perm filename W80.IN[LET,JMC]1 blob
sn#501918 filedate 1980-04-01 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00433 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00041 00002 ∂03-Jan-80 1705 Sandelin at SUMEX-AIM text handling and delivery
C00044 00003 ∂03-Jan-80 2309 LGC Tomorrow @ 2:00?
C00045 00004 ∂04-Jan-80 0756 LGC Tuesday at 2:00?
C00046 00005 ∂04-Jan-80 1109 JK
C00047 00006 ∂04-Jan-80 1302 ROY on TTY20 1302
C00048 00007 ∂04-Jan-80 1424 LES New DM modem
C00049 00008 ∂05-Jan-80 1433 MINSKY at MIT-AI (Marvin Minsky)
C00050 00009 ∂05-Jan-80 1642 FWH via AMES-TIP Program Verification seminars
C00053 00010 ∂05-Jan-80 2040 REM via AMES-TIP CPU TIME
C00055 00011 ∂06-Jan-80 1400 ZM Computer Forum
C00056 00012 ∂06-Jan-80 1728 ME modem interrupts on TTYs 20,21,22
C00057 00013 ∂06-Jan-80 1947 FWH PV seminar -correction
C00058 00014 ∂07-Jan-80 1039 CLT Calendar
C00059 00015 ∂07-Jan-80 1511 Gardner at SUMEX-AIM Re: Your contract example
C00063 00016 ∂07-Jan-80 2143 MINSKY at MIT-AI (Marvin Minsky)
C00064 00017 ∂07-Jan-80 2234 LGC via SU-TIP
C00065 00018 ∂08-Jan-80 0841 REG Your Imlac
C00066 00019 ∂08-Jan-80 0842 REG LOTS, etc.
C00067 00020 ∂08-Jan-80 1339 BTH FYI
C00068 00021 ∂08-Jan-80 1326 LGC via SU-SCORE
C00069 00022 ∂08-Jan-80 1340 CG reference
C00071 00023 ∂08-Jan-80 1440 LGC I'm still alive
C00072 00024 ∂08-Jan-80 2019 JRA lisp talk
C00073 00025 ∂09-Jan-80 1102 JMC*
C00074 00026 ∂09-Jan-80 1915 ADMIN.BOSACK at SU-SCORE Characterization of Editing Users
C00078 00027 ∂10-Jan-80 0741 Feigenbaum at SUMEX-AIM (Response to message)
C00079 00028 ∂10-Jan-80 0803 Feigenbaum at SUMEX-AIM lunch with Jon Doyle
C00080 00029 ∂10-Jan-80 0928 JRA lisp conf
C00081 00030 ∂10-Jan-80 1100 JMC*
C00082 00031 ∂10-Jan-80 1058 JRA lisp conf
C00083 00032 ∂10-Jan-80 1124 JMC dugan
C00084 00033 ∂10-Jan-80 1129 JMC
C00085 00034 ∂10-Jan-80 1309 SEK
C00087 00035 ∂10-Jan-80 1357 LGC
C00088 00036 ∂10-Jan-80 1513 BTH Concurrent Programming Seminar
C00089 00037 ∂11-Jan-80 0903 REG
C00096 00038 ∂11-Jan-80 1402 FWH PV+A Seminar
C00098 00039 ∂11-Jan-80 1403 TOB terminal system
C00099 00040 ∂11-Jan-80 1750 ADMIN.BOSACK at SU-SCORE
C00102 00041 ∂12-Jan-80 1459 LES Datamedia
C00106 00042 ∂13-Jan-80 2031 JK
C00109 00043 ∂13-Jan-80 2217 LGC Monday at 2pm?
C00110 00044 ∂14-Jan-80 0000 JMC*
C00111 00045 ∂14-Jan-80 1006 REG
C00112 00046 ∂14-Jan-80 1111 JMC*
C00113 00047 ∂14-Jan-80 1200 JMC*
C00114 00048 ∂14-Jan-80 1413 LGC
C00115 00049 ∂14-Jan-80 1719 CSD.MCCARTHY at SU-SCORE barrett,kahn
C00116 00050 ∂14-Jan-80 1811 ME running FIXIML from another TTY
C00117 00051 ∂15-Jan-80 0904 Feigenbaum at SUMEX-AIM an interesting terminal for CSD editing
C00120 00052 ∂15-Jan-80 0904 Feigenbaum at SUMEX-AIM Re: Barclay Terminal
C00122 00053 ∂15-Jan-80 0950 JK
C00124 00054 ∂15-Jan-80 1100 JMC*
C00125 00055 ∂15-Jan-80 1233 CLT tonight
C00126 00056 ∂15-Jan-80 1236 TOB
C00127 00057 ∂15-Jan-80 1338 JMC michie address
C00128 00058 ∂15-Jan-80 1555 FWH PV+A Seminar
C00130 00059 ∂15-Jan-80 1902 TOB
C00131 00060 ∂16-Jan-80 0228 Mark Crispin <Admin.MRC at SU-SCORE> Dialnet mail
C00133 00061 ∂16-Jan-80 0842 Feigenbaum at SUMEX-AIM Jon Doyle
C00134 00062 ∂16-Jan-80 0914 TOB
C00135 00063 ∂16-Jan-80 0949 MP John Doyle
C00136 00064 ∂16-Jan-80 1550 Mark Crispin <Admin.MRC at SU-SCORE>
C00138 00065 ∂16-Jan-80 1810 LES
C00139 00066 ∂17-Jan-80 0148 LES DMs
C00140 00067 ∂17-Jan-80 0813 Sandelin at SUMEX-AIM misc
C00142 00068 ∂17-Jan-80 0824 REG
C00143 00069 ∂17-Jan-80 0825 REG
C00144 00070 ∂17-Jan-80 1149 CLT
C00145 00071 ∂17-Jan-80 1349 BTH Concurrent Programming Seminar
C00148 00072 ∂17-Jan-80 1601 BTH concurrent programming seminar
C00149 00073 ∂18-Jan-80 0000 LES
C00150 00074 ∂18-Jan-80 0008 LES
C00151 00075 ∂18-Jan-80 0856 LGC
C00152 00076 ∂18-Jan-80 1359 Denny Brown <CSD.DBROWN at SU-SCORE> Faculty meeting/Gray Tuesday announcement
C00157 00077 ∂18-Jan-80 1522 CSD.BETTY at SU-SCORE Gorin alary Review
C00159 00078 ∂18-Jan-80 1539 Denny Brown <CSD.DBROWN at SU-SCORE>
C00160 00079 ∂18-Jan-80 1627 LES State of things
C00164 00080 ∂18-Jan-80 1642 LES Display makers
C00165 00081 ∂18-Jan-80 2225 JMC
C00166 00082 ∂19-Jan-80 1307 LES
C00167 00083 ∂19-Jan-80 1332 LES Baskett
C00168 00084 ∂20-Jan-80 0642 REM at MIT-MC (Robert Elton Maas)
C00171 00085 ∂20-Jan-80 1400 Mark.Sherman at CMU-10A Datacomputer going away
C00175 00086 ∂20-Jan-80 1955 LGC New Developments
C00181 00087 ∂21-Jan-80 0000 JMC*
C00182 00088 ∂21-Jan-80 1024 Zenon.Pylyshyn at CMU-10A (X320ZP51) Meeting
C00183 00089 ∂21-Jan-80 1252 TOB chern visit
C00184 00090 ∂21-Jan-80 1520 BTH Concurrent Programming Seminar
C00185 00091 ∂21-Jan-80 1606 TOB chern visit
C00186 00092 ∂21-Jan-80 1904 LGC 1. Emendations 2. E-machines
C00188 00093 ∂22-Jan-80 1232 TOB
C00189 00094 ∂22-Jan-80 1321 SELFRIDGE at BBN-TENEXD C and PASCAL
C00191 00095 ∂22-Jan-80 1307 PJH naive physics
C00193 00096 ∂22-Jan-80 1734 TOB returning your call
C00194 00097 ∂23-Jan-80 1900 CLT*
C00195 00098 ∂24-Jan-80 0832 ADMIN.MO at SU-SCORE (Maurice Bizzarri) Re: visit to Grinnell today
C00196 00099 ∂24-Jan-80 1117 ADMIN.MO at SU-SCORE (Maurice Bizzarri) Re: meeting today
C00197 00100 ∂24-Jan-80 1141 FWH PV+A Seminar
C00201 00101 ∂24-Jan-80 1200 JMC
C00202 00102 ∂24-Jan-80 2247 LGC Tomorrow at 2pm?
C00204 00103 ∂25-Jan-80 0224 PMF
C00205 00104 ∂25-Jan-80 0900 JMC*
C00206 00105 ∂25-Jan-80 1030 Feigenbaum at SUMEX-AIM Re the Barclay terminal--any comments?
C00216 00106 ∂25-Jan-80 1221 TOB
C00217 00107 ∂25-Jan-80 1631 ARK S-1 Meeting
C00218 00108 ∂25-Jan-80 1706 HPM The time has come
C00219 00109 ∂25-Jan-80 1710 HPM that is, I want to HOLD the orals in the second week
C00220 00110 ∂26-Jan-80 0307 JK SET THEORY
C00223 00111 ∂26-Jan-80 1436 RWW
C00224 00112 ∂26-Jan-80 1443 RWW Invitation
C00225 00113 ∂26-Jan-80 1646 JB via OFFICE-2 Hello.
C00226 00114 ∂27-Jan-80 2154 LGC new term
C00227 00115 ∂27-Jan-80 2214 ARK S-1 Meeting
C00228 00116 ∂28-Jan-80 0119 JK ramsey
C00229 00117 ∂28-Jan-80 1032 LGC
C00230 00118 ∂28-Jan-80 1839 LES
C00231 00119 ∂28-Jan-80 1847 LES
C00232 00120 ∂29-Jan-80 0916 JMC*
C00233 00121 ∂29-Jan-80 1011 DOYLE at MIT-AI (Jon Doyle) Your request and offer
C00235 00122 ∂29-Jan-80 1021 MP letter ready for your signature
C00236 00123 ∂29-Jan-80 1043 LGC Eliashberg
C00237 00124 ∂29-Jan-80 1430 BTH Concurrent Programming Seminar
C00238 00125 ∂29-Jan-80 1455 MINSKY at MIT-AI (Marvin Minsky)
C00240 00126 ∂29-Jan-80 1601 DBL AI Candidate 2/29
C00242 00127 ∂29-Jan-80 1704 Lenat at SUMEX-AIM (Response to message)
C00243 00128 ∂29-Jan-80 1707 PJH
C00244 00129 ∂29-Jan-80 1710 GIO Accounts
C00246 00130 ∂29-Jan-80 1903 CSL.FB at SU-SCORE [Don Knuth <DEK at SU-AI>: gls ]
C00248 00131 ∂29-Jan-80 1903 CSL.FB at SU-SCORE Re: gls
C00249 00132 ∂29-Jan-80 1903 CSL.FB at SU-SCORE Guy Steele
C00250 00133 ∂29-Jan-80 1903 CSL.FB at SU-SCORE GLS
C00251 00134 ∂29-Jan-80 1903 Don Knuth <DEK at SU-AI> gls
C00253 00135 ∂29-Jan-80 2106 Baskett at PARC-MAXC GLS
C00254 00136 ∂29-Jan-80 2109 Baskett at PARC-MAXC GLS
C00256 00137 ∂29-Jan-80 2112 Baskett at PARC-MAXC GLS
C00258 00138 ∂29-Jan-80 2134 Baskett at PARC-MAXC Re:
C00260 00139 ∂29-Jan-80 2300 TOB
C00263 00140 ∂30-Jan-80 0900 JMC*
C00264 00141 ∂30-Jan-80 0959 REG
C00265 00142 ∂30-Jan-80 1010 DPB via SU-SCORE list of people in Jacks
C00267 00143 ∂30-Jan-80 1603 Denny Brown <CSD.DBROWN at SU-SCORE> Re: GLS
C00269 00144 ∂30-Jan-80 1851 HPM HPM Orals
C00270 00145 ∂31-Jan-80 0823 REG
C00272 00146 ∂31-Jan-80 0900 JMC*
C00273 00147 ∂31-Jan-80 1009 TOB Chern meeting
C00275 00148 ∂01-Feb-80 0557 HPP.GENESERETH at SU-SCORE MACSYMA
C00277 00149 ∂01-Feb-80 1027 CG another recommendation
C00278 00150 ∂01-Feb-80 1302 BTH concurrent programming seminar
C00281 00151 ∂01-Feb-80 1409 LGC The Frame Problem and Philosophy
C00282 00152 ∂01-Feb-80 1555 LES Doyle
C00283 00153 ∂02-Feb-80 2313 LES display procurement
C00284 00154 ∂03-Feb-80 1002 Feigenbaum at SUMEX-AIM Ullman/VLSI/money
C00286 00155 ∂03-Feb-80 1335 LGC Peter Cheeseman
C00292 00156 ∂03-Feb-80 1347 LGC Eliashberg Talk Tomorrow
C00293 00157 ∂04-Feb-80 1000 JMC*
C00294 00158 ∂04-Feb-80 1000 JMC*
C00295 00159 ∂04-Feb-80 1101 FWH PV+A Seminar: Reminder
C00296 00160 ∂04-Feb-80 1619 PJH
C00297 00161 ∂04-Feb-80 1634 PJH
C00298 00162 ∂05-Feb-80 1021 TOB hello
C00299 00163 ∂05-Feb-80 1346 DWW concurrent programming seminar
C00300 00164 ∂06-Feb-80 1249 FWH PV+A Seminar
C00302 00165 ∂07-Feb-80 1041 HVA Ticket--Dina Bolla
C00303 00166 ∂07-Feb-80 1047 JMC*
C00304 00167 ∂08-Feb-80 1000 JMC*
C00305 00168 ∂08-Feb-80 1213 TOB help
C00306 00169 ∂08-Feb-80 1502 ARK S-1 Meeting Notice
C00307 00170 ∂08-Feb-80 1508 ARK S-1 Meeting Notice
C00308 00171 ∂08-Feb-80 2054 Kanerva at SUMEX-AIM Pictures and Words
C00311 00172 ∂09-Feb-80 0808 Feigenbaum at SUMEX-AIM (Response to message)
C00312 00173 ∂09-Feb-80 1736 CSL.FB at SU-SCORE
C00313 00174 ∂09-Feb-80 1738 ADMIN.MO at SU-SCORE (Maurice Bizzarri)
C00314 00175 ∂10-Feb-80 0928 Feigenbaum at SUMEX-AIM (Response to message)
C00315 00176 ∂10-Feb-80 1220 LES ARPA final report
C00317 00177 ∂11-Feb-80 0445 JK RAMSEY
C00318 00178 ∂11-Feb-80 1104 doyle at MIT-AI (Jon Doyle) Job offer
C00319 00179 ∂11-Feb-80 1120 LGC Revised Paper
C00320 00180 ∂12-Feb-80 1211 MINSKY at MIT-AI (Marvin Minsky)
C00321 00181 ∂12-Feb-80 1312 FWH PV+A Seminar: change of location
C00322 00182 ∂12-Feb-80 1339 TOB cart
C00324 00183 ∂12-Feb-80 1353 MP
C00325 00184 ∂12-Feb-80 1420 MP
C00326 00185 ∂12-Feb-80 1507 MP
C00327 00186 ∂12-Feb-80 1549 MP Reiter's phone number
C00329 00187 ∂12-Feb-80 2313 PLS Seminars at the Institute
C00330 00188 ∂13-Feb-80 1134 Bmoore at SRI-KL (Bob Moore) Workshop on Formal Methods in Artificial Intelligence
C00336 00189 ∂13-Feb-80 1443 ENGELMORE at USC-ISI Review
C00338 00190 ∂13-Feb-80 1452 SYS
C00339 00191 ∂13-Feb-80 1456 RWW
C00340 00192 ∂14-Feb-80 1247 ENGELMORE at USC-ISI
C00341 00193 ∂15-Feb-80 0546 Feigenbaum at SUMEX-AIM Re: Jon Doyle and inner space
C00343 00194 ∂15-Feb-80 0706 JRA LISP CONFERENCE UPDATTE AND REMINDER
C00345 00195 ∂15-Feb-80 0927 DCD
C00346 00196 ∂15-Feb-80 1403 100 : phyllis cost of versatec and alphatype
C00347 00197 ∂15-Feb-80 1451 FWH PV+A Seminar
C00350 00198 ∂15-Feb-80 1640 BS Equipment costs
C00352 00199 ∂16-Feb-80 1029 REG I don't mean to sound too petulant, but
C00353 00200 ∂16-Feb-80 1316 REG
C00355 00201 ∂17-Feb-80 0210 BTH concurrent programming seminar
C00357 00202 ∂17-Feb-80 1407 SELFRIDGE at BBN-TENEXD draft of TAC report
C00373 00203 ∂17-Feb-80 1659 PN User's Guide to S-1 Pascal/Fortran System
C00374 00204 ∂17-Feb-80 1922 SELFRIDGE at BBN-TENEXD your message
C00377 00205 ∂17-Feb-80 2024 LGC Engelmore Visit
C00378 00206 ∂18-Feb-80 0938 JJK Tech reports to Russia
C00379 00207 ∂18-Feb-80 1430 Feldman at SUMEX-AIM Reference
C00380 00208 ∂20-Feb-80 0000 JMC*
C00381 00209 ∂20-Feb-80 0818 JRA longish, but IMPORTANT conf. note.
C00389 00210 ∂20-Feb-80 1351 HVA Dossier for DANIEL MEMMI
C00390 00211 ∂20-Feb-80 1354 DCD
C00391 00212 ∂20-Feb-80 1648 TOB
C00393 00213 ∂21-Feb-80 1008 HVA Daniel Memmi Dossier
C00394 00214 ∂21-Feb-80 1154 HVA Committee for Concerned Scientists
C00395 00215 ∂21-Feb-80 1529 BTH concurrent programming seminar
C00397 00216 ∂21-Feb-80 1522 ROY on TTY66 (at TV-66) 1522
C00398 00217 ∂21-Feb-80 1759 Horning at PARC-MAXC Re: here we go again
C00399 00218 ∂21-Feb-80 1658 BTH here we go again
C00400 00219 ∂21-Feb-80 1630 BTH special seminar
C00403 00220 ∂21-Feb-80 1621 BTH concurrent programming seminar
C00408 00221 ∂22-Feb-80 1133 CSL.OWICKI at SU-SCORE (Susan Owicki) CSL systems candidates
C00418 00222 ∂22-Feb-80 1216 PAT the files on j,pat
C00419 00223 ∂22-Feb-80 1328 BTH SNAFU * 2
C00422 00224 ∂22-Feb-80 1408 Wiederhold at SUMEX-AIM Trip and info
C00423 00225 ∂22-Feb-80 1405 FWH PV+A Seminar
C00424 00226 ∂22-Feb-80 1555 ROY on TTY100 (at TV-125) 1555
C00425 00227 ∂22-Feb-80 1650 RAK via SU-TIP snafus
C00426 00228 ∂24-Feb-80 1436 BTH Concurrent Programming Seminar
C00429 00229 ∂24-Feb-80 1643 HPM
C00430 00230 ∂24-Feb-80 2328 HPM via CMU-10A dynamic support
C00432 00231 ∂25-Feb-80 0124 Mark Crispin <Admin.MRC at SU-SCORE>
C00434 00232 ∂25-Feb-80 1058 Wiederhold at SUMEX-AIM Supplier of C-compiler
C00435 00233 ∂25-Feb-80 1130 PAT pat,jmc
C00436 00234 ∂25-Feb-80 1345 HPM via CMU-10A cross fertilization
C00437 00235 ∂26-Feb-80 0048 LLW Dynamic Skyhook Collaboration
C00441 00236 ∂26-Feb-80 1324 MP
C00442 00237 ∂26-Feb-80 1331 MP Hirsch letter etc.
C00443 00238 ∂26-Feb-80 1538 MINSKY at MIT-AI (Marvin Minsky)
C00446 00239 ∂26-Feb-80 1548 Mark Crispin <Admin.MRC at SU-SCORE>
C00448 00240 ∂26-Feb-80 1657 HPM
C00449 00241 ∂26-Feb-80 2047 MINSKY at MIT-AI (Marvin Minsky)
C00450 00242 ∂26-Feb-80 2047 MINSKY at MIT-AI (Marvin Minsky) Reply to HPM, 26 Feb 1657-PST
C00453 00243 no lunch tomorrow
C00454 00244 ∂27-Feb-80 1730 JMC*
C00455 00245 ∂27-Feb-80 1730 JMC*
C00456 00246 ∂27-Feb-80 1751 FORWARD at USC-ECL Dynamic Compression
C00457 00247 ∂27-Feb-80 1751 MINSKY at MIT-AI (Marvin Minsky)
C00458 00248 ∂27-Feb-80 1751 MINSKY at MIT-AI (Marvin Minsky)
C00471 00249 ∂28-Feb-80 0039 LES Doyle funding
C00472 00250 ∂28-Feb-80 1040 MP phone call
C00474 00251 ∂28-Feb-80 1240 LES
C00475 00252 ∂28-Feb-80 1443 FWH PV+A Seminar
C00476 00253 ∂28-Feb-80 1503 CG
C00477 00254 ∂28-Feb-80 1609 HVA Meeting Friday, Feb. 29: 1-3 p.m. Rm 190, Law School
C00478 00255 ∂28-Feb-80 1614 100 phone message
C00479 00256 ∂28-Feb-80 1617 HVA Tel. Call
C00480 00257 ∂28-Feb-80 1638 DCD
C00482 00258 ∂29-Feb-80 1030 Feigenbaum at SUMEX-AIM earlier message to Forest/with more
C00486 00259 ∂29-Feb-80 1059 TOB
C00487 00260 ∂29-Feb-80 1338 Hazen at SUMEX-AIM Meera BlattnerNSF,here to visit 3/4/80
C00489 00261 ∂29-Feb-80 1631 LES
C00490 00262 ∂02-Mar-80 1144 CLT concert
C00491 00263 ∂02-Mar-80 1753 LLW Cannonball Speed and Dissipation
C00493 00264 ∂03-Mar-80 0000 JMC*
C00494 00265 ∂03-Mar-80 0943 Rosenow at SUMEX-AIM Meera Blatner appointment
C00496 00266 ∂03-Mar-80 1132 CLT stamps
C00497 00267 ∂03-Mar-80 1244 HVA SECRETARY FOR YOU
C00498 00268 ∂03-Mar-80 1257 Rosenow at SUMEX-AIM meeting place
C00499 00269 ∂03-Mar-80 1500 HVA Thursday Mar. 6th--Sect'y
C00500 00270 ∂03-Mar-80 1836 FWH PV+A Seminar
C00502 00271 ∂04-Mar-80 0143 LLW Skyhookery
C00505 00272 ∂04-Mar-80 0206 LLW Request for SAIL User Status
C00506 00273 ∂04-Mar-80 0854 HVA Thursday March 6th--3:00 p.m
C00507 00274 ∂04-Mar-80 1324 ADMIN.MO at SU-SCORE (Maurice Bizzarri)
C00508 00275 ∂04-Mar-80 2158 JK knowledge representation
C00509 00276 ∂05-Mar-80 1114 HVA AUDITRON
C00510 00277 ∂05-Mar-80 1208 Denny Brown <CSD.DBROWN at SUMEX-AIM> CS206 for fall quarter
C00511 00278 ∂05-Mar-80 1331 MP
C00512 00279 ∂06-Mar-80 0200 JMC*
C00513 00280 ∂06-Mar-80 0837 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Admissions (from Gene Golub)
C00514 00281 ∂06-Mar-80 0851 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) [CSD.BECK: cs390]
C00515 00282 ∂06-Mar-80 0856 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Admissions
C00518 00283 ∂06-Mar-80 0900 JMC*
C00519 00284 ∂06-Mar-80 0900 JMC*
C00520 00285 ∂06-Mar-80 0912 BOBROW at PARC-MAXC2 Re: Sloman
C00521 00286 ∂06-Mar-80 0919 PJH
C00522 00287 ∂06-Mar-80 0927 Denny Brown <CSD.DBROWN at SU-SCORE>
C00523 00288 ∂06-Mar-80 0926 JK remarks
C00525 00289 ∂06-Mar-80 1018 DBL Janet Kolodner
C00528 00290 ∂06-Mar-80 1631 MP phone call
C00529 00291 ∂06-Mar-80 1750 LGC Design Reference?
C00532 00292 ∂07-Mar-80 0243 ARK S-1 Meeting Notice
C00533 00293 ∂07-Mar-80 0800 JMC*
C00534 00294 ∂07-Mar-80 0844 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: 4331
C00536 00295 ∂07-Mar-80 0852 BTH Concurrent Programming Seminar
C00537 00296 ∂07-Mar-80 0909 JMC*
C00538 00297 ∂07-Mar-80 0909 JMC*
C00539 00298 ∂07-Mar-80 0909 JMC*
C00540 00299 ∂07-Mar-80 1039 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) IBM machine and use
C00543 00300 ∂07-Mar-80 1038 HVA Curriculum Vita
C00544 00301 ∂07-Mar-80 1048 Hans.Berliner at CMU-10A (C350HB03) Committee for Fredkin Prize rules
C00548 00302 ∂07-Mar-80 1253 FWH PV+A Seminar
C00550 00303 ∂07-Mar-80 1426 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) [Terry Winograd <TW at SU-AI>: Response to note on admissions ]
C00553 00304 ∂07-Mar-80 1435 LES ARPA final report reprise
C00554 00305 ∂07-Mar-80 1639 100 : monic my leaving
C00555 00306 ∂08-Mar-80 1636 MINSKY at MIT-AI (Marvin Minsky)
C00566 00307 ∂08-Mar-80 2304 MINSKY at MIT-AI (Marvin Minsky)
C00567 00308 ∂09-Mar-80 2352 JK knowledge representation
C00568 00309 ∂10-Mar-80 0720 Allen.Newell at CMU-10A (A310AN02) aim 326
C00569 00310 ∂10-Mar-80 1000 JMC*
C00570 00311 ∂10-Mar-80 1000 JMC*
C00571 00312 ∂10-Mar-80 1103 doyle at MIT-AI (Jon Doyle) Job offer
C00573 00313 ∂10-Mar-80 1914 Denny Brown <CSD.DBROWN at SU-SCORE> Proposed new Masters program.
C00576 00314 ∂11-Mar-80 1048 HVA Telephone Call
C00577 00315 ∂11-Mar-80 1152 TOB Dave Marimont
C00578 00316 ∂11-Mar-80 1402 BTH a note from S. Owicki
C00581 00317 ∂11-Mar-80 1641 EGL OUTGOING MAIL
C00582 00318 ∂12-Mar-80 0005 HPM HPM orals are today
C00583 00319 ∂12-Mar-80 0131 ME date and time
C00585 00320 ∂12-Mar-80 1201 LES Frances Larson reference
C00587 00321 ∂12-Mar-80 1342 EGL CORRESPONDENCE
C00588 00322 ∂12-Mar-80 1409 ME date and time
C00589 00323 ∂12-Mar-80 1753 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Courtesy Parking Permits
C00590 00324 ∂12-Mar-80 1855 LLW Skyhooks
C00593 00325 ∂12-Mar-80 2119 MINSKY at MIT-AI (Marvin Minsky)
C00597 00326 ∂12-Mar-80 2339 TOB display system
C00599 00327 ∂13-Mar-80 0000 LLW A Promise of Joy!
C00613 00328 ∂13-Mar-80 0006 LLW Yet Another Try. . .
C00616 00329 ∂13-Mar-80 0139 MINSKY at MIT-AI (Marvin Minsky)
C00618 00330 ∂13-Mar-80 0902 JMC*
C00619 00331 ∂13-Mar-80 0902 JMC*
C00620 00332 ∂13-Mar-80 1545 TOB
C00621 00333 ∂13-Mar-80 1600 JMC*
C00622 00334 ∂13-Mar-80 1628 HVA YOUR CORRESPONDENCE
C00623 00335 ∂14-Mar-80 0253 RMS at MIT-AI (Richard M. Stallman)
C00629 00336 ∂14-Mar-80 1101 EGL mental.xgp[w80,jmc] file
C00632 00337 ∂14-Mar-80 1531 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) our meeting (right now)
C00633 00338 ∂14-Mar-80 1601 JMC*
C00634 00339 ∂14-Mar-80 1636 FWH PV+A Seminar
C00635 00340 ∂15-Mar-80 2032 MINSKY at MIT-AI (Marvin Minsky)
C00636 00341 ∂16-Mar-80 0249 LLW Getting Together
C00638 00342 ∂16-Mar-80 0251 LLW Getting Together, contd
C00639 00343 ∂16-Mar-80 1209 MINSKY at MIT-AI (Marvin Minsky)
C00640 00344 ∂16-Mar-80 2013 MINSKY at MIT-AI (Marvin Minsky)
C00641 00345 ∂16-Mar-80 2128 LGC Wednesday?
C00642 00346 ∂17-Mar-80 0744 Siegman at SUMEX-AIM Symposium paper
C00644 00347 ∂17-Mar-80 1200 Bair at SUMEX-AIM Re: Symposium paper
C00646 00348 ∂17-Mar-80 1632 CSD.ROSENOW at SU-SCORE Margaret Jacks Hall Dedication
C00648 00349 ∂17-Mar-80 1845 JMC
C00649 00350 ∂18-Mar-80 0708 JRA reviewing
C00650 00351 ∂18-Mar-80 0830 JMC*
C00651 00352 ∂18-Mar-80 1036 100 : lyn larson Reports
C00652 00353 ∂18-Mar-80 1101 BS Personnel Requisition
C00653 00354 ∂18-Mar-80 1140 HVA Frances Larson-Candidate interviewed for JMC Secty position
C00654 00355 ∂18-Mar-80 1312 EGL your reports
C00655 00356 ∂18-Mar-80 1401 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: Marchuk
C00656 00357 ∂19-Mar-80 0000 JMC*
C00657 00358 ∂19-Mar-80 0029 LLW Hans for Dinner?
C00658 00359 ∂19-Mar-80 0047 FORWARD at USC-ECL M&M's Dynamic Compression Columns
C00661 00360 ∂19-Mar-80 0834 REG
C00662 00361 ∂19-Mar-80 1825 TOB resume for NSF robotics proposal
C00663 00362 ∂19-Mar-80 2212 ADMIN.BOSACK at SU-SCORE Project writeup
C00664 00363 ∂20-Mar-80 1136 HVA FRANCES LARSON--Reconsideration
C00666 00364 ∂20-Mar-80 1251 Gardner at SUMEX-AIM Re: people interested in formalized legal reasoning
C00667 00365 ∂20-Mar-80 1509 LES
C00669 00366 ∂20-Mar-80 1649 HVA NSF RENEWAL AND SUPPLEMENT
C00670 00367 ∂20-Mar-80 2314 LGC Engelmore Document
C00672 00368 ∂20-Mar-80 2319 HPM
C00673 00369 ∂21-Mar-80 0900 JMC*
C00674 00370 ∂21-Mar-80 0926 Bair at SUMEX-AIM Title and paper
C00676 00371 ∂21-Mar-80 1023 DCD
C00677 00372 ∂21-Mar-80 1038 JRA reviews
C00678 00373 ∂21-Mar-80 1052 Siegman at SUMEX-AIM Symposium paper--getting it to us
C00679 00374 ∂21-Mar-80 1119 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Faculty Meeting, April 1
C00680 00375 ∂21-Mar-80 1413 HPM RSVP
C00681 00376 ∂21-Mar-80 1430 DCD $
C00682 00377 ∂21-Mar-80 1604 Sandelin at SUMEX-AIM symposium dinner
C00683 00378 ∂21-Mar-80 1642 Sandelin at SUMEX-AIM symposium dinner
C00684 00379 ∂21-Mar-80 1732 RWG at MIT-MC (Bill Gosper)
C00685 00380 e21-Mar-80 1736 HPM Here is message that was mis-directed
C00686 00381 ∂21-Mar-80 1937 LLW Saturday Scheduling
C00687 00382 ∂21-Mar-80 2324 HPM
C00688 00383 ∂22-Mar-80 0913 Sandelin at SUMEX-AIM
C00689 00384 ∂22-Mar-80 1219 HVA Key for Haugeland
C00690 00385 ∂22-Mar-80 1405 LLW Scheduling
C00691 00386 ∂22-Mar-80 1545 TOB
C00693 00387 ∂22-Mar-80 2236 Kanerva at SUMEX-AIM Criteria for Usefulness of Computers in Offices
C00696 00388 ∂23-Mar-80 0414 LLW (Non-)Spooling of Skyhook Paper Draft
C00697 00389 ∂23-Mar-80 1147 Kanerva at SUMEX-AIM OFFICE[W80,JMC]
C00699 00390 ∂23-Mar-80 1634 Kanerva at SUMEX-AIM Long Message
C00700 00391 ∂23-Mar-80 2034 Kanerva at SUMEX-AIM OFFICE[W80,JMC]
C00702 00392 ∂24-Mar-80 0651 Kanerva at SUMEX-AIM Thank you very much!
C00703 00393 ∂24-Mar-80 0846 HVA Key to Bldg. 460--M. Jacks Hall
C00704 00394 ∂24-Mar-80 0927 CLT
C00705 00395 ∂24-Mar-80 1031 LES Account request
C00708 00396 ∂24-Mar-80 1400 JMC*
C00709 00397 ∂24-Mar-80 1402 FFL phone call.
C00710 00398 ∂24-Mar-80 1546 FFL Memos to Cohen, Dugan, et al
C00711 00399 ∂24-Mar-80 1712 CET via SU-SCORE grades
C00712 00400 ∂25-Mar-80 0129 Mark Crispin <Admin.MRC at SU-SCORE>
C00713 00401 ∂25-Mar-80 1307 Allen.Newell at CMU-10A (A310AN02) Re: AI Memo 326
C00718 00402 ∂25-Mar-80 1330 FFL PHON CALL
C00720 00403 ∂25-Mar-80 2200 JMC*
C00721 00404 ∂26-Mar-80 1148 HVA Telephone Call from Brian Randall
C00722 00405 ∂26-Mar-80 2113 MINSKY at MIT-AI (Marvin Minsky) skyhookery
C00727 00406 ∂27-Mar-80 1013 DCD
C00728 00407 ∂27-Mar-80 1124 DCD
C00729 00408 ∂27-Mar-80 1600 JMC*
C00730 00409 ∂27-Mar-80 1607 FFL phone call
C00731 00410 ∂28-Mar-80 1151 BERN at MIT-DMS (Bahram Niamir) The home information terminal
C00738 00411 ∂28-Mar-80 2156 AS challenge
C00745 00412 ∂29-Mar-80 1116 MINSKY at MIT-AI (Marvin Minsky)
C00746 00413 ∂29-Mar-80 1514 DCD travel reimbursement
C00747 00414 ∂29-Mar-80 1603 AS continued
C00748 00415 ∂29-Mar-80 1608 AS apology
C00750 00416 ∂29-Mar-80 1757 MARIMONT at SRI-KL Grad School
C00751 00417 ∂29-Mar-80 1912 AS hold
C00752 00418 ∂29-Mar-80 1909 LGC
C00753 00419 ∂29-Mar-80 1918 AS Self control!
C00754 00420 ∂30-Mar-80 1513 LES Engelmore visit
C00756 00421 ∂31-Mar-80 0000 LGC
C00757 00422 ∂31-Mar-80 0709 LGC SNAPOV
C00760 00423 ∂31-Mar-80 1100 BERN at MIT-AI (Bahram Niamir)
C00761 00424 ∂31-Mar-80 1654 AS Trip to San Diego
C00762 00425 ∂31-Mar-80 1932 TOB
C00763 00426 ∂31-Mar-80 2210 AS intentionality
C00765 00427 ∂31-Mar-80 2357 LLW More Shyhookery
C00778 00428 ∂01-Apr-80 0000 JMC*
C00780 00429 ∂01-Apr-80 0938 AS
C00781 00430 ∂01-Apr-80 0957 ES Mt. Shasta
C00782 00431 ∂01-Apr-80 1027 MINSKY at MIT-AI (Marvin Minsky)
C00790 00432 ∂01-Apr-80 1408 LES retransmission
C00793 00433 ∂01-Apr-80 2140 LES
C00794 ENDMK
C⊗;
∂03-Jan-80 1705 Sandelin at SUMEX-AIM text handling and delivery
Date: 3 Jan 1980 1706-PST
From: Sandelin at SUMEX-AIM
Subject: text handling and delivery
To: jmc at SAIL
1. IBM - They were to set up meeting regarding 4331. No word. I plan to
bring up with Steve Williams. I will let you know the results. 2. An
intrium report on text handling is now done. It as yet does not have a
definitive strategy, as some research remains to be done. We needed to
put something together for the visiting advisory committee. I would like
to send you the technical stategy portion. The entire report is available
if you wish to see it, but it runs to almost 200 pages and is a bit
disorganized at this stage. 3. After seeing the material, Pentti, Jeannie
Siegman, and I are available to discuss some ideas not as yet surfaced, if
you think it worthwhile. Suppes has seen the technical stategy section
and some interesting ideas are emerging.
-------
∂03-Jan-80 2309 LGC Tomorrow @ 2:00?
Are we still on at 2:00 tomorrow? There has been significant progress since
we last talked, and we will have a lot to discuss. I still hope to get one
more major point written up before we talk, so I'll try to XGP the new
material just before we meet.
∂04-Jan-80 0756 LGC Tuesday at 2:00?
