latter more likely and will do
a good job besides.
Second, is Dialnet going to rent LOTS an auto-dialer and/or
another modem or does all that wait until a second phase of the
Dialnet project?
zzzz
∂06-Jul-77 2138 MRC via AI Dialnet
To: JMC, RDR
I am thankful for RDR's complement. It's nice to know people like my
work (especially as at one time it was effectively my entire life) and
appreciate it.
I don't know what the story is about equipment. I have been gathering
together a few ideas in my head about protocols and implementations,
but still a lot depends on what input I receive from others. Also of
course I have not been officially hired for the project (although I
hope that will be only a formality since I am eager to do it now; I
have fallen in love with the west coast and am willing to give some
things up for this opportunity) so I can't go and make any statements.
I hope that I will have the freedom to do the job well and not merely
be an excuse for somebody else's bad ideas (I have my own bad ideas!),
but everything I have been told says that I will have such freedom. I
expect that it would be a rewarding experience for everybody concerned.
I would think that Dialnet would be valuable for LOTS; but I think that
some measures would have to be taken to ensure some policing of its
use. While I am not advocating instant implementation of protections
or privileges for using Dialnet, some consideration should be given to
the possible problem of Dialnet being used as a 1200 baud link to the
ARPAnet by freshmen to be able to have links with other randoms in
California on certain unprotected systems in Cambridge, MA! [It has
become quite a problem for ML and MC; passwords would be hard to do
since there is no good way of identifying all the authorized users.
It IS a known phenomena that 3-4 randoms are in a party link on ML
or MC all coming from the AMES Tip!] Apparently this problem exists
already on the IMSSS links to SAIL and SUMEX.
Oh well, all this can be discussed later. Just wanted to collect
some input. [By the way, RDR, I think that access problems should be
dealt with as a problem arises, not out of paranoia, so relax!!!]
Hopefully see you soon!
- Mark
∂06-Jul-77 2229 RPG
To: LES, JMC, DCO, DCL
∂06-Jul-77 1316 LES Friends of Maclisp
∂06-Jul-77 1117 RPG GLS/JONL visit
Guy Steele and Jonl White are going to be here from
July 29 thru August 10, and the three of us will be working
on SAIL Maclsp. For this period it would be quite useful for
someone of us to be an emergency user (or at least a super-user
so that the time will be well spent).
-rpg-
[This would be reasonable iff one of our sponsored research projects is
critically dependent on the success of this venture. In other words,
it is necessary to construct a stronger case. - LES]
Then I suggest you speak to DCO,DCL and the rest verification group
Among the major undertakings planned is purifying compiled
code, which is in response to the amazing situation of this system
not being able to run jobs of any appreciable size (AI programs, unlike text
editors, DO tend to be big).
Since they have made the committment to go with Maclisp for the
verifier, and since they have re-written large portions of it to
this end, it would be a pity to have to flush all that work for
a fairly silly reason.
I have spoken with them about the proposed visit, and they are enthusiastic
about the possibility of this modification (and some others) easing
their size crunch.
∂07-Jul-77 1317 JBR
RAK IS USING THE MLISP COMPILER WHICH USES ONLY 2 INPUT BUFFERS
AND THEREFORE DOES 18 DISK OPS PER INPUT. IF THIS PROGRAM GETS ANY
SIGNIFICANT AMOUNT OF USAGE THEN I SUGGEST YOU ENCOURAGE THE VERIFICATION
GROUP TO FIX IT.
∂07-Jul-77 1339 RAK System default buffers
To: JBR, JMC
I think that MLISP just opens the files with whatever is the default
number of buffers. Does that make sense? If so, could the default change
to whatever number Jeff thinks is optimal. Then I can remake MLISP.
Dick
∂07-Jul-77 2110 MRC via AI Call
Sorry I was very tied up today. I am at the MIT AI Lab now; you can call me
there tommorrow sometime.
∂07-Jul-77 2304 LES MRC
I tried to reach MRC several times this afternoon, but got consistent
busy signals (both from WATS and MIT).
∂08-Jul-77 0003 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 8 JUL 1977 0302-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: (@FILE [PRATT;ARTICL LIST]) at MIT-AI
My LISP article is now completed, with a bit of a rush near the
end. I doubt whether I will be able to incorporate any suggestions
at this stage as I am about to send it off. However, feel free to
read it. It is on AI:PRATT;LISP > . My thanks to all of you who
made helpful suggestions.
Vaughan
∂08-Jul-77 1716 WP FOL representation
I'm just reading your paper and found the question on the semantics of
multiterms. I thought about this and found several ways to answer this
question.
One simple way to go is:
- all terms represent sets, we do this by identifying each
element in the domain with the corresponding one-element set,
- if t is a term then t(y) iff yεt
- q<t> is the set {x| ∃zεt.q(z,x)}
This might be worked into something more elegant. Tell me if you'r interested
in some formulization, I'm willing to write something.
Wolf
It would be fine if you would write something. There has to be a theorem
that the formal manipulations produce a schema that is correct.
∂08-Jul-77 2117 FTP:Blohm at SUMEX-AIM Account
Date: 8 JUL 1977 2116-PDT
From: Blohm at SUMEX-AIM
Subject: Account
To: mccarthy at SU-AI
I just thought that you would like to know that you have an account here
at SUMEX. I did a finger on it, and it said that the last time you logged
in was on August 19, 1976. It also says that you have new mail from
somebody named LINDBERG. Since you hadn't logged in for such a long
time, I thought that you might have forgotten about this account....
Jeff Blohm
-------
∂09-Jul-77 2247 GFF Hacking NS.
To: JMC
CC: GFF, MMcM at SRI-KL
Mike McMahon (MMcM%SRI) and myself are interested in doing a few things to
the currenly dormat and unsupported News Service system esp. now that I hear
Martin Frost no long works for the lab anymore.
Much of what could be done is nicely filed away in NS.MSG[1,ME] which is mainly
a collection of suggested `features' from users [many from myself included]
(of which they have gotten back `there is no current support for NS, so...'.
Also in line would be improving NS's disk buffering as in ETV, and hopefuly
improving its speed of displaying stories and overhead on the system; esp. since
the majority of what NS does is just GETTING stories which have been filed away.
and for the Piece Da Resistance, perhaps sometype of Datacomputer interface
whereby instead of just having a 2 week rotating file file, where news older than
2 weeks instead of just being randomly flushed be archived on the Datacomputer
with some type of ability though NS to search for old stories just as if it was
on line, minus the Datacomputer wait and search time.
Would you be willing to support an account for MCM (mainly) for the above
NS hacking? This would be mainly a weekend project, when we visit the AI lab.
[Geoff]
Getting NS improved free of charge sounds great to me, and all the specific
ideas mentioned sound good. The answer is yes, and I will dust off some of
my dormant ideas also. I would like to get together in about a week.
The problem of efficiency should be discussed with JBR.
∂10-Jul-77 0101 TVR
It appears that ZORK exists as SRI-KL:<MDL>DUNGON.EXE, so when DM
is busy... --- Tovar
∂10-Jul-77 1620 DAV MLISP
To: Pratt at MIT-AI
CC: RPG, JMC, DAV
MLISP indeed runs under UCI-LISP. Also INTERLISP. I don't at the
moment have a current tape, or I would just send it to you. Contact
Bill FAUGHT @ SUMEX or Roger PARKINSON @ SUMEX (both on Ken Colby's
staff at UCLA). If they can't help you, contact me again. One of
these days I'll ship fresh LIS 1.6 and UCI-LISP versions off to DECUS
and let them handle the distribution. One of these days ....
Dave Smith
∂11-Jul-77 0243 FTP:Geoff at SRI-KA (Geoffrey S. Goodfellow)
Date: 11 Jul 1977 0244-PDT
Sender: GEOFF at SRI-KA
From: Geoff at SRI-KA (Geoffrey S. Goodfellow)
To: JMC at SAIL
Message-ID: <[SRI-KA]11-Jul-77 02:44:38.GEOFF>
In-Reply-To: Your message of July 11, 1977
Say either:
K #
Or
Logout #
Where # is the job # you want to kill. It will then comeback
with a SYSTAT like print out of the job you propose to flush and
say [Confirm]; if you really mean it, just CR and away it goes...
-------
∂11-Jul-77 1102 FTP:Sherwood at SRI-KL Dial-Net
Date: 11 Jul 1977 1059-PDT
From: Sherwood at SRI-KL
Subject: Dial-Net
To: McCarthy at SU-AI
cc: Geoff at SRI-KA, Sherwood
John, Geoff Goodfellow suggested I contact you directly regarding
your "Dial-Net" project. I'm very curious and would like to learn
more about it. Would you please forward any documentation and or
comments about it ? I'm trying to get a feel for message systems
and would appreciate your help. Thanks, Glenn Sherwood (SRI-KL).
-------
∂13-Jul-77 1100 FTP:PRATT at MIT-AI (Vaughan Pratt) LISP article for Encyclopedia of CS and T (Belzer et al eds)
Date: 13 JUL 1977 1356-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
Subject: LISP article for Encyclopedia of CS and T (Belzer et al eds)
To: (@FILE [PRATT;ARTICL LIST]) at MIT-AI
The LISP article is now in a state of "completion" inasmuch as it has been
sent to the editors. Any further work I do on it will be in response to
people pointing out errors of fact and major errors of exposition, and will
be incorporated into a TR version of the article, as well as into the proof
copy. Thank you all for your many inputs. My apologies to those whose
suggestions I was unable to incorporate, mostly due to pressure of time and
space (the thing was threatening to become a book).
I will leave a version AI:PRATT;LISP UNFRIL on my directory for those who want
the article minus all font-change characters and underlines. LISP > will
retain all the frills if you don't want that info discarded.
∂13-Jul-77 1324 CCG nsf
To: LES, JMC
Talked to Bruce Barnes at NSF about my pending proposal. It is still sitting
in the "approved with no funding" category. He is still looking for scraps
in the 15k to 20k range,but is not hopeful. I emphasized to him our dire
straits. There is also a chance it may get partially funded in october or
november. For planning purposes we shouldn't count on it, but it's not dead
yet.
Cordell
∂13-Jul-77 2111 MRC via RTGT my arrival
To: JMC, LES
i will be arriving in california early friday morning (2 AM ish).
∂14-Jul-77 1350 PAT SUMMER AIR FARE INFO
To: NAMES.TRA[1,PAT]:;
To all summer travelers:
The following is a summary of types of airfares available.
SFO - San Francisco, BOS - Boston, ROC - Rochester, TOR - Toronto
You may make your reservations with whatever travel agent you wish.
The lab has accounts set up with Sequoia Travel on campus - 323-9401
and All-Ways Travel - (9) 964-5366.
If you have questions or need help, please ask me.....Good luck! patte
Regular round trip air fare SFO/BOS $430.00
SFO/ROC $384.00
SFO/TOR $388.80
BOS/ROC $104.00
BOS/TOR $127.44
SFO/ROC/BOS/SFO $459.00
SFO/TOR/BOS/SFO $473.12
Regular night coach SFO/BOS only $344.00
Freedom Fare (excursion)
must book 14 days in advance, must stay a minimum of 7 days at furthest
point on itinerary, maximum of 30 days. Up to two stopovers are allowed.
Round trip excursion fare
SFO/BOS $365.00
SFO/ROC $326.00
SFO/TOR $365.04
ROC/BOS $ 88.00
TOR/BOS $ 75.60
One way fares
SFO/ROC/BOS/SFO $389.50
SFO/TOR/BOS/SFO $402.82
Super Saver
must book 30 days in advance
must stay 7 min 45 max days
flies west coast to only three cities - BOS, NY, Philadelphia
(these are round trip fares SFO/BOS)
midweek fare (flying TWTh) $287.00
shoulder fare (flying on F or M) $307.00
weekend (flying Sat Sun) $328.00
Tour Basing (excursion)
Must purchase $65.00 worth of land arrangements in addition to air
May stay up to 30 days
unlimited stopovers at $15.00 per stopover
PLEASE NOTE: THESE FARES ARE SUBJECT TO A 1 OR 2% INCREASE IF APPROVED
BY THE POWERS THAT BE.
∂15-Jul-77 1257 FTP:Geoff at SRI-KA (Geoff Goodfellow) NS
Date: 15 Jul 1977 1249-PDT
From: Geoff at SRI-KA (Geoff Goodfellow)
Subject: NS
To: JMC at SAIL
cc: MMcM
Would sometime this weekend be a good time to talk about hacking
on NS?
-------
Yes, why not phone me sometime when you see me logged in and active?
∂15-Jul-77 1442 DWW via AMET
To: JJK
CC: PMF, wiederhold at SUMEX-AIM, DPB, feigenbaum at SUMEX-AIM, JMC
Hi Jonathan! This letter is from ejg and dww in response to your
questions about facilities. Sorry we have taken so long to answer
[sigh...].
I. Terminals.
A. Accessability. We are inclined to put this consideration at the
highest priority. Terminals should be available to department PhD
students twenty-four or more hours a day without elaborate prior
arrangements (such as checking out keys), and more than a single
terminal (such as one in his office) should be so available. Taken
by itself, this consideration means that we would prefer a terminal
room to which we have a key over a terminal in our office but no
access to others when that one is busy.
B. Features. Accessability concerns might lead one to a bias for
many cheap terminals over a few more powerful ones. This is
consistent with our suggestion only if not taken to extremes:
features like lower-case letters are absolutely essential, hardware
which will support screen-editing (such as E or TVEDIT) is highly
desirable, and graphics capability would be nice but is hardly
essential.
C. Convenience. This is probably of the lowest priority. Having
terminals in a nearby location is efficient and generally winning,
but having them available is clearly more important. Tradeoffs
which may arise are whether or not there is room for these terminals
in the first place, and whether non-terminal work can be done in an
office with a terminal. Is it preferable (assuming access to all
offices at all times is available) to have a terminal room per floor
at the expense of crowding five people into each ordinary office, or
instead to squeeze a terminal into each four-person office? Could
the problem of working alongside a terminal user be dealt with by
allocating terminals non-uniformly (e.g. 0, 1, or 2 per office, with
the ability to request the kind of office desired) or non-statically
(e.g. movable terminals)?
II. Main Facility.
A. Features.
1. Text editing and document preparation: reports, theses.
2. Fancy output devices: e.g. XGP, plotter, Printronix
(the latter is a low-resolution, low-cost graphics printer)
3. Other graphics capabilities: CRT vs. hardcopy.
4. Department-wide mail facility.
5. High-level language support: e.g. SAIL, ALGOL W.
6. Systems programming language: e.g. C, BLISS.
7. Communication links: ARPA, LOTS, SCIP, SLAC, dialup
B. Power. LOTS tends to have load averages in excess of 20, and
this user community does not consist of sophisticated demanding
users. We consider almost any alternative (even a batch system) to
be preferable to such a heavily-loaded environment. This implies
the need for a very powerful main facility (fast cpu, lots of
memory, etc.).
III. Wild Speculations.
A. Very intelligent terminals with local programmable processor and
memory, disk or cassette storage.
B. Prototype HYDROID: more than 100 interconnected small processors.
Please note that this suggestion is intended as an addition to the
main facility rather than an alternative to it.
We hope that these rather disorganized ramblings give you some idea
of what we consider desirable. Let us know what you come up with.
Erik & David
∂15-Jul-77 1511 FTP:Boyer at SRI-KL A Paper Request
Date: 15 Jul 1977 1510-PDT
From: Boyer at SRI-KL
Subject: A Paper Request
To: JMC at SU-AI
I have seen a reference to a 1977 paper of your, "Representation
of Recursive Programs in First Order Logic". Would you
please send me a copy?
Robert S. Boyer
Computer Science Lab
SRI
Menlo Park, CA 94025
Thanks,
Bob.
-------
∂15-Jul-77 1520 FTP:REM at MIT-AI (Robert Elton Maas) Holofile
Date: 15 JUL 1977 1817-EDT
From: REM at MIT-AI (Robert Elton Maas)
To: LES at SU-AI, JMC at SU-AI, TVR at SU-AI, PBARAN at USC-ISI
CC: REM at SU-AI
.MAIL LES,JMC,PBARAN%ISI,TVR
Subject: Holofile
Type message followed by <CONTROL>Z
I have written to Holofile Industries Ltd., distributor for
the TRW holographic 4" by 6" 200-megabit storage medium, and received
a prompt reply saying that there is no more printed material available
at this time but answering most of my questions.
The basic reader, described in the advertising brochure, costs
$5000 (not $500 as in the brochure). There is no comment on that, so
I don't know whether the brochure or the letter has a typographic error,
although surely one or the other must. If the cost is $5000, it is
too expensive for hobbyists, but reasonable for SU-AI or IMSSS to
replace one disk drive or whatever 200 megabits amounts to.
The writer is about $250,000 but the company will offer the
service of converting magtape to holofile for a very low charge,
available in key cities. [SU-AI MAIL PROGRAM CRASHED HERE, REMAILING FROM MIT-AI]
Delivery expected first half of 1978.
Anyone wishing to read full text of the letter, from James A. Case,
Vice President (of course, like most companies, probably at least 10
v.p.'s, maybe) should contact me (REM).
Major application I envision is storage of system software and
other files that aren't actively being changed but which are being
used (read) online. All files older than a certain date, or 200
megabits thereof, could be copied to magtape and written on holofile,
with extra copies at $.07 each appx., then disk service could be modified to
point from directory to places on this device, and after verification
of correct data the disk version could be purged. Whenever the file
gets edited or otherwise rewritten, the pointer to the holofile could
be scrapped and a copy made on the disk. Of course a study for cost-effectiveness
would have to be done first.
∂16-Jul-77 1354 RPG New Ncomplr
To: MACLSP.DIS[AID,RPG]:;
New NCOMPLR up at SAIL. Bugs → RPG.
Roll back to SYS:NCOMPL.OLD if necessary.
∂17-Jul-77 1353 PAM possible bug in newest ncompl
To: MACLSP.DIS[AID,RPG]:;
I have had some trouble today that appears to stem from the newest ncompl. If
you have similar problems, try ncompl.old[1,3]; it worked for me.
Paul
∂17-Jul-77 2102 TVR
Burning leaves does not bother the bat, nor does it want the .lunch, and
the torch won't light the guano. Maybe the rope or wire may be of some
help. -- Tovar
I tried tying the bat with the wire - with the usual provocative results.
∂18-Jul-77 1439 FTP:TAA at MIT-DMS Saving state
DATE: 18 JUL 1977 1707-EDT
FROM: TAA at MIT-DMS
SENDER: TAA at MIT-DMS
SUBJECT: Saving state
ACTION-TO: jmc at SU-AI
MESSAGE-ID: <[MIT-DMS].54481>
The reason for the current restriction on state-saving is that
the current 'Save' command does a muddle SAVE, which requires on the
order of 87K of disk space. SAVE, by its nature, cannot work over the
net, and we don't have enough disk space here to allow non-local users
to do saves.
One of the locals has (almost) volunteered to write a save
command that would save only the necessary bits, rather than the entire
dungeon (which includes all the code). This still has the problem that
it would create a local file. It might be possible to cause the bits
to come out in ASCII, in which case one could presumably cause them to
be scripted to a file elsewhere, then fed back in; network users would
then be able to save a game on their own systems, which is highly
desirable.
Given the current structure of the user interface, it is (I
think) possible to add macros without too much difficulty; the major
difficulty is random events which go the wrong way: failing to kill the
troll, being robbed of a necessary implement by the thief, etc. You'll
notice that I have not mentioned any mechanism for creating, saving, and
editing macros; I'm not sure what a reasonable user interface would look
like, and I am extremely reluctant to start what would inevitably turn
into an amazing time sink.
∂18-Jul-77 1635 RWW SAMEFRINGGE
PROOF IS COMING ALONG. FINALLY GOT THE IMLAC FIXED. HOPE TO FINISH MOST OF IT BY TONIGHT.
RWW
∂19-Jul-77 0900 RPG NCOMPL
To: MACLSP.DIS[AID,RPG]:;
The new NCOMPLR (see last message) was a loser so r ncompl
gets the old one.
-rpg-
∂19-Jul-77 1039 FTP:Taynai at SUMEX-AIM IBM Fellowship
Date: 19 JUL 1977 1037-PDT
From: Taynai at SUMEX-AIM
Subject: IBM Fellowship
To: McCarthy at SAIL
cc: Feigenbaum
Prof. Feigenbaum received a letter from Ashok Chandra
of IBM notifying us that Christopher Goad has been awarded
an IBM Fellowship for 1977-78. Prof. Feigenbaum will write a letter of
congratulations to Mr. Goad.
Wanted you to know the good news.
Carlyn
-------
∂19-Jul-77 1436 JP
To: MACLSP.DIS[AID,RPG]:;
SAIL MACLSP interprocess communication
A preliminary version of a MACLISP interjob communication facility is
available to those of you lusers subscribing to the (help) autoload
facility. These functions permit arbitrary creation, deletion,
synchronization and intercommunication of MACLISP processes. For more
information read FASFRK.JP[LIB,LSP].
-j
∂20-Jul-77 0808 FTP: Ijcai-77 at CMU-10B Model Theory paper.
Date: 20 Jul 1977 1105-EDT
From: Ijcai-77 at CMU-10B
Subject: Model Theory paper.
To: jmc@su-ai
Sender: IJCAI-77 at CMU-10B
Message-ID: [CMU-10B] 20 Jul 1977 11:05:57 Ijcai-77
John,
We just got a letter from Hayashi saying that you will present
the paper at the conference. Wanted to make sure that you knew about
that.
We had sent a letter to him apologizing for the snafu about
getting the paper in and suggesting that he might want to bring copies
along to be distributed at the conference and also that we would
arrange to get it printed in the SIGART Newsletter. He has not yet
responded to those suggestions (and perhaps he has not yet received that
letter).
Lee
-------
∂20-Jul-77 1612 FTP:SUZUKI at PARC-MAXC concerning the decidability of list
Date: 20 JUL 1977 1613-PDT
From: SUZUKI at PARC-MAXC
Subject: concerning the decidability of list
To: dco at SAIL
cc: jmc at SAIL
I am curious to know how y`ou arrive at the decidability of
LIST with CAR, CDR, CONS, and EQUAL,
but it came to my mind that
since the theory of LIST has no finite models and it is
aleph 0 categorical , and also it is axiomatizable
thus, the theory is decidable. (cf. Mendelsohn).
Do you mean this result or by some other facts.
My decidability comes from more direct
decision procedure as you will guess. Nori
-------
I assume you would include the predicate ATOM as well. My idea that it
is decideable also came from almost pushing through a decision
procedure that involved quantifier elimination. DCO's decideability
is a part of a larger picture which he will doubtless explain. If you
have written yours, I would like to see it.
∂21-Jul-77 0137 REM via AMET PCNet
Protocol specs are now finalizing; I think we know what we
are going to try to implement all the way from the bottom (start+8+stop
UART/TTY encoding of Radix-41) to the top (node and user addresses,
NetMail and FTP services with automatic forwarding). I expect that
within a month we will have working nodes on SU-AI, MTK's personal computer,
MIT-MC, Keith Britton's personal computer in Hayward, and possibly
MGW's personal computer in Boston and my 6502 computer I'll be getting
in about a week.
For actual operations of SU-AI as a host (dialin, as contrasted
with user=dialout) node, there is a question whether the computer dialing
in will log in as [PCN,REM] or as [NET,GUE] or [PCN,GUE] or simply issue
the PCNSRV command which starts up a tempory job like MAIL and WHO etc.
Until such a facility is available, each user who has SU-AI account will
have his node log in here as himself and others won't be able to send
(PC)NetMail to here at all except through one of the REM or MTK nodes.