OK. How about Tuesday at 2:00?
∂04-Jan-80 1109 JK
would like to get together some time; what would be a good time?
∂04-Jan-80 1302 ROY on TTY20 1302
It seems that the picture was too large because the fan had stoped and the
amplifiers were overheating. The fan had a bad connection at the plug
which I think I have fixed. If it behaves like this again let me know and
I'll have Don install a new plug.
∂04-Jan-80 1424 LES New DM modem
Your Datamedia terminal now has the modified modem installed. Please
test it as many ways as you can.
∂05-Jan-80 1433 MINSKY at MIT-AI (Marvin Minsky)
Date: 5 JAN 1980 1733-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
Hi. Local historian of RLE wants to know if we (AI group)
even had space in RLE. As I recall, we started in basement of Bldg. 26.
Right? See u next week.
--marvin
Right. It was 26-007 or perhaps 26-007A or 26-007B.
∂05-Jan-80 1642 FWH via AMES-TIP Program Verification seminars
To: "@SEM.DIS[PV,FWH]" at SU-AI
This is an announcement of the first Program Verification and Analysis Seminar
this quarter. If you prefer not to receive future announcements or know of
someone who would like to be on the mailing list, please send me a message.
Frieder von Henke (FWH at SU-AI)
---------------------------------
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:15 p.m., Tuesday, January 8
SPEAKER: Bob Paige, Rutgers University
TITLE: Plans to implement formal differentiation--a technique for
improving algorithms specified at an abstract level
ABSTRACT:
Formal differentiation is a basic program optimization
technique which generalizes the classical method of 'reduction in
operator strength.' This technique captures a commonly occurring yet
distinctive mechanism of program construction in which succinct
algorithms involving costly repeated calculations are transformed into
more efficient incremental versions.
We have just initiated a project to implement set theoretic
formal differentiation capable of applying a host of program transformations
extending Jay Earley's 'iterator inversion'. The first stage
of our project calls for a semiautomatic implementation supported
by an experimental interactive program manipulation system similar to
ones developed by Standish and Loveman, but using a simplified variant
of SETL as a source language. Use of this initial system is expected
to supply empirical evidence leading to the development of a fully
automatic version which can be installed as part of a conventional
optimizing compiler.
This work is relevant to the development of high level
languages and to program correctness technology.
∂05-Jan-80 2040 REM via AMES-TIP CPU TIME
Oops, I was exciting a bug in MacLISP whereby at end of file the program
goes into a compute loop from which it can't be interrupted by usual
MacLISP mechanisms. While it was in that state, the TIP hung up and/or
the system crashed (I forgot which came first, this was a bad day), and
when the TIP hung up my telephone went out of order for four hours.
Anyway in all that confusion I forgot or was unable to get back to
kill my running job and it accumulated 5 hrs 38 min of cpu time. If
I show up as number one user for the month, it was an accident.
I have now killed it as soon as I noticed it upon login tonight.
∂06-Jan-80 1400 ZM Computer Forum
To: JMC, DCL
I must have the name of your student and the title of his/her talk by tomorrow.
Thanks Zohar
∂06-Jan-80 1728 ME modem interrupts on TTYs 20,21,22
To: ROB
CC: REG, ROY, JMC
The reason JMC's Imlac is getting detached is that it is one of these
three DCA ports, which are seeing lots of modem interrupts, and
interestingly enough, they all see interrupts at the same time, about
every 6 seconds or so, probably in time with some wholine activity. This
suggests indeed that there is a problem with the CD, DSR and RI lines for
these DCA ports. I suggest that those signals should be tied to DTR, as
on the LLL lines. Also, the same signals should be tied to DTR for LLL
ports 40 and 41 (ext clock ports).
∂06-Jan-80 1947 FWH PV seminar -correction
To: "@SEM.DIS[PV,FWH]" at SU-AI
In correction to my previous message and possibly some posted
announcements: the first Program Verification and Analysis Seminar
will start at 2:30 p.m. Thus here are the correct data:
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, January 8
SPEAKER: Bob Paige, Rutgers University
TITLE: Plans to implement formal differentiation--a technique for
improving algorithms specified at an abstract level
∂07-Jan-80 1039 CLT Calendar
We have tickets for the following event:
Thu. Jan. 17 2000 Joffrey Ballet (Memaud)
∂07-Jan-80 1511 Gardner at SUMEX-AIM Re: Your contract example
Date: 7 Jan 1980 1511-PST
From: Gardner at SUMEX-AIM
Subject: Re: Your contract example
To: JMC at SU-AI
cc: gardner
In response to your message sent 06 Jan 1980 2021-PST
Here's the contract problem you asked for:
On July 1 Buyer sent the following telegram to Seller:
"Have customers for salt and need carload immediately. Will
you supply carload at $2.40 per cwt?" Seller received the
telegram the same day.
On July 12 Seller sent Buyer the following telegram,
which Buyer received the same day: "Accept your offer carload
of salt, immediate shipment, terms cash on delivery."
On July 13 Buyer sent by Air Mail its standard form
"Purchase Order" to Seller. On the face of the form Buyer
had written that it accepted "Seller's offer of July 12" and
had written "One carload" and "$2.40 per cwt." in the approp-
riate spaces for quantity and price. Among numerous printed
provisions on the reverse of the form was the following:
"Unless otherwise stated on the face hereof, payment on all
purchase orders shall not be due until 30 days following
delivery." There was no statement on the face of the form
regarding time of payment.
Later on July 13 another party offered to sell Buyer
a carload of salt for $2.30 per cwt. Buyer immediately
wired Seller: "Ignore purchase order mailed earlier today;
your offer of July 12 rejected." This telegram was received
by Seller on the same day (July 13). Seller received Buyer's
purchase order in the mail the following day (July 14).
Briefly analyze each of the items of correspondence
in terms of its legal effect, and indicate what the result
will be in Seller's action against Buyer for breach of
contract.
If this isn't the part of the paper you had in mind, the whole thing
is LAW.1[1,AVG].
When will the lecture be? I'd like to come.
Anne
-------
∂07-Jan-80 2143 MINSKY at MIT-AI (Marvin Minsky)
Date: 7 JAN 1980 2353-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
Meeting changed to 2 days, thus. and fri. Anyone who wants to can
stay another, but only Selfridge plans to. Ed still
in Carribean -- can get phone no. there if you want.
Otherwise call him at 277-4444 in brookline.
∂07-Jan-80 2234 LGC via SU-TIP
Shall we meet tomorrow (Tuesday) at 2:00?
∂08-Jan-80 0841 REG Your Imlac
I found a wiring error that I made yesterday. I have fixed it; hopefully,
your Imlac works now. If not, try to contact me or Roy as soon as
possible.
Sorry for the inconvenience.
∂08-Jan-80 0842 REG LOTS, etc.
1. What axes shall we grind at the Advisory Committee on Computer Science?
They meet Sunday thru Tuesday next week. (1/13 to 1/15)
2. The NSF has a program to grant funds for the upgrade and construction
of facilities for undergraduate science education. Do you think LOTS
would qualify? The NSF program requires matching funds. Should
we bother to write a proposal? I have in mind some money for terminals.
∂08-Jan-80 1339 BTH FYI
To: "@GROUP.DIS[1,BTH]" at SU-AI
Those interested in the Concurrent Programming Seminar may find this
course interesting:
Topics in Theory of Computation - CS256 Winter Quarter
Instructor: Zohar Manna
TA: Pierre Wolper
Time: Tuesday & Thursdays 9:30-10:45 am
DESCRIPTION: Advanced topics in the theory of computation: program
schemes, fixed-point theory, semantics of programming languages, the modal
logic of programming, verification of parallel programs, situational
calculus. (No computability of complexity will be discussed.)
Prerequisite: CS 156 or equivalent.
∂08-Jan-80 1326 LGC via SU-SCORE
It looks as if I won't make it until 2:15 or so, since I am stopping by
MJH on the way over to pick up some hardcopy for us to look at. If you wish
to look at some of the new material on-line before I arrive, it can be found
interspersed throughout pages (not sections) 2 and 4:9 of PRBSLV.TXT[1,LGC].
Much, if not most of the material on these pages is new. By the way, your
on-line comments on the material I had last time were quite helpful. -- Lew
∂08-Jan-80 1340 CG reference
Job hunting time for me has at last actually arrived. So far I've applied to
Cornell, Courant and IBM, and expect to apply several other places as well.
I'm listing you as a reference. IBM and Cornell will contact you themselves,
but Courant asked that I arrange for recomendations to be sent to them.
So, would you send a recomendation to:
The Committee on Visiting Memberships,
Courant Institute
251 Mercer St.
New York, New York 10012
By the way, the Courant Job which I'm applying for is a one-year postdoc.
(All research and no teaching). Thank you very much for doing this for me.
∂08-Jan-80 1440 LGC I'm still alive
The XGP gave trouble; I have the listings now, and will be out as soon
as I can get there.
∂08-Jan-80 2019 JRA lisp talk
woould you consider giving the keynote speech at the lisp cnference
(aug 24, 1980, monday morning)? it seems appropriate since all this
lisp stuff is your fault
john
Yes, I will do it.
Please check your calendar. August 24 seems to be a Sunday.
∂09-Jan-80 1102 JMC*
Tako Protopapas 703-860-0220,Pres.Fuel and Mineral Resources. Inc.
∂09-Jan-80 1915 ADMIN.BOSACK at SU-SCORE Characterization of Editing Users
Date: 9 Jan 1980 1720-PST
From: ADMIN.BOSACK at SU-SCORE
Subject: Characterization of Editing Users
To: JMC at SU-AI
cc: Admin.Gorin at SU-SCORE
After thinking about what information could be gathered about people's
editing behavior, I conclude it is appropriate to start small.
For each program declared 'interesting', we can get histograms (and hence
mean and variance) of time between interactions (think time) and the
runtime per interaction.
I would gather this information by writing a special purpose program using
the SNOOP facility to watch processes as they long term block and unblock.
When a process blocked or unblocked, its declared name would be checked against
our list of interesting programs and if interesting the appropriate data
recored in the appropriate histogram set. This data gathering program would
be usable on any Release 4 TOPS20 system (SCORE and GSB now, LOTS within the
next few weeks).
Once we have the basic program going, a great variety of other data
can be gathered: working set sizes, page faults per interaction,
characters input and output per interaction, etc. I think we should
go for the basic model first. It is very easy to generate reams of
performance data and much harder to understand the data throughly.
Given the first few moments of the think and compute distributions, we should
be able to cons up a decent queuing model for estimating the user capacity
of a given system.
(Interesting Distraction: Kirk Lougheed, the current LOTS maintainer of
EDIT added a bit of code to gather a frequency histogram of editor commands.
Not too surprising, various commands for printing things out were by far the
most popular.)
-------
Suggestions:
1. Include a counter so that statistics are gathered only every
nth time the gatherer is called. This will enable tuning the gatherer
so it won't slow down the system being studied.
2. Will there be any problem in noticing calls from a system
service whose action is in turn instigated by the program
being studied? Such may take more time than the editor itself.
3. Will we be able to identify the inner loops that are using
the time - especially loops involving i-o?
∂10-Jan-80 0741 Feigenbaum at SUMEX-AIM (Response to message)
Date: 10 Jan 1980 0741-PST
From: Feigenbaum at SUMEX-AIM
Subject: (Response to message)
To: JMC at SU-AI
In response to your message sent 09 Jan 1980 1910-PST
Yes, thank you for thinking about the CS reports going to the Soviet
Union. That had slipped my mind. If I take a decisive step unilaterally,
I may create a storm in the faculty. What I propose to do is to institute
a moratorium on mailings and get a faculty vote on the 22nd.
ed
-------
∂10-Jan-80 0803 Feigenbaum at SUMEX-AIM lunch with Jon Doyle
Date: 10 Jan 1980 0757-PST
From: Feigenbaum at SUMEX-AIM
Subject: lunch with Jon Doyle
To: lenat, tw at SAIL, zm at SAIL
cc: jmc at SAIL, winograd at PARC
Will you be able to join McCarthy and me for lunch with Doyle on
Wednesday, Jan. 16 at noon at the faculty club? RSVP
Ed
-------
∂10-Jan-80 0928 JRA lisp conf
excellent! indeed, aug 24 IS sunday; i thought of that as i was driving home.
john
May I assume the conference starts on the 25th then?
∂10-Jan-80 1100 JMC*
copy of 2 puzzles. - maybe Ma's version.
∂10-Jan-80 1058 JRA lisp conf
the sechdule calls for a reception sunday night, aug 24, with the
actual conference monday through wednesday (aug 25-27)
and a banquet tuesday evening, with a panel discussion tuesday afternoon.
∂10-Jan-80 1124 JMC dugan
tell dugan to give prospectus to
Richard Clarke and Grant Horne with
the idea that they should show it to relevant
people at Bechtel
∂10-Jan-80 1129 JMC
Prospectus to Tom Pigford, nuclear eng. Berkeley 642-6469
∂10-Jan-80 1309 SEK
To: "@FOL.DIS[FOL,RWW]"
Dear FOL participant,
I am currently collecting references to significant projects which have
been done in FOL -- either applications or extensions to FOL itself -- as
part of a project to document FOL. (Lest we forget -- right now there is
no one place where this is collected.) Things I'm looking for include
date, author(s), background references, intent/scope, pointers to files,
and explanatory examples. Examples are particularly important, as I would
like to make a series of short on-line demos which show off different uses
of FOL to outsiders. Documentation on FOL projects will accumulate in
file PROJEC[FOL,SEK]. Please feel free to look it over, add information,
mail me pointers to files, suggestions...whatever is appropriate. Project
descriptions don't have to be comprehensive, just enough to be able to
reconstruct what happened. --Scott
∂10-Jan-80 1357 LGC
1. You might be interested in taking a quick look at PRBSLV.TXT[1,LGC]/5P/168L ,
to end of page, which contains an answer to one of the questions you raised
on Tuesday.
2. Would it be of any help to you if you had a brief outline and guide to
the changes in that file that were new as of last Tuesday?
3. The blurb in REASON.ABS[BIB,DOC] on Jon Doyle's talk next Wednesday
looks quite interesting; whether or not he does something really good with
it, he has his hands on a very important problem, which I hope to deal
with in the design of the problem solver.
∂10-Jan-80 1513 BTH Concurrent Programming Seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
The Concurrent Programming Seminar will meet Thursdays at 2:30pm in
ERL237. There will be no meeting Jan 17 due to EE Quals. Toshi Minoura
will speak Jan 24 on his thesis (abstract to follow).
∂11-Jan-80 0903 REG
I'll be around MJH all day on Friday. I expect I'll be around somewhat
on Saturday and Sunday. Jack Alpert (a grad student in school of Ed, and
a LOTS fanatic) is preparing a letter to the Advisory Committee: a
student's view of word processing. I'll see if I can get a copy.
My ideas for what I'd like to see Stanford do over the next couple of
years are enumerated below. These statements need quite a lot of
polishing before sending them off to anyone.
(These are somewhat longer than I intended in this note. My apologies)
1. Provide more service in a LOTS-like environment. Some
courses, notably in the School of Engineering, but also some
advanced Computer Science courses need more computes per
terminal than is presently available at LOTS. A second
system that has a more stringent accounts policy could be
provided in which the ratio of mainframe hardware to terminals
is higher than at LOTS.
2. In the context of a second system, or even as an augmentation to
LOTS, the ability to do graphics, i.e., line drawings on some terminals
would be a great enhancement to the educational uses of the system.
Other engineering schools, Rensselear for example, have a significant
interactive graphics facilities that are used both for instruction
(graduate and undergraduate) and for research.
3. Word processing capabilities for students. This could be
piggybacked on LOTS, or have a system of its own. On a
separate system, word processing could be made self-supporting,
or nearly so, on a fee for service basis. The benefits of
word processing are twofold. First, the productivity of
students can go up; for a given task, the time to complete
may be smaller. In a world where there is more to learn, more
to do, every year, this augmentation of each student's productivity
can provide more time to further his or her education.
Second, the quality of work can be expected to go up: the
effort of retyping an entire page will no longer be a barrier
to the creation of excellent written materials.
4. Networks to attach the various computer facilities together.
Partly, this effort is underway already. The advantages that
come about through networking include better distribution of
written communications, better utilization of existing resources,
less demand for duplication of specialized resources (e.g., laser
printer, mass store). The networking effort should be expanded
to embrace all large computers, most of the terminals, and
the various word processing systems, especially all future
word processing systems.
5. Although microprocessors and personal computers are certainly
in our future, there are several reasons for us to delay our
involvement in this area. When the following criteria are met,
micros will be a much better idea:
a. Micros must be connected via some network to a
central file store. Local file storage on micros is
presently too unreliable and too expensive to be the
only filing system available.
b. The network connection is important for a second
reason. It allows centralized distribution of new
software. Imagine 300 floppies containing PASCAL
scattered all over campus; now think of the problem
of distributing a bug fix. The cost of administering
a fleet of micros can be very high unless a network
is provided.
c. It must be recognized that although micros may be
an adequate tool for many programming tasks, e.g.,
introductory programming, they cannot solve every
programming problem. Therefore, the larger machines
will continue to be necessary. A problem that must
be avoided is the need for total retraining when
a micro user has to move to a larger machine.
We must have a user environment in the micro that
is a subset of the environment of the larger machines.
We must have compatible editors and compatible languages.
There is a large software effort needed before we can integrate micros
into the educational enviroment. LOTS certainly can't do it; perhaps
CSD could do it, but they may not be interested.
∂11-Jan-80 1402 FWH PV+A Seminar
To: "@SEM.DIS[PV,FWH]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, January 15
SPEAKER: David Luckham, Stanford University
TITLE: The Ada Error Mechanism and how to control it.
ABSTRACT:
A method of documenting exception propagation and handling in
sequential Ada programs is proposed.
Exception propagation declarations are introduced as a new
component of Ada specifications. This permits documentation of those
exceptions that can be propagated by a subprogram. Exception handlers
are documented by entry assertions.
Axioms and proof rules for Ada exceptions are given. These
rules are simple extensions of previous rules for Pascal and define an
axiomatic semantics of Ada exceptions. As a result, Ada programs
specified according to the method can be analysed by formal proof
techniques for consistency with their specifications, even if they
employ exception propagation and handling to achieve required results
(i.e. non error situations). Example verifications are given.}
∂11-Jan-80 1403 TOB terminal system
John
I have made a survey of commercial video switches.
There still are some threads to follow up and one
technical question about the possibility of using
telephone reedswitch matrix switches. If that is
possible, we can get quite low cost switches of
various sizes. I will let you know about further
developments.
I suppose the next step is contacting memory suppliers.
Shall I do it?
Tom
∂11-Jan-80 1750 ADMIN.BOSACK at SU-SCORE
Date: 11 Jan 1980 1747-PST
From: ADMIN.BOSACK at SU-SCORE
To: JMC at SU-AI
cc: ADMIN.BOSACK at SU-SCORE
In-Reply-To: Your message of 9-Jan-80 1924-PST
My first guess on how much we will impact the running system goes like this:
Assuming we try for every sample and that every block/unblock is an interesting
program (worst case) and 100 block/unblocks per second (LOTS can be almost
this bad) and around 30 instructions at the breakpoints gives 3000 instr/second
or less than 0.30% of the machine. Running a SNOOP program locks down about
8 additional pages (of over 1900 on SCORE and LOTS).
If we want to use more expensive breakpoints, say 100 instructions, we would
probably have to go to a 1 of N sample scheme.
At the level we will be looking, ALL process activity will be visible - no
matter how deeply nested the system calls get.
Discovering where the bad guys inner loop is can be harder to do. Usually
by looking at a trace of the programs monitor calls in response to a
suspect command one can find the problem. As we get basic data on the
way people use the various editors, we can start looking at the programs'
actual use of the system calls - one sometimes finds realy awful or
just plain stupid things.
-------
∂12-Jan-80 1459 LES Datamedia
To: PJH, JMC
I guess that we should try adjusting the transmission level on the modem,
which is the feature they added. To do this, remove the front panel and
find the adjustable potentiometer that is mounted on the edge of one of
the cards. With the telephone receiver "on hook", adjust the pot downward
(reducing gain) to see if the garble rate reduces. If you go too far,
the modem at the other end will disconnect and the "C" light on the front
will go out.
See if this helps.
∂13-Jan-80 2031 JK
Thanks for the circumscription paper.
Technical comments:
(1) The FOL REFLECT command would give a simple way to use this idea
(2) About the definition of submodel relative to a predicate in section 6:
For the sake of completeness you should also say what happens to
function symbols. Also, as a logician I would be more comfortable
with the word interpretation instead of extension.
(3) I don't understand the parenthetical comment about Goedel-Bernays
set-theory; one doesn't need stronger axioms to give a class model of
set-theory, say the constructible universe L. Circumscription to L
would give the axiom of constructibility.
The remark following this comment concerning model theory seems equally
obscure.
I was probably much too sketchy in describing how to use my decision procedure.
In more detail:
R MMFOL
(DSKIN (DIRECT.LSP))
(FOO)
FETCH DIR.TST;
ASSUME <fla to be tested>;
REFLECT DEC,<the line no>;
The intent about GB set theory was merely that it has a finite number of
axioms. I haven't worked out what happens if one circumscribes a set
of axioms that already uses schemata. Thanks for the comments. I will
try to make the forthcoming AI memo version less cryptic.
∂13-Jan-80 2217 LGC Monday at 2pm?
Shall we meet as scheduled at 2:00 tomorrow? I have found and fixed a number
of potentially confusing bugs in PRBSLV.TXT, including a couple in the passage
I referred you to in my msg of 10-Jan-80, and there is also some other new
material. So I'll try again to have fresh hardcopy for us to peruse.
∂14-Jan-80 0000 JMC*
mints
∂14-Jan-80 1006 REG
The Advisory Committee meets in Durand 450 tomorrow morning
I'm not sure at what time LOTS will be discussed. At 8:30 Linvill
speaks about CIS. LOTS is after that.
∂14-Jan-80 1111 JMC*
israel
∂14-Jan-80 1200 JMC*
Now is the time to make a memo of circum.
∂14-Jan-80 1413 LGC
The hardcopy is in the XGP queue; while we're waiting for it to come out,
you might want to take a look at PRBSLV.TXT[1,LGC]/5P/170L , and the next
page as well.
∂14-Jan-80 1719 CSD.MCCARTHY at SU-SCORE barrett,kahn
Date: 14 Jan 1980 1540-PST
From: CSD.MCCARTHY at SU-SCORE
Subject: barrett,kahn
To: jmc at SU-AI
call them about advice taker support
-------
∂14-Jan-80 1811 ME running FIXIML from another TTY
To: JMC
CC: ROB
∂13-Jan-80 1857 ROB Fowarded from me to ME
∂13-Jan-80 1811 JMC doing it myself
How can I do for myself what you just did? Assume that I log in as myself
on my TI terminal.
ME - Well, you could convince PMF or someone to modify FIXIML to be able
to talk to some other TTY besides the one it is running on. It isn't possible
to duplicate Rob's actions in general.
∂15-Jan-80 0904 Feigenbaum at SUMEX-AIM an interesting terminal for CSD editing
Date: 15 Jan 1980 0900-PST
From: Feigenbaum at SUMEX-AIM
Subject: an interesting terminal for CSD editing
To: jmc at SAIL, les at SAIL, fb at SAIL
Date: 14 Jan 1980 1526-PST
From: Tucker
Subject: Barclay Terminal
To: Feigenbaum, Rindfleisch, admin.Gorin at SCORE
I have contacted the company which which Rand is dealing in an effort to
replace their old Ann Arbor large screed terminals with something of more
modern design. Since they have become accustomed to large format displays
they have avoided the common 80 column by 24 lined units. Rand is due to
get two demo units in "a few weeks" and after trying them out expect to
place an order for 40. The person at Rand in charge of this review is:
Bruce <Borden>@Rand-Unix
213-393-7672
The company involved is:
Barclay International
15541 Beach Blvd. (Suite A)
Westminster, CA 92683
714-972-1008
I talked to Vivian Nininger there.
The terminal involved really represents the low end of a word processor
system. In fact it can be gotten with up to 128k bytes of memory and
a variety of disks and such.
All of this is based on an 8080 microprocessor!
A wide variety of screen widths and heights are available; widths of 80,
136, or 164 characters are possible. A heights of 49 lines is available
with the 164 width and 60 lines with the 136 width.
Prices start under $3000. Fonts are not down line loadable but they
can be custom designed in PROMs. The keyboard has its own microprocessor
and thus keys can be redefined quite easily.
Literature is being sent to me and a member of the company could come
up to talk with us if we so wish.
Bob
-------
-------
∂15-Jan-80 0904 Feigenbaum at SUMEX-AIM Re: Barclay Terminal
Date: 15 Jan 1980 0905-PST
From: Feigenbaum at SUMEX-AIM
Subject: Re: Barclay Terminal
To: Tucker, Rindfleisch, admin.Gorin at SCORE
cc: jmc at SAIL, ;es at SAIL, les at SAIL, fb at SAIL
In response to the message sent 14 Jan 1980 1526-PST from Tucker
To me it looks very interesting. I would like to have one of their
salespeople come up here to talk. I like the idea of the 132chx60li
display. I like all the memory for keeping an editor therein.
RAND is putting the NED editor--incidentally, an extremely good editor--
into the terminal memory. Cassetes are available for local storage.
RAND is paying $2700 each ( I guess with sales tax that must be the $3000
mentioned).
What do you guys think? Should I get a salesperson up here?
Ed
-------
∂15-Jan-80 0950 JK
Sorry, should have been (DSKIN (DIR.LSP)).
Apropos our discussion on Ramseys Theorem, Bulnes and FOL:
I may be wrong about the capacity of FOL to handle such proofs elegantly;
My working hypothesis now is that a sufficiently rich meta theory
should do more for you than goal structuring. The sticky point, of course,
is to prove the correctness of the meta theoretic principles in question,
otherwise they will look like mere operations.
I'll test this thing out next week and see if I can come up with a better
proof of Ramsey given a meta theory of set theory that should have some
general interest.
∂15-Jan-80 1100 JMC*
Barrett, Kahn
∂15-Jan-80 1233 CLT tonight
I think I will check out the seminar that Jane mentioned, tonight at 7:30.
If you wish to go also, you should come home for supper suitably early,
or we can eat afterwards.
∂15-Jan-80 1236 TOB
Dr. Bernard Chern will visit Stanford
for one to two days during the week of Jan 28-31
to meet with investigators. He wants to present
information about funding opportunities with a
new NSF program, Computer Engineering. He also
would like input from investigators on research
directions for the program.
Please let me know about your interest in talking
with Dr. Chern, and when you are available.
∂15-Jan-80 1338 JMC michie address
Michie's current address is:
Donald Michie
Department of Computer Science
222 Digital Computer Laboratory
Urbana, Illinois 6l801
∂15-Jan-80 1555 FWH PV+A Seminar
To: "@SEM.DIS[PV,FWH]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, January 22
SPEAKER: Nori Suzuki, Xerox PARC
TITLE: Analysis of pointer rotation
ABSTRACT:
Pointer manipulation is one of the trickiest operations in
programming and is a major source of programming errors. A great deal
of research effort has been dedicated to making programs with pointers
more reliable. In this paper we will present pointer operations which
reduce conceptual difficulties when writing pointer programs, and
increase reliability of programs. We will analyze theoretically as well
as empirically why these operations are more convenient. Then, we will
define the safety of rotations, and show that it is decidable to check
whether the rotation is safe and also show that it is a good measure of
the correctness of programs. (That is, if the rotation is safe it is
highly probable that the program is correct, and if it is unsafe it is
highly probable that the program is incorrect.)
∂15-Jan-80 1902 TOB
To: "@CHERN.DIS[1,TOB]" at SU-AI
Dr. Bernard Chern will visit Stanford for one to two days during the week
of Jan 28-31 to meet with investigators. He will present information
about funding opportunities with a new NSF program which he will head,
Computer Engineering. The program will include computer architecture and
design aids, VLSI, knowledge engineering, robotics and geometric modeling.
I am arranging a schedule and appointments for him. How are Monday Jan 28
or Tuesday Jan 29?
Please include me in Chern's schedule at whatever afternoon time is
convenient.
∂16-Jan-80 0228 Mark Crispin <Admin.MRC at SU-SCORE> Dialnet mail
Date: 16 Jan 1980 0226-PST
From: Mark Crispin <Admin.MRC at SU-SCORE>
Subject: Dialnet mail
To: JMC at SU-AI
SCORE's ARPANET FTP server will now accept Dialnet traffic for LOTS. It
knows that it has to send it to SAIL to get it there (unfortunately SAIL
doesn't know what to do with it). If SCORE gets a Dialnet connection then
it will process the traffic itself.
This stuff, by the way, isn't really due to anything specifically for
Dialnet, but rather the general multiple-network mail delivery scheme I
mentioned to you over a year ago.
-------
I don't understand from this message how to send mail or files to
and from LOTS.
∂16-Jan-80 0842 Feigenbaum at SUMEX-AIM Jon Doyle
Date: 16 Jan 1980 0843-PST
From: Feigenbaum at SUMEX-AIM
Subject: Jon Doyle
To: jmc at SAIL, lenat, winograd, tw at SAIL, winograd at PARC,
To: zm at SAIL
Lunch is at the Faculty Club at approximately Noon. Downstairs level.
To John McCarthy: are you bringing Jon over to the lunch?
Ed
-------
∂16-Jan-80 0914 TOB
∂15-Jan-80 2030 JMC
What is Kanade's current address?
Prof T.Kanade
Dept of Inform Science
Kyoto Univ, Kyoto, Japan
He may be at CMU already.
He intends to return to CMU. I don't know when.
Dept of Computer Science
CMU
∂16-Jan-80 0949 MP John Doyle
John Doyle is here at the A.I. Lab and wishes to meet with you today.
∂16-Jan-80 1550 Mark Crispin <Admin.MRC at SU-SCORE>
Date: 16 Jan 1980 1530-PST
From: Mark Crispin <Admin.MRC at SU-SCORE>
To: JMC at SU-AI
In-Reply-To: Your message of 16-Jan-80 1314-PST
Well, you see, you can't send mail or files to and from LOTS. There
is no hardware or software at LOTS to do it. HOWEVER, a major hurdle
has been overcome. What the message means is that if you say
MAIL M.MRC@LOTS
to SCORE the mail program will accept it and queue it, and the mailer
daemon will try its damnest to deliver it. It will fail, because as
I said LOTS doesn't have any hardware/software (although it can get
the SCORE software). If that problem is overcome, then either SCORE
needs hardware, or SAIL needs its hardware reconnected and some
additional software work.
File transfer isn't covered in any of this. No progress has been made
on FTP. I don't see any reason to press Ignacio on FTP; he can't do
anything until the hardware is reconnected.
-------
∂16-Jan-80 1810 LES
∂16-Jan-80 1737 JMC
Feigenbaum says I should offer Doyle 25.3K + 1 day a week consulting right.
-----
The offer should make it clear that the 1 day of consulting is in addition
to a full-time effort, or are you talking about an 80% time appointment?
∂17-Jan-80 0148 LES DMs
To: REG at SU-AI, WP at SU-AI, MRC at SU-AI, JMC at SU-AI,
DCL at SU-AI, TOB at SU-AI, ZM at SU-AI
We did some tests this evening using SU-TIP line 326-9252. With an
old Datamedia with UDS modem, there were no garbles in about 10
screenfulls of output. Using the new Datamedia with Product Associates
modem, there was at least one garble on every screenful and often as
many as 5.
I plan to call Martin Hardy of PA today and invite him or his
representative over for a demonstration, and then do some more
negotiating.
∂17-Jan-80 0813 Sandelin at SUMEX-AIM misc
Date: 17 Jan 1980 0814-PST
From: Sandelin at SUMEX-AIM
Subject: misc
To: jmc at SAIL
1. Reminded Steve Williams of IBM about 4331 project..MEntioned interest
in doing student text processing experiment...He was going to discuss with
Bob Creasey at Ibm scientific center 2. Met with James Koerner ofthe Sloan
Foundation.. Told him that two areas where outside monies are needed are
~for student text} processing and to investigate wider collaboration on
what is needed in the area of inter-university cooperation. He invited
submission of a 3 to 4 page letter describing what should be done. We
would like to meet with you and Ralph before deciding if sending such a
letter is worthwhile. 4.A copy of the interium task force report was put
in your mail slot at MJH. Did you getit? -------
∂17-Jan-80 0824 REG
To: ROY at SU-AI, ROB at SU-AI, LES at SU-AI, JMC at SU-AI,
MRC at SU-AI
Roy,
Please plan to have the two 300-baud modems, the dialer, and the Vadic
Modem for dialnet connected to SAIL as soon as possible, and not later
than the end of January.
I suggest that you leave the 300-baud dialer for last; the other
connections will be somewhat simpler.
Ralph
∂17-Jan-80 0825 REG
To: JMC at SU-AI, MRC at SU-AI
John,
Is SCORE supposed to have a modem for Dialnet?
∂17-Jan-80 1149 CLT
try "di lifsch.*[let,jmc]"
∂17-Jan-80 1349 BTH Concurrent Programming Seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CONCURRENT PROGRAMMING SEMINAR
PLACE: ERL 237
TIME: 2:15 p.m., Thursday, January 24
SPEAKER: Toshi Minoura, Stanford University
TITLE: Resilient Extended True-Copy Token Algorithm
for Distributed Database Systems
ABSTRACT:
The resilient extended true-copy token algorithm is intended to
realize consistent operation of a distributed database system in the face
of site crashes and system partitioning. It combines a "true-copy token"
mechanism, a locking mechanism, and a version number mechanism to handle
different aspects of consistency and resiliency control.
The true-copy token mechanism is used to establish the logical
data. These logical data are global entities, and they are not bound to
any particular sites. Each true-copy token designates a physical data
copy that defines the current logical data value. Consistent execution of
transactions is realized by locking over the logical data thus defined. A
version number mechanism is used to properly sequence updates to physical
copies. Also version numbers are essential to regenerate lost true
copies.
The crash recovery problem is solved by switching from the lost
true-copy set to the "recovery set". The recovery set is computed from
the surviving "atomic update sets", which are atomically updated and
closely follow the true-copy set state. This new crash recovery scheme
does not use a log subsystem, and can handle a total site crash.
∂17-Jan-80 1601 BTH concurrent programming seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
Correction: Time should be 2:30 not 2:15.
∂18-Jan-80 0000 LES
∂17-Jan-80 1640 JMC
Would you talk to Suppes?
------
Yes. And perhaps to American Used Computer Corp. But remember that we
need the KA until the Canon or some other XGP replacement is proven.
∂18-Jan-80 0008 LES
To: FWH, DCL, JMC, PJH
SAIL DMs
When I invited Product Associates representatives over today for a
demonstration of the flakiness of their modem compared to the UDS modem,
they discovered that there is another problem with theirs that they
think they can fix -- it has a higher center frequency than the UDS.
They promise another modified modem on Friday afternoon. I plan to test
it upon arrival. PJH can have his terminal back shortly thereafter, with
thanks.
∂18-Jan-80 0856 LGC
Shall we meet today at 2pm? I have several ideas to bring up that we have
not discussed before, some practical, and some theoretical. If we do meet,
I'll be on time.
Let's meet then.
∂18-Jan-80 1359 Denny Brown <CSD.DBROWN at SU-SCORE> Faculty meeting/Gray Tuesday announcement
Date: 18 Jan 1980 1335-PST
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: Faculty meeting/Gray Tuesday announcement
To: jdm at SU-AI, don at SU-AI,
CSD-Faculty: ;
The hard-copy version of this memo should appear in your mail boxes
soon. On-line folks can read it here first.
The first of two annual progress evaluation meetings, dubbed Gray
Tuesday, is scheduled for Tuesday, January 22, 1980.
The meeting will be in Jacks 252 and will start promptly at 2:30.
The meeting will begin with items left over from the January 8
faculty meeting.
\cb{Agenda}
\list0
\item Consulting professorship for Cordell Green.
\item Proposal to limit the size of introductory programming course sections
to 30 students, increasing the number of sections taught. (Floyd, Brown)
\item Proposal for a ``clock-off'' mechanism to be included in the
``reasonable progress'' guidelines.
\item Gray Tuesday.
\normal
\cb{Introductory Courses}
Prof. Floyd has suggested that we move toward reducing the size of the
sections of the introductory programming courses. We now teach two
sections of 105 and two sections of 106. These
sections each have between 75 and 130 students every quarter. During
fall quarter, Bob taught a small section of CS106 in addition to his
normal teaching assignment. This experience prompted him to suggest
that we try to split the courses into many sections of 25 to 30 students.
Financially, we could implement such a plan with a little addition to
our current outlay.
I agree with Bob that this would greatly enhance the quality of
each student's first encounter with computers and Computer Science.
It is not clear, however, that we will be able to staff
12 sections of 105/106 with interested, capable instructors.
At the meeting, Bob and I would like to get the sentiment of the
faculty on this idea. If sufficient faculty members and PhD students
become interested, we could proceed with a trial. It is reasonable
(but maybe optimistic)
to expect that it will be easier to recruit faculty and students
to teach small sections; I'm not sure that we could consistently recruit
enough.
\cb{``Clock off''}
Attached is a copy of a student proposal for a mechanism to legitimize
educational activities which are worthwhile but not directly related to
dissertation work.
\cb{``Reasonable Progress'' Policy}
Attached is a copy of the guidelines for reasonable progress to the
PhD degree.
-------
∂18-Jan-80 1522 CSD.BETTY at SU-SCORE Gorin alary Review
Date: 18 Jan 1980 1522-PST
From: CSD.BETTY at SU-SCORE
Subject: Gorin alary Review
To: McCarthy at SU-AI
cc: CSD.BEtty at SU-SCORE
Ed Feigenbaum asked me to check with you to see whether you have talked with
Gerry Lieberman about a salary review for Ralph Gorin. My records show that
he is to have a salary review on 2/1, and Ed thought that maybe you had
discussed this while he was away last summer and fall.
Betty
-------
I just got a phone call from Queenette, the secretary at LOTS, to the
effect that she a been called by Pat Devaney, saying that Ralph's raise
had been fixed at 10 percent making $38,995 (or something close to that).
I suppose that since I am on leave, Lieberman acted unilaterally. Anyway,
I was delegated to tell him, so I'll stick out may hand and see if
he bites it.
∂18-Jan-80 1539 Denny Brown <CSD.DBROWN at SU-SCORE>
Date: 18 Jan 1980 1535-PST
From: Denny Brown <CSD.DBROWN at SU-SCORE>
To: JMC at SU-AI, DPB at SU-AI
In-Reply-To: Your message of 18-Jan-80 1418-PST
I figured that you would want to stay on the faculty distribution list,
just in case anything caught your eye. I don't expect you to attend any
meeting you don't want to for your own reasons. Do you want me to remove
you from the mailing list so you don't even see what is happening?
No need to respond if you want status quo. -Denny
-------
∂18-Jan-80 1627 LES State of things
I talked to Suppes and promised a memo on what we have & can offer in the
way of cpu, memories, & pager. He would like it by early summer which
might be feasible if we hustled on the Canon, but I am not sure that we
can prove to ourselves that it will work before the time when he will want
a commitment.
I just received a message from Gio saying that he has the only S-1
contract in existence now and it includes just $15,000 for the SAIL
computer. There was to be $26,000 for SAIL in the Baskett proposal,
but it seems that Forest neglected to get his paperwork in on time
and lost in an LLL budget crunch. This means that our cost center
has got to find more support somewhere. Also, we may have to get
tough with the S-1 group on their use of the machine.
The new new Product Associates modem works about the same, i.e.
it does not have an acceptable error rate on some lines. Negotiations
are proceeding, but it looks as though there are no easy solutions.
About the volksdisplay, I have fairly complete specifications on the Grinnell
stuff and I have Panofsky's telephone number. But what I don't have is
a clear model of goals and constraints. In fact, each day for the last
two months I have resolved to do the RFP and get on with it, but have
either been diverted by another mini-crisis or been puzzled about what we
need. Since I see no devices around that look like just what we want, I
guess that the right thing to do is to proceed on a path of successive
refinement -- write a fairly loose RFP, see what we get, iterate once more
and see if we can converge. I would like to have at least one more
conversation with you about what to specify (and not to specify).
I want to talk to Grinell and to Panofsky before writing anything. I
will read their literature, but I want to see if there is any
co-incidence of views that might cause them to go out of their
way to accomodate us. Therefore, I need the names and phone
numbers.
As for S-1, why don't you run SOB, and come up with a position on
the relation between what they are using and what they are paying
for.
∂18-Jan-80 1642 LES Display makers
Grinnell's phone is 408 263-9920. Ted's home phone is 851-3541. Other
potential suppliers include Robert Flexer (Palo Alto: 493-5088) and Chuck
Reiger (now at U. Maryland).
Do you know who is who at Grinnell?
∂18-Jan-80 2225 JMC
To: "@AIPHIL.LST[F79,JMC]" at SU-AI
I propose that we meet as a group Thursday, January 24 at 2:30 and
discuss whether mental qualities should be ascribed individually or
whether there is some minimum level of mentality required to ascribe any.
The date and time are merely to make a definite suggestion, and proposals
for change are welcome.
∂19-Jan-80 1307 LES
∂18-Jan-80 2116 JMC
Do you know who is who at Grinnell?
----
Not really.
∂19-Jan-80 1332 LES Baskett
I am willing to help, but when I offered to assist Forest with some VLSI
stuff awhile back, I got what seemed like a cool reception. An initial
inquiry by you might elicit a more candid appraisal of the situation than
I would get.
∂20-Jan-80 0642 REM at MIT-MC (Robert Elton Maas)
Date: 20 JAN 1980 0945-EST
From: REM at MIT-MC (Robert Elton Maas)
To: SAVE-THE-WHALE at MIT-MC
CC: rem at SU-AI, mrc at SU-AI, jmc at SU-AI
Gazooks. I depend on the Datacomputer heavily and would be willing to pay
a reasonable storage charge to continue that use. I have about 200 megabits
stored, a complete archive of all my currently-not-obsolete programs and
data files since I started working at SU-AI. Although my node is stored
under <<<SUAI>, it includes both my SU-AI files and my MIT-MC files, and
a bunch of files that don't have any particular home since I work on them
at both places and/or they are programs that work at both sites without
change. I use DFTP as my major storage medium, using the disk at SU-AI
and MIT-MC merely as staging areas for files that I'm actively working on
during a particular month. Having all my files available online, without
the hassles of finding somebody to mount a magtape and worrying about if
the magtape drive has a misaligned head this month, allows me to keep
track of my files, merge updates into consolidated files, extract LISP
functions from old files to put with new stuff in new programs, ...
something I was never able to do in the days of DECtapes and magtapes.
Why on Earth would anyone decide to scrap the Datacomputer? Although it
is down much of the time, when it is up it is valuable, and I can adjust
my accesses to take advantage of its uptime usually. I don't want to see
the Arpanet return to the days of manually-mounted storage media!
∂20-Jan-80 1400 Mark.Sherman at CMU-10A Datacomputer going away
Date: Sunday, 20 January 1980 1658-EST
From: Mark.Sherman at CMU-10A
Subject: Datacomputer going away
To: S1 @ SU-AI
Message-ID: <20Jan80 165858 MS40@CMU-10A>
As per the following notice, the Datacomputer facility at CCA is going
away. All those with files stored there should retrieve them in the
next couple of weeks and move them to mag tape (assuming you want to
still save them). While some people are making efforts to keep the
Datacomputer on net, I don't think they will succeed. The S1 root is
named, appropriately, "S1".
-Mark Sherman
Date: 20 JAN 1980 1124-EST
From: JZS at CCA (Joanne Z. Sattley)
Subject: Scheduled termination of Datacomputer Service
I have been requested by ARPA to distribute the enclosed message.
***** begin enclosure
To All Datacomputer Users:
The TBM-based Datacomputer has become too expensive to operate.
We are faced with a combination of increasing maintenance costs
and a funding deficit which was created when a major user community's
need for the service ended. The Ampex TBM hardware is obsolete,
and no compatible follow-on product is planned. Therefore, with
much regret, we have decided that Datacomputer service on the ARPANET
must be terminated.
We want to make the transition off the Datacomputer as painless
as possible for existing users. On the other hand, we want
to phase out the service as quickly as possible to avoid
needless expense.
Our plan is to stop accepting any additional files for storage,
effective immediately. The Datacomputer will continue to operate
for retrieval only during the transition period. We would like
users to retrieve their private files by March 15, 1980. DFTP
public files will be saved automatically by the CCA staff.
We apologize for any problems which this decision causes you.
If there is something specific we can do to ease the transition,
we will do our best to help. Send a message to JZS@CCA if you
have questions or need assistance.
Bill Carlson
Program Manager
DARPA/IPTO
***** end enclosure
It has been a genuine pleasure working with all of you.
For the Datacomputer Staff,
-- Joanne
∂20-Jan-80 1955 LGC New Developments
John, much has happened since we talked on Friday:
1. Thinking about your "design precedes construction planning" maxim and
Huberman's thesis, in the light of ideas that I already had, has led me to
several new insights and what appears to be a real blockbuster of an idea
for a program called RGPS -- REALLY General Problem Solver, together with
a related methodology for constructing a powerful general intelligence.
This is the most important single idea that I have had since beginning
work on the advice taker project; it completes an internally connected
series of high-level design ideas going back to last summer.
2. I think you'll enjoy the following short tale, which I quote in full
from p. 13 of my 6-year-old daughter's copy of AESOP'S FABLES
[Grosset & Dunlap, 1947; 1977 printing]. It leads to an important
qualification of your "design precedes construction planning" maxim:
--------------------------------------------------------------------------
THE MICE IN COUNCIL
For many years the mice had been living in constant dread of their
enemy, the cat. It was decided to call a meeting to determine the best
means of handling the situation. Many plans were discussed and rejected.
At last a young mouse got up. "I propose," said he, looking very
important, "that a bell be hung around the cat's neck. Then whenever the
cat approaches, we always shall have notice of her presence, and so be able
to escape."
The young mouse sat down amidst tremendous applause. The suggestion was
put to a motion and passed almost unanimously.
But just then an old mouse, who had sat silent all the while, rose to
his feet and said: "My friends, it takes a young mouse to think of a plan so
ingenious and yet so simple. With a bell about the cat's neck to warn us we
shall all be safe. I have but one brief question to put to the supporters
of the plan -- which one of you is going to bell the cat?"
Application: It is one thing to propose, another to execute.
13
----------------------------------------------------------------------------
The qualified version of your heuristic maxim goes something like this:
"Design and construction planning go hand in hand at appropriately
coordinated levels of abstraction." Suitably fleshed out, this principle
will become an integral (and explicit) part of the problem-solving
methodology of RGPS.
3. Finally, in the light of point 1, I feel it is my duty to offer you a
certain piece of general advice, preceded by comments on the nature of the
advice, and on the spirit in which it is offered. First, the advice truly
is general; careful scrutiny is required to determine whether it actually
applies in any given case, and if it does apply, further design and planning
is required to implement it. I offer only the advice, and not directions
for applying it. Secondly, the advice is offered in much the same spirit in
which a member of the President's staff might offer him advice on an
important matter: The advice is mine, but the decisions and the
responsibility for them are yours. The advice, expressed in proverbial
idiom, is this: "Cast not our finest pearls before potential adversaries of
our country; rather, cast them before carefully selected knowledge engineers
in our own department."
I am prepared to discuss any or all of these matters with you at any time
(including this evening) either by phone or in person, should you wish to
do that. -- Lew
∂21-Jan-80 0000 JMC*
Technology Essays to Stuart Miller
∂21-Jan-80 1024 Zenon.Pylyshyn at CMU-10A (X320ZP51) Meeting
Date: 21 January 1980 1203-EST
From: Zenon.Pylyshyn at CMU-10A (X320ZP51)
Subject: Meeting
To: John McCarthy <JMC at SU-AI>
Message-ID: <21Jan80 120318 ZP51@CMU-10A>
In-Reply-To: John McCarthy's message of 19 Jan 80 01:25-EST
Sound OK to me.
∂21-Jan-80 1252 TOB chern visit
To: "@CHERN.DIS[1,TOB]" at SU-AI
The schedule is still being arranged. It appears
as though Dr. Chern will be here Thurs all day
and Friday until noon. It may be possible to arrange
a few appointments on Monday afternoon or Tuesday.
Definite schedule will follow when known.
∂21-Jan-80 1520 BTH Concurrent Programming Seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CONCURRENT PROGRAMMING SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Thursday, January 31
SPEAKER: Susan Owicki, Stanford University
TITLE: Temporal Logic and Parallel Programs
ABSTRACT:
An introduction to the basics of temporal logic as applied to the
specification and proof of parallel programs. Long time fans of this
seminar may find much of this talk a review.
∂21-Jan-80 1606 TOB chern visit
To: "@CHERN.DIS[1,TOB]" at SU-AI
Note that in my previous message, the tentative
schedule is for next week, Jan 28-31, Thurs all
day and Friday am. Definite schedule available
probably tomorrow Tuesday.
∂21-Jan-80 1904 LGC 1. Emendations 2. E-machines
1. I've noted two somewhat misleading passages in the message I sent you
yesterday, and would like to straighten them out.
a. RGPS is my name for a major aspect of the problem solver design, but it
is not my proposal for a name of the entire program. The latter will be
related, I presume, to the proposed long-term methodology for developing the
program -- i.e., interactive goal setting and advising.
b. In discussing the new version of your heuristic maxim, I formulated the
qualification only, and not the full qualified version, as I had intended.
The full version goes like this: "Design and construction planning go hand
in hand at appropriately coordinated levels of abstraction, with design
preceding construction planning at each level."
2. Victor Eliashberg of Varian is tentatively scheduled for Monday, Feb. 4,
at 4:15pm in MJ252. Publicity will be arranged.
∂22-Jan-80 1232 TOB
To: "@CHERN.DIS[1,TOB]" at SU-AI
Dr. Bernard Chern, Computer Engineering, NSF
Thursday all day, Jan 31
Friday am, Feb 1
possible other appts Monday pm or Wed
He will present information about funding opportunities with a new NSF
program which he will head. The program will include computer
architecture and design aids, VLSI, knowledge engineering, robotics and
geometric modeling.
Tom Binford, 72797
TOB@SU-AI
∂22-Jan-80 1321 SELFRIDGE at BBN-TENEXD C and PASCAL
Date: 22 Jan 1980 1616-EST
From: SELFRIDGE at BBN-TENEXD
Subject: C and PASCAL
To: jmc at SU-AI
cc: ef at MIT-AI, minsky at MIT-AI
Yes, John, C was mentioned and almost instantly dismissed,
I'm not quite sure why, except that it seems not to be
the wave of the present. In fact, Dennis Ritchie tells me
that there are more versions around than he likes (that
is, more than 1) (Dennis is my wife's cousin), but it is
true that it has been used successfully in some large
systems, which is more than can be said for PASCAL at
present. By control applications, did you mean real time?
It seems to have no special real time features, and
probably needs them -- as does PASCAL.
I intend to circulate our draft before submitting it
formally to AF. You should get it over the net in a week
or so.
Oliver
-------
My impression from a recent IEEE issue on communications was that C
is used or planned to be used in the telephone company's control
applications which probably have as many real time requirements as
do III's.
∂22-Jan-80 1307 PJH naive physics
good idea. The area that i have been thinking about
and would like to explore is paths and trajectories
of movement from place to place. I think this will
have some interesting and fairly rich connections with
the stuff already done on boundaries and containment
(eg. there is no WAY OUT of a closed container) and
also might be a good way to get to grips with the concept
of a place, which i think is important and basic.
If you would like to work in puzzle mode, Id be intersted
in redoing the blocks world using the histories
ontology as a start, and then perhaps seeing whether
one could extend it naturally to include other kinds
of object, or introducing friction, more precise
dimensional information, etc..
but if you have other ideas on where to start, suggest
some.
pat
∂22-Jan-80 1734 TOB returning your call
John
Called after faculty meeting and found your line busy,
then found you gone.
What's up?
Tom
∂23-Jan-80 1900 CLT*
CALL KARL COHEN
∂24-Jan-80 0832 ADMIN.MO at SU-SCORE (Maurice Bizzarri) Re: visit to Grinnell today
Date: 24 Jan 1980 0831-PST
From: ADMIN.MO at SU-SCORE (Maurice Bizzarri)
Subject: Re: visit to Grinnell today
To: JMC at SU-AI
In-Reply-To: Your message of 23-Jan-80 2338-PST
No problem. Where will we meet?
Mo
-------
∂24-Jan-80 1117 ADMIN.MO at SU-SCORE (Maurice Bizzarri) Re: meeting today
Date: 24 Jan 1980 1110-PST
From: ADMIN.MO at SU-SCORE (Maurice Bizzarri)
Subject: Re: meeting today
To: JMC at SU-AI
In-Reply-To: Your message of 24-Jan-80 0956-PST
OK, I know exactly where that is. I happen to live in
that old house near that gate...
Mo
-------
∂24-Jan-80 1141 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
There will be no seminar session Tuesday, January 29.
The next seminar is as follows:
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, February 5
SPEAKER: David Jefferson, Carnegie-Mellon University
TITLE: Relative Decision Procedures by Type Reduction
ABSTRACT:
The verification conditions for simple programs usually involve
terms from a number of data types, some of which will be scalar
types (char, integer, real) and others higher-level types
(vector-of-real, multiset-of-int). Let T1,...,Tn be types, and
let L(T1,...,Tn) be an interpreted formal language having terms
of those types. Then TYPE REDUCTION is a method for organizing
a decision procedure for L(T1,...,Tn) relative to L(T2,...,Tn)
whereby an arbitrary formula wεL(T1,...,Tn) is transformed into
a finite set of formulae {w1,...,wm}⊂L(T2,...,Tn) such that
w1,...,wm contain no terms of type T1 and such that w is valid
if and only if w1,...,wm are all valid. The method seems most
useful when the eliminated type T1 is one of the higher-order
types, so that both the number and level of the types in question
is reduced by this process.
Type reduction is illustrated by giving a decision procedure for
an unquantified theory called BITZV. BITZV has five types:
Boolean, Integer, T (a totally ordered type), Z (finite
zsets-of-T, like multisets except that negative multiplicities
are permitted) and V (finite vectors-of-T). It comes equipped
with a rich enough set of functions and predicates so that the
weak correctness verification conditions for a wide variety of
programs can be expressed in it, including both the ordering and
permutation properties for binary insertion sort, merge sort and
quicksort (but only the permutation property for heap sort).
The decision procedure for BITZV is a composition of two type
reductions, first from BITZV to BITZ, then from BITZ to BIT.
In the absence of multiplication or other nasty functions the
theory BIT (unquantified) is decidable using decision procedures
for unquantified Presburger arithmetic (BI) and the unquantified theory of
total order (BT).
∂24-Jan-80 1200 JMC
To: "@AIPHIL.DIS[W80,JMC]", RCM
BODEN.NS[W80,JMC] is a revieew of her new book on Piaget.
∂24-Jan-80 2247 LGC Tomorrow at 2pm?
Shall we meet tomorrow (Friday) at 2pm? Reflecting on your "Design
precedes construction planning" maxim in the context of the emerging
advice taker design has been very fruitful for me in several different
ways, and has led to a number of new ideas worth our discussing. For
example, design does NOT necessarily precede construction planning (or
achievement planning, as I now prefer to call it) at each abstraction
level, and I now understand (and can explain) why.
Also, we may wish to disuss whether it would be wise to give Victor
Eliashberg some advice concerning which parts of his 120-page paper would
be most likely to interest a Stanford AI audience, and where he can expect
some objection to some of his claims on the possible significance of his
technical developments.
∂25-Jan-80 0224 PMF
The lab was evacuated. Only hardware lost was one disk drive.
∂25-Jan-80 0900 JMC*
John Gould 914 945-2237
∂25-Jan-80 1030 Feigenbaum at SUMEX-AIM Re the Barclay terminal--any comments?
Date: 25 Jan 1980 1028-PST
From: Feigenbaum at SUMEX-AIM
Subject: Re the Barclay terminal--any comments?
To: csl.fb at SCORE, les at SAIL, jmc at SAIL
Date: 24 Jan 1980 1151-PST
From: Tucker
Subject: More on the Barclay terminal
To: Rindfleisch, Feigenbaum, Admin.Gorin at SCORE
The following remarks are the result of further discussions
about the Barclay terminal with Bruce Borden at RAND and Vivian Nininger,
the president of Barclay International. Included is additional
information about the the terminal in its expanded form, RAND's plans for
it, and some considerations related to its use.
Though Rand had initially planned to just use the device as a terminal
as we use the Datamedia, they now expect to get at least a couple of
them with some program (in addition to refresh) memory and either hard or
floppy disk storage. Dave Yost, who is associated with both Rand and
Barclay, is going to supply a version of NED for use as a local editor on
this larger configuration. This editor will be used in a stand-alone
mode rather than in some hybrid fashion combining main frame and
terminal editor features.
The plot thickens, however, in that Rand has ordered about a dozen PERQ's;
the commercial version of the Alto, by Three Rivers. Dave Yost is also
planning an interface between the Barclay terminals and the network
which contains the PERQ's. The claim is that this is a rather straight
forward thing and that the Barclay will be able to deal with the 10 megabit
rate of the network. This sounds rather ambitious when one keeps in mind
the use of the Z80A as the CPU. However, they have planned the equipment
with the expectation of the Z8000 and that will be utilized when it becomes
available.
There is, of course, the need to distinguish plans from reality. Only one
terminal has been delivered thus far. Evidently this is a demonstration
model and includes external storage and such. It will be demonstrated
at a UNIX conference which will be held next week (28 Jan. and following?)
at Boulder, CO. The price currently listed for the system with 64K bytes
of program memory, two serial I/O ports (up to 600K baud, sync), and a
variety of other features (but not the floppies) is $4400. A stripped
model for use as a Datamedia replacement seems to be in the $2500 to
$2900 range (without a modem).
The potential use of such a device at Stanford falls into several catagories
depending on the a configuration purchased. One quickly sees a lot of
possibilities. I list some of them below starting with the simplest.
1. A Datamedia replacement.
In its simplest form the terminal would have no programmable memory
and would thus be just a big Datamedia with a screen size of 60 lines
by 132 characters or 49 lines by 164 characters. The latter opening
the possibilty for a two text page format. It is likely that both
EMACS and TV could easily be modified to use the larger screen though
a two page arrangement would obviously be a major change. (Evidently
EMACS is used with large screens at MIT.
2. Datamedia type use but with scrolling.
This would involve additional memory in the terminal. People talk
about this sort of feature around here occasionally but implementing
such a feature compatible with the editors (or making editor changes)
would involve considerable work as in the case of two page feature
mentioned above. Of course, one could argue that with the additional
screen size the scrolling becomes of less benefit.
3. Some degree of local editing.
Some sort of scheme might be set up in which local editing would be
done and the commands fowarded periodically to update the file at the
CPU. Or perhaps the changed lines would be sent back to the CPU at
certain points in the editing. This second approach would be totally
impractical when used with our 1200/150 dial-ups since the return
speed is so slow. In either case there would be no feedback concerning
transmission errors; a serious problem with such an approach.
4. Complete local editing.
In this system the Barclay would become just a bottom dollar Alto
in which files would be transfered to the local storage, edited and
returned to the main system. Either the terminals would have to be
directly connected to a network or connected in clusters, via something
like a PDP-11, to the network. The need for local storage means we
would have lots of floppies or lots of small hard disks around. This
tends to be much more expensive on a bytes per dollar basis than the
large central storage systems and probably a real maintenance nightmare.
Conclusions.
The only really unique feature of the Barclay equipment is the large
screen format at a fairly reasonable price. Though similiar terminals
may be available elsewhere, I have not checked around. The existance
of an editor as good as NED, on a small system, maybe unique also, I don't
know. And , of course, there is the possibility that they will come up
with a network interface that might be useful to us.
However, when considering the purchase of terminals with local storage and
editing capability there are lots to chose from; dozens of word processing
systems large and small. Probably the most negative feature going for
Barclay is the fact that they have built only one terminal, the president
of the company is also the salesman (and I think works half time somewhere
else), and their programmer may be a part timer also (along with handling
demos).
The terminal is worth keeping an eye on. The question of local editing
involves lots of things apart from the terminal.
Bob
-------
-------
I would argue fiercely that we don't need more high priced local editing
terminals now but rather a display system for the department, but it seems
that action isn't being advocated now, so I'll do no more than just
restate that position. I have started talking with Grinnell about 200 to
300 terminal system.
∂25-Jan-80 1221 TOB
To: "@CHERN.DIS[1,TOB]" at SU-AI
Mon Tues Wed Thurs Fri
Levinthal
9 avail Berkeley HPP travel
10 travel HPP
11 Xerox Dutton
12
1 Hennessy
2 Binford SRI McCarthy
3 SL,BIS Baskett
4 Luckham
5 Earnest
∂25-Jan-80 1631 ARK S-1 Meeting
To: S1 at SU-AI
CC: pickering at SUMEX-AIM
The next S-1 meeting will be Thursday, 31 January at 11:00 am in room 301
Margaret Jacks Hall. Progress reports due.
Arthur
∂25-Jan-80 1706 HPM The time has come
To: JMC at SU-AI, TOB at SU-AI, LEVINTHAL at SUMEX-AIM
You may remember that long ago you foolishly agreed to be on my
thesis orals committee. I would like to schedule the orals
in the second week of March. Do I hear any objections?
Thanks.
∂25-Jan-80 1710 HPM that is, I want to HOLD the orals in the second week
of March To: JMC at SU-AI, TOB at SU-AI, LEVINTHAL at SUMEX-AIM
∂26-Jan-80 0307 JK SET THEORY
I have a version of "heavy duty set theory" on my file
(which seems to work quite nicely in practice, provided one
can wade one's way through the many lovely bugs of FOL) on
SET.FOL. I think I can prove Ramsey's Theorem in under 100
lines given the lemmas on page 9, which take additional 75 lines.
I have already a half of the proof done, I will finish it up
some time next week.There really is no need to introduce
meta theory at all, except for the fact that it provides a handy
way for getting around the inadequacies of FOL.Likewise, goal
structuring is irrelevant. If FOL could handle lambda terms
and the MONADIC command worked properly, one could do the
whole thing without meta. One must say that any proof given in FOL
is phony since the sorting machinery is a joke.
A philosophical issue: It seems to me that the ideal
proof checking system should have a very strong re-writing system
and in addition a "meta theory"-not in the sense of FOL with
indiscriminate attachments on the meta level but a machinery and a
language that allows one to discuss correct proof procedures and
implement them.For example, one should be allowed to write arbitrary
programs in this language. The semantics of this language should
rigged up so that any atomic act translates to a validity preserving
transformation-this will guarantee correctness. Moreover, this should
be more than a mere macro facility.
Comments on the set axioms: Other than the lambda scheme
one doesn't need say anything about APPLY; there is no need to worry
about undefined values or domains etc. Also, it is useful to define
ordered pairs as an atomic notion rather than the usual {{a} {a b}};
in fact they probably should arise from the LISP axioms, which don't
exist yet.
∂26-Jan-80 1436 RWW
JMC to RWW
Victor Eliashberg, an emigre from Russia now employed at Varian, has
some AI ideas he would like to talk about. Could you call him,
493-4000 x2544 and arrange a small seminar?
JOHN, I have not yet done this but I am about to. Do you still want me
to do it?
∂26-Jan-80 1443 RWW Invitation
OK Do you want me to check and do it if he's not?
Richard
∂26-Jan-80 1646 JB via OFFICE-2 Hello.
Hello, John. I remember I owe you a copy of my thesis.
I would like to bring it in person and say hello at the same time.
What time woul be good for you?
∂27-Jan-80 2154 LGC new term
1. Here is a new entry for Daniel Dennet's philosopher's lexicon:
cognosciente (noun, plural: cognoscienti) An expert practitioner of
cognology (sometimes called cognitive science).
Example of use: "He's not only a good knowledge engineer, but is also
among the true cognoscienti."
2. Do you plan to introduce V. Eliashberg when he speaks on Monday,
Feb. 4, at 4:15pm in MJ252? Also, have you any advice that I might
pass on to him concerning the planning of his talk?
∂27-Jan-80 2214 ARK S-1 Meeting
To: S1 at SU-AI
CC: pickering at SUMEX-AIM
The next S-1 meeting will be Tuesday, 29 January at 11:00 am in room 301
Margaret Jacks Hall. NOTE DATE CHANGE. Progress reports due.
Arthur
∂28-Jan-80 0119 JK ramsey
I have a crude 168-line proof of Ramsey's theorem on RAM.PRF
I expect to be able to clean it up and shave about 30 lines off.
∂28-Jan-80 1032 LGC
Will you have a few minutes to meet privately with me this afternoon after
the session on the frame problem is finished? I have a specific practical
question to discuss concerning the March conference on AI & Philosophy at CASBS.
∂28-Jan-80 1839 LES
∂28-Jan-80 1833 JMC
Till when is LLL use paid for?
-----
I believe that the new contract runs for a year from start, which is
probably about now.
I assume that contract covers use by Wiederhold and cohorts. Does it
also cover use by the LLL group directly?
∂28-Jan-80 1847 LES
∂28-Jan-80 1841 JMC
I assume that contract covers use by Wiederhold and cohorts. Does it
also cover use by the LLL group directly?
-----
I haven't seen the contract, but Gio indicated that was his intent.
∂29-Jan-80 0916 JMC*
Call Hersch.
∂29-Jan-80 1011 DOYLE at MIT-AI (Jon Doyle) Your request and offer
Date: 29 JAN 1980 1311-EST
From: DOYLE at MIT-AI (Jon Doyle)
Subject: Your request and offer
To: jmc at SU-AI
CC: DOYLE at MIT-AI
Your request for a probability of accepting your offer makes me very
uncomfortable. All I can say is that I am very interested in it. But
so far, it is the only concrete offer I have recieved, as I expect to
hear from the other places sometime in the next month or later. I can't
predict how I will react to their offers, or to Bell Labs which I visit
in two weeks.
But, again, I am very interested in your offer.
By the way, could you please arrange to have your offer sent to me in
a letter mentioning the specifics of the position, such as responsibilities,
vacations, consulting priviledges, etc? If there is some standard page
or two governing these things from Stanford's policy book I would appreciate
that.
Thank you very much for your offer, your hospitality, and your interest.
I truly enjoyed my visit with you.
Jon Doyle
∂29-Jan-80 1021 MP letter ready for your signature
The letters you asked about are in your office. Also, per the message that you
sent re ACM Information Cassettes release form--I haven't found it. I guess the
thing to do is to wait until they send another one, as you suggested.
∂29-Jan-80 1043 LGC Eliashberg
It turns out that I have until Wednesday noon to set the wheels of publicity
in motion for his talk. So if you wish to talk either with me or with him
about it before I go ahead, there's still time. I'm sufficiently busy right
now that I may not attend his talk (unless you want me to take on that
responsibility), and I fear that many others may feel the same way about it.
I'll wait until I hear from you (however briefly) before proceeding.
∂29-Jan-80 1430 BTH Concurrent Programming Seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CONCURRENT PROGRAMMING SEMINAR
DATE: Thursday, February 7
There will be NO seminar due to the Computer Forum. Technical Session II
of the forum (Thursday, 10:55 - 12:15) is on verification, program
derivation, and testing.
∂29-Jan-80 1455 MINSKY at MIT-AI (Marvin Minsky)
Date: 29 JAN 1980 1740-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
Hi. Alex Rich wonders whatc you might know about Uri Marchuk, since he is
to be at meeting with same, by and by. He'll try to call you but I
thought this system might work better.
Assuming that Rich is referring to Gurii Ivanovich Marchuk, the current
head of the Siberian Division of the USSR Academy of Sciences, Marchuk
is an energetic affable intelligent man and a rather G.I. communist.
When he was Director of the Academy Computation Center in Novosibirsk,
he allowed Ershov to refrain from firing someone who had applied to
emigrate, so that the latter kept his job till he left Novosibirsk.
This counts as a point in someone's favor, but it could be Ershov,
Marchuk, Lavrentiev or even the Akademgorodok party boss.
I was hoping you would show up at the Behavioral while you were out here.
∂29-Jan-80 1601 DBL AI Candidate 2/29
John,
Janet Kolodner, a graduate student of Schank's, will be interviewing
here on Thursday and Friday, February 28-9. In particular, she will
give a talk at noon on Friday in the Chemistry Gazebo. I would ap-
preciate your attendance, as the senior AI faculty member of the dept,
and your feedback afterwards. If her talk is good, we will probably
offer her an R.A.; if it is spectacular, we may try to CONS a faculty
appointment, joint with Psychology, out of affirmative action funds.
Her thesis deals with memory models, from a first-person viewpoint; it
is embodied in a program that models the memory of Cyrus Vance; despite
(!) this, talking with her reveals some solid thinking and creativity.
If you want to schedule an individual appointment, let me know. What
I suggest is that you leave some time Friday afternoon, and if there
is any chance of our offering her a position following her talk, then
you may decide to meet with her then.
Thanks.
Doug
I think I met Kolodner at IBM last summer. I may be available on
the 28th and 29th to attend the talk, but since I am on leave, I
want to preserve my options.
∂29-Jan-80 1704 Lenat at SUMEX-AIM (Response to message)
Date: 29 Jan 1980 1704-PST
From: Lenat at SUMEX-AIM
Subject: (Response to message)
To: JMC at SU-AI
In response to your message sent 29 Jan 1980 1620-PST
Fair enough. You're around more than many people NOT on leave,
so I tend to forget that you're supposed to be. Kolodner did
spend the summer at IBM Yorktown, so if you have any recollections
pro or con let me know.
Regards,
Doug
-------
∂29-Jan-80 1707 PJH
To: "@AIPHIL.LST[F79,JMC]" at SU-AI
some remarks on the frame problem are in FRAMEP[1,pjh]. Feel free
to add to it if you feel inclined.
∂29-Jan-80 1710 GIO Accounts
To: LES, JMC
I looked at S-1 detail accounts and with costing being near I think some
rearrangements are in order. I see many people with S-1 accounts whom I
don't even know. I see people that are definitly not doing S-1 stuff
(e.g., Kaplan = SJK). I see people who are doing some S-1 stuff but also
other stuff now (e.g., ARK). There are undoubtedly also people who are
half and half and no mechanism for splitting up accounts seems to exist.
Perhaps people should be encouraged to log in on different projects of the
same programmer with the accounting based on projects instead of account.
In the meantime I'd like to make a pass dividing people into CSD project,
Kosher Livermore, and unknown.
Should we do this together or do you want a list?
Gio
Ed:
The following is symptomatic of the fact that carrying student
computing on project accounts is likely to be difficult. If it is to be
done at all, there will have to be departmental effort.
∂29-Jan-80 1903 CSL.FB at SU-SCORE [Don Knuth <DEK at SU-AI>: gls ]
Mail-from: ARPANET site SU-AI rcvd at 28-Jan-80 2108-PST
Date: 28 Jan 1980 2104-PST
From: CSL.FB at SU-SCORE
Subject: [Don Knuth <DEK at SU-AI>: gls ]
To: GLS at MIT-AI
cc: dek at SU-AI, fb at SU-AI
Remailed-date: 29 Jan 1980 1824-PST
Remailed-from: CSL.FB at SU-SCORE
Remailed-to: jmc at SU-AI
Guy: Just to let you know we really do care. --Forest
---------------
Mail-from: ARPANET site SU-AI rcvd at 28-Jan-80 2021-PST
Date: 28 Jan 1980 1115-PST
From: Don Knuth <DEK at SU-AI>
Subject: gls
To: EAF at SU-AI, SSO at SU-AI, FB at SU-AI
Has anybody talked to Guy Steele yet, or are you assuming that I was carrying
the ball?
I think we should act now to find out if he is at least 5 per cent interested,
and fly him out here. But I guess I assumed that one of you would handle
it, and maybe you assumed that I am... Anyway I haven't done anything
yet, but am glad to be of help. (on the other hand I don't really have time
to be the organizer of this, my system will be overloaded the next two weeks
while Zapf is here)
---------------
-------
∂29-Jan-80 1903 CSL.FB at SU-SCORE Re: gls
Mail-from: ARPANET site SU-AI rcvd at 28-Jan-80 2119-PST
To: FB at SU-AI
Date: 28 Jan 1980 2102-PST
From: CSL.FB at SU-SCORE
Subject: Re: gls
To: DEK at SU-AI
cc: feigenbaum at SUMEX-AIM, sso at SU-AI, fb at SU-AI
In-Reply-To: Your message of 28-Jan-80 1115-PST
Remailed-date: 29 Jan 1980 1825-PST
Remailed-from: CSL.FB at SU-SCORE
Remailed-to: jmc at SU-AI
8I sent Guy a net note to see if he was interested. No response yet. I will
forward your note as reinforcement.
Forest
-------
∂29-Jan-80 1903 CSL.FB at SU-SCORE Guy Steele
Date: 29 Jan 1980 1823-PST
From: CSL.FB at SU-SCORE
Subject: Guy Steele
To: JMC at SU-AI
We are attempting to recruit Guy Steele. I have forgotten to tell you. I
assume you approve. I am remailing several recent messages on this topic to yo.
Forest
-------
∂29-Jan-80 1903 CSL.FB at SU-SCORE GLS
Date: 29 Jan 1980 1827-PST
From: CSL.FB at SU-SCORE
Subject: GLS
To: jmc at SU-AI
The messages I don't have handy are two. One is from Guy saying he is interested.
The other is a reply from me asking him to come out for a visit soon and asking
when soon could be.
Forest
-------
∂29-Jan-80 1903 Don Knuth <DEK at SU-AI> gls
Mail-from: ARPANET site SU-AI rcvd at 28-Jan-80 2021-PST
Date: 28 Jan 1980 1115-PST
From: Don Knuth <DEK at SU-AI>
Subject: gls
To: EAF at SU-AI, SSO at SU-AI, FB at SU-AI
Remailed-date: 29 Jan 1980 1824-PST
Remailed-from: CSL.FB at SU-SCORE
Remailed-to: jmc at SU-AI
Has anybody talked to Guy Steele yet, or are you assuming that I was carrying
the ball?
I think we should act now to find out if he is at least 5 per cent interested,
and fly him out here. But I guess I assumed that one of you would handle
it, and maybe you assumed that I am... Anyway I haven't done anything
yet, but am glad to be of help. (on the other hand I don't really have time
to be the organizer of this, my system will be overloaded the next two weeks
while Zapf is here)
∂29-Jan-80 2106 Baskett at PARC-MAXC GLS
Date: 29 Jan 1980 9:06 pm (Tuesday)
From: Baskett at PARC-MAXC
Subject: GLS
To: JMC@Sail
cc:
I found it. --Forest
------------------------------------------------------------
Mail-from: Arpanet host MIT-AI rcvd at 25-JAN-80 1359-PST
Date: 25 JAN 1980 1642-EST
From: GLS at MIT-AI (Guy L. Steele, Jr.)