There's no hurry since for at least the next month only MTK and REM will
reasonably want to run SU-AI server. But you might think about what
you want done. I am leaning toward PCNSRV command starting up temp job
running PCNSRV program, and DIALNET or DIANET command starting up temp
job with DialNet server when it comes into existance next year. This
of course assumes you want SU-AI to be formally on the PCNet at all.
I hope so since it will allow SU-AI to perform FTP and MAIL with
universities whose computers and/or dialup lines aren't advanced enough
to support DialNet but which can and will support PCNet.
∂21-Jul-77 1313 BPM
To: JMC, LES
"Computers"
by FRANK VOGL
dispatch of The Times, London
distributed by the New York Times News Service
12:34 pm, 21 July 1977
WASHINGTON - We are surrounded by computers, but most of us
understand very little about them. Computers are no longer merely
business machines, but pieces of equipment that are increasingly
being used in all manner of contexts, from the kitchen to the school
to the hospital to the airport. The extent to which the computer has
come to play a key role in our everyday lives is not widely
appreciated and to most laymen there is something frightening about
the idea of machines of such complexity and versatility.
Some American social scientists are already talking about the
dawning of a second industrial revolution where the machine now
replaces human brain labor in much the same way as it replaced manual
labor in the first industrial revolution. acomtuters appear all the
ening when scientists talk about the development of
machines that are as intelligent as human beings.
A great deal of research is currently being done to produce just
such machinery, and Dr. John McCarthy, the director of Stanford
University's Artificial Intelligence Laboratory in California,
believes that present computers are capable of human level
intelligence.
McCarthy is swift to point out, however, that so far nobody has been
able to write the program that are needed to produce highly
intelligent machines. He noted recently that computers can be
programmed to play chess better than 99 per cent of the population,
but the computer has no general intelligence whatsoever. In fact, he
asserted that ''the machines are dumber than almost the dumbest human
beings.''
Moreover, it looks like being years before the scientists find means
of making intelligent machines. ''If people ask me how far away is
artificial human intelligence,'' said McCarthy ...
''and five Faradays and 0.3 of a Manhattan Project and it
would be best to have the Einstein before the Manhattan project.''
The computer is so far away from being an intelligent machine that
the only frightening thing about it is the general ignorance about
the way it works, the way it can be used to best advantage and the
future practical developments that can be made in computer
technology. This ignorance, according to many computer scientists, is
a barrier to greater use today of the computer.
A typical example is in the field of education. Engineers and
teachers have worked for years on the development of teaching
computers. Some of the most advanced work has been done under the
direction of Dr. Donald Bitzer at the University of Illinois.
Computer technology and programs have been developed ...
establishments and which would certainly raise educational standards.
That these computers and programs are being used by only a small
number of institutions reflects the widespread fears of ignorant
teachers and administrators of education. Some teachers are scared
that they will be replaced by the computer and there are also deep
fears that the computer will serve to stultify the imagination of
students and turn our children into walking robots.
The experts at the University of Illinois stress that the computer
must be seen as a valuable teaching tool and not as a substitute for
the teacher. They argue convincingly that the computer can help the
teacher to explain complex concepts, that it can also help the
student to solve problems and learn facts and theories. Surveys of
students who have used this computer system show that the machinery
is considered valuable and teachers using the computer claim that the
machinery has definitely helped to raise academic achievement.
The computer can free the educator from much of the routine work he
presently has to do and so enable him to spend more time in answering
questions from students and in discussing philosophical and general
problems with his classes. The computer, as demonstrations at the
University of Illinois showed, can be not merely of immense value in
the teaching of the pure sciences, but it can also play a useful role
in teaching such assorted subjects as foreign languages and even
music.
Bitzer's ''Plato'' computers are now capable of printing words and
numbers, drawing pictures of all types, speaking directly, playing
music and displaying color photographs. The communication skills of
the computer are being constantly improved upon. At McCarthy's
laboratory in California, for example, computers have been programmed
to reproduce human voices and the tones of musical instruments almost
to perfection.
The gradual decline in the manufacturing costs of computers is the
most compelling of all arguments in favor of their rapidly increasing
general useage. Miniaturization of components has been the most
sensational of computer developments in recent years and has led to
the production of computers on a massive scale at tiny cost. New
computer developments are being made now at a dramatic pace that will
lead to still greater miniaturization and still further cost
reductions.
The range of common applications of computers is reflected in the
wide array of products that are either now in the shops or that will
shortly be widely available. Computer base wrist watches and pocket
calculators are inexpensive and common place. Watches have been
developed that contain alarm systems, means of telling the time in
differing zones and that can also do highly sophisticated
mathematical calculations.
Computer-based kitchen ovens, refrigerators and other home
appliances have been ...
car engines are now being ...
enable the car engine ...
Computer diagnostic machinery is widely used in hospitals. Travel
agents take for granted the computer booking and reservation systems
they use daily. Supermarkets now use computer cash registers and
scales.
The computer has already become a valuable tool of civil servants
and it is likely to become much more widely used before long.
Computer information banks and systems are now being developed that
enable public administrators to make much more efficient use in
policy planning of the masses of raw economic, sociological,
environmental and demographic data that is constantly being collected.
The use of the computer is rapidly becoming so pervasive that it is
bringing greater changes in our living styles than most of us
realize. It is also, of course, bringing with it new practical and
philosophical problems. However, the computer is something that
everyone should learn more about and which undoubtedly will be
increasingly used as more technological break-throughs are achieved,
as the general level of public ignorance about ...
threatened by what is merely a lump of metal and silicon.
∂21-Jul-77 1319 FTP:SUZUKI at PARC-MAXC (Response to message)
Date: 21 JUL 1977 1320-PDT
From: SUZUKI at PARC-MAXC
Subject: (Response to message)
To: JMC at SAIL
cc: SUZUKI
In response to your message sent 20 Jul 1977 2106-PDT
I was very busy writing other papers and did not get to writing up since I met you
at Carnegie. Have you solved the case with quantifier? Oppen said
he only worked for unquantified theory. Nori
-------
∂22-Jul-77 0512 FTP:CERF at USC-ISI dungeon
Date: 22 JUL 1977 0506-PDT
From: CERF at USC-ISI
Subject: dungeon
To: jmc at SAIL
cc: cerf
John,
I hear you have become an expert in Dungeon -
would you mind revelaing the secret of entering Hades?
I have taken bell, book, and candle there;
I have prayed for entry, I have waved garlic in
the face of the spirits; i have hollered "hello sailor"
in various places around the dungeon hoping to be
sent to hades for my sin, but so far nothing seems to work.
How is the Dial-net stuff coming?
Cheers
Vint
-------
You have to say EXORCISE at the entrance while posessing bell, book and candle.
The candle must be lit, which will require a match if it has blown out. If you
find out about how to get the ruby or how to get into the section of the dungeon
off the Egyptian room, I would be grateful for a clue. The greatest sin I have
been able to commit is to burn the book.
Mark Crispin has just come to work on Dialnet, so we should have some
definite plans soon.
∂22-Jul-77 1145 RWW STATUS
Samefringe is bogged down in bugs.
fixed tauteq for REF
fixed subst for REF
fixed spool for CLT
system now to slow. also need to fix rewrite bug.
will be back this afternoon.
rww
∂23-Jul-77 1410 FTP:TAA at MIT-DMS ZORK
DATE: 23 JUL 1977 1707-EDT
FROM: TAA at MIT-DMS
SENDER: TAA at MIT-DMS
SUBJECT: ZORK
ACTION-TO: JMC at SU-AI
MESSAGE-ID: <[MIT-DMS].54787>
As far as I can tell, the previous solution to the loud room still
works (unless you've been playing on SRI-KL, in which case there is a bug
which may (now that I think of it) make the thing insoluble). If you've
been playing here, I'd appreciate further details as to how you got it to
lose.
At dms, echo works. At sri-kl I think echo originally worked, but I had
to change it at some time to eecho, and some said that <blank>echo worked.
As of last night, neither worked and neither did any obvious variant.
As to how I got it to lose, I usually take the painting, drug the cyclops
and return, then go to the dam, open it, and then go down for the bar
after picking up the boat and the pump. The files scr.1, scr.2 and scr.3
on [e77,jmc] accomplish this. When sri-kl gets fixed, please let me know.
∂24-Jul-77 1003 RDR via AMET LOTS, Provost & misc.
I am worried that it is being accepted that DEC and Dexmach will be slow
without any fuss.
I am more worried that it is being accepted on Ralph's part that this
will be handled by squeezing users.
I am worried that the mechanism for this squeezing is not best developed
by you, Ralph and the Provost.
I think that with the hardware expansion the past load plus the essentially
unavoidable new load (CS105/6) can be handled so that restrictions out to
be more on NEW use than on old.
Most of all, there is a lot to be gained by avoiding tangible forms of resource
allocation in favor of vaguer inuendos. These are less difficult
and more flexible than funny money or counting hours. Only time would
tell the extent of the cost of enforcing whatever might be done, but clearly
it is not desirable to do what so often is done and add bureaucracy,
paperwork, wasted computer time, and STAFF! Also, whatever
formula is developed need not affect 90% of the users
using 10% of the resource--i.e. basically undergraduates shouldn't be
squeezed out of whipping up a program fro their Bio 1 homework or
to calculate a few baseball averages or to text edit just beacause
a certain number of users from a different category need
to be limited. Never should restrictions be made on machine use when it
would otherwise sit idle... Finally, the contribution to the low overhead
made by certain students who devote a lot of time to LOTs should not
be overlooked so that they are forbidden to devote this time.
[What is it, 50% of the SCIP machine that goes to SCIP internal
users who ARE ALSO GETTING PAID more money?]
If you have any suggestion about what I could do about this or if you
want to assure me that all will be well I would be grateful.
∂24-Jul-77 1519 WP
The following outlines a possibility to establish the relation between
subgoal assertions and fol formulae. The multiterm in its original flavour
has disappeared. But still there is a straightforward conversion of
recursive function definitions into functions on sets. The multiterm
now degenerated into <h>, meaning the relation corresponding to the function h.
I did not include formal proofs, they are very trivial, most things follow
by known theorems anyway.
Wolf.
Definitions
Given a functionl F constructed of if-then-else, composition, and some
set of base functions we construct a function R, which maps sets into sets.
The definition proceeds recursively according to the possible structure of
the functional. We write F→R to indicate that R is corresponding to F in
the above sense. A shorthand notation for R is provided in brackets.
(1) λf.λx.h(x) →
λr.{(x,y)|h(x)=y}
[ <h> ]
(2) λf.λx.F1(f)(F2(f)(x)) →
λr.{(x,y)|∃z.(z,y)εR1(r) ∧ (x,z)εR2(r)} where F1→R1 and F2→R2
[ R1(R2(r)) ]
(3) λf.λx.if p(F(f)(x)) then F1(f)(x) else F2(f)(x) →
λr.{(x,y)|(x,y)εR(r) ∧ if p(y) then (x,y)εR1(r) else (x,y)εR2(r)}
where F1→R1, F2→R2 and F→R
[ if p(R(r)) then R1(r) else R2(r) ]
Remarks:
(4) All relationals constructed in the above way are monotonic functions.
(5) The relationals preserve functionality:
If ∀x∃1y.(x,y)εr then ∀x∃1y.(x,y)εR(r).
(6) The transformations defined are homomorphisms in the sense that
if F→R then R(<f>) = <F(f)>.
All three statements can be shown by induction on the structure of R.
Fixed points (let ∞ stand for omega, also ⊂ is not necessarily proper)
(7) Clearly Y(F) is the limes of ∞,F(∞), F(F(∞))... .
(8) Because of (6) 0, R(0), R(R(0))... converges, let r* be the limes.
Since 6 holds componentwise we have r*=<Y(F)>. Since we did not show
continuity of R we must not conclude that r*=Y(R) !.
(9) By a theorem by Park the monotonicity of R is sufficient to get
r* ⊂ Y(R) .
(10) Since Y(R)= ∩{r|R(r)⊂r} we conclude that:
R(q) ⊂ q implies Y(R) ⊂ q therefore by (9) and (8) <Y(F)> ⊂ q.
(11) Writing R(q) ⊂ q explicitely is equivalent to the conditions for
q to be a subgoal assertion. Therefore we have shown that for any
functional F <Y(F)>⊂q holds for any subgoalassertion q.
∂24-Jul-77 2210 RAK Time to go home
Maybe we can get together tomorrow; as it is now 10 pm and
maybe we missed each other.
Dick
∂25-Jul-77 1149 RWW samefringe
writeup coming along need to go to SF about my passport
will be back later
rww
∂25-Jul-77 2352 DCO
To: suzuki at PARC-MAXC, JMC
∂20-Jul-77 1612 FTP:SUZUKI at PARC-MAXC concerning the decidability of list
Date: 20 JUL 1977 1613-PDT
From: SUZUKI at PARC-MAXC
Subject: concerning the decidability of list
To: dco at SAIL
cc: jmc at SAIL
I am curious to know how y`ou arrive at the decidability of
LIST with CAR, CDR, CONS, and EQUAL,
but it came to my mind that
since the theory of LIST has no finite models and it is
aleph 0 categorical , and also it is axiomatizable
thus, the theory is decidable. (cf. Mendelsohn).
Do you mean this result or by some other facts.
My decidability comes from more direct
decision procedure as you will guess. Nori
-------
[It is not correct to say that the theory has no finite models - it depends on
what assumptions you are making. The theory I am interested in also has the
predicate ATOM. I have three proofs that the full quantified theory is decidable;
all three are constructive; one uses quantifier elimination, one finite automata
(I kid you not); one uses game theory (I still kid you not). The first one I
am writing up now since it generalizes nicely to other theories.]
∂27-Jul-77 1628 FTP:CARL at MIT-AI (Carl Hewitt)
Date: 27 JUL 1977 1926-EDT
From: CARL at MIT-AI (Carl Hewitt)
To: jmc at SU-AI
Dear John,
Jean Sammet has asked me to be the primary language coordinator
with you for producing a paper on the early history of LISP. I understand
that you are agreeable to producing the paper by early next year and to
working with me as primary language coordinator and Barbara Liskov as
backup language coordinator. I will try to be helpful in generating
suggestions, comments, and quibbles for the paper that you are writing.
To prime the pump I have collected some questions (in no
particular order) that you might consider:
1: What was the influence of FLPL and IPL on your thinking in 1958?
2: How did Church's monograph influence your thinking on LISP? Why is
LISP different from the lambda calculus in using fluid binding instead of
lexical binding?
3: How did the FUNARG problem arise and become understood? What is the
FUNARG problem?
4: What do you consider to be the most important mistakes in LISP 1 AND
LISP 1.5 both in the languages themselves and in their implementations?
5: What is the historical relationship between M-LISP and LISP 1 and what
is your current view on the relationship?
6: How did you try to influence ALGOL to include LISP concepts and to
what extent do you think that ALGOL was influenced by this advice?
7: Why is Polish prefix notation used in LISP 1 instead of traditional
mathematical infix notation?
8: Why were the names "cons", "car", "cdr" chosen instead of more
mnemonic names like "first" and "rest"?
9: What changes were made in going from LISP 1 to LISP 1.5?
10: What is the history of LISP as the basis for an interactive language
and why has it been so successful in this regard?
11: How did the data structure for programs in LISP come to be same as
the data structure operated on by programs?
12: Where did the definition of eval come from and what is the
relationship of eval to concept of a universal Turing machine?
13: Where did the idea of a garbage collector come from? How does it
compare with previous methods of recovering list cells?
14: What is the relationship between LISP and the ADVICE TAKER?
15: Why were RPLACA and RPLACD introduced into LISP?
16: Why was DEFINE introduced into LISP?
17: Why were PROG, GO, and RETURN introduced into LISP given that you
knew that recursive procedures can do equivalent computations? Was it
purely a matter of being worried about stack overflow or was it a
concession to the FORTRAN mentality? Did you know that by implementing
the interpreter using tail recursion that you do not use any extra stack?
Sincerely,
Carl
∂27-Jul-77 2136 FTP:TAA at MIT-DMS Candle bug
DATE: 27 JUL 1977 2347-EDT
FROM: TAA at MIT-DMS
SENDER: TAA at MIT-DMS
SUBJECT: Candle bug
ACTION-TO: JMC at SU-AI
MESSAGE-ID: <[MIT-DMS].54992>
Actually, it's a bug here, too. It will be fixed when I can get
enough cycles together to install a new version.
∂28-Jul-77 1340 CJS summer absence
To: DRB, EK, JP, LIS, JLS, JJ, RDA, RPG, WLS, DBG, RCM, JMC, ARG
A "little bird" told me you people were going to be gone from the lab for a
while. Please let me know when you are leaving and when you intend to return,
as I have to keep a list of these dates for the University. -Connie
∂28-Jul-77 1544 TOB quid pro quo
Les may be leaving soon. I will see what
I can come up with as alternatives and present
them to you.
Tom
∂29-Jul-77 0015 DON
I suspect I know how to use the wire; I haven't played for over two weeks.
For both reasons, am not currently interested in trade. Besides, at this
point I'm not sure I have anything worth trading.
∂29-Jul-77 0324 DON Auto-zork
To: TVR, JMC
I've noted that both of you have set up input files for speeding your
way through the known sections of Dungeon. I figure that if I'm to
do any more exploring/experimenting, I should follow suit. But I'm
having trouble in that, if I use meta-I to ship the input file over
telnet, the input stream gets too far ahead of the program and some
of it gets lost. This happens even if I cut the individual files
down to about 30-40 lines. Has either of you found a better way to
use these package-plan explorations?
My meta-I files rarely work at DM, so I use SRI-KL. I am thinking
about a LISP adventurer's assistant that could be programmed to
go places and get things on the basis of knowledge accumulated so
far.
∂29-Jul-77 0330 DON
P.S.: I found the mirror-switch.
∂29-Jul-77 1032 FTP:Hart at SRI-KL Lynn Quam
Date: 29 Jul 1977 1033-PDT
From: Hart at SRI-KL
Subject: Lynn Quam
To: JMC at SAIL
cc: Hart
John,
Lynn has applied for a position here, and while some of us already
know him I would nonetheless be grateful for your candid evaluation of his
ability and potential.
Thanks for the help.
Peter
-------
I think his ideas are often pedestrian, but there is no doubt about
his ability to accomplish the goals he sets himself.
∂29-Jul-77 1600 DBG
Regarding your comments on my thesis proposal:
2. (Taking pictures with the Cart and then processing them.) This is what
I intend to do. In fact, I already have taken some pictures and am experimenting
with them.
4. (Calibration using large known object.) Hans Moravec is now in the process
of adding to the Cart a device for moving the camera from side to side by known
amounts. This will solve the scale factor problem. The distances to some points
in the scene can be found using this short baseline, and these can be used to
calibrate a longer baseline between two Cart positions from which the same points
are seen. (Presently, I simply measure with a tape measure how far the Cart has
moved between picture-taking positions.)
∂30-Jul-77 0237 ME Clobbered files
The following files of yours were accidentally clobbered Friday night.
It is up to you to restore them from DART tapes where possible.
The file FILES.GON[1,3] contains the master list of all files clobbered.
Files clobbered:
CALEND.OLD[ESS,JMC]
FLAT.STP[W77,JMC]
HIGHFI.LE1[LOT,JMC]
IJCAI.B[E77,JMC]
INSANE.LSP[206,JMC]
MENTAL.NOT[E77,JMC]
PHYSIC.LE1[LET,JMC]
SATO.LE2[LET,JMC]
SCRIPT.ZOR[LET,JMC]
SEP76.OUT[LET,JMC]
SLOTNI.LE1[LET,JMC]
STOYAN.LE3[LET,JMC]
VIETNA.NS[ESS,JMC]
WISEMA.AX2[F75,JMC]
ZOHAR[F75,JMC]
∂30-Jul-77 1542 REM via AMET PCNET protocol committee, progress report.
This last Thursday we have finished cleaning up loose ends
in our first complete version of the PCNet protocol. We are now in
the midst of writin up exactly what we've designed, and will be
implementing it soon (as soon as I have time to update PCNSRV.SAI
to reflect the design of the top half of the protocol and the changes
from version-1 of the bottom half. Estimated date of first node
able to accept local messages from the net via somebody dialing it
up and talking Radix-41 to it, 3 weeks from now. Hopefully by the
same time we'll have MTK's Altair and/or SU-AI able to dial the
SU-AI server (or just run PTY job with server under it) and send the
message, thus allowing actual testing.
I expect that within a week after that we'll have the ITS
version of PCNSRV.SAI working, thus allowing messages to be
reliably send from SU-AI to MIT-* via PCNet when our IMP is down.
sent
Traffic in reverse direction will be difficult until MIT installs
a dialout line (as far as I know they don't have any).
∂30-Jul-77 1925 FTP:TAA at MIT-DMS ZORK
DATE: 30 JUL 1977 2224-EDT
FROM: TAA at MIT-DMS
SENDER: TAA at MIT-DMS
SUBJECT: ZORK
ACTION-TO: JMC at SU-AI
MESSAGE-ID: <[MIT-DMS].55140>
I agree that it's crockish to have 25 tons of rock fall out of a
clear blue sky, but on the other hand it's hard to fix, so for now it's
a feature. (Try causing the explosion in the room the crock drops you
in after you die. I'm not sure what will happen, but it might be amusing.)
Aside from that, all your bugs have (in theory, at least) been
fixed; thank you for reporting them.
∂30-Jul-77 2010 RWW japanese
To: JMC, WHG
Their stuff is on the bunk bed in my room.
∂30-Jul-77 2011 RWW last message
they can be gotten by asking whg or rwg.
rww
∂30-Jul-77 2016 RWW fol
to repeat
semantic attachment →→ CG
syntatic simplifier →→ AMR
general bugs (e.g. function parameters) →→ DSB
general use →→ REF
∂30-Jul-77 2145 RWW samefringe
I am leaving in your mailbox copies of all the proofs. There is
additional information in INFO[jmc,rww]. The informal proofs
are in SAME.CMD[JMC,RWW]
P1 is termination of gopher
P2 is terminatiom of samefringe (your samefringea)
L1 is ∀p.fringe(gopher(p))=fringe(p)
L2 is ∀p.car(gopher(p))=car(fringe(p) and
∀p fringe(cdr(gopher(p)))=cdr(fringe(p))
L3 is ∀x y.equal(fringe(x),fringe(y))=samefringe(x,y)
ie correctness of samefringe.
Not that the number of steps taken and the number of charactors typed
is comperable of those of the informal proofs. I will take copies with me.
we can discuss them in boston.
rww
∂31-Jul-77 1857 MRC via AMET DIALnet
I just realized you haven't mentioned IMSSS and Sumex irt DIALnet. While Sumex
may not be interested, I think IMS↑3 would be.
I'm going to talk to you Monday about some things about implementation
and so on. TED has invited me to a meeting about networks in general (read my
mail for details) which I am going to attend. Maybe you will want to be there
too.
∂31-Jul-77 1903 FTP: Ijcai-77 at CMU-10B Change in session for your ''Model Theory'' paper.
Date: 31 Jul 1977 1656-EDT
From: Ijcai-77 at CMU-10B
Subject: Change in session for your "Model Theory" paper.
To: jmc@su-ai
Sender: IJCAI-77 at CMU-10B
Message-ID: [CMU-10B] 31 Jul 1977 16:56:03 Ijcai-77
John,
Yet another kicker: We are creating a new session
on Thursday Morning (Aug 25), 8:30-10:00, and are moving
Pat Hayes (from Thmn Prvg-3), C. Schwind (from Repr of K-2),
and your paper (with Sato, et al) into it.