Subject: Position
To: Baskett at PARC-MAXC
CC: GLS at MIT-AI
I'm definitely interested. I expect to finish my graduate
work some time between May and August, "depending".
Kepp me in mind.
-- Guy
------------------------------------------------------------
∂29-Jan-80 2109 Baskett at PARC-MAXC GLS
Date: 29 Jan 1980 9:07 pm (Tuesday)
From: Baskett at PARC-MAXC
Subject: GLS
To: JMC@Sail
cc:
More.
------------------------------------------------------------
Date: 28 Jan 1980 10:57 pm (Monday)
From: Baskett at PARC-MAXC
Subject: Re: Position
In-reply-to: Your message of 25 JAN 1980 1642-EST
To: GLS at MIT-AI (Guy L. Steele, Jr.)
cc: Baskett, DEK@SAIL, Feigenbaum@Sumex-AIM
As you should be able to tell by now, we are also definitely interested, too.
We would like to have you come out for a visit soon. Bring your wife.
What dates would you find convenient, assuming you are willing?
Forest
------------------------------------------------------------
∂29-Jan-80 2112 Baskett at PARC-MAXC GLS
Date: 29 Jan 1980 9:11 pm (Tuesday)
From: Baskett at PARC-MAXC
Subject: GLS
To: DEK@Sail, Feigenbaum@Sumex-AIM, JMC@Sail
cc: FB@Sail, DPB@Sail
More from Guy Steele. I will say yes and try to set up a seminar with Guy and
Denny. --Forest
------------------------------------------------------------
Mail-from: Arpanet host MIT-AI rcvd at 29-JAN-80 1217-PST
Date: 29 JAN 1980 1448-EST
From: GLS at MIT-AI (Guy L. Steele, Jr.)
To: Baskett at PARC-MAXC
CC: GLS at MIT-AI
I would certainly like to come out this spring and perhaps give
a talk (job interview?) etc. February is booked up with
one thing and another, so perhaps March would be best.
Can I get back to you next week with a suggested date?
Thanks for all the encouragement. -- Guy
------------------------------------------------------------
∂29-Jan-80 2134 Baskett at PARC-MAXC Re:
Date: 29 Jan 1980 9:33 pm (Tuesday)
From: Baskett at PARC-MAXC
Subject: Re:
In-reply-to: Your message of 29 JAN 1980 1448-EST
To: GLS at MIT-AI (Guy L. Steele, Jr.)
cc: FB@Sail, DEK@Sail, Feigenbaum@Sumex-AIM, JMC@Sail
Yes, call back next week by all means. Job interview is the respectable,
acceptable, and appropriate term. We are, of course, talking about a regular
faculty position in the Computer Science Department with all the rights,
privilages, and pains thereby entailed. We have a regular CS colloquium on
Tuesday afternoons so if you could make that, you could consider coming out
on Tuesday and staying Wednesday (then people you talk to might have heard
your talk) or you could come out on the weekend and stay Monday and
Tuesday. We can put you up at the faculty club with enough advance warning
to get a room reservation there. Otherwise, there is always the Holiday Inn. Or
you could make or have us make any other plan that would be convenient.
Forest
I am also interested in having Guy Steele.
∂29-Jan-80 2300 TOB
∂29-Jan-80 2257 JMC
John
No, not much. It is also like Binford-Horn, including all the recent
things about zero crossings. Marr has a bug about giving credit I think.
Tom
4. I should mention that Marr's primal sketch is quite a lot
like the drawings generated at the Stanford AI Lab in the 1960s
by Manfred Hueckel. Do Marr et. al. refer to Hueckel?
∂30-Jan-80 0900 JMC*
hersch
∂30-Jan-80 0959 REG
re: series 1, etc.
I would agree that the objections you raise are significant.
One approach is to try a limited number of terminals for a while.
If they're winners, there will be increased demand for them.
If they're losers, that should be apparent too.
∂30-Jan-80 1010 DPB via SU-SCORE list of people in Jacks
To: JMC
CC: LES
Two files of interest are JMC.PTY[PER,CSD] and JMCMAP.XGP[PER,CSD].
These were obtained by filtering the PEOPLE file looking for records
with "MJ" appearing in OFFICE tag.
JMC.PTY is a list of person, phone, office, and title.
JMCMAP.XGP is a map containing only those people on this list.
174 names on the list. Missing from the list will be people like
Baskett, Dantzig, McCluskey, etc. who don't have assigned Jacks offices.
Also missing are jungle residents like Mujtaba, Salisbury, Baker, etc.
who live here unofficially.
I'll put a hard-copy in your boxes.
-Denny
∂30-Jan-80 1603 Denny Brown <CSD.DBROWN at SU-SCORE> Re: GLS
Date: 30 Jan 1980 0951-PST
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: Re: GLS
To: Baskett at PARC-MAXC, DEK at SU-AI, Feigenbaum at SUMEX-AIM, JMC at SU-AI
cc: FB at SU-AI
In-Reply-To: Your message of 29-Jan-80 0911-PST
Current Colloq. schedule for late March/early April wrt. Guy Steele:
3/4 Roy Levin (PARC), 3/11 Frances Yao and Leo Guibas, tentative, (PARC),
3/18 and 3/25 No colloq scheduled - Quarter Break.
4/1 open. 4/8 John Shoch (PARC).
So we could move any of the scheduled speakers, or decide to interview
Guy during quarter break. Even during break, I would anticipate a decent
turnout if it's known that he's on the shopping list.
-Denny
-------
∂30-Jan-80 1851 HPM HPM Orals
To: JMC at SU-AI, TOB at SU-AI, LEVINTHAL at SUMEX-AIM, DEK at SU-AI
Since there seemed to be no conflicts, I am scheduling them for
Wednesday March 12. I will have a reasonable draft of the thesis
ready for you very early in March. Thanks.
∂31-Jan-80 0823 REG
To: JMC, REG
LOTS Future
JQ and I are preparing a memo to Lieberman proposing the following
expansions:
1. Upgrade LOTS to 1.5 Megawords, add 16 ports, 16 terminals.
2. Authorize the purchase of system to be used for word processing.
This system to be paid for by the people who use it. Preliminary
estimates suggest that we could sell connect time for $1.00/hour
and come out approximately even. These estimates assume that
the system will be large and popular.
In the event that the word processing system is not sufficiently
popular, it can be turned into the second LOTS machine. I would
want to run the experiment for at least 18 months before coming
to a firm decision about it.
3. Some mention will be made of graphics terminals for engineering
and scientific applications. But I think the push for this should
come from the engineers. Moreover, I don't have a very good idea
yet about what we want.
I value your suggestions.
∂31-Jan-80 0900 JMC*
sarah's ticket
∂31-Jan-80 1009 TOB Chern meeting
John
Meeting with Chern, 2 pm today, Rm 252 Jacks.
Tom
∂01-Feb-80 0557 HPP.GENESERETH at SU-SCORE MACSYMA
Date: 31 Jan 1980 1828-PST
From: HPP.GENESERETH at SU-SCORE
Subject: MACSYMA
To: CSD-Faculty:
Many of you are no doubt aware of the MACSYMA symbolic
manipulation system. Since I was heavily involved in its development,
I am inn a position to import a copy for use on the Stanford machines,
as a number of people have requested. There are advantages and
disadvantages to such a move, and I'd like an indication of interest
from anyone interested in using it locally. The major problem is
that it would overburden our already heavily loaded machines.
Presumably, SCORE would be the primary host, though the VAX (which is
not heavily used just yet) would also be suitable. The advantages are
that it would be more convenient than using the net to gain access at
M.I.T. and that we could do related research, such as TEX interfaces.
I'd appreciate a note from any of you who would like to see MACSYMA
brought up here (or anyone who violently objects).
-- Mike
-------
I would like to see MACSYMA here, although I have no specific plans
to use it. I don't find our machines heavily loaded at present, but
maybe this is because I mainly use it for writing.
∂01-Feb-80 1027 CG another recommendation
Here is (yet) another address where a letter of recommendation for me is wanted:
Josina M. Arfman- Van der Maas, Manager
Corporate Ph.D. Program
Thomas J. Watson Research Center
P.O. Box 218
Yorktown Heights, New York 10598
Thanks.
Monica:
GOAD.RE1 is a recommendation for Chris Goad. Please make another
addressed as below called GOAD.RE2, pub it and print it.
∂01-Feb-80 1302 BTH concurrent programming seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CONCURRENT PROGRAMMING SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Thursday, February 14 & February 21
SPEAKER: David Wall, Stanford University
TITLE: BROADCAST AND SELECTIVE BROADCAST
ABSTRACT:
A loosely-coupled network like the ARPAnet does not lend itself to a
situation in which a node needs to send an identical message to every
other node, or to an arbitrary set of nodes. The obvious ways of
performing such a broadcast result in a depressing amount of redundancy or
overhead. Dalal did some work on this problem, the major contribution of
which was a distributed algorithm for creating and maintaining a minimum
spanning tree for the network in the face of failures and changing costs.
The bulk of my thesis extends Dalal's work in two directions. First,
I generalize Dalal's MST-based algorithm to an algorithm to find a tree
that spans some subset of the nodes; finding the minimum such tree is
NP-complete, so we will be content with an approximation. Second, I
present a new algorithm for broadcasting, called "center-based
forwarding," which can be used like Dalal's MST algorithm but which gives
a small delay rather than a small cost.
This talk lasts two weeks. (Actually it lasts two hours, but it will
be spread out over two weeks.) The first half deals with the
generalization of Dalal's algorithm; the second discusses center-based
forwarding.
∂01-Feb-80 1409 LGC The Frame Problem and Philosophy
You might take a look at PHIL.TXT[EP,LGC]; I am making several copies of it to
bring out with me this afternoon. It might serve as a good discussion starter;
what do you think?
∂01-Feb-80 1555 LES Doyle
Will do, but I guess that I should circulate his biography in the
department first, following the established protocol.
∂02-Feb-80 2313 LES display procurement
To: MO at SU-AI, FB at SU-AI, REG at SU-AI, JMC at SU-AI
Robert Flexer, who makes bit-map display generators for the the home computer
market, will be here Tuesday, Feb. 5, at 4pm to discuss what he could do for
us. I guess that we should meet in the same conference room.
Better check with Feigenbaum or his sec'y about availability of conf. room.
∂03-Feb-80 1002 Feigenbaum at SUMEX-AIM Ullman/VLSI/money
Date: 3 Feb 1980 1004-PST
From: Feigenbaum at SUMEX-AIM
Subject: Ullman/VLSI/money
To: csl.fb at SCORE
cc: jmc at SAIL, feigenbaum
Forest,
Jeff has initiated his work on "VLSI Theory" and wants to make it a main
research interest (have you seen his new paper?). I want to encourage him
to do so and Kahn to fund him (as Kahn funds you, me, and McCarthy in the area).
Clearly it's too late to get money for this fiscal year from Bob, so I am going
to try to scrape together some bucks from my ARPA contract to fund Jeff
until new money shows up. Do you have any bucks that can be scraped together
for this purpose? The total amount we would need until new money shows up
is in the low tens of thousands of dollars (before overhead).
I will send a copy of this message to John McCarthy to see if he has any spare
bucks in his "VLSI Design System" project.
Thanks,
Ed
-------
∂03-Feb-80 1335 LGC Peter Cheeseman
As I mentioned in our phone conversation, Peter Cheeseman seems to me to
be working from the bottom up, in a very interesting and basically sound
way, on many of the same problems and issues that I am working on from the
top down in the advice-taker design. I feel that his knowledge and skills
are very complementary to mine, and that his basic orientation and address
to things is just right for the advice taker work. Therefore, I believe
that he and I would make an excellent team to serve as a nucleus for the
advice taker effort. I suspect that he would find it not only congenial,
but positively exciting, to work within the basic design for the program
that is now taking shape.
Based on what I heard at his HPP talk, his ideas on abstraction in
planning, top-level control structure, and especially representation of
propositional attitudes do not seem nearly as general or advanced as those
that are now being developed for the advice taker, and I heard no mention
of advice taking as a means of knowledge acquisition. However, I believe
that this is due largely to a lack of knowledge and the fact that he has
been working for not very long from the bottom up, and not to any lack of
intelligence, industry, or sound common sense. He has some very
interesting ideas about the learning of useful heuristic knowledge during
problem solving, and is currently quite excited by some techniques he
learned from papers by Boyer and Moore; you should be sure to draw him out
on these issues.
Cheeseman is presently staying at the Cardinal Hotel in Palo Alto, in
room 331. The hotel's phone number is 323-5101. Since he is (I think)
interviewing at SRI on Monday, it would probably be best if you could
interview him either today or Tuesday. He leaves for Australia next
weekend. The announcement of his HPP talk is appended below. If I can be
of any further help in this matter, be sure to let me know.
-------------------------
Date: 29 Jan 1980 1538-PST
From: Hazen
Subject: Dr. Peter Cheeseman of Australia
To: siglunch:
This Friday, February 1, Dr. Peter Cheeseman of Australia will be our
guest speaker at this week's siglunch. It will be held, as usual, in
the Chemistry Department Gazebo. A copy of his abstract is listed
below:
Problem Solving and Knowledge Representation
When an hierarchic planning, parallel goal approach is used
to solve problems expressed as a conjunction of goals, it is found
that the usual context mechanisms for world modelling do not work. A
solution is to use the problem solver itself to do the required
deductions in much the same way as it reasons about actions. This
idea can be extended so that all the internal problems generated by
the problem solver, such as constraint satisfaction, interaction
removal and learning, are treated as new problems to be solved in the
same way, provided the knowledge for that domain is available. To do
this requires an expressive uniform knowledge representation for all
domains with efficient procedures for its use. Such a representation
is proposed and other related representational issues will be
discussed.
-------
∂03-Feb-80 1347 LGC Eliashberg Talk Tomorrow
Provided that I feel well enough, I plan to attend the Eliashberg talk
tomorrow, and will arrange for cookies and juice before the talk. My
phone discussion with him the other day led me to think that at least some
of what he has to say may be interesting.
∂04-Feb-80 1000 JMC*
Teller
∂04-Feb-80 1000 JMC*
Cheeseman
∂04-Feb-80 1101 FWH PV+A Seminar: Reminder
To: "@SEM.DIS[SEM,VER]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, February 5
SPEAKER: David Jefferson, Carnegie-Mellon University
TITLE: Relative Decision Procedures by Type Reduction
ABSTRACT: see message of 24-Jan-80
∂04-Feb-80 1619 PJH
To: "@AIPHIL.LST[F79,JMC]" at SU-AI
there is a provisional program for our march conference in
the file PROGRA[1,PJH]. I would appreciate getting your
comments on it fairly quickly. You can add them to page
3 of the file if you like. Thanks.
pat
∂04-Feb-80 1634 PJH
To: "@AIPHIL.LST[F79,JMC]" at SU-AI
p.s. i have booked (provisionally) 25 rooms for the
nights of 23-26 march in the tiki inn (corner
of stanford av. and el camino). They cost $29 single,
$33 double. Please tell attendees to contact me before
10 march if they want one or more of these rooms, if
you get any queries. Thanks
pat
∂05-Feb-80 1021 TOB hello
John
Dick Schoen who was Chern's boss at Division of Applied
Research, NSF, says hello. I think that he was at
CalTech with you.
Tom
∂05-Feb-80 1346 DWW concurrent programming seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
For logistic reasons I will be reversing the order of my pair of talks in
the concurrent programming seminar. Week one will deal with center-based
forwarding; week two will deal with the extension of the MST-based algorithm.
The overall content of the talks, and their title, will be the same.
David Wall
∂06-Feb-80 1249 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: Margaret Jacks Hall 426
TIME: 2:30 p.m., Tuesday, February 12
SPEAKER: Gerard Huet, IRIA and SRI International
TITLE: The MENTOR Program Manipulation System
ABSTRACT:
MENTOR is a program manipulation system developed at IRIA. MENTOR
manipulates programs represented as abstract syntax trees, using a specialized
programming language called MENTOL. MENTOR is not specialized to any specific
programming language, but has been used primarily to implement a PASCAL
programming environment. MENTOR is entirely written in PASCAL, and is therefore
developed and maintained using MENTOR. The original version runs on a CII IRIS
80 computer, and the PDP10 version was obtained from it completely mechanically
using MENTOR.
This seminar will be a demonstration of the MENTOR system.
** Please note that the location of this seminar session is NOT ERL 237 **
∂07-Feb-80 1041 HVA Ticket--Dina Bolla
They will mail ticket to your home address.
∂07-Feb-80 1047 JMC*
standards, text reading
∂08-Feb-80 1000 JMC*
What about "John didn't see the man" and John didn't see a man".
∂08-Feb-80 1213 TOB help
To: LES, JMC
The load average is now 13. I find the system slow
enough that it bothers me. I want to again urge measures
to limit the load on the machine.
Tom
∂08-Feb-80 1502 ARK S-1 Meeting Notice
To: S1 at SU-AI
CC: pickering at SUMEX-AIM
There will be an S-1 meeting on Thursday, 14 February at 12:00 noon in room
301 Margaret Jacks Hall. Progress reports should be mailed to area mailing
lists (e.g., S1SOPA) with a copy to S1PROGRESS-LOG@SAIL. Thanks.
Arthur
∂08-Feb-80 1508 ARK S-1 Meeting Notice
To: S1 at SU-AI
CC: pickering at SUMEX-AIM
There will be an S-1 meeting on Thursday, 14 February at 12:00 noon in room
252 Margaret Jacks Hall. Progress reports should be mailed to area mailing
lists (e.g., S1SOPA) with a copy to S1PROGRESS-LOG@SAIL. Note room change.
Thanks.
Arthur
∂08-Feb-80 2054 Kanerva at SUMEX-AIM Pictures and Words
Date: 8 Feb 1980 2055-PST
From: Kanerva at SUMEX-AIM
Subject: Pictures and Words
To: McCarthy at SAIL
cc: Sandelin, Siegman
What you said today about a college student possibly passing his tests
even if his textbooks were stripped of pictures reminded me of this:
(From LESS THAN WORDS CAN SAY, by Richard Mitchell--The Underground
Grammarian. Boston: Little, Brown & Co., 1979, p. 183.)
One picture is NOT worth a thousand words. A picture
isn't WORTH any words at all. One picture, one glimpse
of the moon, even one episode of "Mod Squad," may CAUSE
a thousand words in some beholder. Or many thousand.
Or, in some other beholder, none at all. If knowledge
and understanding are to come from watching "Mod Squad"
or the moon, they must come in the form of language,
the only vehicle we have for knowledge and understanding.
Those unfortunate students "who do not read well" may
look forever at the moon and have nothing to say, for
they have not the power of language, which is why they
do not read well. They may have taken something or other
from their years of film- and television-watching, but it
isn't likely to be knowledge and understanding. Further-
more, whatever knowledge and understanding they found
that way came not from the IMAGES they saw but from some
WORDS they might have heard. How many of them, do you
suppose, were sufficiently "image-oriented" to watch
fifteen years of television with the sound turned off?
-------
∂09-Feb-80 0808 Feigenbaum at SUMEX-AIM (Response to message)
Date: 9 Feb 1980 0809-PST
From: Feigenbaum at SUMEX-AIM
Subject: (Response to message)
To: JMC at SU-AI, FB at SU-AI, EAF at SU-AI
In response to the message sent 08 Feb 1980 1634-PST from JMC at SU-AI
John, what's that all about? How can you be refused an account? Who did the
refusing?
Ed
-------
∂09-Feb-80 1736 CSL.FB at SU-SCORE
Date: 9 Feb 1980 0839-PST
From: CSL.FB at SU-SCORE
To: JMC at SU-AI
cc: eaf at SU-AI
In-Reply-To: Your message of 8-Feb-80 1634-PST
Who refused you an account?
-------
∂09-Feb-80 1738 ADMIN.MO at SU-SCORE (Maurice Bizzarri)
Date: 9 Feb 1980 1239-PST
From: ADMIN.MO at SU-SCORE (Maurice Bizzarri)
To: JMC at SU-AI
In-Reply-To: Your message of 9-Feb-80 1044-PST
Gee, I couldnt imagine anyone saying that. Noah Mendelsohn was
the name of the pony-tailed IBM staff member you were talking to
in the basement of Margaret Jacks hall Friday afternoon. Did he
tell you that? I'm sure we can arrange something...
Mo
-------
∂10-Feb-80 0928 Feigenbaum at SUMEX-AIM (Response to message)
Date: 10 Feb 1980 0929-PST
From: Feigenbaum at SUMEX-AIM
Subject: (Response to message)
To: CSL.FB at SU-SCORE, JMC at SU-AI
cc: eaf at SU-AI
In response to the message sent 9 Feb 1980 0839-PST from CSL.FB at SU-SCORE
Forest, I asked the same thing yesterday. What in the hell is going on?
John, please clue us in.
Ed
-------
∂10-Feb-80 1220 LES ARPA final report
To: JMC, TOB, DCL, ZM, RWW
Since we have just finished another ARPA contract, we owe them a so-called
Final Report. As usual, you should try to make it as painless as possible.
If you would like to start with the "accomplishments" part of the last
proposal to ARPA (last Spring), I can provide pointers to the text --
mostly it is in appropriately named files in [p,les].
The writeup need not be voluminous. A one-page overview of a given
project with pointers to published reports and articles could suffice.
Please try to make the list of reports and publications from your group as
complete as possible.
I would like to have all the material by February 25 so that it can be
compiled and go to press by March 1. Thanks for your help.
∂11-Feb-80 0445 JK RAMSEY
I have a 111-line proof on RAM.PRF, SET.FOL has the axioms and assumptions.
A month ago you mentioned the possibility of getting a home terminal-
is this still possible?
I modified the FOL proof file of RAM.PRF by putting in the correct commands
from the command file and trying to pretty print the formulas so that it will
be somewhat readable.
∂11-Feb-80 1104 doyle at MIT-AI (Jon Doyle) Job offer
Date: 11 FEB 1980 1339-EST
From: doyle at MIT-AI (Jon Doyle)
Sent-by: BORIS at MIT-AI
Subject: Job offer
To: jmc at SU-AI
CC: DOYLE at MIT-AI
Dear Prof. McCarthy,
If you have not yet sent me the formal offer of the position (I have
still not recieved it), could you please also let me know in it if
there is any provision for moving expenses in the offer? I will have
quite a number of boxes and stuff from both Cambridge and Houston if
I move to your area.
Thank you very much.
Jon Doyle
∂11-Feb-80 1120 LGC Revised Paper
If you're interested in taking a look at the revised version of my piece on
what the frame and qualification problems have to do with philosophy, it is
now as ready as it's going to be this afternoon. It's in PHIL.TXT[EP,LGC].
∂12-Feb-80 1211 MINSKY at MIT-AI (Marvin Minsky)
Date: 12 FEB 1980 1510-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
Yes, I would like to come. I didn't understand
reference to conference -- when is it?
The conference is monday march 24 to wednesday march 26 here on ai
and philosophy. You were supposed to have been sent an invitation,
and I will make sure.
∂12-Feb-80 1312 FWH PV+A Seminar: change of location
To: "@SEM2.DIS[SEM,VER]" at SU-AI
The location of today's PV+A Seminar (Huet: Demo on MENTOR)
will be Margaret Jacks Hall 408 (large terminal room 4th floor),
not MJH 426 as announced previously.
∂12-Feb-80 1339 TOB cart
John
I propose to request NSF funds for a new cart.
The old one is too large to get around in the
building. Also, we have shown enough results
to redeem our promise to ourselves that we would
get new equipment when we had shown progress
with the old. I propose a small electric cart,
stereo CCD cameras, sensors, and a microprocessor
on board. I plan to buy as much as possible
ready to go to cut engineering. The deadline
is March 1 for an equipment proposal. I want to
request funding for support of students to work
on it, perhaps in a separate proposal.
Will you be PI on the proposal(s)?
Tom
Glad to.
∂12-Feb-80 1353 MP
x11-Feb-80 1553 JMC Raymond Reiter
REITER.LE1 is a letter that I don't remember whether I sent. Do you
know whether it was sent?
The letter was sent l/30 and is in the Jan. chron file.
∂12-Feb-80 1420 MP
∂11-Feb-80 1832 JMC
See if you can find two passport photos in a file somewhere.
I found the photos and they're on your desk.
∂12-Feb-80 1507 MP
∂11-Feb-80 1558 JMC
Could you MAIL me Reiter's phone number? I think you have a letter filed.
I do not seem to find his phone number, although I have found his address.
Please let me know if you would like me to call the operator etc. to locate.
Yes, please locate him.
∂12-Feb-80 1549 MP Reiter's phone number
His phone number is 604-228-4142.
∂12-Feb-80 2313 PLS Seminars at the Institute
I've been out of circulation for quite a while as I've not been well,
and I wondered if there were still seminars happening at the institute.
I'm still not quite recovered, but I'll be around more and would like to
attend if they're still continuing.
Thanks,
Peter.
No seminars are currently scheduled.
∂13-Feb-80 1134 Bmoore at SRI-KL (Bob Moore) Workshop on Formal Methods in Artificial Intelligence
Date: 13 Feb 1980 1134-PST
From: Bmoore at SRI-KL (Bob Moore)
Subject: Workshop on Formal Methods in Artificial Intelligence
To: rbrachman at BBND, pcohen at BBND, doyle at AI,
To: pjh at SAIL, hendrix at SRI-KL, joshi at WHARTON,
To: zm at SAIL, jmc at SAIL, nilsson at SRI-KL, rperrault at BBNA,
To: brian at AI, waldinger at SRI-KL, webber at BBND, rww at SAIL,
To: woods at BBND
cc: stan at RAND-UNIX, bmoore
For some time we have felt a need for more communication among workers
interested in applying formal methods to problems in Artificial
Intelligence. To try to do something about this, we are proposing to
hold a workshop on formal methods in AI sometime in the next few
months. The purpose of the workshop would be to stimulate efforts to
develop more theoretically-based, formal approaches to core issues in
Artificial Intelligence, inluding the representation of practical
knowledge, commonsense reasoning, and problem-solving. We want to
emphasize these topics as opposed to mathematical theorem proving,
program verification, or the theory of heuristic search, which seem
already to be receiving sufficient attention.
The workshop will be by invitation only, with the number of
participants held to around twenty or fewer, to permit an informal
atmosphere and allow greater interaction among participants. A list
of the invitees is appended to the end of this message. We expect the
workshop to last for two days, with a mix of formal presentations and
general discussions. We hope to hold the workshop at the Pajaro Dunes
conference center near Monterey, California. The Rand Corporation has
tentatively indicated a willingness to support the workshop by paying
for the accomodations and meals and assisting with the travel expenses
of participants who do not have their own travel funds.
We are considering three different time periods for the workshop: late
April, early June, or mid August (just before the AAAI meeting in Palo
Alto). Of these three choices, we are leaning towards holding the
workshop in the August slot, with the expectation that participants
would favor consolidating their travel to the workshop and AAAI
meeting. There is a possible conflict with the AIM Conference, but as
far as we know, none of the intended participants will be involved in
that.
We need to know as soon as possible: Do you think you will be able to
attend the workshop, which of the proposed dates do you prefer, are
any of the proposed dates impossible for you, and do you need
assistance with travel funds? We need this information quickly so we
can schedule the workshop and make sure the facilities are available.
We also would like to know whether you have a formal presentation you
would like to give, and whether you have any proposed topics for
general discussion.
Bob Moore
SRI International
Stan Rosenschein
Rand Corporation
Invitees:
James Allen
Ron Brachman
Phil Cohen
Jon Doyle
Pat Hayes
Gary Hendrix
Aravind Joshi
Zohar Manna
John McCarthy
Drew McDermott
Nils Nilsson
Ray Perrault
Ray Reiter
Brian Smith
Richard Waldinger
Bonnie Lynn Webber
Richard Weyhrauch
Bill Woods
-------
I can attend on any of the proposed dates, will have a formal presentation,
and will not need support.
∂13-Feb-80 1443 ENGELMORE at USC-ISI Review
Date: 13 Feb 1980 1443-PST
Sender: ENGELMORE at USC-ISI
Subject: Review
From: ENGELMORE at USC-ISI
To: JMC at SU-AI, RWW at SU-AI, LGC at SU-AI, LES at SU-AI
Message-ID: <[USC-ISI]13-Feb-80 14:43:44.ENGELMORE>
Gentlemen,
I would like to be brought up to date on the Formal Reasoning
work, having not interacted with you for many months now. I will be in
the Stanford area during the week of March 31st thru April 4, and I'm
sufficiently flexible at this early date that you can pick any of those
five days for the meeting. Please let me know your preference ("don't
come" is not one of the choices!). Cheers,
Bob
Monday March 31 is best for me for the meeting with Englemore.
Any conflicts?
∂13-Feb-80 1452 SYS
Queued mail to englemore at USC-ISI -- ok
The following message was received from the remote host:
No such mailbox at this site.
An operator will attempt to deliver your mail.
∂13-Feb-80 1456 RWW
To: JMC, LES
Monday March 31 seems fine with me. Please let me know when it becomes
firm.
∂14-Feb-80 1247 ENGELMORE at USC-ISI
Date: 14 Feb 1980 1244-PST
Sender: ENGELMORE at USC-ISI
From: ENGELMORE at USC-ISI
To: JMC at SU-AI
Message-ID: <[USC-ISI]14-Feb-80 12:44:54.ENGELMORE>
In-Reply-To: Your message of 13 Feb 1980 1453-PST
John, I'll keep Monday the 31st open as a definite maybe. I think the
afternoon would be enough time, but I'm willing to spend the whole day
if you think it's necessary.
Bob
∂15-Feb-80 0546 Feigenbaum at SUMEX-AIM Re: Jon Doyle and inner space
Date: 15 Feb 1980 0548-PST
From: Feigenbaum at SUMEX-AIM
Subject: Re: Jon Doyle and inner space
To: LES at SU-AI
cc: dpb at SU-AI, jmc at SU-AI
In response to your message sent 13 Feb 1980 1639-PST
Les, thank you very much for having the courtesy to ask. Most people
don't.
The Research Associates come very high on the list of our priorities
since the department research is fueled by their activity. So JMC's
hiring of Doyle is a fine idea and we will create space for him, for
example by doing some shuffling related to visitors, or some other
rearranging related to students. I dont know specifically because
Denny keeps the details of this straight. But we will work it out.
Ed
-------
∂15-Feb-80 0706 JRA LISP CONFERENCE UPDATTE AND REMINDER
To: "@BB.DIS[1,JRA]" at SU-AI
A gentle reminder to all that there is about one month to the deadline
for submission of papers to the LISP Conference. --March 14 is it.
Here's an update on the status:
Aug 24, 1980, Sunday late afternoon/evening reception and registration
Aug 25, 1980, Monday morning: keynote address by John McCarthy.
Aug 26, 1980, Tuesday afternoon panel discussion.
evening: banquet
Around these events will be three days (Aug 25-27) of presented papers, demos,
and informal gatherings.
Since the AAAI Conference runs Aug 19-21, we are considering some activities
for the intervening space: LISP tutorials, demos, etc; trips to local wineries,
perhaps. It should be an interesting summer at Stanford! Tell your friends.
Send questions to JRA at SU-AI.
∂15-Feb-80 0927 DCD
Right you are, John.But Hillary might have. A good joke blown.
∂15-Feb-80 1403 100 : phyllis cost of versatec and alphatype
I checked with Luis Trabb Pardo, and he said the Versatec cost approximately
13K and the Alphatype approximately 30K.
-Phyllis.
∂15-Feb-80 1451 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, February 19
SPEAKER: Martin Brooks, Stanford University
TITLE: Demonstrating Correctness by Testing
ABSTRACT:
It is well known that "Program testing can reveal only the presence of
errors, not their absence." It turns out that if one is willing to make
assumptions about the types of errors which might occur in a program, then
this is not true. We will discuss our mathematical framework for program
testing, which relates the selection of test data to types of errors and
types of observations that are made during testing. The theory provides a
description of conditions which are suffient to insure that a set of test
data will reveal all errors in a program. The theory has been specialized
to the case of recursive programs having simple errors, where the
programmer observes a trace of each test run. An automatic test data
generator has been implemented. As input, the test data generator takes a
Lisp program and a description of the types of errors that may occur in
the program. It returns a set of test data. The programmer observes a
trace of each test run. If every trace is as the programmer intended,
then the program must be correct.
∂15-Feb-80 1640 BS Equipment costs
To: JMC at SU-AI
CC: BS at SU-AI
The alphatype cost approximately $30K; the versatec was not purchased--
it is on indefinite loan from Versatec. Luis Trabb Pardo (70481) has
details if you want to talk with him about it.
Betty
∂16-Feb-80 1029 REG I don't mean to sound too petulant, but
If Roy had brought back the test equipment that he's using to debug your
Imlac, then I might have a chance of finding out what's wrong with 0081, etc.
∂16-Feb-80 1316 REG
I found two modems, on 0081, and 0082 being sick. Replaced them both.
Also fixed 0087 which was really the line that would answer and then
hang up. The problem with 0087 seems to have been the modem to DAA
wiring.
I wonder if the problem on 321-4562 might be incorrect speed. The times
that I have tried it, it has been ok.
I eventually found most of the necessary test gear. I thought Roy had left
some of it in his car (beyond the scope that I knew was at your place).
I did set the Datamedia speeds to 300-300, but if you don't mind, I won't
pursue the matter further now. If there is something wrong, no doubt someone
else will complain.
∂17-Feb-80 0210 BTH concurrent programming seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CONCURRENT PROGRAMMING SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Thursday, February 28
SPEAKER: Amy Lansky, Stanford University
TITLE: EVALUATING SYNCHRONIZATION PRIMITIVES
ABSTRACT:
This talk is a review of a Master's thesis put out by Toby Bloom at MIT.
In it, she forms a "soft" measure by which one can evaluate the strengths
and weaknesses of sychronization primitives for concurrent programming
languages. At the basis of this measure is the notion that synchronization
achieves two major goals: exclusion and priority. Each of these goals must
be satisfied according to a subset of possible criteria such as: the time
of a request, the parameters of a request, the history of computation
preceding a request, etc. She compares 3 methods of synchronization
(monitors, serializers (beefed-up monitors), and path expressions). In
doing so she formulates a methodology of comparison as well.
∂17-Feb-80 1407 SELFRIDGE at BBN-TENEXD draft of TAC report
Date: 17 Feb 1980 1706-EST
From: SELFRIDGE at BBN-TENEXD
Subject: draft of TAC report
To: Fredkin at MIT-AI, minsky at MIT-AI, mccarthy at SU-AI
This is pretty late, I know, and my apologies.
If you can get back to me before, say, three on Tuesday,
I will be able to edit and change the report, which I will
bring copies of to the meeting Wednesday.
Oliver
Memo to: Ed Fredkin
from: Oliver G. Selfridge
Subject: TAC meeting 10-12 Jan 80
copies: Al Fenaughty, Marvin Minsky, John McCarthy, OGS
PARTICIPANTS
←←←←←←←←←←←←
III: Don Gerlich T.A.C.: Ed Fredkin
John Hansen Marvin Minsky
Tom McCarthy Al Fenaughty
Bob Waller John McCarthy
Gary Demos Oliver Selfridge
Steve Gray
John Banter
Mary Ellen Krell (limited)
Pat Anderson (limited)
BACKGROUND
←←←←←←←←←←
As a result of discussions at meetings of the Board of Directors,
it was agreed that the Technical Advisory Committee would meet to
consider the following topics:
1. Higher level languages
2. New hardware, including micro-miniaturized
3. Software management
4. Transitions and Compatibility
all within the contexts of a highly individualization market
structure -- each customer wants his own options and so on -- and
a rapidly and continually evolving technology.
In a letter to the TAC, Al Fenaughty provided us with literature
about some of the current III programming languages. He expressed
the hope that we could "arrive at a plan for our next generation
hardware and software architecture."
DISCUSSION
←←←←←←←←←←
EF started with a discussion of languages -- that is, higher
level languages --in order "to come up with some Company policy."
He suggested that representative candidates were BASIC, FORTRAN,
and PASCAL; he then assumed that BASIC and FORTRAN could be
"scratched out," but the assumption, not argument, was never in
←←←
fact argued. That left PASCAL, almost by default, as the sole
candidate. It was suggested that TRIP, whose documentation had
been among the literature distributed by AF, could be modified so
as to be added to PASCAL, or at least be compatible with it, in
some sense.
The nature of PASCAL, said EF, was "more modern -- thought out."
AF reminded us that "the thrust of the choice of software
language must depend on: ease of writing / coordinating /
updating and its performance on a machine.
!
AF expressed the consensus he heard: do the programming in a
PASCAL-type overlay of TRIP on 10's; in particular, a PASCAL with
certain added features, and with a carefully engineered
incremental compiling capability, or something equivalent in
linking loaders. The idea is that PASCAL on the 10 would allow us
to deal with a micro machined terminal.
GD argued that maintenance of PASCAL was currently "very dicey,"
and AF added that no matter what the language, "we won't get a
totally usable facility -- but we'll have to supply/write etc.,"
other features, variations, enhancements. As programmers learn it
and recognize new powers, however, the gains should quickly make
the change very profitable.
Three microprocessors were discussed, the Z80, 8080, and the
M68000. A certain amount of declaiming occurred. The 68000 "seems
completely conventional, a 16-bit bus, no difference from the
Z8000." Motorola's commitment was also questioned.
It is clear that a crucial question is the availability of I/O
controllers. And the center of discussion moved towards the
functions being performed by the hardware. The gist of this not
terribly compelling argument was to assume that the language was
PASCAL, for the time being, and to go on to the microprocessor.
It was pointed out that what was being discussed was on the order
of 25 Mbytes of programs and program-like data. A million words
of basic programs, more or less. "A typical typesetter has 20K
words of code and 10K in fonts," said Bob Waller. In fact, the
ecology of memory access affected all the areas of concern in the
developing technology:
1) the number of terminals supportable in TECS
2) the response times
3) files and file management
4) control and management of bulk storage.
This was especially noted in the way that DEC handles disc
accesses, which was reflected in the III composing programs which
use a separate disk access for each format call. A deal of
discussion arose for software fixes -- for example, keep local
copies of all format statements.
Then back to the microprocessor. AF asked what was appropriate
for page 2 as a subprocessor. Page 2 might be on two cards,using
wire wrap initially, with the 68000, say, providing half the cost
at half the size. Well, 2/3 the cost. Products that could use
this would include platemakers and controllers; furthermore, the
mirror correction control in the (not yet marketed) laser printer
uses the 8748.
In any event, it was agreed that it would be desirable to work
with Motorola and Intel managements (the latter especially) to
coordinate and plan future developments in chips. This effort to
compatibility with manufacturers' plans and products can be
called "system improvements through politics" (AF).
!
EF asserted without much disagreement that we needed a
development facility for the 6801, after a quick "peek" at the
8800. We ought to choose the better of the 8800 vs the 68000. AF
offered "to support your students to explore, say, the 6801. One
to four student projects ... we would send you keyboards and
camera controllers."
* * * * * *
There was some discussion about software problems and the
microprocessor hardware. "What is the likelihood of software
support from the manufacturer?" asked AF. EF and J McC pointed
out that Motorola had a FORTRAN cross-assembler for the 11 and
the 370, although the support for PASCAL was apparently not
committed.
As before, even if the industry produced usable and reliable
software tools, III would probably have to add modifications and
extensions of its own.It would be important to ensure upward
compatibility; J McC said that we should in that regard compare
the Kbox and the 68000. The Kbox is not outperformed by the
68000, said EF, and there is a real need for bench mark tests for
the components and the configurations.
Furthermore, although the need for better programmers' tools like
higher level languages is evident, we should remember that the
programming itself is also shaped by the other resources
available. Memory is so limited through high memory prices and
tradition that the designers "are handicapped' (EF). Programmers
and designers should learn to take advantage of more (vastly
more?) memory, both in core and on disc.
* * * * * *
Questions of file management and access and bulk disk memory led
us to discuss the disk controller. This is III-15 based, using a
simple seek/read on the 2314, with no overlap. The current (that
is, advanced← industry practice is to use an 11 or a NOVA 16 bit
separate seek/read on a 3300, that is, 200 Mbytes to 600 Mbytes
and up.
The obvious options, said MM, are
1. Design our own controller
2. Wait for the 68000 and adapt it
3. Use the 11 or NOVA
with new disks that will clearly be much larger and faster.
The consensus seems to be option 3, building a buss adapter. When
an I/O controller comes for the 68000 (from somebody else), we
should by all means use it to drive the system, moving modules
from the III 15 to the 68000 as feasible, which means as they are
designed, checked, and debugged.
!
* * * * * *
Somewhat later, MM and OGS conducted a survey of maintenance
programmers, who agreed that the primary need was good
documentation, and only then a higher level language. Estimates
←←←←
of times to be saved, etc., proved hard to find agreement on, but
they are probably closer to 50% than to 10 % or 90%. The extra
cost of good documentation would be about 25% of programming
time.
* * * * * *
There were a number of assumptions made that need to be thought
through again. John McCarthy raised one of these later, by
remarking that PASCAL was not necessarily the only language that
ought to be considered. He suggested looking at C, which had in
fact been raised briefly at the beginning of the meeting. C has
some attractive features, and Bell is planing to use it for large
systems, and may, in fact, have already done so; PASCAL has not
been tested with any large system yet, although there are lots of
plans. C will also need to be extended and enhanced for III
applications.
!
-------
I would like to see one additional issue mentioned in the TAC
report. It was remarked in several places that once III got the language
software, many modifications and enhancements would be required. This is
so, but III should make repeated efforts to innovate in language as little
as possible and TO RETURN TO THE FOLD as soon as the required features
were available in the standard systems. This is important in the
"improvements through politics" doctrine. When we first received the DEC
System 20 for student use, we had to modify the disk file system in order
to increase the number of allowed directories from 400 to 3000 so we could
give each student a directory. However, as soon as we succeeded in
politicking D.E.C. into putting the feature into the standard operating
system, we reverted to standard. This kept us in line for future
improvements. In the minutes so far, I see no indication that the need
for a continuing effort to stay with the manufacturer software is
understood.
∂17-Feb-80 1659 PN User's Guide to S-1 Pascal/Fortran System
To: S1 at SU-AI
USERGD.TXT[DOC,S1] contains random sorts of useful information about
using the S-1 Pascal/Fortran system, including
1) Differences between the Pascal accepted by various Pascal compilers
used by the S-1 project
2) An introduction to using the S-1, including an introduction to
FINE and TECO, the two editors currently in use.
Please feel free to add comments where things are inaccurate or incomplete,
or you think of something that might be useful for other people to know.
∂17-Feb-80 1922 SELFRIDGE at BBN-TENEXD your message
Date: 17 Feb 1980 2221-EST
From: SELFRIDGE at BBN-TENEXD
Subject: your message
To: mccarthy at SU-AI, minsky at MIT-AI, fredkin at MIT-AI
You're quite right, and I should have emphasized that. I will
modify the report tomorrow. I'm also not sure that I did your
feelings about C justice. I share many of them, but I wish I felt
clearer about the balance between C and PASCAL, based on so
little evidence that I know aout through personal experience.
Oliver
-------
PASCAL is largely being promoted in the computer science teaching
community. If C is being used for all kinds of real time system
work it is worth looking into. However, this assumes that it
is available. It just occurs to me that I will meet Doug McIlroy
on Tuesday, and I can ask him what the C situation is as far as
outside availability.
∂17-Feb-80 2024 LGC Engelmore Visit
March 31 is ok for me.
∂18-Feb-80 0938 JJK Tech reports to Russia
I've just read the Gray Tuesday minutes as reported by Don Woods, and I
wish to extend my support to your resolution to withold reports from the
USSR at this time.
∂18-Feb-80 1430 Feldman at SUMEX-AIM Reference
Date: 18 Feb 1980 1429-PST
From: Feldman at SUMEX-AIM
Subject: Reference
To: jmc at SU-AI
John,
We are considering Pat Hayes for a Chair of Cognitive Science. Your
recommendation would be most helpful.
Thanks,
Jerry
-------
∂20-Feb-80 0000 JMC*
Send "Two puzzles" to David Kaplan
∂20-Feb-80 0818 JRA longish, but IMPORTANT conf. note.
To: "@TLC.DIS[1,JRA]" at SU-AI
Dear arpa-net committee member:
Below is the edited text of a note sent to the poor unfortunates, not on
the net. The items that are of most importance are marked with ***.
-------------------------------------------------------------------------
Before the deadline for papers draws too near, I would like to discuss a
few other matters concerning the format of our conference.
First, to bring you up to date: we have space reserved on the Stanford
campus for housing, food services, and conference space. The banquet
details for Tuesday night are being completed. Details for publishing the
proceedings are being worked out now. I am trying to arrange for as many
demonstrations of LISP machines as possible. John McCarthy has agreed to
give the opening address on August 25, 1980.
It now turns out that in August 1980 Stanford will be hosting three
LISP-related conferences! AIM (A I in Medicine) Workshop will run August
13 through 15; alas, this workshop is by invitation only. Next, AAAI
(American Association for A I) will hold their first conference August 19
through 21; this will be a very large conference --about 1000 people are
expected. Finally, our conference will begin with a reception Sunday
evening August 24. With the expectation that many people will want to
attend the AAAI and the LISP conference (as well as some of the AIM
people), the housing office will arrange rooms so that people will not
have to move around. One need only advise the housing office of the
intention to stay when the earliest reservation is made.
*** There are three days between the end of the AAAI conference and the
beginning of ours. We may be able to entice more people to stay over if
we offer some social or technical events in the interim. We are
considering a tour through some of the Santa Clara area vineyards (Napa
Valley is a several hour bus trip, and that is not too attractive in
mid-August). I am also planning to arrange some tutorials, August 22-23,
for people wishing to learn more about LISP; any suggested events would be
appreciated.
*** It would also be attractive to have demonstrations --live or
video-tape-- of LISP machines. Any information about possible sources of
LISP-like machine demonstrations would be appreciated; these can range
from micor-computers to mainframes.
*** It is now time to get a fully-fledged publicity campaign underway;
this should include a glossy brochure mailed to anyone and everyone, and
probably an ad in the CACM and related magazines (after much gnashing of
teeth CACM got our call for papers published). So far, I have been footing
the bill (about $300), but I would like to find a "patron" to pick up this
latest expense. Any suggestions?
*** The final issue to be considered is the choice of participants for the
panel discussion. I would like the committee to suggest participants; from
the nominations, I will select about six individuals. The tentative title
for the discussion is "What is LISP?". The idea came about from an
informal discussion a few years ago at an ARPA workshop; the ARPA sponsor
asked the audience basically "what is it about LISP that is attractive,
and what could be done to make ADA attractive to the LISP community". I
regreted not having a tape recorder. Recently, a similar experience
occurred at the January Principles of Programming Lanugages conference,
where Alan Perlis debatted Peter Wegner on the topic "Will ADA make
programming cheaper?" (note: not "less expensive", but "cheaper"). Perlis
argued against the point, and sighted the LISP experience --its
programming style and environments-- as an example that people should
consider when examining the problems of making complex software.
It is important to make more widely known the benefits of the class of
languages and programming styles that LISP represents. I would like the
discussion to try to illuminate these ideas, giving some attention to what
future courses for programming languages might be. I would like the panel
to reflect a broad spectrum of background and opinion about applicative
programming. If you would suggest candidates for this panel, I would be
most grateful.
This summer at Stanford should be an exciting time!
Please, respond soon, and respond often to JRA at SU-AI.
I see no advantage in advertising LISP after 20 years. I also see no
advantage in the glossy brochure. I trust that you will be able to
reimburse yourself for expenses out of the conference registration.
What happened to your discussions with Nilsson about having IJCAI
grubstake the conference?
∂20-Feb-80 1351 HVA Dossier for DANIEL MEMMI
To: JMC, HVA
Marguerite Shaw (Pat Suppes' office), caled to ask if you had the Memmi dossier.
Pls. call her at 7-3111, or leave dossier or message for me and I'll follow
through. (Monica is out this week).
Don't think I ever saw it. Maybe it's in my incoming mail.
∂20-Feb-80 1354 DCD
To: "@AIPHIL[1,DCD]" at SU-AI
conf[1,dcd] has uptodate infor on the conference.
∂20-Feb-80 1648 TOB
To: "@GEVART.DIS[ARP,TOB]" at SU-AI
Visit by Dr. William Gevarter, NASA
Dr. William Gevarter, NASA, will visit SAIL in order to become familiar
with work done in AI at Stanford. NASA is now formulating a program in AI.
This visit is not connected with immediate funding opportunities, but may
help to shape their program.
I am setting up a schedule for Dr. Gevarter, tentatively for Monday, March 3,
in the afternoon. Please let me know quickly if we should reschedule for
Tuesday pm.
Feigenbaum%sumex,Buchanan%sumex,Lenat%sumex,TOB,BIS,ROD,DCL,TW
JMC,LES,RWW,ZM
I guess I had better remain on leave.
∂21-Feb-80 1008 HVA Daniel Memmi Dossier
John--thank you for reply--Carolyn has just glanced through your mail and it
doesn't seem to be here--I have so informed Pat Suppes' office.
∂21-Feb-80 1154 HVA Committee for Concerned Scientists
To: JMC, HVA
There was a telephone call just now and I suggested they try to reach you at
home (which I don't often do, but they had also tried you at the Center). The
question was whether you had rec'd their mailing of 2/6/80--various visas and
papers re your proposed trip mid-April to Soviet Union. The telephone number
is (212)686-8862, should you wish to call back.
∂21-Feb-80 1529 BTH concurrent programming seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CONCURRENT PROGRAMMING SEMINAR
Note that this talk has been postponed one week due to the talk by Reid.
PLACE: ERL 237
TIME: 2:30 p.m., Thursday, March 6
SPEAKER: Amy Lansky, Stanford University
TITLE: EVALUATING SYNCHRONIZATION PRIMITIVES
ABSTRACT:
This talk is a review of a Master's thesis put out by Toby Bloom at MIT.
In it, she forms a "soft" measure by which one can evaluate the strengths
and weaknesses of sychronization primitives for concurrent programming
languages. At the basis of this measure is the notion that synchronization
achieves two major goals: exclusion and priority. Each of these goals must
be satisfied according to a subset of possible criteria such as: the time
of a request, the parameters of a request, the history of computation
preceding a request, etc. She compares 3 methods of synchronization
(monitors, serializers (beefed-up monitors), and path expressions). In
doing so she formulates a methodology of comparison as well.
∂21-Feb-80 1522 ROY on TTY66 (at TV-66) 1522
Status of dial-in lines. Lines 0081 and 0082 still get wedged if they don`t get
carrier asserted in time and abort the call. They don`t wedge once they have
detected carrier independent of which end hangs up. If you find them in a wedged
condition (apparently half way between off-hook and on-hook) you can force them
to go on-hook by typing tty ttyx hang to the monitor where x=1 0081 and x=2 for
0082. These lines will be repaired as soon as possible.
Lines 0085 and 0087 have been repaired and were operational at the beginning of
this message. Roy.
∂21-Feb-80 1759 Horning at PARC-MAXC Re: here we go again
Date: 21 Feb 1980 5:46 pm PST (Thursday)
From: Horning at PARC-MAXC
Subject: Re: here we go again
In-reply-to: BTH's message of 21 Feb 1980 1658-PST
To: Brent Hailpern <BTH at SU-AI>
cc: "@GROUP.DIS[1,BTH]" at SU-AI
???
Maybe this was intended for some other distribution list?
Jim H.
∂21-Feb-80 1658 BTH here we go again
To: "@GROUP.DIS[1,BTH]" at SU-AI
CC: BTH at SU-AI
The text of the previous message was a little garbled:
Both Purdue and Cornell asked me to make sure that you have sent
your recommendations for me. Im really sorry about bugging you
about these details (there must be more important things to
bug you about!)
Thanks much,
Brent
∂21-Feb-80 1630 BTH special seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CONCURRENT PROGRAMMING SEMINAR
special monday session
PLACE: MJH 402
TIME: 3:30 p.m., Monday, February 25
SPEAKER: Loretta Guarino Reid, Carnegie-Mellon University
TITLE: Control and Communication in Programmed Systems
ABSTRACT:
This talk develops a general model of communicating programmed entities to
describe programmed systems. Such a system is needed to provide a language
in which to present and evaluate proposals for communication mechanisms
such as coroutine structures, message systems, monitors and shared memory
schemes. The model describes systems suitable for architectures ranging
from traditional uniprocessors to data flow machines to networks.
Communication mechanisms or abstractions are described as restrictions on
the order of certain actions. Such abstractions can be satisfied by a
variety of implementations meeting the restrictions. Within the model, we
can address such issues as the effect of naming abilities and binding
requirements, properties of communications mechanisms which prohibit
concurrency, and the effect of certain classes of system restrictions on
the power or flexibility of a system.
∂21-Feb-80 1621 BTH concurrent programming seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CC: REP at SU-AI, DEK at SU-AI
CC: JP at SU-AI
CONCURRENT PROGRAMMING SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Thursday, February 28
SPEAKER: Brian K. Reid, Carnegie-Mellon University
TITLE: Scribe: A High-level Approach to Document Production
ABSTRACT:
Computer systems have grown up enough that they are now useful for much
more than mere computation; for example, they make wonderful typewriters.
Perhaps because of this ancestry in the dark world of algorithm-executing
machines, computers seem to lure people into spending entirely too much
time and energy using computers to edit and print documents. Even when
using sophisticated text processing programs, the time that people spend
in developing, formatting, revising, and publishing their documents is
embarrassing.
The observer whose perception hasn't been clouded by the dazzle of Turing
equivalence will note that the best text processing system is still a good
secretary, to whom one communicates as little information as possible and
from whom one expects and usually gets perfect output.
I assert that the essential characteristic of the "Secretary" text
processing system is the information that is not communicated from the
author to the secretary, and which the secretary provides from his own
knowledge and experience. It follows that a computer system designed to
provide to an author the same kinds of support as would a secretary should
be controlled by the same style of high-level non-procedural language in
which one communicates with a secretary, and that a compiler for this
non-procedural language should fill in the details from its own knowledge
and experience.
The Scribe project at Carnegie-Mellon was a modest attempt to
Design a high-level non-procedural language for the specification of
documents, one that does not permit the specification of unnecessary
details;
Build a compiler for that language in which the details that a secretary
might customarily provide are filled in by the compiler from a knowledge
base;
Implement an "author's workbench" support system based on that compiler;
Extend the system into the realm of actual publishing.
I will cover some of the language and system design issues involved in a
document production system, and the particular set of design choices made
for Scribe, why they were made, and what has been learned. I will
conclude with an outline for the design of a next-generation
knowledge-based document system.
∂22-Feb-80 1133 CSL.OWICKI at SU-SCORE (Susan Owicki) CSL systems candidates
Date: 22 Feb 1980 1129-PST
From: CSL.OWICKI at SU-SCORE (Susan Owicki)
Subject: CSL systems candidates
To: CSD-Faculty:
CSL is interviewing Loretta Guarino on Wednesday, Febrary 27 and Brian Reid on Thursday Feb 28. Loretta is a student of
Habermann's, working in operating systems and programming languages. Brian
is interested in knowledge-based systems and is the designer of the
SCRIBE document-preparation system. I appended the abstracts of their talks
below. If you are interested in talking to either or both of them,
Leslie Hertzig (71175) can schedule an appointment.
Sue Owicki
CONTROL AND COMMUNICATION IN PROGRAMMED SYSTEMS
LORETTA GUARINO REID
ABSTRACT
This talk develops a general model of communicating programmed entities
to describe programmed systems. Such a system is needed to provide a
language in which to present and evaluate proposals for communication
mechanisms such as coroutine structures, message systems, monitors and
shared memory schemes. The model describes systems suitable for
architectures ranging from traditional uniprocessors to data flow
machines to networks. Communication mechanisms or abstractions are
described as restrictions on the order of certain actions. Such
abstractions can be satisfied by a variety of implementations meeting
the restrictions. Within the model, we can address such issues as the
effect of naming abilities and binding requirements, properties of
communications mechanisms which prohibit concurrency, and the effect of
certain classes of system restrictions on the power or flexibility of a
system.
SCRIBE: A HIGH-LEVEL APPROACH TO DOCUMENT PRODUCTION
BRIAN K. REID
CARNEGIE-MELLON UNIVERSITY
ABSTRACT
Computer systems have grown up enhey are now useful for much
more than mere computation; for example, they make wonderful
typewriters. Perhaps because of this ancestry in the dark world of
algorithm-executing machines, computers seem to lure people into
spending entirely too much time and energy using computers to edit and
print documents. Even when using sophisticated text processing
programs, the time that people spend in developing, formatting,
revising, and publishing their documents is embarrassing.
The observer whose perception hasn't been clouded by the dazzle of
Turing equivalence will note that the best text processing system is
still a good secretary, to whom one communicates as little information
as possible and from whom one expects and usually gets perfect output.
I assert that the essential characteristic of the "Secretary" text
processing system is the information that is NOT communicated from the
author to the secretary, and which the secretary provides from his own
knowledge and experience. It follows that a computer system designed to
provide to an author the same kinds of support as would a secretary
should be controlled by the same style of high-level non-procedural
language in which one communicates with a secretary, and that a compiler
for this non-procedural language should fill in the details from its own
knowledge and experience.
The Scribe project at Carnegie-Mellon was a modest attempt to
- Design a high-level non-procedural language for the
specification of documents, one that does not permit the
specification of unnecessary details;
- Build a compiler for that language in which the details that a
secretary might customarily provide are filled in by the
compiler from a knowledge base;
- Implement an "author's workbench" support system based on that
compiler;
- Extend the system into the realm of actual publishing.
I will cover some of the language and system design issues involved in a
document production system, and the particular set of design choices
made for Scribe, why they were made, and what has been learned. I will
conclude with an outline for the design of a next-generation
knowledge-based document system.
-------
∂22-Feb-80 1216 PAT the files on j,pat
belong to you and should be transferred to one of your areas...thanks
∂22-Feb-80 1328 BTH SNAFU * 2
To: "@GROUP.DIS[1,BTH]" at SU-AI
To all of you that were correctly confused, no I do not need your
recommendations. (Thanks for the offers, though.) That comes from
having more than one distribution list on my area with non-descript
names like GROUP. SORRY!
Next, Susan Owicki just informed me that the special monday seminar
has become a special wednesday seminar. She indicated no change of
room but the time is now 2:15 on Wednesday, February 27.
Sorry for the mixups.
CONCURRENT PROGRAMMING SEMINAR
special wednesday session
PLACE: MJH 402
TIME: 2:15 p.m., Wednesday, February 27
SPEAKER: Loretta Guarino Reid, Carnegie-Mellon University
TITLE: Control and Communication in Programmed Systems
ABSTRACT:
This talk develops a general model of communicating programmed entities to
describe programmed systems. Such a system is needed to provide a language
in which to present and evaluate proposals for communication mechanisms
such as coroutine structures, message systems, monitors and shared memory
schemes. The model describes systems suitable for architectures ranging
from traditional uniprocessors to data flow machines to networks.
Communication mechanisms or abstractions are described as restrictions on
the order of certain actions. Such abstractions can be satisfied by a
variety of implementations meeting the restrictions. Within the model, we
can address such issues as the effect of naming abilities and binding
requirements, properties of communications mechanisms which prohibit
concurrency, and the effect of certain classes of system restrictions on
the power or flexibility of a system.
∂22-Feb-80 1408 Wiederhold at SUMEX-AIM Trip and info
Date: 22 Feb 1980 1351-PST
From: Wiederhold at SUMEX-AIM
Subject: Trip and info
To: jmc at SAIL
Sorry that I left so hurriedly at the Airport, the Bus to the Biltmore was
just pulling out.
I was trying to get the Whitesmith address for the C-compiler, but the
person who has it has only one phone and seems to be glued to the terminal.
I should have it this weekend.
Gio
-------
∂22-Feb-80 1405 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
There will be no Program Verification and Analysis Seminar
on Tuesday, February 26.
Speaker and topic for March 4 will be announced next week.
∂22-Feb-80 1555 ROY on TTY100 (at TV-125) 1555
John the Imlac now loads from sail but it behaves strangly. I'll check out its
communication on the lines to see if things are being transmitted and received
properly Roy
∂22-Feb-80 1650 RAK via SU-TIP snafus
To: "@GROUP.DIS[1,BTH]" at SU-AI
Susan,
We have just got to stop having seminars on broadcast protocols.
BTH takes them far too seriously!
Dick
∂24-Feb-80 1436 BTH Concurrent Programming Seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
CC: JLH at SU-AI
CC: REP at SU-AI
Would you believe?
Loretta Guarino's talk will not be in MJH but back in ERL237. It will
be held in cooperation with the Programming Language Seminar. The final
(I hope) announcement is:
CONCURRENT PROGRAMMING SEMINAR
(in join session with the Programming Language Seminar)
PLACE: ERL 237
TIME: 2:15 p.m., Wednesday, February 27
SPEAKER: Loretta Guarino Reid, Carnegie-Mellon University
TITLE: Control and Communication in Programmed Systems
ABSTRACT:
This talk develops a general model of communicating programmed entities to
describe programmed systems. Such a system is needed to provide a language
in which to present and evaluate proposals for communication mechanisms
such as coroutine structures, message systems, monitors and shared memory
schemes. The model describes systems suitable for architectures ranging
from traditional uniprocessors to data flow machines to networks.
Communication mechanisms or abstractions are described as restrictions on
the order of certain actions. Such abstractions can be satisfied by a
variety of implementations meeting the restrictions. Within the model, we
can address such issues as the effect of naming abilities and binding
requirements, properties of communications mechanisms which prohibit
concurrency, and the effect of certain classes of system restrictions on
the power or flexibility of a system.
∂24-Feb-80 1643 HPM
∂24-Feb-80 1557 JMC
Rumor has it that you are at CMU now.
The rumors are correct, Reddy hired me as a "Research Scientist" in
his new interdisciplinary "Robotics Institute". I'm physically
at CMU this month, am returning to Satnford for the month of
March to do orals, and try to finish decorating thesis.
∂24-Feb-80 2328 HPM via CMU-10A dynamic support
To: JMC
CC: HPM
Thes are my first reactions: since existing materials
have strengths giving characteristic lengths of several
hundred kilometers, you should have
no problem making them up to a 1000 km long with taper.
The weak link, if any, is the deflectors. If you make them with
mechanical wheels, of course, as you once pointed out to me, the
spinning wheels will have to be just as strong as a full length
skyhook. In fact, my analysis of free space skyhooks shows that the
taper and mass of a free skyhoo is entirely independent of its
length, it has to be much fatter as it gets shorter to handle the larger accelerations.
The taper and mass go up exponentially with the square of the delta v
imparted.
I'm pretty sure you can't do much better with electromagnetic magic,
though you can keep the overall size of the deflector down by using
it repeatedly with small particles. I'll think about it.
∂25-Feb-80 0124 Mark Crispin <Admin.MRC at SU-SCORE>
Date: 25 Feb 1980 0122-PST
From: Mark Crispin <Admin.MRC at SU-SCORE>
To: JMC at SU-AI
In-Reply-To: Your message of 25-Feb-80 0113-PST
SD does not take a host name as an argument; rather it takes a
one or two character host id and ignores the rest of the command
line. The only recognized host id's are A (for AI), ML, MC, D (for
DM), and S (for SAIL). The reason it is this way is that SUPDUP
was written before NETWRK and other such advanced technology
existed, and those 5 hosts are the only hosts which support SUPDUP
servers anyway.
You have to use TELNET or DTN to get to SCORE.
-- Mark --
-------
Is there a good way to access SCORE if I want to use EMACS there?
∂25-Feb-80 1058 Wiederhold at SUMEX-AIM Supplier of C-compiler
Date: 25 Feb 1980 1053-PST
From: Wiederhold at SUMEX-AIM
Subject: Supplier of C-compiler
To: jmc at SAIL
is
Whitesmith Ltd.
127 East 59th street
New York 10022
-------
Thanks for the Whitesmith address.
∂25-Feb-80 1130 PAT pat,jmc
contains the following files
biojmc [curriculum vitae]
elepha.lst
jmc.bio [short biography]
kyoto.pub
there is no confidential material in these files, recommendations are
all on let,jmc.
∂25-Feb-80 1345 HPM via CMU-10A cross fertilization
I mentioned your idea to Bob Forward, and he pointed
out that the essence of your invention is a substitute for
a compressive member. He suggested a simpler variant;
a short anchord skyhook is supported by masses tossed
back and forth between its ends. I visualize a dual
mass driver at each end.
For your information.
∂26-Feb-80 0048 LLW Dynamic Skyhook Collaboration
To: JMC at SU-AI
CC: LLW at SU-AI, minsky at MIT-AI
∂25-Feb-80 1710 JMC Ballistic Support of Space Stations
To: LLW at SU-AI
CC: minsky at MIT-AI
Thanks for your help last night. Would you and/or Rod be interested
in being a coauthor of a paper on the concept? We especially need
a better discussion of the deflector than our own knowledge can supply,
but I suppose you would also be interested in configurations and
applications. Since we can all communicate by computer, I don't
envisage a face-to-face meeting.
[ Dear John:
Rod and I would both be interested in working on this with the two of you.
I agree that we should be able to get the necessary work done by computer
mail. Will you post a (very general, if appropriate) outline, and assign
authorship of the first draft by sections, so we will mesh well without a
great deal of overlap; also, would you designate a file area (with a
strong preference for SAIL, as neither Rod nor I speak foreign site
languages) where the master draft will be kept?
I will be on travel until Thursday evening, but both Rod and I should be
able to get our portions done in first draft this coming weekend (or so it
appears from here).
Lowell]
∂26-Feb-80 1324 MP
∂20-Feb-80 2018 JMC
Did my letter to Sloman go out?
That letter was sent 2/l3.
∂26-Feb-80 1331 MP Hirsch letter etc.
I have sent the letter by special delivery today.
∂26-Feb-80 1538 MINSKY at MIT-AI (Marvin Minsky)
Date: 26 FEB 1980 1833-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, jmc at SU-AI, llw at SU-AI
Sorry if this is a duplicate:
A message from HPM to JMC noted that Robert Forward suggests
a version of our ballistic-supported skyhook in which an
anchored platform is supported by vertically fired pellets.
Two bugs:
1. The vertical "orbit" is not straight because of earth
rotation, while the skyhook cable is. So line deflector is
needed. It would work at the poles, though.
2. Atmospheric resistance kills this scheme, but If the
sky-hook is hollow, then the atmospheric resistance could be
zero-ed. The cost of vacuumizing the interior of the
skyhook is hard to estimate: aspects
1. The extra structure is at the bottom, where it costs
the most.
2. The mechanics to resist vacuum is COMPRESSIVE; the
cable i under tension. It might be possible to get the
compression with small cost. It seems to me that the cost
should really be zero, practically. The tensile strength of
the super-strong fibres should not be much compromised by a
cross-compression of a mere 1 bar. Do any of you know how
to manipulate the appropriate stress-tensor?
3. If this is not the case, then the pressure-vessel
needs some extra structure and, unfortunately the
weight/meter to support the vacuum is proportional to the
area of the vacuum, so scaling up doesn't make it go away.
4. Of course, the size depends on the density x speed of
the particle stream. Clearly, two pipes are needed.
∂26-Feb-80 1548 Mark Crispin <Admin.MRC at SU-SCORE>
Date: 26 Feb 1980 1542-PST
From: Mark Crispin <Admin.MRC at SU-SCORE>
To: JMC at SU-AI
In-Reply-To: Your message of 25-Feb-80 0127-PST
Well, you can either use a hard-wired terminal or dial up on a Datamedia.
If you are talking about coming in from SAIL, there are several ways of
attack:
. From a DD or a III, use DTN (Datamedia simulating TELNET), then give
the following commands to SCORE:
TERMINAL DATAMEDIA
TERMINAL LENGTH 38
TERMINAL WIDTH 84
These are in a file <ADMIN.MRC>SAIL.CMD, which you can copy to your
directory then just say TAKE SAIL.
. From a DM, use TELNET in transparent mode (META-T) then say
TEMINAL DATAMEDIA
(TE D is good enough).
. From an IMLAC, good luck. You'll need it. Maybe DTN will do the
right thing, probably not.
. From another terminal (assuming SCORE knows it as a display type),
try TELNET in transparent mode (META-T) and tell SCORE the terminal
type you are using with TERMINAL xxx.
. Otherwise give up.
-- Mark --
-------
∂26-Feb-80 1657 HPM
To: minsky at MIT-AI, JMC at SU-AI, LLW at SU-AI,
forward at USC-ECL
Objection 1 is easy to counter; don't shoot straight, but at
a slight angle to counter coriolis force. Since particle
speed is probably very high, angle should be small.
∂26-Feb-80 2047 MINSKY at MIT-AI (Marvin Minsky)
Date: 26 FEB 1980 2235-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
Did you get a message (or several) from me about vertical
particle pipe? I created an automatic mailer, and am not sure it
worked.
I received one at 1538 only.
∂26-Feb-80 2047 MINSKY at MIT-AI (Marvin Minsky) Reply to HPM, 26 Feb 1657-PST
Date: 26 FEB 1980 2113-EST
From: MINSKY at MIT-AI (Marvin Minsky)
Subject: Reply to HPM, 26 Feb 1657-PST
To: HOOK at MIT-AI
With simple tubular vacuum vertical skyhook, we can avoid
all exponential growth of weight, because can intercept
fraction of particles at different heights. I envision it
as two tubes with occasional U-shaped superconducting
detours to provide lift. At the bottom is a huge deflector
that handles all particles. Perhaps tubes are lined with
superconducting rails, to simplify alignment problems.
The interval between ladder rungs is interesting. If they
are close enough, then there need be no extra material for
exponential growth of tension; only enough strength to hold
the vacuum in at atmosperic altitudes. Plus a constant
strength in addition -- to support whatever mass is to be
hauled up. But each "rung" entails the cost of maintaining
particles for its own deflector, and I suppose these might
be heavy.
And note that the latter cost does not have to be multiplied
when we are hoisting that load in each section, for any
loaded section can be given additional lift by deflecting
more particles at that level.
The earth station, then, would also contain a "particle
storage ring" to hold in reserve a supply of particles for
lift purposes.
This scheme -- of building tower-ladder vertically -- seems
easy to launch, too, since you just hoist up sections, bit
by bit.
no lunch tomorrow
∂27-Feb-80 1730 JMC*
call Barrett 202 632-5743
∂27-Feb-80 1730 JMC*
Linda Frenkel, Rutgers Provost's office, 201 932-7193
∂27-Feb-80 1751 FORWARD at USC-ECL Dynamic Compression
Date: 26 FEB 1980 2127-PST
From: FORWARD at USC-ECL
Subject: Dynamic Compression
To: moravec at CMUA
cc: minsky at MIT-AI, jmc at SU-AI, llw at SU-AI, forward
We can get lost in the detailed dynamics of each potential
application. The important thing to recognize (and get published
on other than the ARPA-net) is the M&M method of replacing static
compression columns with dynamic compression columns (using particles,
bb's, or M&M's, depending upon scale).
Bob
-------
∂27-Feb-80 1751 MINSKY at MIT-AI (Marvin Minsky)
Date: 27 FEB 1980 1108-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, FORWARD SUGGESTS JUST WRITING FAST at MIT-AI
To: SMALL PAPER at MIT-AI, McCarthy at SU-AI
without orking out dynamics, etc.
∂27-Feb-80 1751 MINSKY at MIT-AI (Marvin Minsky)
Date: 27 FEB 1980 1148-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
For ease of transmission, here is current state of
draft paper:
.require "memo.pub[let,jmc]" source
.at "qDv" ⊂"%AD%2v%1"⊃
.FONT B "ZERO30";
.TURN ON "↑"
.at "qm" ⊂"%B.%2m%1"⊃
.cb A SKYHOOK SUPPORTED BY ORBITING PARTICLES
.cb "Marvin Minsky, M.I.T."
.cb "John McCarthy, Stanford University"
This essay suggests some ways to support
"geo-stationary" space platforms at lower than synchrouous
orbit altitudes. The idea is to use streams of massive
ballistic particles. The platforms deflect these from one
orbit to another, and the difference in momentum flows
provides supporting forces.
Conceptually, the simplest version is one in which a stream
of particles is hurled up from earth, supoorting a platform
in space. The particles react elastically and return to the
source, where they are again deflected upwards. Such a
system is concervative; once the energy is put into the
particles, no more is consumed, if the collisions are all
elastic.
One application is for building a "skyhook" -- a platform
hovering, say, 300 kilometers above the ground. It is
feasible to then build a cable that hangs down to earth.
Previos skyhook proposals proposed cables that were much
longer, and posed impossible strength-of-materials problems.
There are several variants of the idea:
1. Each particle is in an equatorial elliptic orbit with
perigee just above the atmosphere. Soon after perigee and at an
altitude of a few hundred miles above the top of the atmosphere, the
particle encounters a deflector which deflects it towards the earth
but only about 20 degrees and not changing its speed and hence its
energy. The deflector is accelerated away from the earth by the
reaction but is constrained by skyhook cable anchoring it to the earth
This reaction is what keeps the cable and skyhook up. The particle
goes around the earth again on an orbit congruent to the previous one
but precessed relative to it. By the time it reaches the place in the
new orbit corresponding to the place where it was deflected before,
the deflector anchored to the earth has rotated with the earth to the
new place and deflects it again.
A new particle reaches the deflector each tenth of a second
so that the deflector and skyhook are continuously supported. Each
successive particle is in a new orbit congruent to the others but
rotated with respect to it by the amount the earth rotates in a tenth
of a second.
2. The second variant involves ⊗n skyhooks in a regular
%2n%1-gon around the equator. The particles travel at many times
orbital velocity and closely follow the %2n%1-gon. The higher
the velocity, the less mass is invested in the particle stream.
The mathematics for a hexagon is as follows:
⊗v is the velocity of the particles.
qDv is the change in velocity at each deflection.
⊗g is the acceleration of gravity.
⊗r is the distance to the deflector from the center of the earth.
⊗T is the time of flight of a particle between deflectors.
⊗M is the mass of a skyhook including deflector.
⊗m is the mass of the particle stream between two skyhooks.
qm is the rate of flow of mass by a skyhook.
%2k = m/M%1 is the ratio of mass of particles to mass of skyhooks.
The first group of equations are valid for any %2n%1-gon.