I hope this does not cause you any difficulties.
Lee Erman
-------
∂01-Aug-77 1156 CJS absence from lab
To: JP, JLS, RDA, RPG, JMC
I need to know the dates you will be gone from the lab. Please mail me your
departure and return date as soon as possible. Need this info for the Univ.
Thanks, Connie
∂03-Aug-77 0010 WP Implementation of LISP functions
I thought about the problem you posed on proving the implementation of LISP
functions correct. Clearly there is a whole spectrum of problems involved
ranging to the full compiler verification. I attacked the problems at
its more trivial side. The following PASCAL program realizes the functions
car, cdr, cons as pascal procedures. Note that for formal reasons I have to
return the result with an additional parameter. The implementation is
still a very crude one, not considering special cases. The general idea
is to encode sexp's as integers. These are treated as indices in the array
memory. A dotted pair is represented with two consecutive elements in the
array, pointing to car and cdr respectively.
To avoid the translation problem the main program contains
what I define to be the implementation of car(cons(a1,a2)) and
cdr(cons(a1,a2)). The proof then shows that the former is equal to a1 and
the latter is equal to a2.
The program as it stands is verified by our current system.
Clearly there are refinements possible and necessary. I just wonder
if this attempt has something to do with what you had in mind?
Wolf
PASCAL
CONST MEMLEN=1000;
TYPE
MEMTYP=ARRAY[1:MEMLEN]OF INTEGER;
VAR
MEMORY: MEMTYP;
FREE,A1,A2,R1,R2,R3: INTEGER;
PROCEDURE CAR(VAR RES: INTEGER; X: INTEGER);
ENTRY X<FREE;
EXIT RES=MEMORY[X];
BEGIN
RES←MEMORY[X];
END;
PROCEDURE CDR(VAR RES: INTEGER; X: INTEGER);
ENTRY X<FREE;
EXIT RES=MEMORY[X+1];
BEGIN
RES←MEMORY[X+1];
END;
PROCEDURE CONS(VAR RES: INTEGER; X,Y: INTEGER);
GLOBAL (MEMORY,FREE);
ENTRY TRUE;
EXIT MEMORY[RES]=X ∧ MEMORY[RES+1]=Y ∧ RES<FREE;
BEGIN
RES←FREE;
FREE←FREE+2;
MEMORY[RES]←X;
MEMORY[RES+1]←Y;
END;
ENTRY A1<FREE ∧
A2<FREE;
EXIT R1=A1 ∧ R3=A2;
BEGIN
CONS(R2,A1,A2);
CAR(R1,R2);
CDR(R3,R2);
END.
∂04-Aug-77 2311 RAK Mistak
By your comments do I gather that page 3 is it? What directory is it in
again?
Dick
page 3 mistak[e77,jmc]
∂05-Aug-77 0317 MRC DIALnet
To: JMC, LES
Could you please send me a copy of all the mail that you have received irt
DIALnet so that I can stay up on this? I always feel guilty about reading
other's mail and generally don't like to do it.
I have talked briefly to Sumex people and I will get an appointment to talk
to them sometime next week.
I'll dredge out what's relevant. There is no special need for Sumex to
come into it at this time. Please concentrate on getting draft protocols.
∂05-Aug-77 0758 TVR
If you are interested in district heating and the use of non-drinkable water,
please read SPHEAT.NOT[2,TVR]. --- Tovar$
∂05-Aug-77 0858 MRC queue of things for me to do
To: JMC, LES
On page 2 of my mail file is a list of what I currently see as what I have to
do irt DIALnet (not including thinking, but that is less tangible). Feel free
to read and add things that I have missed; the longer the list, the better
actually due to the way I work.
∂05-Aug-77 1132 JBR
You are exceeding your disk quota.
Files that occupy space beyond your quota are subject to purging!
If you don't delete some of your files, the purger will.
Your disk quota is: 2000
Your files occupy 2500
∂05-Aug-77 1322 MFB MTC QUAL
I WOULD LIKE TO HAVE THE EXAM IN LATE NOVEMBER, EARLY DECEMBER.
MARTIN
∂05-Aug-77 1556 RDR via AMET CS105/6 TA Consulting
To: INTRO.DIS[1,DPB]:;
I have observed two related issues I'd like to point out. The relation
is that there has been a serious degradation in TA availability to students
as compared to Pine Hall. The first problem is that PPL TA'S don't consult
for SAIL and vice versa. The second problem is that there is no TA
romm and so students have difficulty in locating a TA.
Without requiring that only one language be used, the TA's
can still consult for both. What happens when they don't is
that Joe Random answers questions, and Joe Random often has never used
PPL, or SAIL either.
I wish I had some creative solution for the identification of TA's.
The only thing I could suggest is that they wear funny hats,
and that's no good. Perhaps a table could be identified and
this could serve as a TA station. This would maybe limit movement.
Well, time will tell what works out.
∂05-Aug-77 1637 MRC
∂05-Aug-77 1042 JMC
I'll dredge out what's relevant. There is no special need for Sumex to
come into it at this time. Please concentrate on getting draft protocols.
[roger wilco]
∂05-Aug-77 1817 MRC [DLN,MRC]
To: DIAL.DIS[P,DOC]:;
This directory contains information on some current issues irt dialnet and
pre-preliminary low-level protocol information.
∂05-Aug-77 2130 DON
I'm well on my way to creating a 357-point input file. One trick I came
up with in the process which you may not have thought of: I was carrying
the limit of 40 (with the boat deflated) and wanted to carry the rope as
well. So I tried putting it in the boat, but it wouldn't fit. However,
the buoy was in the boat, so I put the rope in the buoy. It worked! It
might be a good idea to determine each object's capacity as well as its
weight. Also, I wonder if other objects (e.g. the buoy) return to their
"empty" weights once they are closed? By the way, the coal's weight is 8.
∂05-Aug-77 2324 DON zork[e77,jmc]/7p contains weights.
I know. But it didn't include the weight of the coal. Looking over my
map, I believe the other objects whose weights you lack are:
lunch
rope
wire coil
stick
ruby
statue
listings
coke bottles
I'm guessing at those last two; I haven't been to the shrine yet. Also, I
figure I'll weigh the bottle before and after emptying it, in case it makes
a difference. And, as I said, it might be nice to determine if objects
other than the boat can conceal the weight of objects contained therein.
Also, I want to make a table of capacities. Are they based on size or
weight? Lastly, what is "gunk"? Is it the contents of the tube?
Gunk is one of the official names for the contents of the tube.
∂06-Aug-77 2024 GLS MACLSP on SYS
To: MACLSP.DIS[AID,RPG]:;
While MACLSP development is going on at SAIL (i.e. until Wed. Aug 10)
you should use sys:maclsp.old instead of sys:maclsp.dmp.
- rpg & gls -
∂07-Aug-77 0715 FTP:REM at MIT-MC (Robert Elton Maas) DATA-COMPRESSING
Date: 7 AUG 1977 1015-EDT
From: REM at MIT-MC (Robert Elton Maas)
Subject: DATA-COMPRESSING
To: JMC at SU-AI
By keeping track of the low-order bit (SOS line number in
text files), a text-mode cruncher can handle binary files
without either crashing or munging the file. I have begun
implementing this method of compressing-files using MACLISP.
I now have a pair of functions CRUNCH and UNCRUNCH which
correctly perform this process in both directions, using
a trivial compressing scheme at present. The transformations
at present are:
(1) 5 characters and 1 bit picked up out of word
(2) if bit is 0 ignored, but if 1 then a '205
character (token) is emitted before any of the normal characters
(3) all alphabetics are converted to upper-case,
with case-shift tokens inserted:
201=uppercase 202=lower 203=upercase for one character
(4) <cr> followed by <lf> replaced by 204
(5) at end of file, 200 inserted
(6) for quoting characters (in trivial code all
characters appear quoted) the special tokens 200:205
are mapped down to 141:146 (lower case characters no longer
located there) so that 7 bits are enough for any token.
Next to continue work on some of my advanced on-the-fly
crunching schemes which almost absolutely require LISP
in order to do the complicated maintaining of nested arrays of statistics....
P.S. I'm compiling everything, so it runs really fast,
currently a bit faster than 1000 words per second of realtime.
(That is about a factor of 2 slower than SPINDL.DMP[1,3]
which was written in assembly language, so I'll have to
see why and fix it...)
∂07-Aug-77 1643 REM via AMET
To: JMC, LES
Informal list of what use of SU-AI and PCNet by each other
might occur, in the file PCNET.SU[1,REM].
Since we're not asking for funds from you, I haaven't bothered to formalize it,
but at request of JMC I have put it in "writing" in the file. I'll add to it
later if I think of anything else.
∂07-Aug-77 1825 DON zork
Some additional weights: rope=4, stick=1, statue=3 (!), listings=28,
coke bottles=6. I'll have the remaining info soon. You
might want to look at ZORK[1,DON]/3P, which has my table.
It includes a second table with the objects alphabetised;
very handy for determining how much you're carrying.
∂07-Aug-77 1918 DON
Additional notes:
Lunch=2, wire coil=1, water in bottle=1
Closing the sack on things does not conceal their weight. I suspect
other containers will be equally useless in this regard.
∂07-Aug-77 2308 DON
Only the boat seems to act to conceal the weight of its contents. Also,
the boat has a capacity of 40. (See ZORK[1,DON]/3P for other capacities.)
However, you can fit much more than that into the boat by putting things
inside other things in the boat. But if you do this, you cannot then put
more stuff in the boat. I.e., once the boat is overloaded, you cannot put
more stuff directly into the boat, though you can put more stuff into
things within the boat:
>put trunk in boat
Done.
>put coffin in boat [the coffin contains 14 units of stuff]
Done.
>take trunk
Taken.
>put trunk in boat
It won't fit.
You want more proof?
>look
You are in the kitchen of the white house. A table seems to have
been used recently for the preparation of food. A passage leads to
the west and a dark staircase can be seen leading upward. To the
east is a small window which is open.
There is an inflated boat here.
The magic boat contains:
A gold coffin
The gold coffin contains:
A trunk with jewels
A sandwich bag
The sandwich bag contains:
A matchbook
A newspaper
A jade figurine
A bottle
The bottle contains:
A leaflet
A grail
The grail contains:
A knife
A red buoy
The red buoy contains:
A platinum bar
∂07-Aug-77 2331 RAK 11pm news--in case you missed
We watched 3 channe;ls--
Channel 4--no mention of demo (but 15 mins of sports)
channel 5--near lead, mentioned demo, 51 arrests, canned footage of
Diablo cany9n, no metnion of you guys
Channel 7--Near lead, metnioned demo and arrests, mentioned you guys,
showed Tom Connallyt and also your placards. You got about 1/3 or so
of the total coverage in my estimate.
Of those that mentioned it, you did ok. Channel 7 is the widest audience--
most popular in bay area.
Dick
∂08-Aug-77 1104 FTP:BOBROW at PARC-MAXC Some proposed questions for panel discussion
Mail from host Maxc2 received at 8-AUG-77 1102-PDT
Date: 8 AUG 1977 1102-PDT
From: BOBROW at PARC-MAXC
Subject: Some proposed questions for panel discussion
To: hendrix at SRI-AI, JMC at SAIL, martin at MIT-ML,
To: brian at MIT-ML, newell at CMU-10A, sridharan at RUTGERS-10
cc: Bobrow
The characterizations of our respective systems (approaches)
is contained in the paper which is being published with the
proceedings. Therefore we should not cover the same material in our
discussion. The following are some suggested questions for
our discussion:
What is a problem that is difficult for your system or approach?
Have any other systems or approaches come up with answers?
If so, and you don't like their answer, why not?
If we took no more than five minutes to present an answer to these
questions we would have time for some interesting interchanges.
Comments? Alternative suggestions?
danny
-------
∂08-Aug-77 2111 BPM
University of California
Berkeley 94720 (57) 642-6000
∂08-Aug-77 2134 DON green light
Are you sure about the yellow button? I could swear I once pushed the
yellow followed by brown instead of the other way around, and it didn't
work. Perhaps the brown button shuts off the green light. That would
be a shame; I prefer it as a sequential thing.
According to Tovar, your conjecture is correct. The yellow turns on
the green light, and the brown turns it off.
∂08-Aug-77 2248 DON zork
To: JMC, TVR
I've heard through the grapevine that there is a way to TAKE (!) the
old engravings near the Riddle Room. Comments?
∂09-Aug-77 0231 DON
basket's capacity: 20
∂09-Aug-77 0232 DON
shaft basket, that is.
∂09-Aug-77 0236 DON
I've confirmed that the mirror rooms are really two distinct locations.
I used BRIEF mode and it still gave me the full description the first time
I entered the second mirror room.
∂09-Aug-77 0308 DON
Further notes:
Chalice can be opened, capacity=2
Tube's capacity is 3 (mind you it seems strange to put the statue in
the tube, but it can be done!)
Coke bottles, trunk, and bell are not containers.
"CHEST" = "TRUNK".
I have heard nothing about taking the engravings.
∂09-Aug-77 1414 FTP:CARL at MIT-AI (Carl Hewitt)
Date: 9 AUG 1977 1611-EDT
From: CARL at MIT-AI (Carl Hewitt)
To: HEWITT at MIT-AI, jmc at SU-AI
Dear John,
Here is the current situation: I have talked to Jean Sammet about Stoyan.
She is favorably inclined to having him participate as a discussant. Of course
there would be no commitment for ACM to pay his expenses. Jean plans to take this up with
the program committee in late November when the entire matter of discussants will be
settled.
Carl
∂09-Aug-77 1719 REF
updated proposal on your table. Want me to do anything else to it? bob
∂09-Aug-77 1901 FTP:MRC at MIT-AI (Mark Crispin) What's up
Date: 9 AUG 1977 2159-EDT
From: MRC at MIT-AI (Mark Crispin)
Subject: What's up
To: JMC at SU-AI
the movers have gotten all my stuff here and i hunted and rescued from the
mess the stuff i've been waiting for: my lcs network implementation notes,
my arpanet protocol handbook, and my tenex and twenex documentation. i haven't
even started to unpack all the other stuff yet and probably i will just leave
it that way until i have some place to put it all.
i read rem's stuff on pcnet. i think they are making some rather bad decisions
just to allow microprocessors and ibm machines to win. it doesn't make all
that much sense to me, since on the one hand they are doing things like sending
atsigns and left brackets as protocol messages, and on the other hand, they
are trying to support ibm machines in the easiest possible way with halfduplex
line turnaround considerations. another thing i do not like is the radix 41
representation; it will require mapping on everybody's part and since the
world is going to get a lot faster as far as modems go i don't see what the
point is. of course they are talking about 110 and 300 baud connections, but
that technology is so passe'.
i got from dave moon at mit some of his description of the mit ai lab's chaos
net. i glanced through it, and i personally think we can steal a lot of ideas
from them; in fact probably more from them than from anybody else, except maybe
the arpanet. i talked to dave while he was here (he was one of my collegues
when i was working as a systems programmer at mit) and he is going to try to
keep me up to date on chaosnet and maybe we can have a mutual stealing of
ideas.
i am on the mailing list right now for lcs net notes. ken pogran is really
into that right now. if you like, i can talk to the girl handling this and
get you on the list too. this is the network encompassing the two multix
machines, the archetecture machine, the three lcs pdp-10's (mc, ml, and dm),
several intelligent terminals, etc. from what i heard, they were hoping to
have the three pdp-10's on the lcs network by october; or at least so they
said at the network forum i attended back last april.
chaosnet is ai's project, and includes the ai pdp-10 (and thus indirectly the
pdp-6, but the 6 is hardly ever used any more), the pdp-11's (also including
the xgp and tv pdp-11, but as i remember the exact relationship between that
and the current 10/11 interface wasn't decided; the chaosnet paper is sort of
obsolete), the chess machine, and the lisp machines. there is also some plan
to have a connection to the lcs network; due to political reasons it appears
that it will be a network-network interface rather than the pdp-10 being on the
lcs network. i believe that the lcs pdp-10's are going to be on chaosnet
however.
the chaosnet paper is on CHAOS.NET[DLN,MRC]. i noticed the project name
conflict between my DLN and your DIA; one of us evidently did not know of the
other's choice. i can rename my directory to DIA,MRC if you like. getting
back to the paper, i suggest you use fix20 as a font to print it since the
pages are set up for fix25 on MIT's XGP which appears to have more lines per
page.
i have sent a note to the bbn people about getting a copy of the IMP-host
protocols. i think though that since bbn is normally catatonic about giving
out things easily to people, we'll probably have to go to the ntis. jeff
has a copy of this, but since i will want to hog it and make notes in it and
all it is better to get me my own copy.
mark
∂09-Aug-77 2329 TVR ZORKing
To: DON, JMC
I haven't been able to figure out what to CALL the engravings, since
it doesn't know 'engravings' as a word.
Has anyone figured out what to do with the recepticle? It seems to hold
small things, like keys, garlic, screwdriver, etc. Also, there now are
two ledges visible from the ledge off the Egyptian room. --- Tovar
∂10-Aug-77 0328 DON Blast!
To: DUNGEON at MIT-DMS
CC: JMC
Blast it all! (a) The word is spelled "smithereens". (b) I tried putting
the setup into the basket in the mine, and lowering the basket in my one
turn between lighting and exploding. It still blew me up! Grumble!
∂10-Aug-77 0431 DON input files
I assume you've encountered the difficulty involved in getting the
necklace using prepared input files, namely, you have to keep
entering the Round Room until right exit is taken, and when the
wrong exit is taken you must decide which of several files to use
to get back. Similar problems when trying to get back to the
Living Room with the necklace. I've overcome these problems to
some extent, and am interested in pooling our efforts. That is
to say, I've got something which works, but it might be made to
work better. Interested?
I also tried lowering with the same result. Yes, let's pool.
∂10-Aug-77 0844 REM via AMET PCNet
An annotated version of the first
(and current) version of my proposal
for installing PCNet node on SU-AI,
exists as PCNET.WHY[DLN,MRC] -- the
comments are 95% negative, but then
you and LES must weigh the pros and
cons of each use of SU-AI.
I believe that cutting SU-AI
off from newly-forming networks like
PCNet will tend to isolate SU-AI to
its little community of Arpanet sites
that will be increasingly backward
and narrowminded as the new netorks
gain in coverage. The mainproblem is
if ARPA says SU-AI is forbidden from
accepting even local mail from non-arpa
networks, in which case you're better
with Arpa-alone than with PCNet-alone.
∂10-Aug-77 2050 DON
My basic method for getting the necklace is not going to win any awards for
efficiency, but it does minimise the amount of thinking required on my
part. Basically, I've constructed an input file which will take you from
any of the 6 rooms off of the Round Room (not counting the Old Engravings
room) back to the Round Room. All 6 paths enter the Round Room at the same
time, of course, after which I just go OUT again. Thus, all I have to do
is give it the first input file, which collects almost everything, enters
the RR, and goes out, then, if I'm not at the Engravings, I give it the
second input file until I am. The third input file grabs the necklace and
goes back through the RR. If I find myself back at the Engravings, I go N
and OUT until I'm not, then I give it my fourth file, while takes me from
any of the other 6 locations back to the Living Room and dumps my stuff.
I also have an E macro which will automatically split apart my input files
(which I keep on a single page of my ZORK file) into the 4 separate files
necessary for the above.
What you can contribute to this scheme is the effort required to find/prove
the minimum number of turns necessary to do the path-merging. I've got it
down to 17 turns to get back into the RR (plus 1 to go OUT again), and 21
to get to the Living Room. I'd appreciate it if you could trim these down
a bit. (It's an interesting variation on shortest-path problems; I haven't
decided whether there is a simple linear algorithm for it.)
If you want to see what I've got so far (you might prefer to try it
unbiased), you can look at sections Z.2 and Z.4 in ZORK[1,DON]/5P.
∂11-Aug-77 0201 DON
I shudder to think how large n would have to be to ensure that you've got
the necklace. I think the Engravings are at 2 of the 8 exits from RR, but
if I'm wrong then even 30 attempts only gives you about a 98% chance of
success. (If Engravings ARE 1/4 of the exits, then 14 attempts gives you
about the same (98%) chance.) In any event, you can save a lot of
meandering by not using PRAY but instead using the Mirror Rooms to bring
everyone together. One other point; you overlooked the "Narrow E/W Crawl"
(one of the exits from RR), from which your initial E takes you back into
RR. This obviously won't do. Some of your other paths are also flawed;
e.g. (E W) from Grail does not leave you at Grail; the W does not invert
the E. All in all, I have a feeling that, to get a sufficiently high
probability of success, you would have to do an outrageously long sequence.
I may be wrong, though. But I don't have the time to look into it now,
either.
∂11-Aug-77 0403 FTP:RWG at MIT-MC (Bill Gosper)
Date: 11 AUG 1977 0703-EDT
From: RWG at MIT-MC (Bill Gosper)
To: jmc at SU-AI
you could borrow hellman's
copy of schroeppel's letter and make us some copies.
∂11-Aug-77 0444 FTP:PRATT at MIT-AI (Vaughan Pratt)
Date: 11 AUG 1977 0744-EDT
From: PRATT at MIT-AI (Vaughan Pratt)
To: jmc at SU-AI
Will fix typos when I get the proofs. (The publishers have the MS.)
Did I send you hard copy?
∂11-Aug-77 0447 RPG New MACLSP
To: MACLSP.DIS[AID,RPG]:;
There is a new MACLSP running at SAIL, which has fancy
high segment capabilities such as
1. High segment stored on a separate file
2. Ability to fasload into the hiseg
3. Ability to create pure list space in hiseg
4. Ability to clobber uuolinks (for debuggin)
5. Ability to save high segments anywhere (and get them back under
all sorts of adverse conditions)
The documentation for the above will be available next week. If you
want to use these features, contact RPG.
∂11-Aug-77 1226 REF NSF proposal bibliography -- possible change:
I received the following from Sato.
∂11-Aug-77 1216 MS
Dear REF:
Our paper "On the Model Theory of Knowledge" (McCarthy et al.) will not
appear in the 1977 IJCAI Proceedings, and instead, will appear in the
SIGART newsletter. This is because it took too much time for a letter
from Prof. Reddy, Program Chairman IJCAI-77 at CMU to Mr. Hayashi in Kyoto.
However the paper will be given by Prof. Mccarthy at the Conference.
Sincerely,
MS
Do you want the bibliography of NSF[E77,JMC] to reflect this? Or is it too
trivial to matter?
bob
Yes, cite it as given at IJCAI-77, but to appear in SIGART Newsletter.
∂12-Aug-77 1024 RWW via AI
I am sorry I forgot about bibliography. If you have not yet done it I will do it from
here
please leave me a message as soon as possible and i will work on it. thanks.
will see you in a cuple of days. People were impressed with the samefringe proofs
but we need to write some down about that. I have been writing steadily and by the
time I return I hope to have a lot for you to read. Please bring some copies of first (ie the recurs
sive representation paper) with you. I think the first order way of looking at things is
going to win. We need to talk about it when you arrive. see you soon.
rww
Filman has fixed the bibliography.
∂12-Aug-77 1058 MRC via AMET
i got this wonderful idea for protocols that makes everything
much simpler. i'll show you when i have it written down.
∂12-Aug-77 1131 DCO
Can you give me a reference for your paper on first-order logical
semantics? I would like to reference it in my abstract.
I will probably call my paper `Reasoning about Recursively Defined
Data STructures'. I toyed with list structure, LISP data structures,
shells (Boyer and Moore's terminology), and finally ended up with the above.
Will you be here on the weekend or early next week? I have a few
points I would like your opinion on.