!!a2: %2qmqDv = Mg%1
!!a3: %2qmT = m.
For the hexagon we have
!!a1: %2T = r/v%1
!!a7: %2qDv_=_v%1.
(()) All this gives
!!a4: %2k = gr/v↑2%1,
and if we take %2g = 10%1, %2r = 7 x 10↑6%1 and %2v = 2 x 10↑4%1, all
in mks units, we get
!!a5: %2k = 0.7%1,
i.e. the mass of particles is about the same as the mass of skyhook.
Lower velocity particles would require more mass in the particle
stream.
As in other skyhook schemes, objects and material can climb
the cable on an elevator. However, because this skyhook ends only a
few hundred miles up (how many depends on the strength of available
materials) , the matter is not in orbit, so we have to consider what
to do next, and there are two main possibilities.
The first possibility is that there is another skyhook whose
bottom is at the top of the first one - and so on up to synchronous
altitude. This ladder is like the original skyhook concept except
that the ladder is supported at each rung.
The other possibility is that a mass driver is supported from
one or several skyhooks that accelerates mattter and/or people to orbital
velocities.
We have not determined the optimal size of particle or
whether magnetic or electrostatic deflection is preferable. They
need to have mass enough so that the reaction will support the
cable, so they shouldn't be ions.
More elaborate schemes are possible involving several
deflectors, and these may permit skyhooks above places on the earth
that are not on the equator. These would be optimal for communication
"satellites". For this application the cable might be dispensed with
and the position of the station maintained entirely by reaction. For
example, one could imagine a dodacehedral network of 12 satellites suspended above the earth.
Streams of particles between neoghboring pairs would act
like compressive structural members.
(Question: how few in such a network? Clearly one can do it
with four, in a tatrahedron, fixed in Galilean space. Is
there any problem in making it rotate with the Earth? I
don't think so. But, to obtain the sky-hook feature, there
is a tradeoff between low altitude and many-sided polygons,
and the associated expenses.
LAUNCHING
Because the system would fall if any part is missing,
perhaps the easiet launch method is to assemble the system
at near-geosynchrouous orbit. Then the reaction particles
are inserted at low velocities and the system is lowered
slowly, increasing the number and speed of the particles as
the vertical loads grow. Of course, some rockets will be
needed to correct for the higher natural speed of lower
orbits.
As for the skyhook cable, presumably one begins with a very
minimal one, and then uses it to hoise the materials for
larger ones.
We remind the reader of the final effect of a skyhook that
extends to synchronous orbit or beyond; once anything is
lifted to that point, and constrained by the cable, it
begins to have net outward force. In principle, then, one
can siphon any amount of matter further out with a gain (!)
in energy -- at the expense of the earth's angular velocity.
Inexhaustible.
Problems:
PLANAR ORBITS: there is a problem in supporting an
"unttethered polygon that is planar, because there is no way
to correct for deviations out of the plane.
DEFLECTORS: The scheme depends on being able to change the
angle of a fast -- order of escape velocity -- massive
particle, without energy loss, in a relatively short
distance. Are superconductor "mirrors" feasible for massive
particles?
LAUNCH POWER:
if our system stops at a single stage, e.g., at 800 KM
above the ground, we need further launch capability. One
could use a mass driver on the platform. Power could be
piped up the cable -- in principle -- but that might be too
weighty because the best conductors are not the strongest
tensile materials. Probably better to generate the power
there -- or beam it up via microwave.
PRECISION: any particle launch error will result in a
large spatial dispersion of the stream that strikes the
deflector. This makes an "aperture" problem, if the
"particle mirror" must have a wide receiving area.
This proposal already incorporates suggestions from Lowell Wood, Rod ?, Hans Moravec, and Robert Forward.
∂28-Feb-80 0039 LES Doyle funding
Given that Ben Moszkowski is on your NSF account and Bob Filman has left,
we have $16,000/year to work with in the ARPA budget. Depending on when
Doyle could come, this might be enough for the first year, but we would
have to deal with a squeeze next year.
∂28-Feb-80 1040 MP phone call
Linda Frenkl of Rutgers Univ. Provost Office would like you to call her
collect at 201-932-7193.
∂28-Feb-80 1240 LES
⊂¬28-Feb-80 0829 JMC
I don't understand when the year you refer to begins and ends.
-----
Feb. 1. Thus a start date of 15 June or later would supposedly be OK
provided that enough funding is assembled for the year beginning
Feb. 1, 1981.
∂28-Feb-80 1443 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, March 4
SPEAKER: Richard J. Lipton, University of California, Berkeley
TITLE: On the Safety of Encryption Systems
ABSTRACT:
In recent years there have been a large number of suggestions for
new methods of encryption. We will discuss here the safety of some
of these methods. In particular it will be shown how to "compromise"
some of these methods including the "mental poker" protocol.
∂28-Feb-80 1503 CG
Letter to Courant
Due to a beaurocratic error at the Courant institute, you may recieve a
second request for a recommendation letter from the people there.
I have written them to clear up the misunderstanding, and you
can safely ignore this second request.
Chris
∂28-Feb-80 1609 HVA Meeting Friday, Feb. 29: 1-3 p.m. Rm 190, Law School
To: JMC, HVA
Booz, Allen, Hamilton are presenting final report on Telecommunication Study
and Phyllis, Telecommunication (7-1532), called to ask if you will be present.
Will you?
∂28-Feb-80 1614 100 phone message
Anna Maria 7-3764 asked that you call her to set up an appt between you,
Gorin and Vevaney regarding LOTS and the need for an advisory board
∂28-Feb-80 1617 HVA Tel. Call
Now that's what I call a fast response--thank you! I have so informed Phyllis.
∂28-Feb-80 1638 DCD
To: "@AIPHIL[1,DCD]" at SU-AI
will someone jog those who haven't rsvp'd the conf invitation?
∂29-Feb-80 1030 Feigenbaum at SUMEX-AIM earlier message to Forest/with more
Date: 29 Feb 1980 1027-PST
From: Feigenbaum at SUMEX-AIM
Subject: earlier message to Forest/with more
To: csl.fb at SCORE, jmc at SAIL, les at SAIL
Date: 3 Feb 1980 1004-PST
From: Feigenbaum
Subject: Ullman/VLSI/money
To: csl.fb at SCORE
cc: jmc at SAIL, feigenbaum
Forest,
Jeff has initiated his work on "VLSI Theory" and wants to make it a main
research interest (have you seen his new paper?). I want to encourage him
to do so and Kahn to fund him (as Kahn funds you, me, and McCarthy in the area).
Clearly it's too late to get money for this fiscal year from Bob, so I am going
to try to scrape together some bucks from my ARPA contract to fund Jeff
until new money shows up. Do you have any bucks that can be scraped together
for this purpose? The total amount we would need until new money shows up
is in the low tens of thousands of dollars (before overhead).
I will send a copy of this message to John McCarthy to see if he has any spare
bucks in his "VLSI Design System" project.
Thanks,
Ed
-------
-----------------------------------------------------------------
Dear Forest , John, Les:
My latest calculation says that Ullman needs $20K, including overhead,
for the Spring Quarter and one month in the summer. I propose that we
give it to him, half out of the Feigenbaum-McCarthy ARPA contract for
VLSI and half out of Forest's ARPA VLSI contract. Jeff will have his
own money from ARPA starting Oct. 1.
John and Les: is this OK with you guys? We have spent zero or close to Zero
on our ARPA contract. But we have had to pay for the VAX entirely out of
first years' money (hence $210K is earmarked). This stills leaves asizeable
pot. Do you plan to hire people into your portion soon or otherwise use it
up? If you plan to be slow about this, it should leave plenty of money
to float Jeff Ullman at 10K.
RSVP!
Ed
-------
How do we stand with Ed's request to help Ullman? I think we are getting
started slowly and so can do it, especially if we can perhaps get some
of it back later.
∂29-Feb-80 1059 TOB
∂28-Feb-80 1719 JMC
John
In what regard is this? Gevarter's
visit?
tom
I guess I had better remain on leave.
∂29-Feb-80 1338 Hazen at SUMEX-AIM Meera BlattnerNSF,here to visit 3/4/80
Date: 29 Feb 1980 1336-PST
From: Hazen at SUMEX-AIM
Subject: Meera BlattnerNSF,here to visit 3/4/80
To: faculty list:
Meera Blattner, Program Director for TheoryComputer Science office
National Science Foundation will be visiting Tues. Afternoon 3/4/80 and
will be available to discuss the research supported in her program. She
is also prepared to discuss other Computer Science funding programs at
the NSF. If you want to see her arrange anappointment with<Rosenow@sumex>
Ed Feigenbaum
-------
I would like to see Meera Blattner of NSF on Tuesday.
Any afternoon time would be fine.
∂29-Feb-80 1631 LES
∂29-Feb-80 1033 JMC
How do we stand with Ed's request to help Ullman? I think we are getting
started slowly and so can do it, especially if we can perhaps get some
of it back later.
-----
Right. I spoke with Helliwell again yesterday. He would like to come in
sabbatical mode, which would be free to us. It remains to be seen whether
it can be worked out at his end.
I have no other prospects currently, so supporting Ullman sounds like a
good idea.
∂02-Mar-80 1144 CLT concert
Waverly Consort this coming Tuesday evening--8pm
∂02-Mar-80 1753 LLW Cannonball Speed and Dissipation
To: JMC
CC: LLW, "#JUNK2.MSG[1,LLW]"
∂02-Mar-80 1540 JMC paper
Two questions:
1. What is the feasible range of velocities for the cannonballs? The
faster they are, the less mass is required in a system.
[No upper limit; if energized by ion rocket pods, they can reasonably get
to 1000 klicks. Faster is also better as it stores more energy in a given
amount of cannonball mean momentum flux--the system becomes `stiffer'
against loading of the cables.]
2. What is likely to be the major cause of dissipation, and do
you have any idea how large it would be?
[Eddy current dissipation when the cannonballs interact with the loops--
this can be made (nearly) as small as one pleases, by cooling the metal
shells carrying the eddy currents as the cannonball sweeps through; the
resistence drops exponentially, while the thermodynamic penalty increases
only hyperbolically. Shock absorber losses can likewise be made almost as
small as you want to take the trouble to design for. The system really is
very close to a perpetual motion machine of the second kind.]
∂03-Mar-80 0000 JMC*
Call San Jose water works.
∂03-Mar-80 0943 Rosenow at SUMEX-AIM Meera Blatner appointment
Date: 3 Mar 1980 0941-PST
From: Rosenow at SUMEX-AIM
Subject: Meera Blatner appointment
To: jmc at SAIL
cc: feigenbaum
I have scheduled an appointment for you at 2:00. If you would like to
change this, please let me know.
Jeanie
-------
∂03-Mar-80 1132 CLT stamps
ok
∂03-Mar-80 1244 HVA SECRETARY FOR YOU
To: JMC, HVA
I have a very interesting candidate and hope you can meet her this week (she
will be out of town next week , and Carolyn tells me your calendar is very
full after that). Frances Larson could be available at any time during the
day of Thursday, March 6th--what time would be best for you?
3pm would be most convenient, but any time is possible.
∂03-Mar-80 1257 Rosenow at SUMEX-AIM meeting place
Date: 3 Mar 1980 1255-PST
From: Rosenow at SUMEX-AIM
Subject: meeting place
To: jmc at SAIL
In Ed's conference room--sorry I forgot to mention the place.
Jeanie
-------
∂03-Mar-80 1500 HVA Thursday Mar. 6th--Sect'y
Thank you John--I've called Frances Larson, but she is not at home, so I'll keep
trying and let you know tomorrow.
∂03-Mar-80 1836 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
For those that have access to CS bulletin boards and gotten confused:
The seminar is Tuesday. Here again the announcement:
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, March 4
SPEAKER: Richard J. Lipton, University of California, Berkeley
TITLE: On the Safety of Encryption Systems
ABSTRACT:
In recent years there have been a large number of suggestions for
new methods of encryption. We will discuss here the safety of some
of these methods. In particular it will be shown how to "compromise"
some of these methods including the "mental poker" protocol.
∂04-Mar-80 0143 LLW Skyhookery
To: minsky at MIT-AI
CC: LLW at SU-AI, JMC at SU-AI, "#JUNK2.MSG[1,LLW]" at SU-AI
∂03-Mar-80 1242 MINSKY at MIT-AI (Marvin Minsky)
Date: 3 MAR 1980 1408-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, LLW at SU-AI
ANY FURTHER CONCLUSIONS ABOUT DEFLECTING ESCAPE
VELOCITY PARTICLES? IF ANY OF THE SYSTEM IS ELASTIC
(OR INELASTIC) THERE COULD BE AWFUL LOSSES.
DHE MORE I LOOK AT IT, THE MORE INTERESTING SEEMS THE
U-TUBE IDEA. THE SYSTEM IS TWO MASS DRIVERS,
ONE DECELERSATING UPBOUND PARTICLES, THE OTHER
ACELERATING RETURNING ONES DOWNWARD, WITH A
STORAGE RING AT THE GROUND. QUESTION IS
WHEHER THE THING IS JUST ORDERS OF MAGNITUDE TOO MASSIVE.
[Rod is developing some very interesting numbers on the particular variant
which exploits a ring orbiting more slowly than the stable orbit
corresponding to its diameter, supporting a multiplicity of skyhooks via a
longitudinal slip joint. Using reasonable numbers, the system's mass
reproduction time (the interval over which it can lift its own weight into
space from the ground) is of the order of a day!
I'm collecting all communications on this topic in the file area
JUNK2[1,llw], whenever you're interested in scanning over it--it will
probably serve as the area for the paper draft, as well. (Lest this be
misinterpreted, virtually all my working files belong to the JUNKxx
sequence.)
Lowell]
∂04-Mar-80 0206 LLW Request for SAIL User Status
To: LES
CC: LLW, JMC
Dear Les:
Would you please confer SAIL user status on Roderick A. Hyde (user name
'Rod Hyde';l 50 K word disc allocation; password 'SKYHOK'), a member of my
Group at LLL and a corresponding member of the S-1 Project? Rod will be
using 'E' to do most of the (co-)authoring of a paper with John, Marvin
Minsky and myself, and this will involve enough logged-in time that it
would be markedly inappropriate for him to use someone else's account.
Thanks,
Lowell
∂04-Mar-80 0854 HVA Thursday March 6th--3:00 p.m
To: JMC, HVA
Frances Larson will be here for interview. Thank you.
∂04-Mar-80 1324 ADMIN.MO at SU-SCORE (Maurice Bizzarri)
Date: 4 Mar 1980 1019-PST
From: ADMIN.MO at SU-SCORE (Maurice Bizzarri)
To: JMC at SU-AI
In-Reply-To: Your message of 9-Feb-80 1809-PST
your 4331 account id=mccarthy password=8480898 is open
Mo
-------
Thanks Mo, and please transmit my thanks to whoever co-operated.
∂04-Mar-80 2158 JK knowledge representation
I have some thoughts on the FOL system, meta theoretical reasoning,
and a suggestion for another way of solving the Mr.S and Mr.P puzzle
written up in KNOW.SUG[1,jk]. I would be curious to hear what you think of
them though I am definitely not committed to them.
Incidentally, I had a very interesting conversation with Cohen
about proof-checking; he does have some worthwhile ideas.
∂05-Mar-80 1114 HVA AUDITRON
To: JMC
CC: CLT, HVA
In checking monthly count, I discovered that your auditron is not working, i.e.,
it activates the coping machine but does not record copies--so I've taken it
and will replace it with new one.
∂05-Mar-80 1208 Denny Brown <CSD.DBROWN at SUMEX-AIM> CS206 for fall quarter
Date: 5 Mar 1980 1155-PST
From: Denny Brown <CSD.DBROWN at SUMEX-AIM>
Subject: CS206 for fall quarter
To: jmc at SU-AI, clt at SU-AI
What are the plans for CS206 next fall?
- Who?
- When?
choices 1. MWF 9 (Was done here this year.)
2. TTH 9:30-10:45. (Used to be done here?)
3. Other preference?
Note: TTH 11-12:15 is not a televisable time. We put CS204
in that slot since it is not a televisable course.
We go to press this week. RSVP -Denny
-------
I will teach it. Either time is ok.
∂05-Mar-80 1331 MP
∂04-Mar-80 1712 JMC
Sending the proofs back to North-Holland is urgent. Be sure it's airmail.
Please explain this message-I don't recall sending any proofs.
∂06-Mar-80 0200 JMC*
phone Sloman
∂06-Mar-80 0837 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Admissions (from Gene Golub)
Date: 6 Mar 1980 0832-PST
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: Admissions (from Gene Golub)
To: CSD-Faculty:
-------
∂06-Mar-80 0851 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) [CSD.BECK: cs390]
Date: 6 Mar 1980 0845-PST
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: [CSD.BECK: cs390]
To: CSD-Faculty:
Date: 31 Jan 1980 1934-PST
From: CSD.BECK
Subject: cs390
To: csd.tajnai
I am taking cs390 from mike genesereth - that is section 81, I think. I plan to
do an add/drop tomarrow (fri) on monday, so the department
shouldn't have to do anything about it. is there some technicai]lity involved
in doing that? if so, please let me know. Sorry for the mix-up.
-Micah
-------
---------------
-------
∂06-Mar-80 0856 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Admissions
Date: 6 Mar 1980 0853-PST
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: Admissions
To: CSD-Faculty:
cc: CSD.Tajnai at SU-SCORE
5-Mar-80 16:42:00-PST,00000001100;000000000001
Date: 5 Mar 1980 1641-PST
From: CSD.GOLUB
Subject: admissions
To: csd.Faculty
We are in the throes of making our decisions with respect to admission
to the Ph.D. program for next year. It appears to me that the
applications for next year are unusually good; furthermore, we have
more applicants than ever before. The policy in the past has been to
admit between 20 and 25 students. Last year, we admitted 25 students
and we had several students on the waiting list. Eventually, 3 students
from the waiting list were admitted, and a total of 18 students came
this year.
In view of the fact that we seem to have a bumper crop of applicants,
how many students do you think we ought to aim for for next year?
Recall that we now have a new faculty member--Jeff Ullman, and perhaps
we can increase the number of admittees.
Also, there are several Republic of China applicants. We don't have
much specific information on them, other than that those who are
here are very bright. Should we place them in a special category?
Any advice would be appreciated.
Gene
-------
-------
I think we might increase the number a bit. I assume you meant
"People's Republic of China" which is the mainland not "Republic
of China" which means Taiwan. Republic of China can be treated
in the usual way, I would think. As for the People's Republic,
judging from the case of Yue, it seems to work to admit them to
the Master's program, and then see if they work out there.
I hope there are a few good students who will take an interest
in AI theory - preferably with some background in logic.
∂06-Mar-80 0900 JMC*
manne+moses
∂06-Mar-80 0900 JMC*
curtis
∂06-Mar-80 0912 BOBROW at PARC-MAXC2 Re: Sloman
Date: 6 MAR 1980 0910-PST
From: BOBROW at PARC-MAXC2
Subject: Re: Sloman
To: JMC at SU-AI
In response to your message sent 06 Mar 1980 0059-PST
I phoned Aaron this morning, and he said you had already called
him. He asked me to give you his flight information.
Pan Am PA125 arrving 15:10 on Sunday the 23rd of March
See you at the center, if not before.
danny
-------
Thanks. I asked for a home number, so I could call him during working
hours, but I stayed up late enough to call him at work.
∂06-Mar-80 0919 PJH
sloman: will do. pat.
∂06-Mar-80 0927 Denny Brown <CSD.DBROWN at SU-SCORE>
Date: 6 Mar 1980 0925-PST
From: Denny Brown <CSD.DBROWN at SU-SCORE>
To: JMC at SU-AI, DPB at SU-AI
In-Reply-To: Your message of 5-Mar-80 2012-PST
We go to press (preliminary version) tomorrow. So it's not too late,
but we need the new copy quick. -Denny
-------
∂06-Mar-80 0926 JK remarks
Thanks for your comments. Tuesday is an impossible day for me; how
about Wednesday or Friday? Meanwhile, where can I find out more
about (a)abstract syntax (b)Montagues paradoxes?
(b) "A Paradox Regained" with David Kaplan, Notre Dame Journal 1960 and
Syntactical Treatments of Modality, with Corollaries on Reflexion
Principles and Finite Axiomatizability in Acta Philosophica Fennica 1963.
Both papers are reprinted in his posthumous Formal Philosophy, Yale
University Press, 1974.
(a) Abstract syntax is introduced in my "Towards a Mathematical
Science of Computation" proceedings International Congress on
Information Processing 1962. Its not a deep idea. Instead of
regarding wffs and their parts as strings, we can regard them
as elements of an abstract space with predicates and functions
to test them and taket them apart and put them together. If one
is using FOL, one can define the domains abstractly and not have
to deal with FOL formulas.
∂06-Mar-80 1018 DBL Janet Kolodner
John,
Janet Kolodner interviewed here last week, to mixed response.
If you have any recollections from last summer at Yorktown to
input, please do so. Current feeling is that she is too into
psychology and data bases to fit in well here (I mildly disagree).
Barring some groundswell of support, she won't be offered any
type of position by the department. She had wanted a split
appointment in CS and Psych (which Gordon Bower said he would
support), and we probably could CONS affirmative acion money
for such a slot if there were enough support within our dept to
bring her here.
Doug
I didn't have a strong enough reaction to justify action.
∂06-Mar-80 1631 MP phone call
Jerry Sullivan, an attorney from San Francisco called today at 433-0440
and would like to locate Arlene Bloom. Please call him.
∂06-Mar-80 1750 LGC Design Reference?
Have you published anything that discusses (or even mentions) your slogan
about design preceding construction planning, i.e., anything to which one
might refer in crediting you with the idea?
No. Better just mention a personal conversation.
∂07-Mar-80 0243 ARK S-1 Meeting Notice
To: S1 at SU-AI
CC: Pickering at SUMEX-AIM
There will be an S-1 meeting on Thursday, 13 March at 12:00 noon in room
252 Margaret Jacks Hall. Please send progress reports to area mailing
list (e.g., S1SOPA) with a copy to S1PROGRESS-LOG. Thanks.
Arthur
∂07-Mar-80 0800 JMC*
curtis 632-7346, Kathy Green 714 547-6133
∂07-Mar-80 0844 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: 4331
Date: 7 Mar 1980 0842-PST
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: Re: 4331
To: JMC at SU-AI, EAF at SU-AI, FB at SU-AI
In-Reply-To: Your message of 6-Mar-80 1654-PST
John, I'll have to find out why you're having so much trouble with this.
Friedl or Flatt (both very nice guys, incidentally) have no control over who
we at Stanford put on the machine. So there must be something deeper and
perhaps more insidious going on.
I don't agree with you about the floor space being more valuable than the 4331.
But I'll get back to you when I find out what's going on.
Forest, do you have a clue to this mystery?
Ed
-------
∂07-Mar-80 0852 BTH Concurrent Programming Seminar
To: "@GROUP.DIS[1,BTH]" at SU-AI
PROGRAM VERIFICATION AND CONCURRENT PROGRAMMING SEMINARS
PLACE: ERL 237
TIME: 2:30 p.m., TUESDAY, March 11
SPEAKER: David Stevenson, Stanford University
TITLE: Techniqus for Translation of ADA Multitasking
ABSTRACT:
(none at present)
Note that this is a Tuesday seminar, there will be no seminar on Thursday.
∂07-Mar-80 0909 JMC*
mazda
∂07-Mar-80 0909 JMC*
faculty club 4325
∂07-Mar-80 0909 JMC*
Manne, moses Starr
∂07-Mar-80 1039 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) IBM machine and use
Date: 7 Mar 1980 1037-PST
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: IBM machine and use
To: jmc at SU-AI, csl.fb at SU-SCORE
cc: csd.feigenbaum at SU-SCORE
I just spent an hour on the phone with Paul Friedl, which in itself indicates
that the problem raised is not trivial. Friedl and Flatt are as harrassed
by the hemming-in that the IBM lawyers are doing as we are. (As we have known
all along, the sensitivity rlates to the 1969 consent decree and the pending
anti-trust suit; now that they are opening a door the lawyers are being
cautious). In any event here is the situation:
1. The use of the 4331 has been set up on a "project" basis (similar to
SUMEX but diametrically opposite to SCORE).
2. There will be a project called Exploratory Research and Familiarization,
whose Principal Investigator will be me as Chairman of the Department.
Anyone who like JMC wants an account to "try out some things" can get one
through my secretary (after this gets set up in a few days).
3. We are going to have John Porter, the IBM person in charge of the entire
Joint Study Program idea in IBM, come out to be told about how harrassing
the whole thing has become.
Friedl and I both agreed that if the lawyers set things up so that we cant
easily get done what we want to get done and behave toward each other like
scientists, that we should remove the machine. He seems as frustrated
(angry?) as we are.
Ed
-------
∂07-Mar-80 1038 HVA Curriculum Vita
To: JMC, HVA
John--cld. you provide me with current copy of your cv-or tell me where to
find it? There will be proposals going out (from TOB), in near future and
I'll need the CV. Thks.
∂07-Mar-80 1048 Hans.Berliner at CMU-10A (C350HB03) Committee for Fredkin Prize rules
Date: 7 March 1980 1344-EST (Friday)
From: Hans.Berliner at CMU-10A (C350HB03)
Subject: Committee for Fredkin Prize rules
To: jmc @ su-ai
Message-ID: <07Mar80 134406 HB03@CMU-10A>
John,
As you may have heard, Ed Fredkin has donated $100,000 to IJCAI to establish
the Fredkin prize to be awarded to the first computer program that becomes
World CHess Champion. I have been asked by Raj Reddy to form a committee to
negotiate with FIDE to produce rules that would guarantee both the rights of
human opponents and to prevent computers from arbitrarily being barred from
competition at some future date. In case it is not possible to get unlimited
participation guaranteed by FIDE, we might be interested in special competitions
for an OPEN world championship open to machines as well as humans. In any case,
there are a number of things that need to be determined and rules for play under
the prize specified and conditions for awarding, etc. My question is:
are you interested in serving on such a committee. I expect that there
will be four people on it, when we get done, and there will be some collecting
of suggestions, and possibly some meetings with various officials when things
get to that point. At the moment, I would like to form the committee because
this will be part of the announcement that Raj will make about the prize.
Thanks for your consideration of this.
Hans
In my opinion, offering the prize at a time when AI is so far from being
able to deliver creates expectations that will lead to disappointment
and recrimination. Also the work put into negotiating with FIDE might
better be spent in science. Modest suggestion: put off even the negotiation
till a program reaches the master level. It will take less time then,
because FIDE will take it more seriously then. In any case, include me
out.
∂07-Mar-80 1253 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
and
CONCURRENT PROGRAMMING SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, March 11
SPEAKER: David Stevenson, Stanford University
TITLE: Techniques for Translation of ADA Multitasking
ABSTRACT:
A technique is presented for translating the multitasking constructs of
ADA into the medium level language ADA-MINUS. ADA-MINUS has been
developed by the Verification Group at Stanford as a tool for development
and verification of parallel systems. The multitasking features of
ADA-MINUS are a convenient and useful means to study possible
implementation and compilation techniques for ADA and to examine the
relative efficiencies of various implementations.
∂07-Mar-80 1426 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) [Terry Winograd <TW at SU-AI>: Response to note on admissions ]
Date: 7 Mar 1980 1420-PST
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: [Terry Winograd <TW at SU-AI>: Response to note on admissions ]
To: CSD-Faculty:
Mail-from: ARPANET site SU-AI rcvd at 7-Mar-80 1247-PST
Date: 07 Mar 1980 1247-PST
From: Terry Winograd <TW at SU-AI>
Subject: Response to note on admissions
To: csd.golub at SU-SCORE
CC: csd.tajnai at SU-SCORE
Gene,
I have mixed feelings about accepting more students in a situation
where we are already feeling jammed into the building. Unless there
is an obvious space gain to be made in the next year, I would try
to follow a ZPG policy.
As to the Chinese, I suspect that over the next few years we could fill
the entire department with Chinese students while maintaining the same
level of quality---given a population that size to draw on, the
traditional Chinese emphasis on learning and technology, the current
interest in modernization and the absence of good training schools there.
Although quotas have their problems, we clearly can't just become the
"Peking abroad" institute of computer science, so we will have to have
some policy of limitation. I think that something on the order of 2-3
students per year is reasonable. That leaves the problem of evaluating
records of people from a totally different system, but hopefully there own
government will do some quality control on who gets to apply.
In my continuing work with Mr. Yue, I remain highly impressed, and
regardless of the general policy on Chinese admissions would like to
encourage his acceptance for next year.
--t
---------------
-------
∂07-Mar-80 1435 LES ARPA final report reprise
To: JMC, TOB, DCL, RWW
On Feb. 10 I sent you a request for material to be included in the ARPA
final report, specifying a Feb. 25 due date. I have received just one
section so far, from ZM.
Today I received a second "overdue" notice from the contract monitor.
Please help me wrap this up before it gets bloody. As I indicated before,
the write-up need not be elaborate.
∂07-Mar-80 1639 100 : monic my leaving
Today is my last day at Stanford and I want to say goodbye and I've enjoyed
working at the lab and for you. I have intructed Lyn Larsen my replacement
in your procedures, so I hope the transition is smooth. Today I obtained
the blank film that you requested for the 60 minutes taping and would you
please let Lyn know if you would like it sent. Her ppn is egl.Thank you.
Please send the tape to Deakins at the address given in the letter
I gave Hersche and send a suitable cover letter. Be sure the return
address and a card with the tape identify who sent it to them.
Thanks.
∂08-Mar-80 1636 MINSKY at MIT-AI (Marvin Minsky)
Date: 8 MAR 1980 1935-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
I just got copy of Searle's paper. Enclosed is my theory of
his problem. I really feel this is the trouble, and explains
why our friends can't find the trouble
within the paper. Also applies, I think
to what I was saying to Fodor -- about the concept of
"reference".
here it is:
On John Searle's paper: "Minds, Brains, and Programs"
Marvin Minsky
M.I.T
The trouble is not within this technically brilliant discussion but in
an absurdity it assumes from the start -- that the everyday concepts
of "intentionality" or of "belief" are robust enough to permit such a
precise discussion. Searle says without blushing: "I will simply
assert the following without argument. The study of the mind begins
with such facts as that humans have beliefs, while thermostats,
telephones, and adding machines don't. If you get a theory that
denies this --- the theory is false".
This is so wrong. The study of Mind is not the study of Belief --
despite Searle's career commitment to that position. Instead, it is
the attempt to discover powerful and intelligible concepts -- be they
old or new -- that help explain why some machines or animals can do so
much more than others, e.g., manipulate processes that so effectively
control things in the world, perhaps via producing robustly reliable
"extensible" generalizations, internal models, or whatever.
In the previous century, biological philosophers must have argued
about machines and life much as Searle argues about programs and mind:
"The study of Biology begins with such facts as that humans have
→life_, while thermostats, telephones, and adding machines don't. If
you get a theory that denies this --- the theory is false". As all
readers know, we today understand much of "life" in terms of its
energetics and information-processing structures. The absolute,
precise quality of "self-reproduction", for example, survived quite
accurately in the DNA, but not on the larger scale. Though sexual
animals do not, exactly, "reproduce", the simple-minded ideal concept
served very well to lead us to our current better ways to understand
cell division.
From our pre-scientific heritage we inherit many useful idea-germs,
like "dead-alive", "intellect-emotion", "believe", know", "designate",
"intend", and so forth. Searle presumes these to be exclusively
worthy of study; I see them as flimsy suggestions left from our
failing ancestors. They are useful in practical life, but too crude
to support powerful theories. Now it would be immodest (and
unconvincing) for me to argue why my own new ideas have more
potential; instead I will explain the point via a fantasy in which our
future successors recount our own plight:
.in 5
"The twentieth concept of 'Believe' proved inadequate until replaced
by a continuum in which, it turned out, stones placed near zero,
thermostats scored .072, and some men achieved as high as 67.6.
Acoording to modern concepts, It appears theoretically possible for
something to be believed as intensely as 3600. We were chagrined to
learn that men do not in fact believe very well nor, for that matter,
are they very proficient (on an absolute scale) at intending. Still,
they are comfortably separated from the thermostats."
.in 0
Of course, I'm joking, because I don't think a one-dimensional
continuum -- or indeed, any kind of scale, will do. The next fantasy
explains why, and perhaps also explains Searle's distress about the
condition of the person who incorporates into his mind -- "without
understanding" -- the hypothetical "squiggle-squoggle" process that
appears to understand Chinese:
.in 5
"The most pathetic error of that era was the unwitting adoption of the
ancient formula 'A believes that B means C' -- without questioning the
concept of the believer A itself. Because of this, our philosophical
ancestors persisted in what we today call the 'single-agent fallacy'
-- that inside each mind is a single Knower, Believer, or Meaner 'A'
of whom one can predicate that 'A believes B'. It is strange indeed
how long this philosophical obsession survived even after Freud
published his first clumsy portraits of our complex, inconsistent and
adversary mental constitutions. To be sure, we still find it useful
to preserve in everyday life much of the phenomenology that makes this
seem so natural, because the concept of an 'I' or a 'Self' seems still
indispensable for social purposes. And it will probably remain so,
until our brain scientists become better able to preserve, after
'death' -- and to rearrange and recombine -- those aspects of a
brain's mind's parts that might seem of value to our own successors."
.in 0
I wonder if Searle will see what I am getting at. He talks glibly of
how we can "let the individual internalize all of these elements of
the system" and then complains that "there isn't anything in the
system that isn't in 'him'". My point is that we can get nowhere if
we continue to talk of 'him' as an atom. Perhaps Searle's
correspondents found it hard to explain what they meant because they,
too, fell into the same trap. They should have said that the entire
point of whatever Searle means by "Strong AI" is not, as he seems to
think, to try to preserve and explain certain words like "Self",
"Mean", "Intend" and "Believe". For these are not given "things" that
are really there for us to confront. They are, instead, mere steps
toward better ideas -- and it is with discovering →those_ that concern
me, at least.
To an earlier, naive, "self-centered self" of my own, Searle's
arguments would have seemed the best of their kind. But now they seem
just like those of Searle's own example! They read like Chinese: I
understand the syntax, and how to parse the sentences; I follow the
technical deductions. But the terms and assumptions themselves --
what the words like 'intend' and 'mean' intend and mean -- escape the
new me. They seem only like formally specified symbols, incorporated
into an older part of my mind that is no longer in agreeable contact
with the newer ones. And it is the newer ones that seem to be better
able to untangle the web of belief, using the tools of "strong AI".
∂08-Mar-80 2304 MINSKY at MIT-AI (Marvin Minsky)
Date: 9 MAR 1980 0153-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, jmc at SU-AI
I finished better version of Reply to Searle.
Erase the one I sent and read MINSKY;SEARLE >
If you ahev any comments, I'd appreciate them -- because
I might let it be published along with his paper,
as a "reply".
∂09-Mar-80 2352 JK knowledge representation
Page 4 of know.sug[1,jk] contains an outline of a formal theory that seems to
avoid the paradoxes discussed by Kaplan and Montague. However, I am not
very satisfied with it.
∂10-Mar-80 0720 Allen.Newell at CMU-10A (A310AN02) aim 326
Date: 10 March 1980 1015-EST (Monday)
From: Allen.Newell at CMU-10A (A310AN02)
Subject: aim 326
To: jmc @ sail
CC: Allen.Newell at CMU-10A
Message-ID: <10Mar80 101510 AN02@CMU-10A>
John: For some reason I seem to have missed seeing you "Ascribing mental
qualitites to machines", AIM 326, 1979. At least that's the way Searle
gives it in his BBS thing. I would enjoy seeing a copy. Thanks. AN
∂10-Mar-80 1000 JMC*
mazda
∂10-Mar-80 1000 JMC*
Lindstrom 321-0900.
Larry Donnelly 321-6722
CMS in Menlo 364-0348, Colin Whiteside
∂10-Mar-80 1103 doyle at MIT-AI (Jon Doyle) Job offer
Date: 10 MAR 1980 1402-EST
From: doyle at MIT-AI (Jon Doyle)
Sent-by: JLS at MIT-AI
Subject: Job offer
To: jmc at SU-AI
CC: DOYLE at MIT-AI
I'm sorry to bother you with another question, but
has there been some problem in offering me the position
you indicated, or is there lots of red tape? I confess
I didn't quite expect this much delay.
Again, sorry to bother you, and thank you for your help.
Jon
∂10-Mar-80 1914 Denny Brown <CSD.DBROWN at SU-SCORE> Proposed new Masters program.
Date: 10 Mar 1980 1908-PST
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: Proposed new Masters program.
To: bboard at SU-SCORE, bboard at SU-AI, csd.bbd at SUMEX-AIM,
CSD-Faculty: ;
cc: csd.tajnai at SU-SCORE, csd.rosenow at SU-SCORE, hazen at SUMEX-AIM,
les at SU-AI, tob at SU-AI
The AI faculty is now reviewing an almost final draft for a proposed
new masters program. The "Master of Science in Computer Science:
Artificial Intelligence" (CSAI) is a program patterned after the
CSCE program. It has similar goals: namely, producing highly
qualified engineers for the development part of R&D.
We hope to converge on a final draft for the Courses and
Degrees bulletin this week. Then we will need full departmental
approval before presenting it to the graduate dean for school
approval. If anybody wants to provide input into any of
these processes, contact me (soon.) We will probably want
to do a telephone poll vote so we don't need to call a regular
faculty meeting.