Derek
∂12-Aug-77 1858 FTP:POSTEL at USC-ISIB Dialnet
Date: 12 AUG 1977 1747-PDT
From: POSTEL at USC-ISIB
Subject: Dialnet
To: JMC at SU-AI
cc: postel at ISIB
John:
I finally read the computer faire paper on Dialnet and while i knew about
the idea for some time this prompts me to volunteer to review any protocols
you develop and to try to make suggestion in keeping with your goals.
--jon.
-------
Thanks, I'll take you up on it when we have something. Mark Crispin,
whom you may know, is working on the protocols.
∂13-Aug-77 0054 WD
Is Susie home and awake?
∂13-Aug-77 0524 FTP:CERF at USC-ISI dial-net
Date: 13 AUG 1977 0517-PDT
From: CERF at USC-ISI
Subject: dial-net
To: jmc at SU-AI
cc: cerf
John,
your summary of Dial-net just popped up from a mound of
papers on my desk so I thought I would ask about progress.
What do the protocols look like? have they been tried out on
low speed lines? Have you or your students been working on
the PCNET stuff as well?
Cheers,
Vint
-------
Vint,
We have just hired Mark Crispin to work on the Dialnet project,
and we don't have protocols yet. Our present plan is to try out the
protocols on 1200 bit/sec lines. PCNET seems to be suffering from
committee disease - overelaboration from trying to get everybody's
ideas in. There also seems to be a shortage of actual PCs.
John
∂13-Aug-77 0954 DON
∂12-Aug-77 2305 JMC discovery
I now suspect that I have discovered a bug, not a feature. Putting
the brick in a container and carrying the container somewhere else
and setting off the explosion in the container causes the explosion
to take place in the room in which the brick was placed in the container
rather than in the room where the brick is when the explosion takes
place.
[Oh well.]
∂13-Aug-77 1215 MS
Do you have time to discuss with us this afternoon?
We have moved to Escondido village, 118-F yesterday. The telephone number is
327-2832.
Masahiko Sato and Hiroakira Ono.
∂13-Aug-77 1503 PAM
|Above all else, consider a small problem in depth as opposed to a large one
|with a broad, shallow coverage.
I am focussing on understanding the goals of two players as evidenced in their
dialogue. I am not trying to synthesize plans, or to do any of the hard things
that distinguish understanding a dialogue from being able to produce it. I am
sticking to a fairly simplistic notion of the strategy of Diplomacy and of
bargaining.
|Look into Do-until something as a representation for plans
An implicit form of do-until is imbedded in the representation of likely actions
and achievement tests used in my representation. The achievement test specifies
the until portion, and the likely actions are the things to do. If there is
structure to the actions, i.e. if they embody accomplishing subgoals of the
goal, then these subgoals are supplied as goal states to be attacked seperately.
Thus any goal which includes subgoals to be achieved is attacked by using the
actions and achievement tests of the subgoals until they are achieved, and then
applying the likely actions and achievement test from the original goal. The
use of sequences of subgoals and the practice of breaking goals into simpler
subgoals both simplify the task of knowing what do next. Since my system is
designed to understand the actions of the conversants rather than to actually
converse, it has a much weaker need for information about what to do next; it
only needs enough to produce a reasonably small set of clues concerning what to
look for next.
|Simpler case of the same problem is any variable-length plan, where the steps
|are optional, conditional, etc.
This is in fact the driving force for my implementation. By using sequences
instead of sets I can indicate ordering of goals or actions, and by omitting
certain subgoals from the achievement test of a goal, I can indicate its
optionality.
|The duality of program-to-do-something and subgoals of wanting that something;
|where does it lead one? Improve X until good enough =?= achieve x by
|improving y =?= do-until.
The duality between plans and subgoals is most obvious for systems where the
planner is the only actor. The case which I am modeling is quite different;
there are at least two actors, and the goals of one actor may be sometimes be
accomplished only by actions of the other actor. Because of this "weakness" of
the actors, I represent goals by their subgoals and likely actions, but
determine whether a goal has been accomplished by applying a test on the state
of the world that is designed to be true iff the goal has been accomplished,
whether by the actor who wants it, or by the other actor.
!|Consider what the theoretical limitations on improve-until are, and also what
|the tasks it handles "naturally" are.
|Try not to get into the mire known as goal formation...
I intend to simply recognize "canned" goals, rather than discovering a player's
"true" goals in some mysterious way. That seems to be how humans impute goals to
other humans, so it's clearly good enough for me.
|Diplomacy may be too complicated a case of achieving a goal by getting help. See
|if there is some simpler domain in which similar goal structuring exists.
Diplomacy is complex only in the sense that the players can devise schemes of
arbitrary complexity to influence the behavior of other players. The game is
simple if viewed from the perspective of the actual moves on the board; other
cases of goal achievement by getting help include the dialogues collected by a
TENEX system of users linking to the operator to ask for help. Their requests
range from "please mount tape nnn" to "how are the wife and kids". The domain of
real world knowledge is so huge that I am disinclined to attempt to model such
dialogues.
During one semester as an undergrad, I worked with John Cook, a social-psych
researcher at North Carolina State, whose interests lay in studying the
interactions of players in simple non-zero-sum independent choice games. The
best known of these games is called prisoner's dilemma. The payoffs are
expressed by a matrix, reflecting the results for each player of their mutual
choices. I will illustrate with prisoner's dilemma, but I hasten to point out
that it is only one of nearly a dozen qualitatively distiguishable payoff
patterns that he looked at. If the players each choose "B", then each gets
an equal and substantial payoff. If they each chose "A", each gets an equal but
miniscule payoff. If one chooses "A" and the other "B", the "A" player gets a big
payoff and the "B" player loses money. The game was of course zero-sum, and the
payoffs are arranged so that cooperative behavior will produce the biggest
net gain for the players over the duration of the experiment.
His major interest was in metamorphic matrix games, where the players had the
opportunity to gradually alter the payoff structure of the game. They could use
their change tokens to improve their own payoffs, improve the mutual payoff of
some desirable state, or (as some did) make the payoffs worse for their
opponent. The point of studying metamorphic games was to discover what effect
the initial payoff matrices would have on the sort of matrices players would
create using their limited abilities to change the rules.
John Cook's experiments were conducted primarily in a paradigm where
consultation between the players was either not allowed or severely restricted.
I think that the ideal dialogues for me to be analyzing would be a set collected
by a computer link between a series of players pursuing various flavors of
metamorphic matrix games. The domain is simply the game and the payoff
adjustments, but the richness of the linguistic behavior between the players can
include most forms of dealing that occur in Diplomacy, without the chaff
introduced by playing a war game.
I have chosen not to use metamorphic matrix games simply because there is a
considerable amount of overhead involved in collecting the dialogues. I don't
have the time (or the substantial sums of money used to give realistic payoffs
to the players) to consider actually doing such experiments at this time. John
Cook was more interested in statistcal measures of the players' game behavior
than in the content of their discussions, so I can't lift a set a dialogues from
his work. I have Diplomacy dialogues, collected on-line from real games, and
their very availability is the strongest pragmatic argument for Diplomacy as a
domain.
!|Do not get into writing parsers... Forget the English input trap. JMC suggests
|micro-planner..... I still like LISP; KRL is too Xerox dependent. JMC's
|objection to Conniver-type languages is that he doesn't believe exotic
|communication between different program levels is a well structured way to build
|representations; when such things are used, they could be better represented by
|a different data or program structure that eliminates the need.
|I feel that with RPG's help, a useful micro-KRL can be whipped up in a month.
|Since we both need something to handle units and descriptions with a process
|agenda, this seems the best bet; provided we don't spend any time on it until we
|have a firm idea of exactly what we need.
Despite my own admonitions above, I have already spent two weeks writing a
version of KRL that runs in Maclisp. I decided that a miniature of my
proposed system would help clear my thinking about the problem and help
sell my proposal to a prospective reading committee. It did show me (as of
about the first of August) that I hadn't given enough thought to my
representational formalism. By determining a less hand-wavy representation,
I was forced to recognize and deal with some of the sloppier thinking that
was embedded in my earlier notions. I still believe that writing a working
a system is the best enforcer of rigorous thinking in this project.
The control sructure I am using is a relatively unexplored one; I feel that
using resource limited computations is at the heart of modeling human
understanding. While this is one more potentially huge pitfall, it is the
source of much of my interest in this project.
I am implementing this paradigm by using a form of cooperative timesharing
between a number of instances of a small set of processes, each applied to
a single node in a graph structure. The instances of the processes
communicate ala actors/beings, but only in a much simpler manner. They
communicate only to report success or failure of their individual problems,
and to redistribute resources based on these successes or failures. I think
this control structure can be made tractable, but I recognize that it
contains the germ of a potentially explosive set of problems.
|Do not get into parallel actions; people can be assumed to do one thing at a
|time and that is clearly hard enough to keep me entertained.
I am expecting the players to be pursuing only one goal and its subgoals at
any time. Thus I might decide that an actor stepped into a phone booth to
make a call and then consider why he might want to make a call; I would not
be able to pursue simultaneously the notion that he might be simply
escaping a rainstorm. This seems to be a reasonable restriction.
!|Consider what would be different in a Diplomacy dialogue if there were
|enforcable deals. Would people still do the same things; would dishonesty sneak
|in as weasel worded deals? Suppose we could enforce the motivation rather than
|actions of a player?
In the first case of including enforcable deals, we can consider two
possibilties based on the form of enforcement chosen. If we establish some
standard pnishment for lying, then the behavior of the players could be
expected to similar to the behavior of police when they are threatened with
punishment for obtaining evidence by illegal means. A number of subterfuges
are possible, but the primary instance of cheating seems likely to be cases
where the payoff for the crime exceeds the penalty. Thus the player who can
see that he has more to gain by being dishonest will choose to do so, just
as police tend to be overzealous if the penalty seems minor compared to the
possible conviction.
The second possible means of enforcement parallels the law concerning
illegally obtained evidence. The punishment for cheating could be made to
be so severe as to overwhemany possible gain from it, effectively insuring
that it is never "worth it" to cheat and be caught. This is the same sort
of reasoning that specifies dropping charges as the punishment for police
gaining evidence illegally. In this case, there would probably be no
dishonest deals in the sense that players would no longer promise things
that they did not intend to deliver.
In either of these situations, the players could still do the snooping form
of dishonest dealing. Snooping consists of discussing possible deals, not
with the intent of reaching an agreement, but rather simply to gain
information about the other player's plans.
The most interesting plan for removing dishonesty would be to alter the
motivation of the players. If the rules were ammended to allow two (or
more) players to form an unbreakable partnership, then the motivation of
the members of such a partnership should be purely cooperative. The
permanence of the partnership is only one way to reduce the tendency of
partners to arrange things so as to gain an advantage in the event of a
split up. Arranging to divide the players mutually-owned tokens by some
fair (or random) procedure at the time of a split up would also insure
fully cooperative behavior within a partnership.
Such an arrangement would take away much of the structure that currently
exists in Diplomacy dialogues. The partners would no longer be looking for
a deal which gave roughly equal benefits to each player, but rather would
be seeking a mutual decision on the deployment of their jointly owned
troops to greatest mutual advantage. Modeling the dialogue of two partners
would be be more akin to model the decision process of a single player as
he considered possible strategies.
!|What can we see in plans that have the goal of altering another person's
|motivation; if for instance there were a way to make someone a partner so he
|would be motivated to help you in any way possible?
The only instance of such a plan to alter motivation in Diplomacy under the
present rules seems to occur in the case of establishing a third player as
the mutual enemy of the conversants. This situation crops up in two forms.
In the first case a player argues to another player that the third party is
such a serious threat to both of them that they must cooperate to defeat
him. This sort of thing happens often at the point where one player is
apparently becoming very powerful, either locally or globally. The second
case is an attempt by a player to split up an existing partnership between
two other players, not by offering one of them a better deal, but by
emphasizing the dangers of betrayal by one's partner. This is a fairly
complex argument since it is suggesting in essence that the player should
betray his partner now in order to avoid being himself betrayed later.
The first line of planning is common and straightforward enough to be
included as a global strategy in the model. The second is more subtle, and
I haven't yet thought of a way to properly represent it.
|What are the implications of making the other players believe that you are
|vengeful; i.e. that if you are forced to lose the game, you will do all in your
|power to hurt the person who forced you into the situation? Is this the best
|strategy for Diplomacy sessions?
The idea of using threats of vengence to enforce honest dealing is an old and
effective one in the real world. The form of vengence where any hurt is avenged
by a greater hurt is more common in the real world, but let us first consider
the special form of vengence as defined above.
This form of vengence only applies to the action that is the immediate perceived
cause of the vengeful player losing. The vengeful player must determine that he
is doomed to lose, and from that point forward play a strategy of pure vengence.
Such a strategy requires identifying the player that caused the situation to
become hopeless and then pursuing those actions by which the greatest possible
harm is done to the player who caused the inevitable loss.
Diplomacy as it stands is not very well adapted to this sort of strategy, for a
number of reasons. One obvious reason is that it is non-trivial to determine
when one's game is hopeless, that is to discover that one can no longer
realistically entertain hopes of winning. Since the determination of the winner
is based on the number of troops (or equivalently, occupied supply centers) that
a player owns, a player who can determine that he is doomed to lose has very
little ability to hurt another player. It is also important to note that in
Diplomacy there are seven players, each pursuing his own goals to the detriment
of the other players. Deciding which of the opponents was the proximal cause of
a players defeat is difficult, if not impossible. The ambiguities involved in
determining the point of inevitable loss and in recognizing the villain combine
to make this specific vengence scheme only marginally useful in Diplomacy.
There are specific instances of Diplomacy games in which the situation is right
to use this special type of venegence. If there are only three players who have
a good chance to win, one of them can threaten the others with vengence in the
event that he is forced out, and it will be a credible and useful threat. The
problem becomes harder to analyse if all three adopt this stance, since it seems
clear that one of them must be edged out first.
Consider now a different specific form of vengence that also falls under the
term's common usage. Let us define a strategy of vengence such that in any case
of betrayal, the vengeful player will work to cause more harm to the betrayer
than the betrayal did to him, effectively temporarily suspending his pursuit of
the other goals of the game until he has avenged himself.
How well would this strategy work? It draws a clear line between harm caused by
a "battle" on the board with another player, and harm caused by dishonest
dealing. As such, it does not require the vengeful player to be constantly
attacking other players to avenge their strategic moves. Neither does it
require a player to assess his chances of winning the game; he need merely be
able to unambiguously recognize betrayal.
It is trivial to recognize the simplest form of betrayal; an ally fails to
perform the moves he promised. A more subtle form of betrayal occurs when an
ally divulges one's plans in order to aid a third player in thwarting them, but
conceals his complicity by going through the promised motions. This sort of
betrayal leaves three clues available to the betrayed player; the cheater must
talk to the third party after his consultation with the betrayed player, the
third player must give some concession to the dishonest dealer to motivate the
betrayal, and the efforts of the betrayed player will be miraculously futile
against the third player. When such a patten of clues occurs more than once,
the player can safely guess that he is being betrayed.
It seems that avenging betrayal is an effective stance to avoid betrayal, but it
is irrelevent in the face of the other sorts of harm that regularly befall a
Diplomacy player. Thus the two special cases of vengence that we have looked at
fail as a general stance for Diplomacy; the first applies only to relatively
rare cases and the second is defined in such a way as to exclude all forms of
harm except betrayal.
Determining an optimum strategy for Diplomacy can be approached by at least two
methods. The easiest is to rule out strategies that suffer from easily
recognized weaknesses, and consider the less tractable ones as containing any
possibly existent optimum strategy. The other approach is to pick a likely
strategy and try to demonstrate its superiority over all others. Since these are
both informal techniques, proving one's conclusions is yet harder.
I have not discovered any strategy that even seems likely to be optimal in
Diplomacy, but I believe that some can be demonstrated to be superior to others,
and that determining an optimum strategy more exactly than that is beyond the
scope of my interests and my abilities.
!|What about the players other than the two in the session? Should I care about
|the possibilities of deals which use them? The old "I can get a better deal down
|the street" or " I'll give you more than you could get from that other player"?
It is clear that the relative strengths of the players involved in a deal affect
the resulting deal; a weaker player can ill afford to ask too much, and a
stronger player can make a variety of deals spanning the spectrum from an equal
trade to outright extortion.
What effects can we expect from the players other than the two involved in the
dialogue? The kinds of possible interactions are many. In the actual game, we
can observe players arguing an incredible variety of cases that depend on other
players. A less than exhaustive listing of these cases follows.
There are a number of general arguments based on the observed or inferred
personal characteristics of other players. "Don't deal with Russia; he'll eat
you alive" or "Never trust Italy; he burned me last turn". These sorts of
arguments are more or less independent of the board situation and rarely have
any supporting evidence other than the speaker's word to back them.
There are arguments based on the geography of the board and the current
positions of the players. "Germany and Austria are neighbors; they must ally or
fight it out." "Turkey is facing your flank; make a good deal with me or I'll
talk to him." The truth of a board-dependent assertion about other players is
at least partially observable, so these arguments tend to carry more weight.
There are arguments based on the relative strengths of players at a given time.
"You must not deal with England; he is getting to be too strong." "We must band
together against France or he will do us both in." While the global and local
strength of a player is easy enough to assess, the implications that are claimed
to follow from a player's strength are less obvious; perhaps is to someone's
advantage to ally with a stronger player.
There are arguments which rely primarily on the player's option to deal with
anyone he wants. "If you won't support me into Denmark, I know Germany will."
The strength of such an argument rests on the evidence from the board and from
the past dealing history of the game.
The other side of this coin appears when players include in their deals
restrictions against conversations with other players. "Okay, it's a deal, but
don't talk to Turkey this round." This is an obvious way to protect against
betrayal, since whether two players talk is much more readily observable than
whether they make a deal of some particular kind.
From these few examples, it is clear that the role of other players in a
dialogue can be important. It also seems clear that effects are quite complex. I
am avoiding consideration of this part of the problem for the time being. I
think that some of these interactions can be modeled as bargaining strategies,
so that there is at least a hook in the initial model to hang these ideas on as
they develop.
It is important to point out that the dialogues I have were collected in a
fairly haphazard manner; the system used to collect them merely collected the
dialogues; it did not keep any of the other sorts of information that could be
useful in understanding a Diplomacy dialogue. This extra-conversational
information could usefully include who the conversants are, the current state of
the board, the time of the conversation in terms of game rounds, and the
previous conversations by the same players. Thus the information available to an
understander program is exactly the dialogue; this forces the understander to
accept or deny the claims about other players without the help potentially
available from other sources of information.
To simulate the availability of this outside information, I can guess what it
might be and add the appropriate "facts" to the information base of the system
before presenting a particular dialogue, but the very fact that I have
manually filtered the information for the relevant portions makes this whole
approach inherently suspect. At this time, it seems that the working system will
remain relatively ignorant of the other players.
!|Be sure to discuss any and all simplifications from reality to thesis; it is
|essential to show awareness of all the real problems, and to answer the question
|of why one has been omitted. The reason had better not be that I didn't think of
|it, but rather that it was too hard or too peripheral for the following reasons.
|The program may implement only a fraction of the area that the paper explores,
|but this is a proper way to do research that is not "dispensable" research.
The previous question is an outstanding example of an area of the problem that I
know in advance that the program will handle poorly. I am aware of the dangers
inherent in getting so wrapped up in an implementation that the implemented
portions of the problem tend to become "the problem." The need to show progress
on the implementation cannot be allowed to stop all efort on those portions of
the problem that are not being implemented.
|Suppose A likes/dislikes B. Can we consider this as a possible motivation? If
|so, what about the motivation of B to get A to like him? What about the
|motivation of A to get B to do something to try to win his favor? etc....
This is another area of the representation that is beyond the scope of my intial
ideas for an implementation. I am representing some of these phenomena in a more
game related way by using the notion that accepting a deal with unequal
immediate payoffs creates a debt of a favor between the players. The fact that a
player owes a favor can be motivation to help the creditor of that favor. While
this simplification clearly leaves out many of the possible forms of "likes"
motivation, it is closer to the game and seems much more tractable.
A crude form of the "dislikes" motivation can be included as a response to a
previous harm caused by another player. By removing the emotional component, I
can represent the phenomenon in terms of Diplomacy actions; I lose some of the
richness of the more emotional truth of the matter, but I gain a significant
simplification.
Where do these simplifications lead to mistaken conclusions? Since the degree
one is willing to help on the basis of "like" or hurt on the basis of "dislike"
may be quite different from the actual "debt" in terms of Diplomacy moves, there
can be cases where a human player would behave quite differently from the model.
Since this variation could vary in either direction from the model
"predictions", no simple fix short of actually modelling some of the emotions
will improve the model.
|The CMU voice gang has a production package and Dendral has some sort of prod
|system either of which might be profitably cannibalized to save some time in
|building the whole system.
One of my major interests in this thesis is actually modelling a resource
limited computation. In order to do so, I am compelled to use some formalism
which allows explicit accounting of the resources used and dynamic reallocation
of them. I can build such a system using cooperative time sharing between
proceses of a KRL agenda style scheduler. I do not think that this can be
implemented as naturally on other existing systems, so I am forging ahead with
the KRL approach.
Despite the differences in style between production systems and the KRL scheme
of things, I think that some important formal equivalences exist. One of the
things I want to do is to consider the theoretically equivalent processing
paradigms that could be applied to resource limited calculations. This seems
like a good thing to postpone until I have my thoughts in better order on the
system I am actually building, but a bad thing to postpone forever.
∂13-Aug-77 2011 RWW via AI mtc
we telnetted first.new here. there is considerable interest. I have spent a good part of
today talking about it and people want to see it. If you don't want it distributed in its present form
let me know. If you do please bring a pubed copy for xeroxing.
rww
I'll bring a copy on Tuesday.
Could the following have been intended for you?
∂14-Aug-77 1111 100 : Cuthbert Breakfast
Cheese blintzes with Elizabeth.
∂14-Aug-77 1742 PAM
To: winograd at PARC-MAXC, JMC, CCG
Reading Committee
I am interested in having you on my reading committee. I am in the
process of preparing both a formal proposal for my research and a
working miniature of my thesis program. The abstract below is included
to give you an overview of the project. The file PROGRE.SS[RES,PAM] is
a running account of what I am doing. I am not asking for close
supervision, as I have determined that this research is something I
must do largely on my own initiative. I have a formal deadline of
October 1 from the department; by that date I need a reading committee
and an accepted proposal. I am working for the next two weeks on my
miniature system, and at that time switching my efforts to composing
the formal proposal. I would like to talk to you about this whole
project, and the purpose of this note is to arrange scheduling of such
talks without crowding my deadline. Please read the material
referenced in the progress file and get in touch with me.
Abstract
A resource limited understander for Diplomacy dialogues is proposed. The
dialogues are two person bargaining sessions collected from computer
mediated games of Diplomacy. The dialogues are presented to the system
already parsed to the level of speech acts. The understander determines the
goals being pursued by each player, using a graph of connected goals and
actions. The goals and actions cover three related domains: conversation,
bargaining, and Diplomacy. Resource limited computations are associated
with each recognized action by a player and with each goal currently
imputed to a player. These processes run in a cooperative timesharing
environment, attempting to establish a complete connection from high level
goals (e.g.win the game) to the low level observable linguistic behavior of
the players. The allocation of computational resources among the processes
is determined by evaluation of the degree of local success so far and the
global importance of an answer from a given process. The resources are
measured in terms of search effort expended to improve the fit of a match
between a goal and an action. The system uses the description and
scheduling machinery of nano-KRL, a subset of Xerox PARC KRL-0 implemented
in MacLisp at SAIL.
∂15-Aug-77 1716 JBR
You have exceeded your disk quota.
The files listed below have been purged to reduce your disk
area to your quota of 2000
Before purging, your files occupied 2529
BACKUP.TMP[F76,JMC]
PPSAV.TMP[F76,JMC]
BACKUP.TMP[W77,JMC]
QQPUB.RPG[W76,JMC]
QQPUB.RPG[F75,JMC]
QQPUB.RPG[W77,JMC]
QQPUB.RPG[F76,JMC]
QQPUB.RPG[S77,JMC]
SIGART.LST[LET,JMC]
ENERGY.LST[LET,JMC]
GTREE.LST[206,JMC]
LEADER.LST[LET,JMC]
AIQUAL.LST[W77,JMC]
NEWELL.LST[LET,JMC]
DIALNE.LST[DIA,JMC]
CRYPT.DMP[ 2,JMC]
CODE.DMP[ 2,JMC]
SOURC2.LAP[ 1,JMC]
2P.LAP[W77,JMC]
REPRES.PRO[ 1,JMC]
FUNS[ 1,JMC]
EPIS[ 1,JMC]
P1[ 1,JMC]
PART[ 1,JMC]
PERM[ 1,JMC]
PATH2[ 1,JMC]
PATH[ 1,JMC]
ANTIN[ 1,JMC]
SYLL[ 1,JMC]
MEET[ 1,JMC]
COMPIL[ 1,JMC]
TIMES[ 1,JMC]
TESTA.SAI[ 1,JMC]
TESTC.SAI[ 1,JMC]
TESTB.SAI[ 1,JMC]
TESTD.SAI[ 1,JMC]
DADDA[ 1,JMC]
ORDIN[ 1,JMC]
PROB1[ 1,JMC]
ROTAT.SAI[ 1,JMC]
ROTA.SAI[ 1,JMC]
ROTB.SAI[ 1,JMC]
ROTC.SAI[ 1,JMC]
SEMAA[ 1,JMC]
SEMAB[ 1,JMC]
INTER[ 1,JMC]
CONVER.SAI[ 1,JMC]
RELREP[ 1,JMC]
MC[245,JMC]
FWGC[245,JMC]
PATH[245,JMC]
PATH2[245,JMC]
DEMO[ 1,JMC]
LISPAD[ 1,JMC]
TESTE.SAI[ 1,JMC]
ICOM[ 1,JMC]
TRANS1[ 1,JMC]
SIMUL[ 1,JMC]
COMPU2[ 1,JMC]
SOURCE[ 1,JMC]
SOURC2[ 1,JMC]
PUZZ.SAI[ 1,JMC]
TCLUB[ 1,JMC]
TCLUBA[ 1,JMC]
NEWDOC[ 1,JMC]
BOARDS.SAI[ 1,JMC]
REFER.ENC[225,JMC]
OUTLIN[206,JMC]
LIB.RLS[206,JMC]
ECHO.FAI[206,JMC]
MCCRAC.LET[ 1,JMC]
PATHS.RLS[225,JMC]
SYMFUN.RLS[206,JMC]
PARKER[ 1,JMC]
GRPDAT.RLS[225,JMC]
GRPALG.RLS[225,JMC]
GROPER.RLS[225,JMC]
DIMEN.RLS[225,JMC]
CHAR.RLS[225,JMC]
PERMU2.RLS[206,JMC]
LARRY2[ 1,JMC]
COMPU2.LET[ 1,JMC]
DEG5.IN[225,JMC]
REPRES.RLS[225,JMC]
S4.REP[225,JMC]
DEG6.IN[225,JMC]
R42.IN[225,JMC]
PUZZ.SAI[225,JMC]
PUZZA.SAI[225,JMC]
PANIC.SOS[225,JMC]
ANNOUN[225,JMC]
PUZZB.SAI[225,JMC]
ENCYC1.PRO[ESS,JMC]
PUZZE.SAI[225,JMC]
MINE.DIR[ AI,JMC]
PTCH2.DIR[ AI,JMC]
DRIVE.DIR[ AI,JMC]
KNOW2.AI[ESS,JMC]
ARCH.PRO[ESS,JMC]
ARTNA1.ART[ESS,JMC]
STUD[206,JMC]
RECOG.LET[ 1,JMC]
MEMMTC.QUA[ESS,JMC]
LANG1.AI[ESS,JMC]
MTC71.QUA[ESS,JMC]
S5.QUA[ESS,JMC]
R1301.ART[ESS,JMC]
R1303.ART[ESS,JMC]
N30[ 1,JMC]
QA.PRO[ESS,JMC]
ARTNAT.ART[ESS,JMC]
SEMAN.AI[ESS,JMC]
FRANK.STA[ESS,JMC]
COMPUT.STA[ESS,JMC]
MTCSYL.QUA[ESS,JMC]
AISY69.QUA[ESS,JMC]
FILEC.FAI[ESS,JMC]
LETTVI.REV[ESS,JMC]
CH4A[206,JMC]
R1300.ART[ESS,JMC]
LCFMEM.MTC[ESS,JMC]
PEARL.NOT[ESS,JMC]
PUZZ.RLS[226,JMC]
PUZB.SAI[226,JMC]
PUZZ.SAI[226,JMC]
TRANS.DOC[226,JMC]
BLISET.PRF[226,JMC]
BLISET.COM[226,JMC]
IMPROV.DOC[226,JMC]
SETRUL.DOC[226,JMC]
PCHECK.M2[226,JMC]
COND.PRF[226,JMC]
BACKSL[ 1,JMC]
SYSTEM.HAC[226,JMC]
SORTS.DOC[226,JMC]
TH1.PRF[226,JMC]
TH1.COM[226,JMC]
PCHECK.M1[226,JMC]
ADD.COM[226,JMC]
ADD.PRF[226,JMC]
NETJO2.PRO[ESS,JMC]
FALSE.PRF[226,JMC]
DI[226,JMC]
SCWORL.AX2[226,JMC]
ONE.PRF[226,JMC]
IPT73.REP[ESS,JMC]
NSFX.9[ESS,JMC]
NSFX.2[ESS,JMC]
NSFX.6[ESS,JMC]
NSFX.5[ESS,JMC]
NSFX.7[ESS,JMC]
NSFX.8[ESS,JMC]
NSFX.3[ESS,JMC]
NSFX.4[ESS,JMC]
HAYES.FRM[ESS,JMC]
ADTEL2.NOT[ESS,JMC]
PLATT.NOT[ESS,JMC]
REPRES.BIB[226,JMC]
ONTO.NOT[226,JMC]
TASK1.MAR[ESS,JMC]
TASK2.MAR[ESS,JMC]
LETTER.HE2[ESS,JMC]
XLET[ESS,JMC]
ZF.AX[ESS,JMC]
FOO73A[ESS,JMC]
FOON.PR1[ESS,JMC]
ROTD.SAI[ 1,JMC]
ROTE.SAI[ 1,JMC]
ROTF.SAI[ 1,JMC]
SWOPSI.RE[ESS,JMC]
HART.ME1[ESS,JMC]
GGGG[ESS,JMC]
PRO[ESS,JMC]
OUT2[206,JMC]
OUTPUT[206,JMC]
SHUFF2.RLS[206,JMC]
SORT.RLS[206,JMC]
SORTIT.RLS[206,JMC]
M29AUG.MES[ESS,JMC]
SYMPOS[ESS,JMC]
CH1.BH2[ESS,JMC]
CLOCK.FAI[ESS,JMC]
FONTS[ESS,JMC]
TASKS.206[206,JMC]
SLIDES[ESS,JMC]
FACMIN.MEM[ 1,JMC]
W74.PLN[226,JMC]
CLASS.REG[226,JMC]
GAME.CHK[226,JMC]
FOO1[ESS,JMC]
FOL2.MEM[226,JMC]
FOL1.MEM[226,JMC]
WUTHER.QUE[ESS,JMC]
ROY[LET,JMC]
FORDH.LET[LET,JMC]
TOLLES.LET[LET,JMC]
AD.NET[ESS,JMC]
KNUTH.SAI[225,JMC]
NEWRUL.DOC[226,JMC]
COMPUT[ 1,JMC]
TOLLES.LE1[LET,JMC]
TUCKER.LET[LET,JMC]
TUCKER.LE2[LET,JMC]
TCLUB.MEM[ESS,JMC]
DRIVE[ AI,JMC]
AIBIB[ AI,JMC]
MINE[ AI,JMC]
NOTES2[ AI,JMC]
PUTCH3[ AI,JMC]
RUSS[ AI,JMC]
CHYLIN.LE1[ AI,JMC]
GEN1[ AI,JMC]
DRIVE3[ AI,JMC]
MPRO[ AI,JMC]
PUTCH[ AI,JMC]
PUTCH.DIR[ AI,JMC]
BOWDEN.LE1[LET,JMC]
SLAMEC.LE2[LET,JMC]
MATCH.RLS[206,JMC]
CHERN.LE2[LET,JMC]
III[206,JMC]
SIMP.PRO[206,JMC]
CARTES.RLS[206,JMC]
FUNS.MLS[206,JMC]
BASIC.OLD[206,JMC]
MAY.ME[LET,JMC]
NSFX.10[ESS,JMC]
MC.AI[ESS,JMC]
HENRY[ESS,JMC]
BIRKHO.LE1[LET,JMC]
GAME.OLD[206,JMC]
ERIK.LE6[LET,JMC]
FLANAG.LE1[LET,JMC]
BARAB[ 1,JMC]
TAMBOV[ 1,JMC]
HARPER.LE1[LET,JMC]
PUZZE.F4[225,JMC]
BAJCSY.LE1[LET,JMC]
ERIK.LE7[LET,JMC]
STARKM.LE1[LET,JMC]
BRIT.LET[LET,JMC]
MT[ 1,JMC]
IRANI.LE1[LET,JMC]
BOWEN.LE1[LET,JMC]
LCOM5.RLS[206,JMC]
JAF.STA[ESS,JMC]
RECUR[ESS,JMC]
TERMIN.BLA[ESS,JMC]
ANDERS.RE1[ 1,JMC]
PEOPLE[226,JMC]
LCOM4.RLS[206,JMC]
OCT.ME[LET,JMC]
OLSEN.LE1[LET,JMC]
ERIK.LE0[LET,JMC]
SUBCOM.LE1[LET,JMC]
MICHIE.LE2[LET,JMC]
NEWSCI.LE1[LET,JMC]
AHLHAU.LE1[LET,JMC]
MEMOS.CUR[LET,JMC]
PUBINT.LE1[LET,JMC]
HORNIN.LE1[LET,JMC]
FELDST.LE1[LET,JMC]
CHECK.DEP[LET,JMC]
TUCKER.LE1[LET,JMC]
SCANDU.LE1[LET,JMC]
MELTZ.LE1[LET,JMC]
PIASET.LE1[LET,JMC]
MILNER.REC[LET,JMC]
FLESS.LE1[LET,JMC]
TRESSL.LE1[LET,JMC]
PIASET.LE2[LET,JMC]
LICKL.LE1[LET,JMC]
COOP.LE1[LET,JMC]
LICKL.LE2[LET,JMC]
POE.LE1[LET,JMC]
JMILLE.LE2[LET,JMC]
IGARAS.LE1[LET,JMC]
ITO.LE1[LET,JMC]
CLARK.LE1[LET,JMC]
FLESS.LE2[LET,JMC]
MICHIE.LE3[LET,JMC]
TAVARE.LE1[LET,JMC]
SCIENC.LE1[LET,JMC]
SCRIB.LE1[LET,JMC]
ATKIN.LE1[LET,JMC]
KOCHEN.LE1[LET,JMC]
GREER.FAC[LET,JMC]
CU.LE1[LET,JMC]
RALSTO.LE2[LET,JMC]
MESSAG.SCI[LET,JMC]
MCCRAC.LE2[LET,JMC]
RALSTO.LE3[LET,JMC]
CHRON.LE1[LET,JMC]
LOGIC.QUA[LET,JMC]
SHIGER.LE1[LET,JMC]
PAMELA.LE1[LET,JMC]
TAUGNE.LE1[LET,JMC]
MORGAN.LE2[LET,JMC]
FROOK.LE1[LET,JMC]
LOMONA.LE1[LET,JMC]
APR.ME[LET,JMC]
LOGIC.PUB[LET,JMC]
CONSUL.LE1[LET,JMC]
VERE.LE1[LET,JMC]
PAWLAK.LE1[LET,JMC]
TSUDA.LE1[LET,JMC]
NOLL.LE1[LET,JMC]
ROBERT.LE1[LET,JMC]
RILI.LE1[LET,JMC]
RALSTO.LE1[LET,JMC]
MORGAN.LE1[LET,JMC]
MIYA.LE1[LET,JMC]
HEALTH.LE1[LET,JMC]
WILKS.LE1[LET,JMC]
LAST.LE1[LET,JMC]
SEDEL.LE1[LET,JMC]
CAMPB.LE1[LET,JMC]
ARDEN.LE1[LET,JMC]
BURNS.LE1[LET,JMC]
SHOCK.LE1[LET,JMC]
CU.LE2[LET,JMC]
DAVID.LE1[LET,JMC]
SCHOEL.LE1[LET,JMC]
CALHOU.LE1[LET,JMC]
FINCH.LE1[LET,JMC]
LYKOS.LE1[LET,JMC]
JMILLE.LE1[LET,JMC]
MITCH.LE1[LET,JMC]
JERRY.LE1[LET,JMC]
MCQUIL.LE1[LET,JMC]
PAM.LE1[LET,JMC]
DRLUK.LE1[LET,JMC]
FORSHA.LE1[LET,JMC]
STORK.LE1[LET,JMC]
COMPT.LE1[LET,JMC]
R1302.ART[ESS,JMC]
ASTRO2.ART[ESS,JMC]
FRIEDM.LE1[LET,JMC]
FF[ESS,JMC]
AIRLIN.MTC[ESS,JMC]
JFOL.USE[ESS,JMC]
MEARS.LET[LET,JMC]
GRIMM.BKP[ESS,JMC]
SCIAM.LE1[LET,JMC]
NYT.LE1[LET,JMC]
LISPN[206,JMC]
LISPN[ 1,JMC]
TECH[ 1,JMC]
NOTES1[ AI,JMC]
NAVROZ.LE1[LET,JMC]
TXP1.LSP[ 1,JMC]
STARK.LE1[LET,JMC]
IGARAS.LE2[LET,JMC]
REPRES[LET,JMC]
PAT.LE1[LET,JMC]
MDAVIS.LE1[LET,JMC]
FLESS.LE0[LET,JMC]
KNOW.NOT[ AI,JMC]
GLUGG[LIT,JMC]
GLUGG2[LIT,JMC]
KNOW.ART[ AI,JMC]
JAF.LE2[LET,JMC]
KILPAT.LE1[LET,JMC]
PIERCE[ 1,JMC]
NAS.NOT[ESS,JMC]
TELLER.LE1[LET,JMC]
LARRY[ 1,JMC]
YOUNG.LE1[LET,JMC]
JAF.STA[LET,JMC]
AAAS.LE1[LET,JMC]
MEIR.LE1[LET,JMC]
PERSON[ 1,JMC]
MCCRAC.LE1[LET,JMC]
CRYPT.FAI[ 2,JMC]
MICHIE.LE1[LET,JMC]
WEISSK.LE1[LET,JMC]
ONEILL.NOT[ESS,JMC]
3D.NSF[ESS,JMC]
AAAS.PRO[ESS,JMC]
ISRAEL.ART[LET,JMC]
ISRAEL.LE1[LET,JMC]
CORBAT.LE1[LET,JMC]
CHENEY.LE1[LET,JMC]
OCTOBE.OUT[LET,JMC]
ONEILL.LE1[LET,JMC]
TES1.RLS[206,JMC]
TES1.MLS[206,JMC]
EFOO[ESS,JMC]
CALHOU.LE2[LET,JMC]
COMNEE[ 1,JMC]
COMTOP.LIS[ESS,JMC]
MICHIE.LE4[LET,JMC]
EVANS.LE1[LET,JMC]
EVANS.LE2[LET,JMC]
INFO[ 1,JMC]
LIB.PRO[ 1,JMC]
NAMES[ESS,JMC]
NET1[ 1,JMC]
NETDOC.PLN[ESS,JMC]
AIMEM2.JMC[226,JMC]
ALUMNI[ 1,JMC]
INFO.NEE[ESS,JMC]
SUFFIC.FRE[LET,JMC]
OCT.USE[ESS,JMC]
OLIVER.LE1[LET,JMC]
OBVI.RLS[206,JMC]
SOLIT.RLS[206,JMC]
SOLIT.WRU[206,JMC]
INTRO.AI[ 2,JMC]
COMMON.NS[LET,JMC]
DRELL.LE1[LET,JMC]
ADMIS.LET[LET,JMC]
SIGART.LE2[LET,JMC]
SIGART.LE1[LET,JMC]
SIGART.BUL[LET,JMC]
BOWEN.LE2[LET,JMC]
RUSSOF.NOT[ESS,JMC]
SAFE1[206,JMC]
NOTE.MJH[206,JMC]
SOLF73[206,JMC]
INVOIC[LET,JMC]
ACM.LE1[LET,JMC]
ACM.LE2[LET,JMC]
FERRIS.LE1[LET,JMC]
JOSHI.LE1[LET,JMC]
CARGO.LIT[LIT,JMC]
NYT.LE2[LET,JMC]
SOCIAL[ 1,JMC]
AILAB.AIL[ESS,JMC]
BDAY.LIT[LIT,JMC]
BLOWIN.LIT[LIT,JMC]
1421.LIT[LIT,JMC]
ALGOL.SNG[LIT,JMC]
NONTER.PRB[258,JMC]
CORREC.ABS[258,JMC]
BACKUP.MCP[258,JMC]
HOMOMO.NOT[258,JMC]
BRACEW.LE1[LET,JMC]
MCP.PR1[258,JMC]
COND.PRF[258,JMC]
PQR.PRF[258,JMC]
SCHEM.WRU[258,JMC]
GREEK.JMC[258,JMC]
MCCUS.PRO[258,JMC]
FOLSCO.MEM[258,JMC]
EXTFOR.MEM[258,JMC]
ACTOR0[258,JMC]
FEMANO.LE1[LET,JMC]
NAUR.LE1[LET,JMC]
MTC.GRP[ESS,JMC]
PETERS.LE1[LET,JMC]
JAN.ME[LET,JMC]
LEGUIN.CRI[LIT,JMC]
SCIFI.CRI[LIT,JMC]
JAN14.DBX[258,JMC]
SUM.CAR[ 1,JMC]
ESSEX.LE1[LET,JMC]
GELLMA.LE1[LET,JMC]
SHELDO.NS[LET,JMC]
FOLMAN.MOD[258,JMC]
APPEND.CON[258,JMC]
LOOKUP.SPE[ 1,JMC]
HOBBS.LE1[LET,JMC]
BREMER.LE1[LET,JMC]
CONSUL.POX[ESS,JMC]
AILET[LET,JMC]
REYNOL.LE1[LET,JMC]
ACM.LE3[LET,JMC]
LEAST.CM1[258,JMC]
NORBYE.LE1[LET,JMC]
JAN.OUT[LET,JMC]
DEC74.IN[LET,JMC]
HAND.2[258,JMC]
JO.NOT[258,JMC]
MTC.BLA[258,JMC]
LOGIC.QUA[258,JMC]
OUTLIN.256[258,JMC]
MANNA.PR2[258,JMC]
MANNA.CM2[258,JMC]
HARDPR.OOF[258,JMC]
OKAMOT.LE1[LET,JMC]
MANNA.WRU[258,JMC]
DEMO.PR1[258,JMC]
HAND3.W75[258,JMC]
HAND.W75[258,JMC]
HAND1[206,JMC]
UNSOLV[206,JMC]
HAND.73[206,JMC]
EXX.PUB[206,JMC]
CH10[206,JMC]
MKMAP.RLS[206,JMC]
PROJEC.WRU[206,JMC]
VIS.CAP[ 2,JMC]
NNN[ 2,JMC]
TTY[206,JMC]
SYNTAX.PUB[206,JMC]
VITASK[ 2,JMC]
CHRI.LE1[ 2,JMC]
VIS[ 2,JMC]
PERM.EX[206,JMC]
VIS1[ 2,JMC]
FRAGA.LE1[LET,JMC]
GORDON.LE2[LET,JMC]
FIELDS.ME2[LET,JMC]
ENERGY.LE1[LET,JMC]
GORDON.LE1[LET,JMC]
FEB.ME[LET,JMC]
TASK[LIT,JMC]
DEC.ME[LET,JMC]
ADHOC.THE[258,JMC]
SAKAI.LE1[LET,JMC]
COMPLR.PRO[ WD,JMC]
LICK.ME1[LET,JMC]
LICK[ESS,JMC]
LICK.LE1[LET,JMC]
PARTIT.LSP[ 1,JMC]
S4.PAR[206,JMC]
HAND.1[258,JMC]
PERMUT.RLS[206,JMC]
GAME.WRU[206,JMC]
BIBLE[245,JMC]
USER[245,JMC]
IDEAS[225,JMC]
SUZUKI.REC[LET,JMC]
SUSIE[ WD,JMC]
SAKAI.LE2[LET,JMC]
MUFTIC.LE1[LET,JMC]
VOR1[ 1,JMC]
INDUC.AX[258,JMC]
INTEGE.AX[258,JMC]
COMPIL.AX[258,JMC]
ADMIND.AX[226,JMC]
SCOTT.AX[226,JMC]
SET3.AX[226,JMC]
SET2.AX[226,JMC]
BLISET.AX[226,JMC]
ZF.AX[258,JMC]
EQUAL.AX[258,JMC]
EXTFOR.AX[258,JMC]
COND.AX[226,JMC]
IGR.AX[226,JMC]
SCINT.AX[226,JMC]
ORD1.AX[226,JMC]
LIST.AX[226,JMC]
MCP.AX[258,JMC]
MCP2.AX[258,JMC]
MCP3.AX[258,JMC]
REST.AX[226,JMC]
TERMS.AX[226,JMC]
SET.AX[226,JMC]
INT.AX[226,JMC]
SCWORL.AX[226,JMC]
INTEG2.AX[258,JMC]
COND.AX[258,JMC]
W0.AX[226,JMC]
EQUA.AX[226,JMC]
ORD2.AX[226,JMC]
SEQ.AX[226,JMC]
LISP.AX[226,JMC]
MRHUG.QUE[226,JMC]
WEINER.LE1[LET,JMC]
WEINER.GRP[LET,JMC]
WEINER.ME1[LET,JMC]
ASHLEY.LE1[LET,JMC]
74DEC1.AJT[LET,JMC]
WEINER.ME2[LET,JMC]
SMITH.LE1[LET,JMC]
GAREY.LE1[LET,JMC]
BERG.LE2[LET,JMC]
KOREA.2[S75,JMC]
KOREA.POL[S75,JMC]
ENGINE.LE1[LET,JMC]
RUSSIA.LE1[LET,JMC]
GELFAN.LE1[LET,JMC]
MT.PRO[ 1,JMC]
APR75.OUT[MSG,JMC]
MAY75.OUT[MSG,JMC]
JUN75.OUT[MSG,JMC]
JUL75.OUT[MSG,JMC]
AUG75.OUT[MSG,JMC]
SEP75.OUT[MSG,JMC]
JAN75.OUT[MSG,JMC]
DEC74.OUT[MSG,JMC]
MAR75.OUT[MSG,JMC]
∂16-Aug-77 0105 MRC
Using SAIL from ITS as a display
[Well, you can use DM simulator mode to moderate success, but to
completely win, instead of doing :TELNET or :TN, use :SAIL or :SA
(or SAIL↑K or SA↑K).