The current version is <CSD.FILES>CSAI at SCORE and CSAI[INF,CSD]
at SAIL. Anybody interested is welcome to comment on the
proposal.
After the approval of the deans (3/31) we will formally announce
the program, and begin accepting applications.
-Denny (CSD.DBROWN@SCORE)
-------
∂11-Mar-80 1048 HVA Telephone Call
Dave Marimont, SRI, would like you to call him at 326-6200, Ext. 2783.
∂11-Mar-80 1152 TOB Dave Marimont
John
Dave has been doing a project with me for credit.
He also works for Tenenbaum at SRI. He seems quite
serious, works independently, reports frequently,
is quite conscientious. He has attacked problems
analytically, which I strongly favor. Thus far,
the problems have not been exciting to me, but not
bad. I think that he has reasonable judgment in
choosing problems and that he can learn good judgment.
I don't think that he has bad taste, scientifically.
In sum, I think very well of Dave as a student. He
seems really mature.
Tom
∂11-Mar-80 1402 BTH a note from S. Owicki
To: "@GROUP.DIS[1,BTH]" at SU-AI
Susan Gerhart will be here on Friday to talk to anyone interested in
either permanent or summer employment in areas relating to program
verification. (This is at ISI, a research organization that has some
relation to USC that is similar to some stage of the relation between
Stanford and SRI.) If you are interested in exploring this with her,
contact her through the Computer Science department on Friday or by
net mail to Gerhart@usc-isic.
∂11-Mar-80 1641 EGL OUTGOING MAIL
John, I have completed your travel expense report form for NAS.
Since your signature is already on both copies to be returned to
NAS, I will make a copy for your file and send the originals out
on 3/12. Please let me know if this is all right.
The letters of recommendaion
for Mr. Brooks will be completed tomorrow. The XGP was not working
today, and I can read the file and get the letters out.
Lyn Larson
Yes, please send out the travel expense forms. Thanks for the messages.
∂12-Mar-80 0005 HPM HPM orals are today
To: TOB at SU-AI, JMC at SU-AI, LEVINTHAL at SUMEX-AIM, DEK at SU-AI
at 2:30 in room 402 of Margaret Jacks hall. See you there!
∂12-Mar-80 0131 ME date and time
∂12-Mar-80 0000 JMC
I notice that the date in who changes before the time. Can it cause bugs?
ME - Do you mean on the wholine, it says something like Wed 23:59 instead
of Tues 23:59 or Wed 00:00? That seems strange and might have been
caused by some output of the wholine being suppressed on a DM, if that's
where you saw it. Internally, the date is changed precisely when the
time reaches midnight, so there can't be any bugs.
On my Imlac tonight, the date changed to Wednesday, but the time
remained at 23:59 for perhaps 30 seconds. If this isn't just an
artifact of the wholine program, it might cause bugs in the dates
and times attached to files.
∂12-Mar-80 1201 LES Frances Larson reference
To: JMC
CC: HVA
I just talked to Rabbi Sidney Akselrad of Congregation Beth Am in Los Altos
Hills. He was Frances Larson's supervisor for the last 18 years.
Akselrad was extremely positive about her performance. He said that she
the principal administrator there over the last 10 years and that she has
very high standards and showed considerable personal initiative.
She was extremely good at dealing with a wide range of people.
It was his impression that she left partly just to have a change and
partly because her responsibilities were taking a considerable number
of evenings -- that she would prefer a position that did not require
involvement around the clock.
My conclusion is that we should make an offer. Right?
Right.
The offer should explicitly mention using the computer as part of the job.
∂12-Mar-80 1342 EGL CORRESPONDENCE
John, there is completed correspondence on your desk for your signa-
ture. Also, the XGP is still down. Lyn Larson
∂12-Mar-80 1409 ME date and time
∂12-Mar-80 0223 JMC
On my Imlac tonight, the date changed to Wednesday, but the time
remained at 23:59 for perhaps 30 seconds. If this isn't just an
artifact of the wholine program, it might cause bugs in the dates
and times attached to files.
ME - That must be indeed just an artifact of wholine output being
cut off. The date isn't advanced internally until the time reaches
midnight (0000).
∂12-Mar-80 1753 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Courtesy Parking Permits
Date: 12 Mar 1980 1454-PST
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: Courtesy Parking Permits
To: CSD-Faculty:
cc: CSD.Betty at SU-SCORE
Professor Feigenbaum asked that you be informed that we have one-
day Courtesy Parking Permits for visitors. They cannot be used on
vehicles with Stanford permits. The permit must be typed with the
date and license plate number. The copy is to be returned to the
Police Dept. We purchased a few for $1 each; they are in my office.
Carolyn
-------
∂12-Mar-80 1855 LLW Skyhooks
To: JMC at SU-AI, minsky at MIT-AI
CC: RAH at SU-AI, LLW at SU-AI, "#JUNK2.MSG[1,LLW]" at SU-AI
We haven't heard anything new for some time. Wie gehts?
We have analysed the case of an orbiting cable supporting anchored
skyhooks; our secretary will be keying this into POXable form over the
next several days. Obviously a magnetic slip joint is required. Once you
have climbed the skyhook to the cable height, you can achieve orbit by
grabbing hold of the cable with a dragging slip joint. As with the mass
driver-in-orbit of Kingsbury, some means of momentum restoration is needed
to maintain the cable's orbit. The required cable mass is relatively
large, due to its length and the small angle it makes with the vertical at
the skyhook sites. Depending on how the slip joint mass scales, you can
start small and use the cable-skyhook to grow itself; the self-replication
time is of the order of a day.
The vertical cable supported by bb acceleration up (and down) its core
appears feasible and quite attractive; obviously you want a dv/dt = -v
type of bb brake on the up leg (& then reaccelerate them down). Unless
you extend it to synchronous orbit, the problem of giving orbital speed to
payload remains.
Rod
∂12-Mar-80 2119 MINSKY at MIT-AI (Marvin Minsky)
Date: 13 MAR 1980 0015-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, llw at SU-AI, jmc at SU-AI, rah at SU-AI
Hi. I7ve been trying to find out if anyone has designed a lossless
deflector. Because I think
orbital BB7s might be easier than cable, exactly because of sharp bend.
Of course, bend can be made over couple of kilometers. Cable
has stability rpobelms, I suppose.
I have talked to O'neill and, today, Kolm.
I had a little trouble communicating, because they
have invested so much design into
the mass driver, which has severe energy losses, e.g,
over a percent. Of course, they don't care,
but we do. Trouble is that they are used to
reaction via eddy current in conductor, etc.
But Kolm agreed that one could make an all-superconductor
reflector that has close to elastic
collisions. Any of
you expert at that? I agree that cable is essentially the same thing, provided
flexibility is no problem, and that the
orbit is circular
enough that there is no velocity change to
cause length changes.
I found a slight bug in the vertical elevator scheme. If the thing is
alike a pair of lossless mass drivers, decelerating on the way up,
and accelerating on the way down,
then we can't assume constant loading per vertical meter.
Reason; if the momentum transfer per meter is constant, still
the particles have much larger V at the bottom. In fact,
the power transfer between the linked drivers
increases with V, so the superconductors are
bigger at the bottom. result is exponential growth, consequently.
However, it is possible that the larger currents could still be handled
with relatively reasonable components. It boils down to how many times
its weight a superconducting repulsion thing can support, I think. Anyone
know? I have a feeling that it might be pretty bad for my vertical
elevator.
∂12-Mar-80 2339 TOB display system
I talked with Don Knuth and realized a consideration
for the department display system. In order to guarantee
that the display system is available for vision I want to
limit severely other use of it. We had strong handicaps
in working without graphics and image display when we
could not get Data Disc channels. On the other hand,
XGPSYN seems like an important cause, particularly with
work on TEX, etc. Probably everyone will want that
capability. Single bit plane graphics may be ok, but
gray scale works better, since resolution is about twice
better with gray scale. Presumably, most of the department
will want to use it if XGPSYN is good enough.
I think it may be worth including that consideration in
the department display system. Otherwise, I may push for
an equipment proposal to support a system for general use.
Tom
∂13-Mar-80 0000 LLW A Promise of Joy!
To: minsky at MIT-AI, JMC at SU-AI
CC: LLW at SU-AI, RAH at SU-AI, "#JUNK2.MSG[1,LLW]" at SU-AI
∂12-Mar-80 2118 MINSKY at MIT-AI (Marvin Minsky)
Date: 13 MAR 1980 0015-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, llw at SU-AI, jmc at SU-AI, rah at SU-AI
Hi. I've been trying to find out if anyone has designed a lossless
deflector. Because I think orbital BB's might be easier than cable,
exactly because of sharp bend. Of course, bend can be made over couple of
kilometers. Cable has stability problems, I suppose.
[Rod has done a detailed stability analysis which indicates that the cable
is stable against growth of small amplitude, low order modes. Indeed, the
'bend' at the skyhook attachment point need not be at all sharp. A major
variable in this analysis is whether the cable flows through the bending
region at a velocity which is sub- or super-sonic, relative to its sound
speed; it is stable in both, quite different regimes. Amusingly enough,
the stressed cable approach had been suggested to Rod a couple of months
back by Jordin Kare (whom you may have known at MIT as a recent applied
physics undergrad of uncommon intelligence, Marvin), but its potential
interest for skyhook purposes had been overlooked until you and John
suggested its discretized analog, the particle stream, in our first
conversation on the subject--Jordin and Rod had been discussing it
primarily as an orbital insertion mechanism for pop-up systems.]
I have talked to O'Neill and, today, Kolm. I had a little trouble
communicating, because they have invested so much design into the mass
driver, which has severe energy losses, e.g, over a percent.
[Their entire approach, specifically including their mass drivers, is an
oxcart one, compared to what we're presently considering. Don't waste
effort (and spoil the surprise) trying to convert them now--let them be
stunned without further warning by the advent of what's coming from
advanced skyhook-ery.]
Of course, they don't care, but we do. Trouble is that they are used to
reaction via eddy current in conductor, etc. But Kolm agreed that one
could make an all-superconductor reflector that has close to elastic
collisions. Any of you expert at that?
[Both Rod and I have done extensive 'paper' engineering with
superconductors for a variety of applications, and it was the use of Type
I superconductors (or Type II ones, at magnetic stress levels below the
lower critical field intensity)--the ones which exhibit the full Meissner
effect--which we proposed to use for cannonball deflection the evening we
first discussed this whole business with you; eddy current losses implicit
with use of normal conductors are indeed rather unacceptably high.]
I agree that cable is essentially the same thing, provided flexibility is
no problem, and that the orbit is circular enough that there is no
velocity change to cause length changes.
[Flexibility is indeed no problem. Orbital eccentricities of a few
percent can certainly be accommodated within the elasticity limits of the
cable material.]
I found a slight bug in the vertical elevator scheme. If the thing is
alike a pair of lossless mass drivers, decelerating on the way up, and
accelerating on the way down, then we can't assume constant loading per
vertical meter. Reason; if the momentum transfer per meter is constant,
still the particles have much larger V at the bottom. In fact, the power
transfer between the linked drivers increases with V, so the
superconductors are bigger at the bottom. result is exponential growth,
consequently.
[This bug is illusory. Rod and I have made a lot of progress on the
vertical elevator scheme recently, and it's presently looking quite nice
(albeit not perfect), even after numbers are inserted (so many beautiful
schemes die hideously on such occasions).
In order to rather exactly balance the gravitational momentum flux all
along the elevator (so that it need have very little intrinsic strength),
you must have a dv/dt = -v type of deceleration of the up- and down-going
pellets. This results in the pellets being relatively widely spaced at
the bottom of the tower, and relatively close-packed at the top-- this is
an inevitable consequence of shedding momentum at a constant rate to the
tower (ignoring the decreasing gravitational loading at altitudes
comparable to/larger than an Earth radius), while conserving pellet mass.
Thus, the pellets are decelerating against both the gravitational loading
on them *and* against supporting the tower on the upward leg, and
accelerating against both of these loadings on the downward leg. An e.g.,
linear induction motor at tower bottom 'catches' them and off-loads the
momentum they've picked up on their trip. They are they fed into a
parallel-situated twin, which flings them back up another column of the
tower, thereby completing the cycle (and permitting the necessary
variability as the tower is 'grown up' to synchronous orbit. A large
reactive power flows between the two motors, but only relatively very
small real power input need be made to keep this reactive power flow
going.
This approach, with the vertically oriented inductor motor pair anchored
in terra firma, into which the entire gravitational momentum flux incident
on the tower is conducted and sunk by the rising-and-falling pellet
stream, has the *exceedingly* important feature that the tower can be
continuously built *on the ground* and then *lofted* into space, out to
synchronous orbit and beyond, by a literal skyhook constituted of the
rising-and-falling pellet stream. This stream (which apparently cannot be
made into a continuous tape, for the 'flow' stagnation reason noted above)
is tightly analogous to a VandeGraaff generator belt; it transports
momentum instead of electric charge, however, thereby increasing (or, in
steady state, maintaining) the tower's gravitational (rather than
electrical) potential. In principle, the tower need have no mechanical
strength, *at all*--the Clarke-Moravec-.... super-material problem is
completely obviated in this approach; you could 'pipe' water to the moon
(or L-5, if O'Neill will pay the utility bill) with this scheme--without a
pipe!
It presently appears to us that such a tower might be constructed with
*present*, industry-standard materials and technology for a total cost of
the order of a few billion dollars, on a few year time scale, and could
thereupon loft thousand ton lumped loads, spaced almost bumper-to-bumper
along the tower, into trans-synchronous orbit!!! Indeed, it is amusing
that the cost of the electrical power plant needed to keep the pellet
stream-manipulating motors running at the tower's bottom is a
non-negligible fraction of the total system's cost--the facility on the
ground which 'grows' the tower from its base could apparently be a fully
automatic one, which needs be fed only power, petroleum (or other
carbonaceous feedstock), and a smidgen of aluminum!]
However, it is possible that the larger currents could still be handled
with relatively reasonable components. It boils down to how many times
its weight a superconducting repulsion thing can support, I think.
[Two-to-three orders of magnitude, with optimal use of the best present
superconductors. However, as just noted, this is irrelevant--the
gravitational momentum flux onto the entire tower can be sunk back into
the Earth via the rising-and-falling particle stream, through the
mountings of the motor pair at the tower bottom--the tower can literally
be made weightless, both integrally *and* differentially!]
Anyone know? I have a feeling that it might be pretty bad for my vertical
elevator.
[Rejoice! The foregoing is A Promise of Joy. Arthur Clarke's 'Bridge to
the Stars' can be erected in the next two decades, a couple of centuries
earlier and at four orders of magnitude lower cost than he imagined!!!
We'll try to deliver draft documentation of the salient features,
equations and numbers by the end of this coming weekend.]
[Lowell]
∂13-Mar-80 0006 LLW Yet Another Try. . .
To: JMC at SU-AI, minsky at MIT-AI
CC: RAH at SU-AI, LLW at SU-AI, "#JUNK2.MSG[1,LLW]" at SU-AI
SAIL died as this was being mailed out early this past AM, and apparently
no one got it; it died again earlier this evening as a second attempt was
being made, and some got it while others didn't. Please pardon any
redundancy associated with this third attempt.
∂12-Mar-80 1855 LLW Skyhooks
To: JMC at SU-AI, minsky at MIT-AI
CC: RAH at SU-AI, LLW at SU-AI, "#JUNK2.MSG[1,LLW]" at SU-AI
We haven't heard anything new for some time. Wie gehts?
We have analysed the case of an orbiting cable supporting anchored
skyhooks; our secretary will be keying this into POXable form over the
next several days. Obviously a magnetic slip joint is required. Once you
have climbed the skyhook to the cable height, you can achieve orbit by
grabbing hold of the cable with a dragging slip joint. As with the mass
driver-in-orbit of Kingsbury, some means of momentum restoration is needed
to maintain the cable's orbit. The required cable mass is relatively
large, due to its length and the small angle it makes with the vertical at
the skyhook sites. Depending on how the slip joint mass scales, you can
start small and use the cable-skyhook to grow itself; the self-replication
time is of the order of a day.
The vertical cable supported by bb acceleration up (and down) its core
appears feasible and quite attractive; obviously you want a dv/dt = -v
type of bb brake on the up leg (& then reaccelerate them down). Unless
you extend it to synchronous orbit, the problem of giving orbital speed to
payload remains.
Rod
∂13-Mar-80 0139 MINSKY at MIT-AI (Marvin Minsky)
Date: 13 MAR 1980 0415-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, llw at SU-AI, jmc at SU-AI
Fantastic.
I had figured most of that; the problem I mentioned is
precisely that one you mentioned of the reacti load.
Perhaps you can reassure me, but what i was afraid
of was that, although the momentum-flow per meter is
constant, and exactly supports the
weight, the reactive currents
would increase with the V at which the momentum flows.
Then, what I worry about, is that the mass of superconductor has to be
proportional to this reactive current, which I fear so increases.
See what I mean? (I'm only an intuitive physicist) The energy flow across between the
two columns is higher at the bottm, because
the energy/momentum ratio is -- so
it might need thicker conductors.
Tell me I'm wrong.
By the way, did you derive that the velocity is proportional to the
(negative) height? And that the acceleration
is constant per meter (instead of per second)?
If so, that's pretty cute.
∂13-Mar-80 0902 JMC*
nsf renewal and arpa report
∂13-Mar-80 0902 JMC*
Alm,bbs,miro
∂13-Mar-80 1545 TOB
do you want to reopen buying an OCR device?
Yes if you are in a position to look for one.
∂13-Mar-80 1600 JMC*
call Miro
∂13-Mar-80 1628 HVA YOUR CORRESPONDENCE
To: JMC
CC: EGL, HVA
John--Lyn was not in today (so I cld not gain access to any messages you may
have sent her), but I have picked up (from your desk), and mailed the following:
1)Bill to III
2)Memo to Doug Murray
3)Letter to Alan Bundy
4)Letter to Merrell Patrick
I have made copies of all outgoing corresp. and they are attached to incoming--
on Lyn's desk.
∂14-Mar-80 0253 RMS at MIT-AI (Richard M. Stallman)
Date: 14 MAR 1980 0551-EST
From: RMS at MIT-AI (Richard M. Stallman)
To: jmc at SU-AI
Hello. I am very eager to take you up on your invitation for me to
come visit. I would especially like to be there starting in
September, because, arriving at the start of the term, I'd be able to
combine my visit with something that I would love dearly: to study
Gamelan at Berkeley.
Lately I have been working here mostly on EMACS, which isn't very
deep. I'm also working on the Lisp machine file server, but right now
that is on the back burner because of trying to get an EMACS version
and the EMACS manual ready for DECUS. It's partly because I think I'd
find it easier to get above things like this that I'd like to visit at
Stanford some more.
On the side, I'm reading Drake's book on set theory, because I have
developed a hunger for large cardinals. So far I haven't reached
the things I don't know, but the book describes the concepts and
motivations much more clearly than other books I've seen.
Also, I've been thinking about whether "I know that we jointly know"
is implied by "We jointly know". I was inspired by the analogy
between joint knowledge and the ordinal omega, to wonder whether there
were anything to correspond to larger ordinals. If we add more people
in, we can get omega+n, and up to omega*2: "He and I jointly know that
you and I jointly know", and with infinitely many people up to omega
squared. Unfortunately it looks like 2 people can't reach omega*2.
I'm not sure whether they can do omega+n, which is what leads to the
question at the start of this paragraph. It's interesting because it
reduces to the philosophical question of whether "(n) I know f(n)" is
equivalent to "I know (n) f(n)", where f(n) is up to n levels of back
and forth I-know-that-you-know. I haven't been able to answer this
for myself.
I also thought for a while about the difference between accessibility
functions and explicit knowledge modalities, and why it could possibly
make things easier. I came to the conclusion that accessibility
allows one to avoid the need for anything like
((A => B) and (I know A)) implies (I know B)
because if all the worlds accessible to me have A, and A => B, then
those worlds have B. I wanted to look over a proof to see whether
this was indeed ever being used, but since the papers had only the
axioms and not the proofs, I had to try to work them out and got bored
with the detail. However, if this is true, then it shouldn't be an
insuperable disadvantage for explicit modalities, and I would still
rather use them because they are a more intuitive concept.
I am offering a paper to the Lisp conference about a very simple stack
management idea to replace spaghetti stacks.
I still disagree with Dennett on the subject of mental images, but I
haven't written down what I think, because I am daunted by the fear of
being misunderstood, of not writing so as to rule out every possible
misinterpretation, every way to slide between what I say. This fear
is triggered by the way his essays seem to try address all possible
objections. Does he still believe what he wrote about mental images
in Brainstorms? If I know for certain that he does, I'll know that
writing down my ideas is worth while.
For more variety, I got my pilot's licence, and I'm about to start
teaching folk dancing. I am starting to believe that I can be the
renaissance man that I have always hoped I could be.
I'm assuming that your invitation means you can offer me funding! or that
MIT would keep paying me. Perhaps if I knew academic customs better,
I would know to take it for granted that you were offering, or take it
for granted that you were not.
∂14-Mar-80 1101 EGL mental.xgp[w80,jmc] file
I accidentally reformated the file using E.
Please recreate the file and I will spool it again
Lyn Larson
∂14-Mar-80 1531 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) our meeting (right now)
Date: 14 Mar 1980 1527-PST
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: our meeting (right now)
To: jmc at SU-AI
John, looks like we aborted on the 3pm meeting with Betty to discuss
sail administrative matters. How about rescheduling for 2pm on Monday the
24th of March (I'll be out of town most of next week).
Ed
-------
∂14-Mar-80 1601 JMC*
look for time-sharing memo
∂14-Mar-80 1636 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
There will be no PROGRAM VERIFICATION AND ANALYSIS SEMINAR
on March 18. The seminar will start again at the beginning of
Spring quarter.
∂15-Mar-80 2032 MINSKY at MIT-AI (Marvin Minsky)
Date: 15 MAR 1980 2325-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, jmc at SU-AI, llw at SU-AI
Hi. I'm planning to appear out there some time late Friday.
Lowell -- when would be a good time to
see you? Apparently Danny, Marg, DLW are planning
planning dinner at Louie's Friday evening.
Any chance for JMC, LLW and MINSKY to get together
between friday and Monday?
How about you all coming to dinner Saturday night?
∂16-Mar-80 0249 LLW Getting Together
To: minsky at MIT-AI, JMC at SU-AI
CC: LLW at SU-AI, RAH at SU-AI
∂15-Mar-80 2032 MINSKY at MIT-AI (Marvin Minsky)
Date: 15 MAR 1980 2325-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, jmc at SU-AI, llw at SU-AI
Hi. I'm planning to appear out there some time late Friday.
Lowell -- when would be a good time to
see you? Apparently Danny, Marg, DLW are planning
planning dinner at Louie's Friday evening.
Any chance for JMC, LLW and MINSKY to get together
between friday and Monday?
[I'll be happy to see you next weekend, which looks pretty open at
present. DLW told me that Danny and Marg would be visiting the Project
sometime next weekend, and that I would be talking with Danny on various
matters. Rod and I would be game for Louie's Friday evening, if that's
convenient for the two of you. Alternatively, you could meet us over here
somewhere. What's your pleasure? Lowell]
∂16-Mar-80 0251 LLW Getting Together, contd
To: JMC at SU-AI
CC: LLW at SU-AI, minsky at MIT-AI, RAH at SU-AI
∂15-Mar-80 2056 JMC
To: minsky at MIT-AI, LLW at SU-AI, RAH at SU-AI
How about you all coming to dinner Saturday night?
[John: I just saw this, after responding to the last note. This
particular option is fine with me (and presumably Rod)--it would also
afford us an opportunity to get some coordination work done on the paper,
either before or after. Lowell]
∂16-Mar-80 1209 MINSKY at MIT-AI (Marvin Minsky)
Date: 16 MAR 1980 1507-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
By the way I want to talk to someone named Van de Shell
who is visiting SAIL from Schlumberger. Do you know how
to find him -- or what his login name is?
I don't know any Shell and neither does the system. Do you know whom
he might be visiting? By the way, how about dinner Saturday night
as per previous message?
csd.Vanderschel@score
∂16-Mar-80 2013 MINSKY at MIT-AI (Marvin Minsky)
Date: 16 MAR 1980 2311-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
Saturday night sounds fine with me. Probably should get
together in middle afternoon to work things out? OK
to drop in at your home Friday night, or should I look
elsewhere?
Hmmm. I guess he's visiting feigenbaum. I'll ask him.
∂16-Mar-80 2128 LGC Wednesday?
We should get together sometime this week to discuss the meeting with
Bob Engelmore on the 31st. Would sometime Wednesday be ok?
∂17-Mar-80 0744 Siegman at SUMEX-AIM Symposium paper
Date: 17 Mar 1980 0740-PST
From: Siegman at SUMEX-AIM
Subject: Symposium paper
To: bair, mccarthy at SAIL
cc: sandelin
John, Jim passed me a copy of "The Real Computer Revolution" which I
have just glanced at. Before it goes into the collection that will
be reproduced for attendees at the symposium next week, I've asked
for three minor changes:
1) p. 11, "...IBM which has generously invited me to this symposium."
delete rest of sentence following I"IBM".
2) p. 12 "...Lewis Branscomb gave an interesting example this morning..."
delete "this morning."
3) p. 13-- delete questions and answers following presentation
(from earlier talk)
Jane Marcus is assembling the outlines and/or papers the speakers have
sent--I have told her to make these changes with liquid paper as
a last resort. If you could regenerate p. 11 and 13 minus
those references and send via ID mail to Jane at Cedar Hall, that
would make for much cleaner look. Thanks - Jeannie Siegman
-------
∂17-Mar-80 1200 Bair at SUMEX-AIM Re: Symposium paper
Date: 17 Mar 1980 1123-PST
From: Bair at SUMEX-AIM
Subject: Re: Symposium paper
To: Siegman, mccarthy at SAIL
cc: sandelin, bair
In response to the message sent 17 Mar 1980 0740-PST from Siegman
Dear John:
I was under the impression that the paper you sent me was for background.
I wasexpecting a new presentation, not too different perhaps, for the
symposium. We need that this week in order to include in the Conference
Proceedings.
Jim Bair 494-3942 ext 690
-------
My impression agreed with yours, and I was preparing a different
paper. When I received the Siegman message I was confused and
supposed I was being let off easy. It will be ready soon.
∂17-Mar-80 1632 CSD.ROSENOW at SU-SCORE Margaret Jacks Hall Dedication
Date: 17 Mar 1980 1618-PST
From: CSD.ROSENOW at SU-SCORE
To: CSD-Faculty: ;
To: CSD Faculty
CC: Affiliated Faculty
Subject: Margaret Jacks Hall Dedication
This Spring is the appropriate time (at last) for the dedication of
the Computer Science Department building. I would like the dedication
to be scientifically substantive as well as formal, and to that end I have
a plan that I would like to discuss at the faculty meeting on April 1.
I want to alert you now, however, to the date that I am currently
discussing with the university administration: Monday, June 23.
This will put it out of the range of the Spring Quarter ratrace into
the quiet time, and increase the chances that colleagues from around the
country could attend. I would like to hear from those who may have
specific objections to that date as soon as possible.
Ed Feigenbaum
---------------
-------
∂17-Mar-80 1845 JMC
call martha (preferably before mdnight)
∂18-Mar-80 0708 JRA reviewing
john,
when i asked you to be on the program committee, you mentioned that you
didn't want to play a day-to-day role in the goings on. how do you feel about
reviewing papers? if you'd rather not, or would accept a reduced load, i will
respect your wishes.
john
∂18-Mar-80 0830 JMC*
alm,weather
∂18-Mar-80 1036 100 : lyn larson Reports
I would like your help in getting the reports out of your
file area that you would like sent out.
Will you be stopping by your office this week and can you
offer me your assistance. Thank you -----Lyn Larson
Tomorrow, I hope.
∂18-Mar-80 1101 BS Personnel Requisition
To: JMC at SU-AI
CC: csd.betty at SU-SCORE
John, I am holding the personnel requisition for a Secretary III (your
secretary) which Hersche gave to me last Friday morning. You asked me
not to send it through yet. Is it o.k. to process it now, or do you want
me to hold it longer?
Betty
Please hold it a bit longer. I'm going to Bakersfield today, and I'll
try to take care of deciding between Lynn Larson and Frances Larson
tomorrow.
∂18-Mar-80 1140 HVA Frances Larson-Candidate interviewed for JMC Secty position
To: JMC
CC: LES, HVA
Frances Larson called me this morning to say that, after considerable discussion
with her husband, she has decided not to accept a full-time, permanent position.
She asked me to thank you both for your time and for considering her seriously
as a candidate.
∂18-Mar-80 1312 EGL your reports
thank you --that sounds very good ---lyn larson
∂18-Mar-80 1401 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: Marchuk
Date: 18 Mar 1980 1347-PST
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: Re: Marchuk
To: JMC at SU-AI, feigenbaum at SUMEX-AIM
In-Reply-To: Your message of 16-Mar-80 2240-PST
John, I happened to see it by accident in the NY Times...Ed
-------
∂19-Mar-80 0000 JMC*
expense report for Bakersfield in Current events file
∂19-Mar-80 0029 LLW Hans for Dinner?
To: JMC
CC: LLW
I heard a bit ago that Hans Moravec is still in the area. Would he be a
suitable invitee to the Saturday Afternoon Session? to dinner? Lowell
∂19-Mar-80 0047 FORWARD at USC-ECL M&M's Dynamic Compression Columns
Date: 19 MAR 1980 0044-PST
From: FORWARD at USC-ECL
Subject: M&M's Dynamic Compression Columns
To: hpm at SAIL, minsky at MIT-AI, jmc at SAIL, llw at SAIL,
To: rah at SAIL
cc: forward
Am glad to hear that progress is being made on my idea of a
vertical skyhook using the M&M dynamic compression column concepts.
I too was afraid that after the numbers were inserted that the
beautiful scheme would die a hideous death, pilloried on the
rigid frame of the conservation laws.
Fortunately, Rod and Lowell's work on superconductors has given
some engineering substance to what was originally an ephemeral
concept. Certainly more work needs to be done, but may I have
permission to transmit some of this conversation to Arthur
C. Clarke? I am sure he would be interested.
Last week I was thinking about the topology of the various
versions of dynamic compression structures. The original M&M
concept of late February was a 2-D hexagonal structure around the
earth. I then suggested the vertically oriented 1-D structure, with
its differential up-and-down problems. There is obviously a 3-D
extension to the original M&M dynamic compression column concept;
a tetrahedral, cubical, or higher order 3-D shape structure
surrounding the earth and supplying nodal stations at unique points.
(These should have an even number of sides at each
vertex to conserve particles.)
I can't think of anything useful yet for this idea. (If I could
extend it to 4-D, I might have something.)
Bob Forward
You're welcome to inform Clarke as far as I am concerned. The name
"dynamic compression" also occurred to me, but it reminded me of
the "dynamic tension" of the Charles Atlas ads, so perhaps we should
inform the ghost of Charles Atlas.
∂19-Mar-80 0834 REG
John,
So far as I know all necessary hardware and software to support Dialnet
exists and functions at SAIL.
Tim Eldredge at HP is presently having trouble with his modem and thus
is unable to test any of the TOPS-20 dialnet software.
Is there anything that you want me to do about Dialnet?
I suppose we should wait for Tim, although we could offer to lend
him equipment until his gets fixed.
∂19-Mar-80 1825 TOB resume for NSF robotics proposal
John
I have your resume in hard copy for the stereo
proposal. Would you tell me the name of your
file if accessible or ship me a copy if not accessible?
Tom
∂19-Mar-80 2212 ADMIN.BOSACK at SU-SCORE Project writeup
Date: 19 Mar 1980 2207-PST
From: ADMIN.BOSACK at SU-SCORE
Subject: Project writeup
To: JMC at SU-AI
!I have left a copy of my project description in your mailbox at Jacks
hall. I didn't include a bibliography, as most of the references would be
to various manuals. There is not much about the program that gathers the
data, in that the gritty little details of how the problems were solved
seemed rather less interesting than the results.
Questions, comments, suggestions, etc. gladly accepted.
Regards,
Len Bosack
-------
∂20-Mar-80 1136 HVA FRANCES LARSON--Reconsideration
To: JMC at SU-AI
CC: LES at SU-AI, BS at SU-AI, HVA at SU-AI
As we know, LYN LARSON has accepted another job, to start Mon. 3/24, and I
needed someone else to assume the duties Lyn has carried. SO I talked to
FRANCES Larson and asked if she would be interested in filling the position
on a temporary basis (I told her the JMC Secty. position was being posted
and it would be a few weeks before the job could be filled on a permanent
basis--and she just might reconsider a permanent job)--the good news is that
Frances Larson will appear on Monday, March 24th.
∂20-Mar-80 1251 Gardner at SUMEX-AIM Re: people interested in formalized legal reasoning
Date: 20 Mar 1980 1250-PST
From: Gardner at SUMEX-AIM
Subject: Re: people interested in formalized legal reasoning
To: JMC at SU-AI
In response to your message sent 19 Mar 1980 1538-PST
Thank you for the message. I'll try to get in touch with them.
--Anne
-------
∂20-Mar-80 1509 LES
Berliner trip for DEW's orals
DPB says that CSD will resist paying for Berliner's travel expense on the
simple grounds that they have no money -- it is all tied up in SCORE.
The available alternatives now seem to be:
1) take it up with EAF, requesting that CSD pay;
2) add a statement that "Berliner did consulting in support of the Formal
Reasoning research program" and return it to Surra with a straight face;
3) transfer the expense to one of our unrestricted funds.
Attached below, for your information, is the message that caused me to
approve this expense in the first place.
--------
∂17-May-79 1419 DEW your approval
Hans Berliner flew out for my oral at JMC's request and at the lab's expense.
JW says you should send her a note to approve it.
Let's try (1), (2) and (3) in that order.
∂20-Mar-80 1649 HVA NSF RENEWAL AND SUPPLEMENT
To: JMC, HVA
Everything has been typed and I am putting papers on your desk for signature.
I checked with Betty Scott re your salary for 1980-81; the recommended figure
is $41,700., and I have based salary in budget on that amount. Another thought:
since you have included justification for supplement in text for renewal, I
think we should add a sentence and the amount on cover sheet. O.K.?
OK and thanks, Hersche.
∂20-Mar-80 2314 LGC Engelmore Document
The file SNAPOV.TXT[DGI,LGC] contains the material I am developing for
Bob Engelmore. Some of it is still in rough form, and some of it is
still unwritten, but enough is there (especially the introductory
section and other section heads) for you to get an idea of where it's
going. Naturally, I'd like to have your reactions to various aspects
of it, and any suggestions you may have for improving either the content
or the presentation. I'll be on campus tomorrow in case you'd like to
discuss any of it in person. The relevant section of last year's ARPA
proposal can be found in FR[P,LES]/15P . The two timed milestones for
the problem solver from that section are as follows:
⊗ Preliminary study completed and direction decided - April 1980.
⊗ System design for problem solver and preliminary experiments - August 1980.
Accordingly, SNAPOV deals not in details, but with main ideas and general
direction.
In case you've misplaced it, the password for ACCESSing DGI,LGC (which I'd
like to keep confidential) is 'psyche'.
∂20-Mar-80 2319 HPM
∂20-Mar-80 2237 JMC
RSVP on Saturday dinner invitation.
[Thanks for the invitation. May I give you my answer tomorrow?
I find out about possible complications then.]
∂21-Mar-80 0900 JMC*
alm 495-1322 and faculty club 4325 and cASbs
∂21-Mar-80 0926 Bair at SUMEX-AIM Title and paper
Date: 21 Mar 1980 0926-PST
From: Bair at SUMEX-AIM
Subject: Title and paper
To: mccarthy at SU-AI
cc: sandelin, bair, siegman
John, Please send us your paper today if at all possible.
It has to be reproduced and collated into approx. 200 notebooks.
A detailed topic outline (self evident) is fine.
If you can't , send us a title so we can make the program and I can
prepare my transitions and intros.
Thanks, Jim Bair
-------
∂21-Mar-80 1023 DCD
Ringle at Vassar College, Poughkeepsie, NY Box 525, zip 12601
∂21-Mar-80 1038 JRA reviews
john,
did you get my msg about your availability for reviewing papers for the
lisp conference?
Yes, I guess I'm too busy for reviewing, and I always do it badly.
∂21-Mar-80 1052 Siegman at SUMEX-AIM Symposium paper--getting it to us
Date: 21 Mar 1980 1051-PST
From: Siegman at SUMEX-AIM
Subject: Symposium paper--getting it to us
To: Mccarthy at SAIL
cc: Bair, Siegman
John, it won't work to mail it to Jim Bair. As soon as it's ready, would you
call Jane Marcus at 7-3831 to have it picked up. Or if you want to drop it
off, we are in Cedar Hall. Today would be best; Monday 10:00 am is the
absolute deadline. May Thanks, - Jeannie siegman
-------
I will call Jane Marcus. It should be ready today - probably by noon.
∂21-Mar-80 1119 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Faculty Meeting, April 1
Date: 21 Mar 1980 1113-PST
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: Faculty Meeting, April 1
To: CSD-Faculty:
cc: CSD.Tajnai at SU-SCORE, JDM at SU-AI, sgn at SU-AI, csd.betty at SU-SCORE
Please mark your calendars.
Faculty meeting -- Tuesday, April 1, 2:30 p.m., Jacks 252
Agenda will follow next week.
First item will be to approve degree candidates.