Running either SA or SAIL will get you a TELNET with not only the
Datamedia display protocol set up, but also will set up bucky bit
mode and tell SAIL you are a simulmedia. In this mode, the Knight
keyboard will appear as a funny SAIL keyboard, with the following
changes: [NOTE α means CONTROL and β means META as usual]
To get [CALL] you must type β[CALL].
To get α[CALL] you must type αβ[CALL].
To get [ESCAPE] you must type β[ESCAPE]. [BREAK] does not need metizing.
To get alternate holding and unholding you use β[BACK NEXT] instead
of α[CLEAR] and α[BREAK]. If people request I will make these do
the toggle function too.
β[FORM] is the TELNET command escape; to input β[FORM] you must do
it twice. A complete list of commands is displayed by the `H'
command; the only really useful ones are:
I change the command escape character (INTERCEPT)
Q close connection and kill the TELNET (QUIT)
∂16-Aug-77 0919 FTP:BOBROW at PARC-MAXC Panel Time
Mail from host Maxc2 received at 16-AUG-77 0919-PDT
Date: 16 AUG 1977 0918-PDT
From: BOBROW at PARC-MAXC
Subject: Panel Time
To: Jmc at SAIL
According to the program I received, our panel session is Wednesday
Aug 24 at 1:30PM-3:30. Locations were not given so you'll have to find out there.
danny
-------
∂16-Aug-77 2217 FTP:PAM at SU-AI (Paul Martin) mail file
Date: 16 Aug 1977 2217-PDT
From: PAM at SU-AI (Paul Martin)
Subject: mail file
To: jmc at MIT-AI
John .. I opened your mail file to remove a huge file that I had sent you
(the file JMC on [res,pam]. I switched to the index page to find it, found
a huge number of messages, and got into snooping one about how a LISP
micro-manual should be written. I had been talking to David Shaw when he
got a phone call, and Iopened your file your file while he was on the
phone. When he got off, I went to chat and left the file open an
inordinately long time. The time though is not really the point, since
what I was doing wwas a serious breach of courtesy. While I didn't see
anything wrong with retrieving my file for more editing, it was clerly
wrong to read anything else in your mail file. All I can do is offer you
my apology, and say that I hadn't really felt like I was snooping until I
got your message. Then I felt like a thief, and I think that is the more
appropriate reaction.
-------
∂16-Aug-77 2312 PAM
New Version of questions file
The file JMC[res,pam] has a few changes from the version I mailed to
you earlier. When you have time to read it, please read the latest version
directly from [RES,PAM]. I want to apologize again for fiddling with your
mail file; I should have just sent you a message instead of editing it.
Paul
∂17-Aug-77 1432 RAK FROM HERE TO THERE....
According to a U.S atlas on the desk of the manager of our apartment
building there are 2 ferries at different points across the Mississippi
from Missouri to Kentucky.
Dick
∂18-Aug-77 1047 FTP:BOBROW at PARC-MAXC The russians have cancelled
Mail from host Maxc2 received at 18-AUG-77 1039-PDT
Date: 18 AUG 1977 1042-PDT
From: BOBROW at PARC-MAXC
Subject: The russians have cancelled
To: jmc at SAIL
I just got a call from Bill Grieb (the west coast escort)
and he says the russians have cancelled there trip for now. They may come in
Novemember. So Tuesday the 30th is free again.
Thanks
danny
-------
∂19-Aug-77 0953 FTP:RSMITH at RUTGERS-10 copies of your unpublished papers
Date: 19 Aug 1977 (Friday) 1252-Est
From: RSMITH at RUTGERS-10
Subject: copies of your unpublished papers
To: jmc at SU-AI
Do you have, on-line, versions of the followin
memoranda:
"First-Order Theories of Individual Concepts"
"Minimal Reasoning--A Way of Jumping to Conclusions"
"Ascribing Mental Qualities to Machines"
I have read two of these, but don't have the
copies at present.
Thanks,
Bob Smith
∂20-Aug-77 2040 FTP:Lots at SRI-KL LOTS Resource allocation
Date: 20 Aug 1977 2040-PDT
From: Lots at SRI-KL
Subject: LOTS Resource allocation
To: McCarthy at SAIL
Please try to get Ralph to favor doing something which will
cut big users a lot, moderate users a little, and light users not at all.
With three times the machine and a prospect of less than double the
load, The need for resource allocation schemes is less than last Spring.
If allocation schemes work too well, the result could be the loss of
the budget for use by academic users in excess of that already committed
to LOTS -- a chance to cut the computing budget by straining LOTS.
What may be easiest, most efficient and most beneficial is to simply place
relatively light restrictions on some of the heaviest users. This is
because compute cycles have been at a surfeit because of the off-balance
system configuration. Now these users would not be soaking up what
would otherwise be system overhead, but would rather be
taking them from other lesser users.
The possibility for remaining simply organized with respect to resource
allocation should not be lightly thrown away. A way to make University
administrators realize this optimizes the resource would be ideal.
-------
∂20-Aug-77 2042 FTP:Lots at SRI-KL correction
Date: 20 Aug 1977 2043-PDT
From: Lots at SRI-KL
Subject: correction
To: McCarthy at SAIL
Make that "relatively light restrictions aimed at some of the
heaviest users."
-------
I am inclined to agree with your message and would like to
know whom I am agreeing with.
∂22-Aug-77 1149 MJL CS 206 classnotes
hi John, what progress is being made with classnotes for 206 this Autumn?
Will they be sold in the Bookstore? If you can give me some idea as to
what the status is, I would appreciate it. Thanks, Moira.
Carolyn Talcott (CLT) has been helping revise them. We want to
sell them and the Maclisp manual in the bookstore. I'll
be back Friday night and review progress. By when must they
be available? Some chapters have been revised.
∂23-Aug-77 1110 FTP:MORRIS at PARC-MAXC A debate on program verification
Date: 23 AUG 1977 1108-PDT
From: MORRIS at PARC-MAXC
Subject: A debate on program verification
To: McCarthy at SAIL, Luckham at SAIL, Manna at SAIL,
To: Oppen at SAIL
Next Thursday (25th) we are staging a debate (Perlis and Lipton vs. Horning and Owicki). In order to reduce the bombast we are hoping that
members of the audience will ask some good questions of the various sides. Please come
if you can.
Jim
May the truth win. I'm still at IJCAI.
-------
∂23-Aug-77 1438 PAT PHONE CALL
TODOROVICH CALLED AND WOULD LIKE YOU TO CALL HIM BEFORE FRIDAY ...patte
∂23-Aug-77 1517 REG via AMET Replacement for Roode
I'd like to hire Mike Byron to replace Roode. Mike
graduated in June, and has been working here this summer. I think that
he is qualified, and suitable, and willing.
I want to change the specifications of the job to pay 80% time all year,
instead of the present silliness (50% for three quarters) and
100% for summer). The majority of the work occurs during the three quarters.
I'd divert money from the "student assistants" budget to pay for
the increased percentage of time worked (from 62.5% to 80.0).
If this is agreeable, i'd like to get underway with this ASAP. Or do you want
to wait and interview Byron?
I trust your judgment entirely in this matter.
On another subject, I didn't mean to suggest that the Director would not
be involved in any resource allocation decision. However, whatever is decided
will have to be done by the A-board, with guidance from as, of course.
I assume that we will not be faced with a situation where the A-board seeks
to adopt a policy that we would feel is inappropriate. I assume that
you and I will agree on what to do before we do it.
As to the need, I keep thinking that the system, even augmented as it will
be by 30 more terminals and memory and disks will be insufficient for
the 400 cs105/106 users plus all the other classroom users. Given a "free"
resource like LOTS, various people are inspired to attempt things that they
wouldN7t have doneif they'd been paying for it. This applies to class
assignments and to research problems. We should have a visible and continuous
feedback mechanism to show who the resource hogs are (This is a separate
notion from an allocation system). One more thing: the next time we want something
from Miller, he'll want to know what we've done about resource allocation.
Enjoy Boston.
∂25-Aug-77 0350 REM via AMET CRUNCHING SOFTWARE, DFTP
Thanks to DFTP allowing me to keep all my files online
(with suitable delay of access of course) I have been able to
regain organization of my many programs and other files.
In particular, tonite I have restored my UNPACK.SAI program,
written in 1971 or 1972, and modified it so it knows about the
new (1975) CRU2.FAI and CRU3.FAI=SPINDL programs. I have also
modified CRU2.FAI so it doesn't flush input unnecessarily, thus
allowing the user or the UNPACK program to type ahead at it.
Given a crunched and/or spindled file of unknown source, it is
now possible to check it to see if there is any known way to
uncrunch it, by running UNPACK[UP,REM]. If the file was crunched
with any of the known programs, it checks to be sure the uncruncher
program is on the disk and if so it sets up all the necessary
commands and swaps to it. In the case of the spindle program,
it isn't defined what you really want done so it types an illegal
command to get the spindle program to type "type ? for help",
otherwise it completely uncrunches the file and returns to the
monitor where it does a directory to show you where the original
(crunched) and uncrunched files are located (the uncrunched file
has a funny name so this is necessary) and how large they are.
Everything including the new version of CRU2 seems to work ok.
∂25-Aug-77 1619 MJL
CC: JMC, CLT, MJL
∂24-Aug-77 2127 JMC
To: MJL
CC: CLT
Carolyn Talcott (CLT) has been helping revise them. We want to
sell them and the Maclisp manual in the bookstore. I'll
be back Friday night and review progress. By when must they
be available? Some chapters have been revised.
John, We should have them over to bookstore a week before registration
which means about 3rd week of Sept. Allow another week for copying.
If all could be ready by week of Sept 12 - it would be super - but
we can always squeeze if necessary. Let me know of progress, thanks, Moira.
∂26-Aug-77 1304 FTP:Moore at SRI-KL Processing at function definition time
Date: 26 Aug 1977 1306-PDT
From: Moore at SRI-KL
Subject: Processing at function definition time
To: JMC at SU-AI
cc: BOYER
I thought I should explain to you exactly what our theorem prover
does vis a vis "understanding" the termination and recursion
of a fn at definition time. What I told you in answer to your
question after my IJCAI talk was not, at that time, literally
true. It is now. Let me explain...
As of last week, when the user of our system defined a function
the system would collect all of the recursive calls, e.g., (FOO (D1 x))
and (FOO (D2 x)), and would search its lemma base for lemmas that
implied that under some conditions, P(x), some measure, M, was going
down in all those calls, e.g., P(x) -> M(D1(x)) < M(x) and
P(x) -> M(D2(x)) < M(x). If it found such a P and M, then it stored
an "induction principle" to be associated with FOO, namely, to prove
Q((FOO X)) try proving ~P(X) -> Q((FOO X)) and P(X) & Q((FOO (D1 X)))
& Q((FOO (D2 X))) -> Q((FOO X)).
This is a sound induction principle, since it has nothing to do with
FOO. Its just an induction on the well ordering established by the
lemma that says P(x) -> M((D1 x)) < M(x).
If the system failed to find any "plausible inductions" then it printed
a warning message. Furthermore, the structure of the system
is such that if a recursive fn had no plausible inductions stored with
it, the system would infinitely recurse when trying to simplify
expressions involving the fn. This is what it would have done
for GOPHER (because it recursed on (CONS (CAAR X)(CONS (CDAR X)(CDR X))))
had we defined before proving the lemma that said the leftcount of that
CONS is smaller than that of X, under sufficient conditions on X.
Of course, just because the system finds some plausible
inductions, there is no assurance the fn was total,
since the system did not actually check to see whether
the tests performed in the definition of the fn were sufficient to
imply the conditions under which the measure decreased.
Thus, the system would not actually guarantee that FOO was
total, merely that it was plausible for it to recurse the way it did,
and that the system could imagine sufficient tests to make a fn
recurse like that.
We didn't bother to check totality, since we were willing to just assume
the user understood that all bets were off in the face of non-total fns.
Of course, checking totality is easy, given the rather elaborate
analysis we went through to find plausible measures that were going down.
Bob explained the above analysis to Richard Weyrauch on Thursday, the week
before my talk. Afterwards, Bob and I decided it would make our
explanations easier if we went ahead and made the theorem prover prove
totality during its definition time processing.
Thus, the current version of the system does the analysis above, but then
proves (using only its lemma driven simplifier) all the theorems of
the form P'(x) -> P(x), where P' is the conjunction of conditions
tested in the definition before each recursion. Of course, since the
induction analysis may find many such P's (some fns recurse on more than
one arg, and any one of them might be sufficient to insure termination),
the theorem prover checks them all and throws out any "plausible induction"
scheme that it couldn't prove accounted for the termination of the fn.
This has two nice effects. FIrst, of course, it means the system
insures totality. Second, it weeds out induction principles it
previously thought were "plausible" (for example, if you had a fn that
kept one arg constant and counted another up by 1, one of the
plausible inductions stored was the one that said you were
counting the second arg up to the first. This was stored
even if you manifestly weren't doing that, i.e., you didn't have a
test in the fn watching for when the max had been exceeded).
J
-------
∂26-Aug-77 2017 CLT chapter 5
The current state of chapter 5 exists as LSPCH5.XGP (or .PUB for tinkering).
There are a few gaps. The first couple of pages are notes to myself on what
remains to be done and will disappear eventually. I have not done much proof
reading yet, other than to look for pubbing errors.
∂27-Aug-77 1627 TVR ZORK
To: JMC, DON
I have heard a rumor that there exists a balloon.
If so then the obvious thing to try is burning the leaves in the basket.
∂27-Aug-77 2208 DON
To: TVR
CC: JMC
∂27-Aug-77 1627 TVR ZORK
To: JMC, DON
I have heard a rumor that there exists a balloon.
[Aha! Wouldja believe a basket with a cloth bag? I'll try it tonight...--DON]
Yes, it answers to the name of balloon and can be boarded.
∂27-Aug-77 2214 DON
Indeed. I had ENTER BASKET on my list of things to try, but haven't
ZORKed at all lately and so never got a chance to try it. Or doesn't
ENTER work?
No. Enter doesn't work.
∂27-Aug-77 2215 DON
Hmm. Nasty of them.
∂28-Aug-77 0106 GFF
I think you might find SCIENC.PAP[1,GFF] interesting reading, by Baran & Farber.
∂29-Aug-77 0041 MRC via AMET finger problems
To: JMC
CC: RWG, ES
I will see the doctor later today and find out what is going on. thanks
for the assistance.
i'll continue my research on other networks and how they solve their problems
so i can get something productive done. i think after a lunch date on
tuesday i will talk with ralph about modems, unless the lunch date becomes
extended (you never know with women after all,...). hopefully i'll be able
to reasonably type soon.
at least now i have a good idea of what all of the low level stuff is going
to look like; higher level protocols are going to be more of a pain since
they are less objective and more of a "taste" thing. to just get things
going though, might even use a-net protocols in simple form for high
level stuff and then branch out; we'll see.
∂29-Aug-77 0311 DON Balloon
Have you gotten anywhere with the balloon? I tried burning the leaves
in it to inflate it, with no success. Also, are you still able to
get 357 points? My input files only get 352, and I can't think of
anything I'm failing to do, unless there's some obscure location
which I didn't realise is worth 5 points for entering.
I haven't tried getting 357 points recently. Did you perhaps forget to
put something in the trophy case? I have tried burning the leaves and
the listings. My current idea, which I have not tried, is see if the
coal gas can be used, or, even more far fetched, to see if one of the
sources of water is really a source of oil. It would be cute if the
solution were to burn the brick rather than explode it.
∂29-Aug-77 1317 FTP:DCROCKER at USC-ISI Social Statistic of Computer-based Human Communication
Mail from USC-ISI rcvd at 29-Aug-77 1304-PDT
Date: 29 AUG 1977 1242-PDT
Sender: DCROCKER at USC-ISI
Subject: Social Statistic of Computer-based Human Communication
From: DCROCKER at USC-ISI
To: Header-People at MIT-MC, [ISI]<MsgGroup>Mailing.List:
Message-ID: <[USC-ISI]29-AUG-77 12:42:58.DCROCKER>
Redistributed-To: JMC at SAIL
Redistributed-By: GEOFF
Redistributed-Date: 29 Aug 1977
Last Thrusday evening, at about 5pm Pacific time, I sent a note
to approximately 130 people around the country. The response
statistics are, to me, a little scary:
Even though I sent the note at the end of the day, I had 7
responses within 90 minutes ((6 west-coasters and one
mid-westerner). Within 24 hours, I had received 28 responses, or
approximately 20 per cent of the people I polled.
This sort of two-way connectivity is incredible. Since I know of
no other hard data on the subject, I thought you might like to
have the above available for citing.
Dave.
-------
∂29-Aug-77 1446 FTP:PSZ at MIT-ML (Peter Szolovits) LISP commemorative conference (?)
Date: 29 AUG 1977 1746-EDT
From: PSZ at MIT-ML (Peter Szolovits)
Subject: LISP commemorative conference (?)
To: JMC at SU-AI
CC: PSZ at MIT-ML
Dr. Egbert Lehmann from East Berlin is visiting here at MIT
this week after the IJCAI conference. He would like to know, on behalf
of one of his colleagues, Dr. Stoyan, whether and when there will be a
conference to commemorate the creation of LISP. If you reply to me
before the end of this week, I will relay the response to Dr. Lehmann
and he to Dr. Stoyan. Thank you.
Peter Szolovits
Unfortunately, no-one seems to be organizing such a conference. There is
a conference on the history of programming languages to be held in Boston
next June and LISP is one of the languages included. Dr. Stoyan should
have received an invitation to that conference.
∂29-Aug-77 2329 DON
I'm pretty sure I put everything in the case (even the burned-out torch).
Maybe taking the listings and coke bottles causes you to LOSE points?
I get the feeling that something has to be put in the receptacle to get
the balloon aloft. Do you know enough about balloons to be able to guess
what the receptacle is? (We can also draw clues from the capacity of the
receptacle, which is between 1 and 3. If we find out exactly how much it
can hold, chances are that what goes into it is exactly that size.)
I forget whether I told you that I am thinking of trying to collect
coal gas, but don't give it much of a probability. The sack is described
as a tall brown bag, and it is non-flammable.
∂29-Aug-77 2359 DON
What has the sack to do with anything? Did you think I was referring
to that when I asked what sort of "receptacle" might be on board a
balloon? Or are you still hoping to find something else to do with
the sack? By the way, here's something which might prove interesting
and/or useful: I tried closing the basket, which is permitted, and
looked to see what things were now hidden. (Of course, due to the
way they implemented closed objects, things are hidden even if you
are inside the basket.) I got some surprising results, particularly
the last line:
>CLOSE BASKET
Closed.
>DISEM
You are on your own feet again.
>LOOK
You are at the bottom of a large dormant volcano. High above you
light may be seen entering from the cone of the volcano. The only
exit here is to the north.
There is a very large and extremely heavy wicker basket with a cloth bag
here. Inside the basket is a metal receptacle of some kind. Attached to
the basket on the outside is a piece of wire.
>BOARD
You are in the basket.
>LOOK
You are at the bottom of a large dormant volcano. High above you
light may be seen entering from the cone of the volcano. The only
exit here is to the north.
You are in the basket.
The cloth bag is draped over the the basket.
>OPEN BASKET
Opening the basket reveals a , a cloth bag, a braided wire, and a receptacle.
I wonder what the "invisible object" is?
I suppose the "invisible object" is just a grammatical goof. Had you
left an object in the basket, it would have appeared in that place.
As for the sack, it fits in the receptacle even if it contains the
brick which will not fit in the receptacle by itself. My far-fetched
idea is to collect coal gas (lighter than air) in the sack. Alas, I
can't guess what the receptacle is. I would guess a torch holder if there
were a way of not using up the torch or restoring it, but
perhaps it's a valve for gas, turnable with screwdriver or wrench.
TAKE VALVE might elicit a response if there is one.
∂30-Aug-77 0031 DON
Being able to leave things in something does not NORMALLY cause
that grammatical goof. E.g., opening the receptacle does not
cause ANY message to be printed, since the receptacle is initially
empty. Also, though the sack can hold more stuff than just the
lunch and garlic, it doesn't "leave room" for it. In a list
processing language it's absurd to think that fixed amounts of
room would be allocated. No, I'm sure there's something strange
going on there, and it's almost certainly an object with no
description. Whether we can deduce anything from this is another
question entirely.
As I've pointed out with regard to things in the boat, containers
can contain their normal capacity even when they are themselves
contained in something else, so it's no surprise that the brick
will fit in the sack which is in the receptacle which normally
wouldn't hold the brick. But I consider this more of a bug than
a feature (not that I want it fixed!), and am fairly certain that
we do not have to take advantage of it to win. So whatever is
supposed to go in the receptacle almost certainly fits of its
own accord.
∂30-Aug-77 0221 DON thief
Sounds like more trouble than it's worth (just writing a program to
interface with ZORK is probably more than I care to do at the moment!).
I'm sure it's possible to solve the thing by more realistic methods.
Can you get me a xerox of the article referred to.
∂30-Aug-77 1534 MS
I and Mr. Ono have something to say about non-knowledge. Do you have
time to discuss this afternoon? Do you, by any chance, have a copy of
the Journal of Philosophy, vol. 72, no. 19, Nov. 6, l975, pp 690-716?
This issue contains an article by Kripke
∂30-Aug-77 1629 JJK CSD Research Reports
To: MAIL.CRR[CSD,JJK]:;
Once again an edition of CSD Research Reports is being assembled.
We need a current statement of your research; a pointer to a file
is sufficient. Besides the usual collection of faculty and associate
research descriptions, a departmental brochure will be put together
which will draw on the material you submit. Thank you for your
cooperation. -- Jonathan King
p.s. Computer mail replies to JJK@SU-AI; hard copy to 214 Polya.
∂31-Aug-77 0641 FTP:EF at MIT-AI (Edward Fredkin)
Date: 31 AUG 1977 0941-EDT
From: EF at MIT-AI (Edward Fredkin)
To: jmc at SU-AI
i finaly got out by
going up the chiminy i the studioroom (i am talking about dongouns of course)
so i don't need to go to through the maze .
Michael Fredkin
Unless you know something I don't, you can't carry treasures, except the
painting, up the chimney.
∂31-Aug-77 0808 RDR
To: JMC, REG
I think LOTS' participation in optimizing use of the system could
include provision of a TA dsk in the CERAS lobby. I said
this before but I was overruled or ignored. It's still a good idea.
∂14-Aug-77 1444 FTP:Wiederhold at SUMEX-AIM Re: CS105/6 TA Consulting
Date: 14 AUG 1977 1438-PDT
From: Wiederhold at SUMEX-AIM
Subject: Re: CS105/6 TA Consulting
To: RDR at SU-AI
cc: dpb at SAIL
In response to your message sent 5 Aug 1977 1556-PDT
Ta's should have a (n attractive office) so that they will be
encouraged to be there, and a terminal with a TA acount so that
they can be expeditously reached from students working at
remote sites. The camderie of the comon keypunch and
card-reader queues may be dfficult to emulate in adistributed
system, but some compensation should be made.
I do agree that all TA's can be expected to have at least some
knowledge of other languages being used in parallel courses,
after all we hope that the problems we present are problems
challenging the intellect rather than knowledge of computer
languages. Gio
-------
∂31-Aug-77 1049 RDR constructive proposal
To limit the hogs (few though they may be) at LOTS next year and make
the resource be thereby allocated.
This idea is drawn from SUMEX, where they do the same thing,
only mess around determining in advance groups to
which users belong and each group's share of the machine.
There idea doesn't depend on this determination, which I think
is avoidable and desirable to avoid -- politics, people in more than
one group, bureaucracy, red tape, overhead, hassle.
On the oter hand, the SRI scheduler that is halfway up at LOTS
accomplishes what SUMEX does, but they DO depend on the preestablishment
of groups and group priorities.
The idea is simply to discriminate agaoinst people
who use more than a certain threshold of CPU time during
prime time at LOTS. Many parameters can be adjusted so I'll
assume some now.
The priority bulit in to the DEC supplied SCHED and PAGEM si utilized,
as opposed to being ignored. A detachd system job sets processes'
prorities to a predetermined level unless they have used their
one minute of CPU time during the current day's prime
period of 1:20pm till 5:30pm and 7:30 pm till 12:30 the next morning.
Over limit processes are set to have no priority. Thus, it takes them
longer to get anything done. At 12:30am a new day begins and the
system recycles.
It is possible that the performance of the DEC system could
be improved (as I believe they do at SUMEX) by setting processes
other than the limit-execeded ones to priority zero on a rotating basis.
The effect of this is to logically reduce the number of jobs on the system
to something the scheduler is better able to deal with.
At peak periods, the threshold of full-priority
crunching is reduced to encourage people to work
when the system is less loaded if possible,
without forcing them too. this is announced in advance. it also
mitigates the effect of this delayed crunching on othr users of the
system.
I think this would be easy to try and worth trying.
∂31-Aug-77 1107 FTP:Wiederhold at SUMEX-AIM Re: TA office space
Date: 31 AUG 1977 1107-PDT
From: Wiederhold at SUMEX-AIM
Subject: Re: TA office space
To: JMC at SU-AI
In response to your message sent 31 Aug 1977 1057-PDT
that should help. The TA's may be consulted on what they think they need
there to work effectively. Gio
-------
∂31-Aug-77 2106 FTP:MF at MIT-AI (Michael H. Fredkin)
Date: 31 AUG 1977 2339-EDT
From: MF at MIT-AI (Michael H. Fredkin)
To: jmc at SU-AI
no i dont but once you go up the chimney
the trap door stays open .
michael fredkin
∂01-Sep-77 0000 JMC*
talk to mark about extendability, parentheses
∂01-Sep-77 2052 MRC SUPDUP
Sorry, SUPDUP only works on DD's for the time being. I'll make it run on
DM's sometime; but it's sort of low priority now (or rather, it is normal
priority, but DIALnet is on high priority now).
∂02-Sep-77 1120 FTP:PBARAN at USC-ISI Computers and communications
Date: 2 SEP 1977 1119-PDT
From: PBARAN at USC-ISI
Subject: Computers and communications
To: jmc at SU-AI, farber at RAND-UNIX
cc: PBARAN
Hi John:
Always good to hear from you. I may not always agree
with everything you say, but I always enjoy your observations.
Dave Farber would like to enter the discussion as well
so I suggest that I moderate. Dave had the opening
remarks. John countered in baran.re1[e77,jmc] so it is now
Dave's turn.
I'll sit back and learn.
-------
∂02-Sep-77 1128 FTP:Farber at Rand-Unix Re: Computers and communications
Date: 2 Sep 1977 at 1126-PDT
Subject: Re: Computers and communications
From: Farber at Rand-Unix
To: PBARAN at Usc-Isi
cc: jmc at Su-Ai, Farber at Rand-Unix, PBARAN at Usc-Isi
Message-ID: <[Rand-Unix] 2-Sep-77 11:26:06.Farber>
In-Reply-To: Your Message of 2 SEP 1977 1119-PDT
I will be off net for the next week on vacation. I will take my
turn at that time.
Have a happy
Dave
∂02-Sep-77 1818 FTP:TAA at MIT-DMS Dungeon
DATE: 2 SEP 1977 2114-EDT
FROM: TAA at MIT-DMS
SENDER: TAA at MIT-DMS
SUBJECT: Dungeon
ACTION-TO: JMC at SU-AI
MESSAGE-ID: <[MIT-DMS].56829>
Sorry about the slow response; I was on vacation, and no one else
has the gall to modify the source. The actual installation of some bug
fixes may be further delayed due to compiler bugs, but that's something
else again.
1) It is indeed a bug that putting the brick in something and blowing
it up affects the old location of the brick, rather than the current
location.
2) The coke bottles and listings aren't meant to be useful. We're
entitled to some narcissism in our own tomb, certainly.
3) The thief will no longer accept bombs with burning fuses attached.
As a result, you will still blow up.
∂02-Sep-77 2341 WD
Thought this might interest you.
Whit
n064 1332 01 Sep 77
AM-AGNEW 1stadd
ANNAPOLIS: years.
* * * *
The theory of the taxpayers' class action lawsuit -upheld thus far by Judge
Williams - is that if kickbacks were paid for Maryland engineering contracts,
the added public cost of the bribes placed an extra burden on every tax paying
citizen of the state which may now be recovered from the defendants by court
order. The legal theory is that the bribes have been held by the defendants ''in
trust'' since they were paid.
* * * *
0901 1631ped
***************
∂03-Sep-77 0855 FTP:PDL at MIT-DMS The Record
DATE: 3 SEP 1977 1121-EDT
FROM: PDL at MIT-DMS
SENDER: PDL at MIT-DMS
SUBJECT: The Record
ACTION-TO: jmc at SU-AI
MESSAGE-ID: <[MIT-DMS].56844>
Sorry, but there used to be a way to kill it in one move.
Unfortunately I've forgotten what it was.
∂03-Sep-77 1201 BPM
∂03-Sep-77 0003 JMC
"Put air in X" where X is a container kills the program.
[Good grief!!!!]
∂03-Sep-77 1657 REM via AMET
You have a detached E idle more than 999 minutes, and a WHO idle 355 minutes...
∂03-Sep-77 1729 FTP:Lynch at SRI-KL Re: Crowding on your machine.
Date: 3 Sep 1977 1732-PDT
From: Lynch at SRI-KL
Subject: Re: Crowding on your machine.
To: JMC at SU-AI
cc: LYNCH
In response to your message sent 3 Sep 1977 1710-PDT
John,
Your observations are on target. Once one gets used to a decent
TV terminal, hardcopy is a drag. We are not in the class where we
can afford enough terminals to have one at home as well as one at
the office of the Datamedia type. We do have enough terminals for
most of us to have one at home and one in the office and we have not bought
a hard copy terminal in about 2 years, so we are heading
in the right direction. Of course, 1200 baud modems are still not cheap
yet! (We have bought about 30 Datamedias in the last 2 years.)
Another factor is that more than half of my current users are NLS
clients of the Government type who never get
to take home terminals -- even those who want to do so have to hurdle
"discrimination" regulations of the type as follows:
"you can't give a special resource to certain employees only because
that will enhance their careers at the expense of those employees
who have not been given the same opportunity". Yes, John, the
government has really got its lawyers thinking overtime.
Dan
-------
∂03-Sep-77 2053 REF
Sorry. I seem to be disappearing for the week. Perhaps after that.
∂03-Sep-77 2101 REF birthday
By the way, happy birthday.
bob
∂03-Sep-77 2356 MRC
Read PROTOC.OLS[DLN,MRC] for current set of bad ideas.
∂04-Sep-77 0148 MRC
I should have wished you happy birthday. Sorry, I didn't know.
∂04-Sep-77 0800 JMC*
Bring in Prior.
∂04-Sep-77 0800 JMC*
No grievances.
∂04-Sep-77 1453 FTP:Raphael at SRI-KL Re: Datamedias
Date: 4 Sep 1977 1445-PDT
From: Raphael at SRI-KL
Subject: Re: Datamedias
To: Lynch, Hart, Jones
cc: RAPHAEL, JMC at SU-AI
In response to the message sent 3 Sep 1977 1734-PDT from Lynch
How about trading some computing resources with SU? Our chronic excess of
night and weekend resources could have value to the University types,
and they may have some mass storage, or(a smaller amount of) prime time
capacity, or.... for us. I know it's impractical for administrative reasons
as well as incompatible systems, but maybe if we keep it in mind
someday we can all win..
Bert
-------
∂04-Sep-77 2233 FTP:Boyer at SRI-KL Crowding on the SRI-KL
Date: 4 Sep 1977 2233-PDT
From: Boyer at SRI-KL
Subject: Crowding on the SRI-KL
To: JMC at SU-AI
In your note to Lynch on the subject of Crowding on our machine
you suggest that the lack of reasonable home terminals is a reason
that our load is not more spread out. Until
a year ago, everyone at SRI was happily using a TI in his office. Now that
almost everyone has a Datamedia in his office, there is a plethora of
TI's available for home use. Since I personally prefer a TI for
home use, I have to believe that the availability of reasonable home
terminals is not a serious factor in the uneven loading of our system.
I have looked fairly carefully at your IJCAI paper "Epistemological
Problems of Artificial Intelligence". Through that paper and the
Bobrow chaired panel discussion I have become convinced of the
importance of developing a logic based epistemology as a foundation
of AI. I would be interested in an opportunity to participate in
a seminar or course along the lines of your paper (and the three
forthcoming referenced papers).
Incidentally, I believe there is a typographic error on page
1041 at the top of column one of your IJCAI paper. The last implication
sign in 3) should be reversed, I believe.
-------
∂06-Sep-77 2033 FTP:Raphael at SRI-KL Review
Date: 6 Sep 1977 1807-PDT
From: Raphael at SRI-KL
Subject: Review
To: JMC at SU-AI
Don't bother reviewing the Jeffrey paper ("A new paradigm for AI)-- I have
rejected it on the basis of other reviews.
Bert
-------
∂07-Sep-77 1232 MRC
There should be a non-PUB version of DIALNE someplace for lineprinter
or just normal console examination. I will generate this if you don't have the
time.
Thanks, please do it.
∂07-Sep-77 2140 DGR via NBST DIALNET
What is happening with Dialnet.....last communication from you was that you
were holding up all decisions and details until you had someone on board
to manage things from your end. Many peoiple from DECUS and NIH are
losing restraint in working out some joint protocols and there already
seem to be quite a few ad-hoc propositions and implementations even.
Have you been able to find someone yet? When can design commence?
. . . . Glenn
The design has started. Mark Crispin has been working on the project
since early in July. He is called MRC here and has already prepared
some documents. I suggest you introduce yourself by a message to MRC
here. He is logged in at the moment.
∂08-Sep-77 0258 MRC DIALnet paper LPT version
A lineprinter version of the DIALnet paper, with a 60-character
margin, now exists as DIALNE.MEM[DLN,MRC]. I have altered all
the pointers in all my files to point to this file; I haven't
altered anything in your directory.
I converted it to RUNOFF, which is the only text formatter I
know (I have this deep seated prejudice against text formatters;
they are all too hairy for my needs) and then generated the MEM
file, courtesy of Harvard (collecting computer accounts DOES have
a good use!!!) and FTP'd the whole mess back here.
It's even reasonable to read it with E!!!
∂08-Sep-77 0927 FTP: Lee Erman at CMU-10B Re: Book with invited addresses
Date: 8 Sep 1977 1224-EDT
From: Lee Erman at CMU-10B
Subject: Re: Book with invited addresses
To: JMC at SU-AI (John McCarthy)
CC: reddy@CMU-10B
Sender: LEE.ERMAN at CMU-10B
Message-ID: [CMU-10B] 8 Sep 1977 12:24:42 Lee Erman
In-Reply-To: Your message of September 5, 1977
John,
Raj tells me that he still plans on publishing a book with the
invited papers and thus requests that you do NOT submit it for publication
elsewhere.
Since Raj will be handling this (I believe), you should probably
send future inquiries about this to him.
Lee
-------
∂08-Sep-77 1342 LES Dialnet blurb & Bell modem
To: JMC
CC: MRC
On the short Dialnet blurb, if it intended for external publication, I think
the sections that use first person plural should be changed to third person.
If you would like me to hack it along those lines, give me a pointer.
I heard from the Telco modem guy, who is Jim Burnett, (56) 291-5143.
Their 1200 baud modem is called the 202S and is available, on 3 weeks
notice, at a cost of $75 to install plus $45/month. I think we would also
want the reverse channel feature ($5.50/mo.) and the 801 Calling Unit ($26/mo.).
There is no additional installation charge if these are on the initial order.
So it looks like the cost is $75 installation plus $76.50/month.
Please improve the short Dialnet blurb as you proposed. Let's wait on the
modem till we see the Vadic literature which is being mailed, but let's
plan to go ahead with one or the other promptly. We also need new
dialers at each end. Vadic sells them and Telco rents them. I want
a separate dialer for Dialnet, because the program should be a package,
and the using the other dialer would mean that the program would have
to take into account other uses of the dialer.
∂08-Sep-77 1801 FTP:Raphael at SRI-KL Ad
Date: 8 Sep 1977 1803-PDT
From: Raphael at SRI-KL
Subject: Ad
To: Bobrow at PARC, Fikes at PARC, Les at SU-AI, JMC at SU-AI
Are you or anyone you know interested in sharing a family-oriented
condiminium at Alpine Meadows for the ski season? Walking distance from
the lifts. About $400 for five months' unlimited family use. Send me
a message or call 941-0102(eves.).
Bert
-------
∂09-Sep-77 0014 DON zork
To: JMC, TVR
The following excerpts indicate something somewhat interesting about the
newly-discovered word, AIR, namely that it is considered to be BREATH.
(On the other hand, BREATH isn't in the vocabulary. Very strange.)
I also note that GRUE is a known word, with some amusing messages
associated. Some are shown below.
>TAKE AIR
Not a prayer.
>FIND AIR
I can't see that here.
>AIR
What should I do with it?
>PUSH AIR
Pushing the breath is not notably useful.
>BREATH
I don't know that word.
>HOT AIR
I don't know that word.
.
.
.
>UP
It is pitch black. You are likely to be eaten by a grue.
>WHAT GRUE
The grue is a sinister, lurking presence in the dark places of the
earth. Its favorite diet is adventurers, but its insatiable
appetite is tempered by its fear of light. No grue has ever been
seen by the light of day, and few have survived its fearsome jaws
to tell the tale.
>LIGHT LAMP
The lamp is now on.
You are in the attic. The only exit is stairs that lead down.
There is a square brick here which feels like clay.
A large coil of rope is lying in the corner.
On a table is a nasty-looking knife.
>FIND GRUE
There is no grue here, but I'm sure there is at least one lurking
in the darkness nearby. I wouldn't let my light go out if I were
you!
>EAT GRUE
I don't think that the lurking grue would agree with you.
>TAKE GRUE
Chomp, Chomp.
>PUSH GRUE
Pushing the lurking grue does not seem to do anything.
>BREAK GRUE
Munging a lurking grue doesn't appear worthwhile.
∂09-Sep-77 1733 FTP:Blohm at SUMEX-AIM LOTS Student Coordinator
Date: 9 SEP 1977 1734-PDT
From: Blohm at SUMEX-AIM
Subject: LOTS Student Coordinator
To: jmc at SAIL
9 - AUG - 77
Professor McCarthy,
As a student who has used LOTS extensively for the past three quarters,
I feel that I have earned the right to express my opinions regarding the
operation of LOTS. [I also have come to expect that all reasonable thoughts
should be given serious consideration.] Specifically, I feel that the
position of "Student Coordinator" as it is being advertised this year,
is totally UNREASONABLE! By making the position an almost full time
job (80%), Mr. Gorin is essentially making it impossible for ANY student
to take the job. For this reason, the job of Student Coordinator will
probably fall to a non-student! [This year, it will go to Mike Bryon,
a person who showed NO interest in LOTS until Mr. Gorin hired him for
the summer (he graduated last June).]
I feel that the Student Coordinator should be required to be a student.
A non-student just does not have the contact with the general student
populace that a student does. Furthermore, there is a way that LOTS can
get twice as much for their money. By hiring two people at 40% time,
Mr. Gorin will get two different viewpoints on any one issue. Second,
by paying two people 40% each, he will probably get about 110% time,
PROVIDED that two interested people are chosen. I have discussed this
possibility with a number of students who were interested, and they all
agreed that while they could not handle an 80% job; they could handle
a 40% job. As to the type of people who are desirable to have for
coordinator, it would probably be best to have one science/engineering
student and one social science student.
What do you think?
Jeff Blohm
-------
How do you envisage the job of student co-ordinator? It seems to me that
we don't need two people just to express viewpoints. I would like your
view on how much co-ordinating there is to do. I am not sure of Ralph's
exact view on how much co-ordinating he needs as compared to programming,
but it's certainly clear that in the immediate future we have a lot of
tuning of the system to do. I suspect his view of the matter is influenced
by the fact that David Roode got into an adversary position almost
immediately. By the way, I assume you have sent a message of similar
content to Ralph.
∂09-Sep-77 1753 FTP:Creary at SUMEX-AIM Request for Papers
Date: 9 SEP 1977 1755-PDT
From: Creary at SUMEX-AIM
Subject: Request for Papers
To: McCarthy at SU-AI
I am currently reading your IJCAI-77 paper "Epistemological Problems
of AI", and would like very much to have copies of three papers that you
refer to there:
1. "Minimal Inference - A New Way of Jumping to Conclusions"
2. "First Order Theories of Individual Concepts"
3. "Ascribing Mental Qualities to Machines".
Do you have on-line versions of these papers that you'd be willing to
let me GET with FTP? If so, I'll appreciate knowing the appropriate file
I'll appreciate receiving copies. Thanks very much.
Lew Creary (@SUMEX-AIM)
Heuristic Programming Project
-------
The following are pubbable versions of these papers. A note to PAT here
will get you printed versions: MINIMA[S77,JMC], MENTAL[F76,JMC], CONCEP[E76,JMC].
∂09-Sep-77 1808 FTP:Creary at SUMEX-AIM Oops!
Date: 9 SEP 1977 1810-PDT
From: Creary at SUMEX-AIM
Subject: Oops!
To: McCarthy at SU-AI
It appears that a line is missing in the final paragraph of my message
to you of a few minutes ago. The paragraph should read as follows:
Do you have on-line versions of these papers that you'd be willing to
let me GET with FTP? If so, I'll appreciate knowing the appropriate file
names. Also, if any of these papers are available as CS Reports or preprints,
I'll appreciate receiving copies. Thanks very much.
Lew Creary
-------
∂09-Sep-77 1828 FTP:Blohm at SUMEX-AIM
Date: 9 SEP 1977 1829-PDT
From: Blohm at SUMEX-AIM
To: jmc at SAIL
What I feel the duties of the student coordinator should be:
1) Serve as a user to staff interface for complaints, etc. The fact is,
some users just can't converse with Ralph, and find it much more
convenient to talk to a fellow student. Also, a lot of students are
afraid of Mr. Gorin.
2) He should make life "easier" for the users. This entails making
various system programs, such as "OPEN" easier for the user to
understand.
3) He (or she, in fact, it would be rather nice to have a female
coordinator) should keep the users informed. Actually, this is Mr.
Gorin's job, but he never seems to do it. Users should be given
access to concise information describing the state of the system
(this does not necessarily have to be in the form of a system message,
a constantly updated file would work fine). Most users (myself
included don't understand such explanations as "continuing REV10
upgrade".
4) He should keep in touch with other systems in the country, bringing
updated versions of programs to LOTS. Again, this should be done
by Ralph or Mo, because the student coordinator's job is transitory
in nature. For instance, Dave is constantly bringing over revisions
of SOS; and by his request, I am working with the author of HG in
order to make it compatable to the LOTS environment and bring it
over.
Undoubtedly, there are more tasks that are the SC's responsibility, but
these are the one's that come to mind. As for LOTS's present situation,
certainly, what we need now is more programming. However, the job
description for SC states that the candidate should have a general
knowledge of computing languages, NOT that he be a hacker! Of course,
if a candidate is a hacker, this should be counted as an asset.
If you don't mind, I'd like to come back to my suggestion of having one
person from "hard" science and another from the social sciences. There is
no shortage of people who could both program and work with the users. For
instance, in the social science's we have such people as JQJOHNSON and
KO (who has indicated an interest in the position, IF it was 40%). Both
of them are excellent PROGRAMMERS as well as students. Knowing the way
people divide their time, by hiring two people, you would get more than
you would get if you hired one person at 80%.
A final thought, I think that student's should have some kind of input
in the decision of who is chosen as coordinator. After all, he will be
their representative! In addition, it precludes the possibility of the
manager choosing a "patsy", who will always side with him on all issues.
Also, if a coordinator the students don't like is chosen, LOTS could very
well lose the services of a number of programmers (they could/would vacate
to such places as SCIP,SAIL,SRI,or SUMEX).
-------
I think the following shows that you need to improve your relations
with the students. The main requirement is more frequent and more
informative notices of what is going on. Brief messages dictated to
Queenie would fill the need, and the co-ordinator could do much of
the work.
∂10-Sep-77 0133 FTP:MRC at MIT-AI (Mark Crispin) We should consider this!
Date: 10 SEP 1977 0433-EDT
From: MRC at MIT-AI (Mark Crispin)
Subject: We should consider this!
To: JMC at SU-AI, LES at SU-AI
CC: MRC at SU-AI
MOON@MIT-MC 09/09/77 23:20:31 Re: PDP-11 front end
MRC@MIT-AI 09/09/77 12:24:00 Re: PDP-11 front end
To: MOON at MIT-AI
Does MC's KL have a 20-style front end running RSX-11M? It would be nice if
you could give me some hints about hacking it if it does; they want the
DIALnet implementation on LOTS to be done in the front end.
One of the first decisions we made was to flush that program completely.
I can't even begin to describe how hard and long it sucks, but you
probably already know it anyway.
Our front end pdp11 runs a combination of KLDCP and a program I wrote
(SYSTEM; IOELEV >) which is what does the I/O.
∂10-Sep-77 0153 MRC LOTS wheel who abuses his privileges
There has been a lot of this lately. Maybe there should be a tightening
up when the new accounting system happens.
∂10-Sep-77 0510 MRC DIALnet & DECUS
I got a note from a DECUS type at NBS about DIALnet. Apparently we're
going to be working with them on joint protocols or something. Could
you fill me in?
-- Mark
Was it Glenn Ricart at NIH who uses the NBS TIP. They are Dialnet
enthusiasts but have been transmitting information between PDP11s
by telephone. We should listen to them and try to meet their needs
if we can, but we can't commit ourselves to compatibility with their
existing implementation.
∂10-Sep-77 0612 MRC
To: LES, JMC
I have gobbled down Moon's IOELEV program. It would be worthwhile
to look at it at least; but if we want to make the DEC types happy I guess we
are stuck with DEC's bad ideas...sigh.
Sigh!! We have more than keeping D.E.C. happy for sticking with
their bad ideas in IOELEV. LOTS needS compatibility with their future
system changes.
∂10-Sep-77 0907 MRC
[MRC - Yes it was Glenn.]
∂10-Sep-77 0943 FTP:RYLAND at RUTGERS-10 guest account
Date: 10 Sep 1977 (Saturday) 1241-Est
From: RYLAND at RUTGERS-10
Subject: guest account
To: les at SU-AI
cc: Ryland at RUTGERS-10, jmc at SU-AI
Les,
I hope to be in frequent contact with Mark Crispin,
concerning the DialNet project, and wonder if it would be
possible to get a guest acount at Sail for that purpose.
Another question I have is about SAIL (the language):
we got the DECUS tape, and I am having problems installing it
on our DEC-20. Who is the best person (on the ARPANet) to
contact for help with installation?
Thanks,
-Chris
∂10-Sep-77 1513 RDR MRC
He raises a lot of hackles among the LOTSites because although he is
new he thinks he understands LOTS perfectly, he expresses himself
forcefully and he tosses in threats about logging people out and
killing thier directory if they do things of which he disapproves.
He thinks I think he "supports" Ralph and don't like him
therefore. I don't think he "supports" Ralph, I don't dislike him,
but he is a perfect example of a hacker type who needs someone
around to mediate in some fashion or disaster will strike!
∂11-Sep-77 0855 MRC Person from Columbia University
To: JMC, LES
By the way, I forgot to mention that this person from Columbia
called the other day asking about some SAIL problems and about
DIALnet. I couldn't help him in his SAIL problems (assembly
errors and catatonia when he runs it), but he seemed to have talked
to you about DIALnet and indicated that Columbia's 20 would be
willing to be a test site.
He also asked about getting an account at SAIL; I told him that
he'd have to talk to you about that.
∂11-Sep-77 0937 MRC System notice REM put up today
To: JBR, JMC
SYSTEM SEEMS TO BE INTERMITTANTLY BROKEN TODAY, HANGING VARIOUS WAYS
(1) WAS IN MODE OF HANGING INDEFINITELY ON ATTEMPTED CONNECTION FROM AMES-TIP BUT
ACCEPTING CONN FROM MIT DURING SAME TIME, (2) LOGIN ON TTY10,11 HANGING FOR 10 MINUTES,
(3) SEVERAL TIMES A SEND COMMAND WOULD HANG AND ↑C INEFFECTIVE FOR 2 MINUTES.
[I flushed this message; the problem is simply moby music jobs tying
up the DSKQ for long periods of time. Can something be done (1) to
discourage REM from putting up such marginal messages, (2) to alleviate
the cause of REM's unhappiness? He is right; the system has been almost
unusable in the past few hours.]
REM's messages would be taken better coming from someone whose relation
to the Lab was less marginal, but the fact is that others prefer to
mutter under their breaths, so his messages serve a useful purpose,
and I don't want to make him stop. After we have two channels, maybe
we can think of some way to confine the long transfers to one channel
so they will mainly compete with each other.
∂11-Sep-77 1321 RDR via AMET
To: JMC, REG, DSB, mo at SRI-KL
∂09-Sep-77 1052 DPB
∂08-Sep-77 1902 RDR via AMET
Has 105/6 picked a language for fall?
SAIL for both sections of 105 for sure (They are mine.) Almost sure
that SAIL will also be choice of 106 also. Herriot and Floyd most
likely instructors, and both are likely to agree with the SAIL choice.
Thanks for the info abbut 105/6. I agree with their choice, mainly
because I don't want a lot of 105/6 alumni trying to do big numerical
and array computations
on LOTS with an interpreter.
∂11-Sep-77 1323 RDR via AMET idea
You know I really think if you would come to a once-montly system meeting
at LOTS you'd have a more reliable perspective of what's going
on AND this would motivate the establishement of said meetings,
which would probably be more of a benefit in itself whatever that means.
I assume that this would include all the student hackers, that they
could add agenda items, and that some sort of notes would be taken.
I assure you that the universal attutude among the student hackers at
LOTS is that the system currently is in inexcusably bad shape
but that Ralph is stubborn and stupid about changing it. The discussion could
extend to other issues such as what paper to use in the lineprinter
(remeber you agreed to democracy in that case), ideas for non-system
improvements to LOTS, the wheel issue such as it is.
This would have to encourage communication and th only bad effect on
your part would be to tie you up for an hour a month, but I think
you'll find yourself tied up more frequently from bad effects
that will result without this meeting. Also, Ralph's only
criticism will be that it will force HIM to plan ahead and I don't think
that's bad at all.
That's a good idea. I'll do it.
∂11-Sep-77 1333 RDR via AMET yes there are a few summer PPL alumni around now
Would you believe blackjack system development?
Fortunatley she is a math grad student with friends using the system
and she wants to learn machine langauge, and she is quite reasonable.
∂11-Sep-77 1802 MRC via AMET REM mail
To: JMC, JBR
Maybe I was a bit strong in refering to this message, but the fact is I told
REM that the disk queue was high due to this moby program getting run.
I just don't like seeing such ill-informed broadcast messages; at the least it
is unesthetic and at worse it gives others bad impressions.
∂12-Sep-77 1643 FTP:Feigenbaum at SUMEX-AIM weizenbaum review
Date: 12 SEP 1977 1644-PDT
From: Feigenbaum at SUMEX-AIM
Subject: weizenbaum review
To: jmc at SU-AI
John, just a reminder to send down a copy of your Physics Today
review. Would you send one to Bruce too? Thanks!
Ed
-------
∂14-Sep-77 0433 CLT Chapter 4
I have been able to get BB forms of the MACLISP versions of LCOM0 and LCOM4.
I have also produced the LAP files for the DROP example using LCOM0, LCOM4
and NCOMPLR. With minor modifications to the text explanations of PDP
instructons Chapter 4 will be converted to MACLISP. What about including
an appendix containing the internal forms (Maclisp version) of LCOM0 and LCOM4?
Also I think it would be good to include the MICRO-MANUAL in an Appendix.
I will be working in my POLYA office today as I have several meetings
during the day. Back at the Lab in the evening.
∂14-Sep-77 2127 LLW via AMET meeting
John,
Friday at 10 is fine for Tom and Curt.
Tom
Excuse my confusing you. My idea was to talk with you about other matters
at 10 and have the meeting at 11. I have told the others to come at 11,
so I assume that will be ok with Tom and Curt. You and I could talk
afterward or some other time if 10 doesn't suit you. My only commitment
is a group of touring Swedish superannuated parliamentarians at 2:45pm.
∂15-Sep-77 1135 HVA TELEPHONE PROBLEMS
The repairman cames to A.I.Lab this a.m. - he could not get your phone to
malfunction when he tried it, but I insisted there were continuing problems
(and there are on my line,too), so he is still trying to locate the source of
trouble.
∂15-Sep-77 1348 EJG via AMET Stanford-1 time-sharing system
PMF mentioned the meeting tomorrow to me. Since I might be
involved in the design and/or implementation, I hope you don't mind
if I attend the meeting also. I have little to say about it myself
at this point, but I am anxious to listen in on the initial discussions.
∂15-Sep-77 1409 FTP:Taynai at SUMEX-AIM CSAC Meeting - Oct. 17/18
Date: 15 SEP 1977 1410-PDT
From: Taynai at SUMEX-AIM
Subject: CSAC Meeting - Oct. 17/18
To: JMC at SAIL, Gorin at SAIL
cc: TAYNAI
The President's Advisory Committee for Computer Science will be
meeting October 16-18.
The CSD presentation will be on Monday. (Oct. 17)
The Cocktail party will be at the Faculty Club from 6-7:30 on Monday evening
The LOTS presentation will be Tuesday morning
Durand 450 8:30 - 9:15.
There will be a continental breakfast from 8:00 to 8:30 and you might
want to go early for that.
An official agenda will be sent to you, plus the cocktail party invitation,
but Prof. Feigenbaum asked me to alert you so you could mark you calendars.
Please let me know if you plan to attend the breakfast so I can order enough
food.
Carolyn Tajnai
-------
∂15-Sep-77 1424 LES Swedish visitors
To: JMC, JC
CC: TED, ARG, BES
We expect the Committee on Education of the Swedish Parliament to visit
on Friday, Sept. 16, beginning about 2:30-2:45 and lasting for about one hour.
I propose the following schedule:
15 min. Walk around the computer room, do a short Hand-eye demo, then proceed
to the Music Lab.
25 min. JMC talks and demonstrates programs using TV monitors tied to the
Data Disk terminal.
20 min. Music demo by JC or somebody (default is LES, if no Music person is
available).
We will try to keep the presentation not too heavy, since this will be the
end of a rather long day for them.
∂15-Sep-77 1451 LCW meeting
To: LCW, TM, PMF, LES, JBR, REG, JMC
CC: LLW
Friday at 11 am is okay for tm and lcw.
∂15-Sep-77 2102 FTP:Geoff at SRI-KA UPI
Date: 15 Sep 1977 2100-PDT
From: Geoff at SRI-KA
Subject: UPI
To: jmc at SAIL, me at SAIL, les at SAIL
Did you know that UTEXAS has UPI wire on their UNIX system? Perhaps some type
of net connection could be cons'd up to allow NS to get at its data?
-------
∂15-Sep-77 2305 LLW via AMET Meetings on Friday AM
I regret that i confuse so readily. I plan to meet with you at 10 am, and
then with everyone else as well at 11 am. Lowell
∂16-Sep-77 0203 LES UPI at UTexas
To: Geoff at SRI-KA
CC: JMC, ME
I believe that UPI would object to an intertie, as would AP and NYT.
We can get UPI whenever we want it--all it takes is $$.
∂16-Sep-77 0334 CLT The state of things.
I have done some minor additional tinkering with chapter 4.
Namely (after reading about NCOMPLR in the old Maclisp manual) I see more
what is going on in the LAP code and have added a couple of additional
comments to the PDP-10 section. Also, the comments you added about
⊗flat were essentially the same as those I made, except mine were
attached to the function ⊗comp. So I moved the definition of ⊗flat
to follow ⊗comp and combined the comments.
Your suggested improvements for chapter 3 have been made and
a new version pubbed. I had to decrease the page size in general
as the XGENLINES kludge failed.
I have incorporated your suggested introduction to the FOL
appendix.
I have assembled all reference sections into a single annotated
bibliography.
The big question is--can I get pub to assemble all of this into
a coherent manuscript??
The Maclisp manual is assembled and ready for the printer I think.
(The ∂ that I thought the Maclisp Editor generated must have been noise.)
∂16-Sep-77 1053 HVA Quarterly Report - NASA
To: JMC
CC: TOB, HVA
Phil Surra called me to say NASA Quarterly Report is due.
∂16-Sep-77 1300 JMC*
call Shoshoni LBL 853-2740 x5171
∂16-Sep-77 1309 ZM LOGIC QUALS
I think it is time to fix the dates for the logic quals. Martin Brooks asked
not to do it before Nov. 20, so what about Sat. Nov. 27. Do you prefer an
oral exam, a written class exam, or a take home ? Zohar
∂16-Sep-77 2029 CLT LISP notes
The first page of LSPCMT.[1,CLT] is an annotated directory for all of the
files relevant to the LISP notes.
∂16-Sep-77 2128 LLW via AMET JMC letter to trimble/onr
To: PAT
CC: JMC
The revised last paragraph which john left in my mail reads very well from
my standpoint. please send it off, with my thanks to john. lowell
∂17-Sep-77 1011 JRA BOOK
weeeelll, it WILL be out next spring. i'm busily setting
type!!! turns out that the publishing industry is a LOSS.
so i'm styling, copy-editing, and photo-typesetting
the manuscript. i hope mcgraw-hill knows how to sell
books, at least. if not, i'll be going door-to-door.
i should have a reasonable new edition within a week. if you'd
like i'd put a copy on reserve in the library then.
john
That would be fine.
∂18-Sep-77 1630 LES
∂17-Sep-77 2048 JMC
I have seen an unknown named SOL recently.
[He is the Harvard fellow who we decided to let in through November, but I
neglected to put him in the directory until today.
By the way, I note that Peggy Waters is back on the list. Is that for old
times sake, or is she doing a project of some sort? ]
∂19-Sep-77 1059 PAT note from Ringle
You received a note from Martin Ringle as follows:
"I am preparing to send the final manuscript for Philosophical Perspectives
in Artificial Intelligence to the publisher. I have received no word from
you since my letter of 13 July and my note of 24 August.
In the opinion of both myself and several of the editorial consultants for
Humanities Press your paper, "Ascribing Mental...", is extremely interesting
and would certainly be an asset to the anothology. I cannot sign a contract
to include it, however, unless I have a written indication of your acceptance
of the contract agreements. The terms, as mentioned in my note of 24 August,
are 1% of the actual selling price of each book sold for each of the ten
contributors (including myself); 1% of the actual selling price of any
future sales involving reprints, translations, etc.
I would be very grateful if you could drop me a note about this at your
earliest convenience. Sincerely."
I looked for some recent correspondence but didn't find any. patte
∂19-Sep-77 1636 PAM Hooray
To: JMC, CCG
The first draft of my proposal is finished. It is on PROP[RES,PAM];
PROGRE.SS[RES,PAM] gives pointers to other files that are relevant to
the project. I feel that the proposal needs numerous stylistic changes,
but I am presently much more interested in comments on the content.
∂19-Sep-77 1725 CLT
Paula Apsell at NOVA would like you to call her collect. Her office phone
number is 617-783-1214, and her home phone is 617-581-1631.
∂21-Sep-77 0036 FTP:PBARAN at USC-ISI PCNET Advisory Board
Date: 20 SEP 1977 0912-PDT
From: PBARAN at USC-ISI
Subject: PCNET Advisory Board
To: jmc at SU-AI
The PCNET group intends to start correspondence with the major
turn-key computer manufacturers (Commodore, Tandy, etc.). We would
like to encourage them to use reasonable communications
standards/protocols; we think it would be tragic if manufacturers
tried to 'lock in' customers to their own equipment by using
mutually exclusive communication protocols, especially at the
lower levels.
To this end we're going to print up stationary for
'The PCNET Committee'; to create some presently non-existent
reputation we'd like to have an Advisory Board of luminaries
down the left margin. Would you be one of the luminaries ?
Duties are strictly nominal - if you get a phone call from
Commodore you tell 'em that the PCNET committee is a bunch of
good guys trying to do reasonable things. We will use the
stationary sparingly in ways that will be a credit to us all.
Please let me know if you can be an advisory board member;
if not, do you have a nominee ?
Dave Caulkins
(mailbox PBARAN@ISI)
-------
∂21-Sep-77 0306 CLT
I plan to include a paragraph of the following flavor in the fact sheet. It is
basically extracted from earlier fact sheets and seems to agree with current
policy. Any objections?
COURSE WORK:
This course consists of lectures, homework, a midterm
examination, a final examination, and a term project. Projects will be
due on the day of the final examination. It is possible to get a
grade of B+ or Pass in the course without doing a project, and the
grade of INC will be given only under exceptional circumstances.
Therefore, those who aspire to some flavor of A should start their
projects early. A handout with more complete information on the term
project will be out soon.
Proposed paragraph in handout is fine.
∂21-Sep-77 0825 100 : TPS PHONE CALL
Cuthbert Hurd called, message is Noyce no
∂21-Sep-77 1052 FTP:Taynai at SUMEX-AIM Terman Dedication
Date: 21 SEP 1977 1043-PDT
From: Taynai at SUMEX-AIM
Subject: Terman Dedication
To: JMC at SAIL, Wiederhold
cc: TAYNAI
Professors Herriot, Knuth, McCarthy, and Wiederhold,
Prof. Feigenbaum has been invited to attend the Frederick Emmons
Terman Enginering Center Dedication Thursday Oct. 6 at 2 p.m.
The ceremony is scheduled from 2-3, followed by refreshments and
tours. Prof. Feigenbaum would like for you to accompany him to the
ceremony.
Please let me know if you can attend as we must respond to Dean Lack.
Carolyn Tajnai
7-3264
-------
∂22-Sep-77 1039 FTP:Taynai at SUMEX-AIM ( Forwarded Mail )
Date: 22 SEP 1977 1039-PDT
From: Taynai at SUMEX-AIM
Subject: ( Forwarded Mail )
To: McCarthy at SAIL, Wiederhold
Date: 21 SEP 1977 1043-PDT
From: Taynai
Subject: Terman Dedication
To: JMC at SAIL, Wiederhold
cc: TAYNAI
Professors Herriot, Knuth, McCarthy, and Wiederhold,
Prof. Feigenbaum has been invited to attend the Frederick Emmons
Terman Enginering Center Dedication Thursday Oct. 6 at 2 p.m.
The ceremony is scheduled from 2-3, followed by refreshments and
tours. Prof. Feigenbaum would like for you to accompany him to the
ceremony.
Please let me know if you can attend as we must respond to Dean Lack.
Carolyn Tajnai
7-3264
-------
-------
∂22-Sep-77 1353 FTP:Feigenbaum at SUMEX-AIM private guesses from pete hart re interlisp/please dont propagate
Date: 22 SEP 1977 1316-PDT
From: Feigenbaum at SUMEX-AIM
Subject: private guesses from pete hart re interlisp/please dont propagate
To: jmc at SAIL, rindfleisch, rick at RAND-UNIX,
To: anderson at RAND-UNIX
Mail from SRI-KL rcvd at 1-SEP-77 1143-PDT
Date: 1 Sep 1977 1143-PDT
From: Hart at SRI-KL
Subject: Re: interlisp on the kl-10
To: Feigenbaum at SUMEX-AIM
cc: HART
In response to your message sent 1 SEP 1977 1030-PDT
Ed,
Here's an attempt at answers:
1. We run Interlisp on the KL-10.
2. I don't know how many Lisp jobs brings the KL to its knees.
A load average of 18-20 is where things crumble. The KL
probably does considerably better than the dual KI, because
(on our configuration) the 1 megaword primary memory is a big
win. We also have (I just learned) a better garbage collection
algorithm in at the monitor level.
The 2020 is a mystery. Rumors are that it's about .7 KA's, which would
make it more than a one Lisp user machine. I did a very!! rough back of
the envelop calculation some months ago that said that the 2020 was more
effective than a KL in terms of Lisp users/dollar.
Peter
-------
-------
∂22-Sep-77 1357 FTP:Wiederhold at SUMEX-AIM Program verification course
Date: 22 SEP 1977 1335-PDT
From: Wiederhold at SUMEX-AIM
Subject: Program verification course
To: sso at SAIL, MJL at SAIL, DPB at SAIL, JMC at SAIL
Course Description
CS244 Program verification
Formal and informal verification techniques based on asssertions (Floyd/Hoare),Burstall/Manna), functional representation of programs, and computation
histories. Specification methods. Modular design, specification, and verifica-
tion of large systems, including parallel programs. Systematic testing.
Prerequisite: CS156
(Overlap with 258 must be discussed)
-------
∂23-Sep-77 1739 JBR at TTY43 1739
BETTER, I FIXED A SYSTEM CRASHING BUG TODAY. TWO MAIN PROBLEMS
LEFT: 1 WITH THE CHANNEL/DISK PROTOCOLS WHICH WE WILL TRY TO MAKE A KLUDGE
FIX TO TONIGHT, 2 THE DISK INTERLOCK FOR USE OF THE SYSTEM TRACK BUFFER CAUSES
MANY JOBS WAITING IN IOWQ FOR LONG TIME. THE FIX TO THIS IS A DIFFICULT I/O BUFFER
MANAGEMENT SCHEME THAT I HAVE STARTED CODING UP.
∂26-Sep-77 0130 FTP:Hedberg at SUMEX-AIM Credit for consulting at LOTS
Date: 26 SEP 1977 0129-PDT
From: Hedberg at SUMEX-AIM
Subject: Credit for consulting at LOTS
To: jmc at SAIL
I am interested in consulting for credit at LOTS. I asked Ralph
about it and he said I should contact you, since he can't give credit.
Are there any plans afloat of this nature?
Erik Hedberg
-------
No general plan is needed. We can work it out individually.
∂26-Sep-77 0410 DPB List of your advisees
To: DCL, SSO, CCG, JMC, TOB, TW, buchanan at SUMEX-AIM
To: wiederhold at SUMEX-AIM
CC: MJL
The file ADVSRS.PTY[INF,CSD] contains a list of advisors and their advisees.
If you want to obtain your advisees' folders please contact Moira. -Denny
∂26-Sep-77 1049 100 : patte phone call
please call Cuthbert Hurd - 854-1901.
∂26-Sep-77 1421 FTP:Feigenbaum at SUMEX-AIM Congratulations
Date: 26 SEP 1977 1339-PDT
From: Feigenbaum at SUMEX-AIM
Subject: Congratulations
To: winograd at PARC
cc: jmc at SAIL
Terry, I want to express to you my congratulations on winning the
Dean's award for Excellence in Teaching. That makes everyone around
here proud, and I'm sure it made you
feel good.
Best Wishes,
Ed
-------
∂27-Sep-77 0911 MJL LOTS meeting
John, I received a message this a.m. from someone in I.E. about a
LOTS meeting. She spoke with Ralph Gorin and he said it was ok
to cancel another meeting you and he had at the same time, so
the LOTS meeting will be Wednesday, October 5 @ 4:15. they will
get in touch with you regarding meeting place.
∂27-Sep-77 1137 FTP: Bob Sproull at CMU-10A dial-up mail
Date: 27 Sep 1977 1437-EDT
From: Bob Sproull at CMU-10A
Subject: dial-up mail
To: jmc@su-ai
CC: raj.reddy@CMU-10A
Sender: BOB.SPROULL at CMU-10A
Message-ID: [CMU-10A] 27 Sep 1977 14:37:09 Bob Sproull
John,
We are contemplating installing some facilities for dial-up
mail, for reasons similar to those in your
mind (I suspect). Our two main interests are:
1. A facility to promote home computer use of mail by giving
access to the PDP-10 and hence the ARPA net.
2. We are entering into collaboration with Cal Tech
(who have a Dec-20) and want to (a) be able to send them
mail and receive from them and (b) to promote mail-like
services within their environment.
Raj mentioned that you have very similar aims, and are
probably well onthe way to solutions. Other things being
equal, it seems to me we should be compatible with you. This
message is intended to inquire into your thoughts and plans
for the facility. (Just as an aside, I believe that
the interface sould support computer-computer interactions,
and hsould also support, perhaps in a crude way, an individual
who has no more than a terminal and dials into the same
number. Reliability ad encryption are important
--reliability first, encryption later.)
Please let me know the state of your progress, or a reference
to someone I can deal with to avoid bothering you. Also,
I am interested whether whatever you plan for LOTS
would be suitable for Cal Tech's Dec-20.
Thanks for your time.
Bob
-------
Bob,
We have an NSF grant to develop Dialnet - protocols for giving ARPAnet
like facilities to users of telephones. We should have preliminary
protocols in a few months. The initial test machines will be
SAIL and LOTS, a DEC-20 at Stanford for teaching and unsponsored
research. Since it is a standard DEC-20, the programs will be immediately
usable by Caltech if theirs is likewise standard. Various files in
[DIA,JMC] are relevant. The man doing the work is Mark Crispin who is
called MRC here. He has some preliminary documentation in some files.
John
∂27-Sep-77 1438 PAM
To: CCG, JMC
Reading Committee
I want you to be on my reading committee; Terry Winograd has agreed to be my
advisor, but I do not expect him to do all the advising. I have a second draft
of my dissertation proposal on the file PROP[RES,PAM]. I want to meet with you
and discuss this proposal and my expectations for a committee. My deadline is
October 1; I need to meet with you as soon as possible.
Besides the proposal itself, the file SCHED[RES,PAM] contains my schedule for
completion of this research by October, 1978. The file PROGRE.SS[RES,PAM] contains
pointers to my other files that could be of interest.
Paul Martin
∂27-Sep-77 2020 DPB Wagner's degree
To: JMC, TJW, MJL
There is nothing in Todd's file indicating that he satisfied the seminar
requirement. The faculty voted to witthold granting the degree until
evidence exists that the requirement has been met. A message from
JMC (or any other faculty member) that a public seminar (other than orals)
was given, and that the faculty member approved the presentation as
satisfying the requirement, is sufficient. If said message is received
by 1pm tomorrow, Sept 28, then the degree will be granted as of this
quarter. If not, it will have to wait til next quarter. -Denny