Carolyn
-------
∂21-Mar-80 1413 HPM RSVP
Ok, I will come on Saturday. Thanks again for the invite, and your patience.
∂21-Mar-80 1430 DCD $
Ringle will be here for the conference, of course. at Tiki as of Sunday.dan
∂21-Mar-80 1604 Sandelin at SUMEX-AIM symposium dinner
Date: 21 Mar 1980 1602-PST
From: Sandelin at SUMEX-AIM
Subject: symposium dinner
To: jmc at SAIL
There will be a wine tasting of some of california's better wines followed
by dinner at the Faculty Club on Tues evening. Starting time is 5:30 for
wine tasting and about 7 for dinner. Symposium speakers, respondents, and
hosts are invited. I need a rough final count for the Faculty Club on
Monday. Please let me know if you are unable to attend...Thanks
-------
∂21-Mar-80 1642 Sandelin at SUMEX-AIM symposium dinner
Date: 21 Mar 1980 1641-PST
From: Sandelin at SUMEX-AIM
Subject: symposium dinner
To: jmc at SAIL
dinner is Tuesday, March 25...
-------
∂21-Mar-80 1732 RWG at MIT-MC (Bill Gosper)
Date: 21 MAR 1980 2034-EST
From: RWG at MIT-MC (Bill Gosper)
To: MLB at MIT-MC, jmc at SU-AI
SHAH'S SPLEEN TO IRAN NOW!
e21-Mar-80 1736 HPM Here is message that was mis-directed
∂16-Mar-80 1207 MINSKY at MIT-AI (Marvin Minsky)
Date: 16 MAR 1980 1505-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, jcm at SU-AI, rah at SU-AI, llw at SU-AI
Saturday night sounds fine with me. Probably should get
together in middle afternoon to work things out? John: OK
to drop in at your home Friday night, or should I look
elsewhere?
∂21-Mar-80 1937 LLW Saturday Scheduling
To: JMC
CC: LLW, RAH
At what time should Rod and I show up at your place tomorrow?
Lowell
∂21-Mar-80 2324 HPM
∂21-Mar-80 1927 JMC
Thanks, but how did you come upon the message?
I originally saw it in Marvin's MIT mail, but forwarded it to you
from JCM's SAIL mail.
6:30 is fine.
∂22-Mar-80 0913 Sandelin at SUMEX-AIM
Date: 22 Mar 1980 0912-PST
From: Sandelin at SUMEX-AIM
To: jmc at SAIL
You are of course invited to have a glass of wine with us if you can
squeeze it in before your banquet....Faculty Club gold room starting at
5:30 on Tues evening...I have heard your paper is excellent...I look
forward to reading it and to your presentation...Thanks very much for
helping us in this venture.
-------
∂22-Mar-80 1219 HVA Key for Haugeland
To: JMC, HVA
I'll get one Mon. 3/24. Do you want to take it to him, or will be be
coming here soon? I'd rather not send key via I.D. Mail.
You can MAIL JCH when it is available, but I don't see that a building
key is too special to send by ID mail.
∂22-Mar-80 1405 LLW Scheduling
To: JMC
CC: LLW, RAH
∂21-Mar-80 2230 JMC
To: LLW, RAH, HPM
Please come at 6:30pm if convenient.
[Fine with Rod and me. Lowell]
∂22-Mar-80 1545 TOB
∂21-Mar-80 1810 JMC Alcoa unrestricted
John
The money is in an unrestricted account. It is received for support
of computer aids to mechanical design. I thought of using it
for equipment which we need.
I like the storekeeper project and wire wrapping/envelope stuffing.
I think that we can bootleg the people time for that. I intend
to start some work on them.
Tom
If the ALCOA $5K or maybe even $10K is truly unrestricted, then
I would propose using it for the storekeeper project or for an
application like wire wrapping or envelope stuffing. I believe
a little bread thrown upon this particular pond will be returned
manyfold eventually.
∂22-Mar-80 2236 Kanerva at SUMEX-AIM Criteria for Usefulness of Computers in Offices
Date: 22 Mar 1980 2235-PST
From: Kanerva at SUMEX-AIM
Subject: Criteria for Usefulness of Computers in Offices
To: McCarthy at SU-AI
cc: Kanerva
Jane Marcus asked me to let you know if I found any typos in your
paper on computers in offices.
My wife Dianne (she is Suppes' editor) read the paper and made
suggestions and corrections, about 30 in all. As it is cumbersome to
give them in a message, I could bring a marked copy to your house on
Sunday (I need an address), or if you rather have us make the changes,
please send the file over--PUB commands and all--we will change it and
send it back for your approval.
Jane would like to have a corrected copy printed by Monday noon.
Is this possible? When and from where could she come to get it?
We (Committee on Office Systems and Technology) are preparing
a section on text systems for the final report of the Task Force
on the Future of Computing. Your paper would provide an excellent
perspective, and I would like to see it in the report. How would
you feel about its being included in the COST section?
- Pentti Kanerva
-------
The file is OFFICE[W80,JMC], you can ftp it to any
convenient computer. I would like a copy back with
each correction annoated with some special character, say &.
I don't mind it being included in the COST reprt. Les Earnest
and I may want to publish an expanded version. We can certainly
print a revised version Monday am, but for safety, a version
should be printed tomorrow just in case the XGP fails Monday.
∂23-Mar-80 0414 LLW (Non-)Spooling of Skyhook Paper Draft
To: JMC at SU-AI, minsky at MIT-AI
CC: LLW at SU-AI
Marty Frost confirmed that the skyhook paper draft can't be spooled
through the XGP by the KA-10, at least in its present form. I'm going to
munch on it a bit more later today, and try to get it into a form which
will spool either over there or locally (LLL).
Lowell
∂23-Mar-80 1147 Kanerva at SUMEX-AIM OFFICE[W80,JMC]
Date: 23 Mar 1980 1146-PST
From: Kanerva at SUMEX-AIM
Subject: OFFICE[W80,JMC]
To: JMC at SAIL
The revised file is <KANERVA>OFFICE @ SUMEX, ready for FTP. All
comment lines begin with an "&".
I don't know how to transfer the file back to SAIL, apart from
sending it as a message, nor how to print it on the XGP. I plan to be
on campus later today and will try to find someone in Margaret Jacks
to help me. I will check my mail at Sumex in case you have further
comments.
Thank you for your permission to include the paper in the COST
report. I hope that you and Les will have the time to publish an
expanded version.
- Pentti Kanerva
-------
∂23-Mar-80 1634 Kanerva at SUMEX-AIM Long Message
Date: 23 Mar 1980 1632-PST
From: Kanerva at SUMEX-AIM
Subject: Long Message
To: McCarthy at SAIL
Simple methods work.
COMMENT ⊗ VALID 00007 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00002 00002 .if false then begin
C00004 00003 .require "memo.pub[let,jmc]" source
C00006 00004 1. Office computing at the Stanford Artificial Intelligence Laboratory
C00011 00005 2. Conclusions from our experience
C00018 00006 3. Some more controversial contentions
C00024 00007 4. A step further out
C00029 ENDMK
C⊗;
∂23-Mar-80 2034 Kanerva at SUMEX-AIM OFFICE[W80,JMC]
Date: 23 Mar 1980 2033-PST
From: Kanerva at SUMEX-AIM
Subject: OFFICE[W80,JMC]
To: McCarthy at SAIL
I found two more marks to add to OFFICE[W80,JMC].
On page 6, about line 30:
#. Idiot-proofing programs is often a bad idea. It is easy
. . .
charges. However, it has several disadvantages.
& ↑ [punctuation missing]
On page 7, about line 30:
The 1200 baud limitation of present Dialnet is important for
. . .
for a 9000-byte message will be only a little more than the price of a
& ↑
stamp.
We now have a copy of the file on the IMSSS system and can print it
on the Diablo printer. But XGP output would be nicer. Jane or I could
come for it Monday morning. Please advise.
- Pentti Kanerva
-------
∂24-Mar-80 0651 Kanerva at SUMEX-AIM Thank you very much!
Date: 24 Mar 1980 0649-PST
From: Kanerva at SUMEX-AIM
Subject: Thank you very much!
To: JMC at SAIL
-------
∂24-Mar-80 0846 HVA Key to Bldg. 460--M. Jacks Hall
To: JCH
CC: JMC, HVA
Per JMC's request, I am sending key to you via I.D. Mail.
∂24-Mar-80 0927 CLT
∂23-Mar-80 1740 CSD.MCCARTHY at SU-SCORE liquid chalk board
Date: 23 Mar 1980 1735-PST
From: CSD.MCCARTHY at SU-SCORE
Subject: liquid chalk board
To: clt at SU-AI
What's your opinion of replacing the office blackboard with a 4 by 6 foot
liquid "chalk board"? Susan Hill offers.
-------
it is your office
∂24-Mar-80 1031 LES Account request
Opinion?
∂23-Mar-80 1832 IRWIN.MARIN at CMU-10A Obtaining authorization for an account at SAIL
Date: 23 Mar 1980 2128-EST
From: IRWIN.MARIN at CMU-10A
Subject: Obtaining authorization for an account at SAIL
To: les at SU-AI
Dear Mr. Earnest:
I would like to,have permission to use the system at SU-AI
so to be able to access and run KRL and FOL in connection with
my research in the representation of knowledge.
The project which I require this access for is entitled:
Generative Epistomology (cf: SIGART #70 p17 for details).
Its goals are the development of mathematical dialects for representing
different forms of knowledge.
For more details of some prliminary results see (encyclopedia of Computer
Science and Technology vol 9 pp1-27)
I expect the language delelopment phase of the project to last 1-2 years.
In the event I am given permission to use the system for my research
I would like to have the account designation to be "1,ICM". I have checked
these initials out for uniqueness with the finger program and found them OK.
If any further information is required mail can be sent to my network
address : MARIN@CMUA .
Thank you for considering my request and your anticpated prompt response.
ICM
∂24-Mar-80 1400 JMC*
grade for Bosack
∂24-Mar-80 1402 FFL phone call.
roj reddy called. left no number
∂24-Mar-80 1546 FFL Memos to Cohen, Dugan, et al
To: JMC, FFL
Do you want a cover note sent with these memos, or will the JMC heading at top
be sufficient?
They may not recognize JMC. If you put "from John McCarthy" on it, that
will suffice.
∂24-Mar-80 1712 CET via SU-SCORE grades
Plase turn in your grade sheet tomorrow a.m. I must walk them over at
noon. Thanks, Carolyn
∂25-Mar-80 0129 Mark Crispin <Admin.MRC at SU-SCORE>
Date: 25 Mar 1980 0124-PST
From: Mark Crispin <Admin.MRC at SU-SCORE>
To: JMC at SU-AI
In-Reply-To: Your message of 24-Mar-80 1338-PST
Your mail is now forwarded to SAIL. Your old SCORE mail has been
renamed to <CSD.MCCARTHY>OMAIL.TXT (do not rename it back to MAIL.TXT!).
-- Mark --
-------
∂25-Mar-80 1307 Allen.Newell at CMU-10A (A310AN02) Re: AI Memo 326
Date: 25 March 1980 1601-EST (Tuesday)
From: Allen.Newell at CMU-10A (A310AN02)
Subject: Re: AI Memo 326
To: John McCarthy <JMC at SU-AI>
CC: Allen.Newell at CMU-10A
Message-ID: <25Mar80 160108 AN02@CMU-10A>
In-Reply-To: John McCarthy's message of 12 Mar 80 00:17-EST
John: The memo arrived, much thanks. It appears to be still a draft
though it claims to be published in a book as of July 79. What am I
therefore to believe? I enjoyed it. I think some of parts of it are
on the right track, especially the approximate theory bit. I am less
convinced that the second order ploy is necessary, but I think I
understood what was driving you toward it.
I have, it turns out, a developing view towards how to treat
knowledge, which shares some of the view in your memo. I don't really
have a write up of it, but a paragraph or two is in the paper on
Physical Symbol Systems which I gave at the LaJolla conference (and
which is hung in an almost-done state). I'll send you a copy
presently. A central notion, which does not really appear there even,
is that there is another system level, the knowledge level, which is
approximate (vide your notion). It is constructed by the observer of
an agent so as to satisfy the principle of rationality: If an agent
has goal G and knows K and K implies that action A will attain G,
then the agent will take action A. This must be shot though with
ceteris paribus qualifications, but that is the central idea. Actual
agents often cannot be predicted too well by imputting knowledge to
them with this sort of link between knowledge, goals and action. But
then it is necessary to use models of the agent that include some
processing notions, eg, "The agent won't do that because, though he
learned K, he won't remember it"; or "He remains blind to the fact
that his love for his country is inconsistent to always telling the
truth". To predict behavior in such situations requires not an
enhanced notion of knowledge (ie, that permits inconsistent
knowledge), but an explicit notion of processing, data structures,
etc. Thus, representation becomes "knowlege plus access", where the
access is a processing notion.
AN
The book is indeed s said to be published, although I haven't received
my copy. The version that was printed as an AI Memo was a draft
printed by mistake. Therefore I tried and hope I succeeded in
having you sent the version that I hope was printed in the book.
I'll reply to your substantive comments later.
Now that I have had time to read your note fully all I can say is that
I look forward to reading the paper.
∂25-Mar-80 1330 FFL PHON CALL
Please call Mr. Okner at 327-1160.
∂25-Mar-80 2200 JMC*
A general interpretd modal.
∂26-Mar-80 1148 HVA Telephone Call from Brian Randall
He's leaving now for IBM YOrktown, where he will be for next two days; his host
there is Dr. C.V. Freiman (914)252-5905. Wld. you pls. call him there--in about
an hour.CORRECTION: the IBM No. is (914)945-2813 (sorry, I was looking at another
number).
∂26-Mar-80 2113 MINSKY at MIT-AI (Marvin Minsky) skyhookery
Date: 27 MAR 1980 0012-EST
From: MINSKY at MIT-AI (Marvin Minsky)
Subject: skyhookery
To: MINSKY at MIT-AI, jmc at SU-AI, rah at SU-AI, llw at SU-AI
To: hpm at SU-AI
That was fun, last Saturday.
To finish paper on skyhook, we should decide what is needed.
simplest thing is to give the physical analysis, and then
briefly discuss the bugs and problems. We don't try to
solve the problems, only invite others to invent.
I have some worries.
1) If the particles must sustain several g's for thousands
of kilometers, then isn't there an invariant, approximately,
in which Length x Acceleration = constant,
for the above- and below- ground parts? Then,t looks like
underground we need thousands of g's for kilometers. My
question is then, is it really practical to decellerate a
heavy projectile at a thousand g's for any length? (If not,
the borehole has to be many kilometers, which seems
impractical. But with the high Gee, don't we encounter the
sorts of high breaKDOWN voltages, etc., that arise in your
paper on the ballistic fusion idea? And, if we do have very
high Gee, then don't we risk either passing the
superconducting field limit, or dissipation in mechanical
strain in the pellet?
2) An amusing bug occurred to me while talking to Sussman
about the idea. If you have an escape-velocity projectile
in a tube, and it hits one air molecule, then that molecule
will make a micro-crater -- because it has enough energy to
vaporise few atoms from the pellet7s surface. These atoms
will then strike the tube wall, now with escape velocity
w.r.t that -- hence will release a number of "secondary"
particles. these may have velocity low w.r.t the tube wall,
but high with respect to the projectile -- or the next one
So, it seems to me we have a photomultiplier-type avalanche
or chain reaction!
(Proposed cure is to build two-stage system, with big
tower, with slow pellets, up to top of atmosphere. Then the
second stage needs no continuous tube, and this bug wno't
hurt.
3). At least for my own edification, where can I get brief
tutorial on probable dissipation of the all-superconductor
system, and of the system in which the BBs are super, but
the tower coils are not. You mentioned cryogenic aluminum.
Is it that this has low enough resistance to keep losses
below the .00001 or so we need to survive?
4) Although we should say something about stability --
e.g., will a wind-pulse on the side propagate upwards into a
disastrous whip-like wave -- I have faith, and I think so
does JMC, that even if we cannot find a "strong-focussing"
geometry, that there must be a computational way to
stabilize the system. After all, a centrigugal rope is very
stab+e -- and this one, with all its directed momentum, must
really be pretty stiff. But it needs a brief discussion.
Again, my view is that we might actually do better by
leaving all fine points for others to write papers about.
This gets them involved, and more likely for someone
cleverly to invent a variant that is really more practical.
--- Marvin
∂27-Mar-80 1013 DCD
To: "@AIPHIL[1,DCD]" at SU-AI
see cleanu[1,dcd] (not "cleannu") for conference tidbits.
∂27-Mar-80 1124 DCD
To: "@AIPHIL[1,DCD]" at SU-AI
does anyone have the Stanford library copy of Zenon's book?
∂27-Mar-80 1600 JMC*
Ask about time-sharing memo.
∂27-Mar-80 1607 FFL phone call
To: JMC
CC: FFL
Man called from MIT (home is Newcasttle upon Tyne) asking if you will come to
an AI Conference in Newcastle Sept. 2-6, l980. Had sent a letter earlier.
Checking to see if you had received it.
∂28-Mar-80 1151 BERN at MIT-DMS (Bahram Niamir) The home information terminal
Date: 28 Mar 1980 1450-EST
From: BERN at MIT-DMS (Bahram Niamir)
To: jmc at SU-AI
Cc: bern at MIT-DMS
Subject: The home information terminal
Message-id: <[MIT-DMS].141807>
I was very happy to read your article on the prospects of a future
community information system (sail:hoter.ess[ess,jmc]). Here are
a few points. Please respond to those you choose:
1- As to the publishing capability of such a system: The Information
system that you describe will effectively remove 1- the heavy initial
capital investment, and 2- the economy of scale factors in the publishing
and broadcasting industries. This will undermine the current position
of monopolies that dominate TV/Radio broadcasting and the
mass circulated daily papers. Making a copy of an article and
sending it to a subscriber will cost the same as the unit cost of
making 100,000 copies and distributing it to 100,000 recipients. It will
no longer be true that the bigger you are the easier it is for you to
print and distribute stuff.
2- There is a problem with the ownership of the timesharing services.
Since the media industry, advertising industry, and the business and
especially the political community will correctly perceive such a system
to be a threat to their interests, they will try to influence the
development and breadth of such a service by owning and controlling
the timesharing services themselves. They will set rules as to who
can use a certain service (such as broadcasting) and in what capacity.
For example, polling of users of the baldness lotion will be outlawed
and anybody submitting an article for general consumption will have
to pass a strict editorial control board that would eliminate critical
articles. Worse of all, the discussion of such practices will be
forbidden.
Jmc: It seems to me that you exaggerate the effects of self-interest
on people's behavior, perhaps through not understanding the mechanism
by which self-interest affects behavior. Time-sharing services are
not a natural monopoly, so for your horror story to come to pass,
you have to figure out how someone is to develop the self-righteousness
and political power to make it a monopoly. It is easy to see that
a dictatorship of either the right or left could find reasons for
preventing misleading publication, but I can't see it happening
in a society that otherwise permits uncensored publication.
The polling problem is more complicated than it seems, because
the question arises of who gets to formulate the issues on which
polling takes place. The Swedish nuclear referendum is a case
in point where the parties mixed the issue of capitalism vs.
socialism with the nuclear issue, and also left no option for
the unabashed nuclear enthusiast.
I would like to be on the mailing list, and I am pleased to have
people read the paper which is one of many "position papers" that
I have in the computer. It would be nice to receive a message
when someone FTPs the paper just to get an estimate of readership.
3- A community information system is the natural mechanism for polling
the opinion of the community about certain policies, and most important
of all, the natural setting for voting on proposals of importance to
the community. However, are the politicians going to relinquish
their authority back to the constituents? Then there is the problem
of unscrupulousness: Passwords will ensure that elections will not
be rigged; but what about those who will sell their votes?
4- There is a (fairly large) network wide mailing list (HUMAN-NETS@MIT-AI) set
up for discussion of topics such as those covered in your paper. Let me know
if you would like to have your name added to the list. Back issues of
messages may be found in (MIT-AI:DUFFEY;HUMNET ARCHIV).
5- I would like to put a mit-cmu wide system message telling users
about your article and where on mit they can find it. Let me know
if you agree with this.
6- What conference is it that you mention at the end of your paper?
Bahram (bern@mit-dms)
∂28-Mar-80 2156 AS challenge
To: JMC, JCH
I"ve tidied up my message and put it into file CHALL[1,as].
...I"m beginning to get on top of the editor.
Aaron
∂29-Mar-80 1116 MINSKY at MIT-AI (Marvin Minsky)
Date: 29 MAR 1980 1359-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, jmc at SU-AI, rah at SU-AI, llw at SU-AI
To: hpm at SU-AI
Still worried about underground accelerator. Any ideas? I
can imagine that it is practical to build a 5 kilometer
vertical accelerator inside a mountain, for a couple of
billion dollars.If it could sustain 1000G, then we'd have
5000 G-Kilometers. . Something like that would be enough to
get to a low-←erigee stable orbit, wouldn't it? But if
1000G isn't practical, we need an additional idea to make
thing work.
ht
∂29-Mar-80 1514 DCD travel reimbursement
To: "@AIPHIL[1,DCD]" at SU-AI
I am slowly gathering the travel reimbursement requests from speakers and
participants. Shirley expects all that to funnel through me, so people will
just have to wait till I get back and submit the whole batch to her. But I
suppose if anyone is DESPERATE one of you can shepherd it through Shirley and
get a check made out. Also, don't forget to turn in (to me) party expenses
for Sunday night. Have a nice week. Dan
∂29-Mar-80 1603 AS continued
oops typed wrong key again.
I can"t find a combination which has the same effect as CTRL-B has
in non-DM mode, i.e. making output pause. I gather there"s meant to be
a key to do this. Do these CASBS terminals have it?
Aaron
PS. Dan has left your copy of my book on his desk.
∂29-Mar-80 1608 AS apology
I think the first half of that last message must have got lost. Guess
I should have typed CONT not MAIL.
It was about there being no clear distinction between what is and
what isn"tt a conceptual problem. Also I feel that having to invent
the idea of two personalities in one person is not too different in
kind (thought very different in detail) from having to invent the
idea of a line-like structure in something that isn"t a line,
or having to invent a way of looking at the counting process
which makes the result clearly independent of the order of counting,
or seeing the same structure twice over (the second time somewhat
distorted) in my example of seeing an axis of symmetry.
In all the cases there are problems about how the newly discovered
view of the world should be represented, and problems about what
cues could indicate a need for the new kind of representation,
and what mechanisms could make it possible to devise and use the
new representation.
Perhaps lots of examples of understanding of metaphors make
similar points.
Anyway, I"ll think some more about your comment. (Still can"t find a
key for apostrophe!)
Aaron
∂29-Mar-80 1757 MARIMONT at SRI-KL Grad School
Date: 29 Mar 1980 1757-PST
From: MARIMONT at SRI-KL
Subject: Grad School
To: JMC at SU-AI
cc: Marimont
I have been admitted to the PhD program in Electrical Engineering. Thanks for
your help.
Dave
-------
∂29-Mar-80 1912 AS hold
have at last found what to do about lack of "hold" key:
use ctrl-break to hold and ctrl-clear to unhold. But it doesn't
seem to work al the time. E.g. works with DIR and TY, but not WHO.
I've been browsing in documentation. The facilities here are very
powerful, but I think the human engineering has not been of the highest
calibre......
Your are right about <ctrl>break, etc. It works to control typeout
but not to control display programs.
∂29-Mar-80 1909 LGC
I'd like to phone you to discuss the schedule for Engelmore's visit, the
present status of SNAPOV.TXT[DGI,LGC], and other related matters. Would
sometime this evening (say, 8:00 or 8:30) be ok?
∂29-Mar-80 1918 AS Self control!
To what extent should I worry about telephone charges when using
this system? I've been logged in quite a long time now, finding out
how things work.
No need to worry about telephone charges, but we aren't supposed
to use the phones in the office for anything but calling the
computer since most offices don't have phones. Local calls
do no real harm, but the phone in the library is available on
weekends, and you should get a key to it on Monday.
∂30-Mar-80 1513 LES Engelmore visit
To: JMC, RWW, LGC
Reminder: my calendar shows Bob Engelmore coming Monday, March 31 to
review formal reasoning work.
∂31-Mar-80 0000 LGC
It just occurred to me that if you're now in MJH and would be willing
to pick up two XGP listings of SNAPOV which I'll make, it would save me
a round trip from San Jose to Palo Alto this evening -- that I would
appreciate!
∂31-Mar-80 0709 LGC SNAPOV
I have (and had) no intention of distributing SNAPOV.TXT in its present form.
In the first place, I am aware of some rough and questionable spots in it
which I want to improve. For example, Section VIII on the programming
formalism is tentative, and I want to formulate concrete examples of all the
points made as a soundness check before letting it out. In the second place,
I wouldn't want to distribute something of this sort until I had discussed
its content with you in detail, and taken appropriate account of any comments
or criticisms you might have to offer. Thirdly, some parts of SNAPOV are
still too incomplete and undeveloped for any but the most limited
distribution.
My original intention was to let Bob Engelmore take some version of SNAPOV
(possibly a cut down and revised one) back to Washington with him when he
leaves the Stanford area, and there may still be time (a few days) to make
the required revisions and/or excisions. A more immediate problem is how
much of it to reveal to him this morning -- I would like him to have a
document in his hands to direct the discussion and give him something to
raise questions about. This, too, may still be possible, but is something
I want to discuss with you (at least by telephone) before he arrives. If
you're up early enough, you might give me a call before 8:00 at 272-2908.
I don't want to make a presentation to Engelmore without the benefit of at
least a brief discussion of SNAPOV with you. -- Lew
∂31-Mar-80 1100 BERN at MIT-AI (Bahram Niamir)
Date: 31 MAR 1980 1358-EST
From: BERN at MIT-AI (Bahram Niamir)
To: JMC at SU-AI
You have been added to the HUMAN-NETS mailing list.
You may find back issues in MIT-AI:DUFFEY;HUMNET ARCHIV.
The archive is approximately 54kw. You may find the
last half of it (i.e, chronologically older messages) of more
interest.
∂31-Mar-80 1654 AS Trip to San Diego
Geoff Hinton phoned to ask if I'd be willing to go to San Diego Thursday
returning Friday. Is this OK with you? Most other people seem to be going
away with kids this week, so they'll not miss me!
Aaron
∂31-Mar-80 1932 TOB
∂30-Mar-80 1556 JMC
John
E.M.Riseman
Tom
What is first initial of Riseman at U.Mass.?
∂31-Mar-80 2210 AS intentionality
In INTENT[1,as] there's the beginnings of a discussion of intentionality
which is relevant to both your paper on first-order theories and the
papers Dan has been writing about beliefs and other intentional states.
My criticism of Dan is that he has not taken computing concepts seriously
enough. The relevance to your paper is that I want to stress the
importance of distinguishing a model of human mental processes from
a modelel of human thinking about human mental processes. Your formalism
is an attempt to deal with the latter. (Of course it's a subset of the
former.)
Aaron
PS - did you get my enquiry about whether it would be OK for me to
go to San Diego Thursday, returning Friday?
∂31-Mar-80 2357 LLW More Shyhookery
To: minsky at MIT-AI
CC: JMC at SU-AI, LLW at SU-AI, RAH at SU-AI
∂26-Mar-80 2114 MINSKY at MIT-AI (Marvin Minsky) skyhookery
Date: 27 MAR 1980 0012-EST
From: MINSKY at MIT-AI (Marvin Minsky)
Subject: skyhookery
To: MINSKY at MIT-AI, jmc at SU-AI, rah at SU-AI, llw at SU-AI
To: hpm at SU-AI
That was fun, last Saturday.
To finish paper on skyhook, we should decide what is needed.
simplest thing is to give the physical analysis, and then
briefly discuss the bugs and problems. We don't try to
solve the problems, only invite others to invent.
[I agree in general, but Rod and I feel strongly that at least one
solution must be *sketched* to each of the major problems, lest the basic
idea be dismissed out of hand; also, doing so is probably a prequisite to
publication in a refereed journal. Others won't really dig into improving
the sketched solutions unless they think the whole scheme is basically
sound, and few have the breadths of expertises and backgrounds that the
four of us do--it really seems incumbent on us to make the basic
feasibility arguments, in a space-spanning (albeit not -filling) fashion.]
I have some worries.
1) If the particles must sustain several g's for thousands
of kilometers, then isn't there an invariant, approximately,
in which Length x Acceleration = constant,
[Indeed--it's the energy (integral)!]
for the above- and below- ground parts? Then,t looks like
underground we need thousands of g's for kilometers. My
question is then, is it really practical to decellerate a
heavy projectile at a thousand g's for any length? (If not,
the borehole has to be many kilometers, which seems
impractical. But with the high Gee, don't we encounter the
sorts of high breaKDOWN voltages, etc., that arise in your
paper on the ballistic fusion idea? And, if we do have very
high Gee, then don't we risk either passing the
superconducting field limit, or dissipation in mechanical
strain in the pellet?
[We are planning on sustained accelerations of roughly 100,000 g's over
linear motor lengths of about 1 km. These are comparable to those
routinely endured by artillery shells which have far fancier gadgetry in
them than will our little shells. Strong materials don't crush
significantly until accelerations of the order of 100 giga-gees, and the
engineering of sophisticated structures for 0.1 mega-gee acceleration
environments is a well-developed (if slightly arcane) technology. No,
high voltage breakdown (electrical or magnetic) is *not* a
limitation--recall that the acceleration of our ballistic fusion pellets
was about an order of magnitude greater than will occur in the motors, so
that peak electrical field strengths are down by three-fold (which is huge
in the regime in which we are working) in this pair of accelerators.
Likewise, mechanical dissipation is very small--really no problem.]
2) An amusing bug occurred to me while talking to Sussman
about the idea. If you have an escape-velocity projectile
in a tube, and it hits one air molecule, then that molecule
will make a micro-crater -- because it has enough energy to
vaporise few atoms from the pellet7s surface. These atoms
will then strike the tube wall, now with escape velocity
w.r.t that -- hence will release a number of "secondary"
particles. these may have velocity low w.r.t the tube wall,
but high with respect to the projectile -- or the next one
So, it seems to me we have a photomultiplier-type avalanche
or chain reaction!
[Tell Sussman to concentrate on flogging GLS to thesis completion, and
leave the molecular engineering to us! Since the sound speeds (i.e.,
atomic/molecular bond frequencies) in the materials of our (very strong)
cannonballs shells and tower are nearly an order-of-+magnitude greater
than that of any atmospheric component in the solid state, even fairly
high energy collisions of the cannonballs with individual gas molecules
are relatively elastic for everyone but the gas molecules, which will be
dissociated into atoms by collisions with the cannonballs, thereby
inelastically soaking up much of the collision energy. Any metal which is
sputtered from the cannonball surface will leave with a kinetic energy
small compared to its latent heat of vaporization per atom/molecule, and
therefore will condense with near-unity likelihood on the tower wall (I
would expect him to recall this from chip fabrication technology, where
sputtering deposition under comparable conditions is frequently employed).
Cannonball atoms scattered highly forward will be slowed by collision with
residual gas atoms well before they 'see' the tower wall the first time,
and will thereupon bind. Highly reactive gas atomic species will then be
very efficiently 'gettered' by the tower interior, so that the tower will
be self-cleaning, not self-dirtying--this trick is used ubiquitously in
magnetically confined plasma experiments. The 'chain reaction' is
markedly sub-critcal!]
(Proposed cure is to build two-stage system, with big
tower, with slow pellets, up to top of atmosphere. Then the
second stage needs no continuous tube, and this bug wno't
hurt.
3). At least for my own edification, where can I get brief
tutorial on probable dissipation of the all-superconductor
system, and of the system in which the BBs are super, but
the tower coils are not. You mentioned cryogenic aluminum.
Is it that this has low enough resistance to keep losses
below the .00001 or so we need to survive?
[Landau and Lifshitz, Electrodynamics of Continuous Media (Addison-Wesley)
have excellent treatments of both superconductors magnetostatics and of
moving conductors in magnetic fields. Superconductor dissipation physics
is semi-empirical above the lower critical field, when the bulk Meissner
effect becomes imperfect; it's gets acutely messy for AC fields, which are
of interest to us--we'll send references along with the completed paper
text, if you're really feeling masochistic. Cryogenic aluminum is
magical, in that it not only has a low temperature resistance equaled only
by copper, but it has also negligible magnetoresistence, as well, which is
of course crucial to us.]
4) Although we should say something about stability --
e.g., will a wind-pulse on the side propagate upwards into a
disastrous whip-like wave -- I have faith, and I think so
does JMC, that even if we cannot find a "strong-focussing"
geometry, that there must be a computational way to
stabilize the system. After all, a centrigugal rope is very
stab+e -- and this one, with all its directed momentum, must
really be pretty stiff. But it needs a brief discussion.
[Being MIT-trained from boyhood to PhD (i.e., late boyhood), Rod does
multi-dimensional stability analyses before breakfast--it would be a real
struggle to keep one such out of this paper. As for strong focussing and
all related issues, we are determined to have a system which is
comprehensively single-point failure-proof; strong focussing is
preferable, in that it is passive, for some failure modes, but we'll go
active, as you suggest, just as often as necessary to accomplish this
failure-immunity goal.]
Again, my view is that we might actually do better by
leaving all fine points for others to write papers about.
This gets them involved, and more likely for someone
cleverly to invent a variant that is really more practical.
[Fine points, certainly. Completeness of exposition should be our business,
however.]
--- Marvin
[Lowell]
∂01-Apr-80 0000 JMC*
pay house tax
∂01-Apr-80 0938 AS
OK thanks. I'll be back Friday evening.
Using the editor to look at mail and send replies is very nice.
Would you recommend the seminar on 'flavourful (?) objects'
advertised for this afternoon? Or is it something to do with
April fool's day?
∂01-Apr-80 0957 ES Mt. Shasta
Are you still interested in climbing Mt. Shasta? The end of April would be
a good time to go.
Unfortunately, my leg is still out of order, and I'm out of condition
from lack of running. I doubt I'll be climbing before summer.
∂01-Apr-80 1027 MINSKY at MIT-AI (Marvin Minsky)
Date: 1 APR 1980 1325-EST
From: MINSKY at MIT-AI (Marvin Minsky)
To: MINSKY at MIT-AI, rah at SU-AI, llw at SU-AI, jmc at SU-AI
To: hpm at SU-AI
Great. You have propped up the Promise of Joy again. A few
more points.
1. I still think that people will be skeptical of 100,000G,
particularly because of anchoring problems. We should
select a "standard model", the middle line in a table of
possibilities, as in your earlier draft. Even as low as 2G
in the tube could be considered, which would need about
12,000G for 1 kilometer.
2. The Coriolis force when lifting a load. The tower will
bend Westward as the load is raised. Then, both the stiff
particle stream and the earth's centrifugal force conspire
to pull it straight; these torques are basically unopposed
except by inertia. So the column is set into an inverted
pendulum swinging mode. Right?
It would then be possible to time things so that the load
has a substantial Eastward velocity when it gets to the top!
If this is right, it means that the load can get a free
extra orbital injection velocity. I haven't computed it
exactly, but I estimate thatif the tower were to oscillate
under the natural centrifugal load over a wide angle, its
top would move at a couple kilometers/second. This could be
exploited to make the tower quite a lot lower, and still
obtain orbital velocitis at the top, but maybe its a bad
idea. At the extreme of the swing, it could approximate an
evolute of the earth's surface. The stream would have to
modulated temporally to keep the system balanced. As I
said, pobably a bad idea, but could reduce costs by a modest
factor, but gives up free synchronous orbit capability.
How to damp out the oscillation -- assuming that we don't
want it? Well, I presume that "natural" damping will be
essentially nonexistent. But we can simply send up loads at
time-intervals of one-half cycle -- and cancel the effect.
I haven't estimated the cycle time, which comes from the
quotient of beam stiffness and column density. Is it an
invariant, or does the period decrease with faster beam
speeds? It would be nice to have it come out around a few
days or so.
In any case, this means that the system needs careful
dynamic load programming. Also, there will be substantial
lateral forces on the stream (in addition to normal balanced
Coriolis) when loads are raised.
3. Elevators. At first I thought of a platform that uses
electric motors to lift itself, using traction on the
column, and deriving power from its own ccoils. But
perhaps it can be made to levitate itself nicely: it has
its own set of retarding colis which react directly against
the magnetic stream in the tubes. Thus, the platform
doesn't even touch the tower. It consists of a coil and a
beam. The payloads hang, balanced from the beam. The
elevator coil reacts against an ascending particle, stores
energy, and returns it to a descending particle; three coils
may be necessary to aviod dead spots. Is there enough
external field? I should think that if the particles are
any fraction of a meter in length, then there will be enough
magnetic dipole field, a meter or two away from the stream,
to induce plenty of lift.
4. Vulnerability. When I suggested leaving bugs for
others, I didn't mean omitting any fundamental physics -- I
agree entirely that unless we complete a physically
practical skeleton, no one will believe. What I meant was
to list clearly the known bugs, and to propose a solution to
each, but leave an invitation to improve. For instance,
Meteors. If the column is a meter wide and 30Km high, it
has a cross-section of 30 square kilometers. What is
puncture rate?
5. Finally, I suppose we should discuss Reliability. Here
is where inventions of others could be welcomed. he idea of
deflecting slots is fine, but means that the retarding coils
must be very far apart, etc. It's fun to imagine the worst
case disaster, of losing the whole system -- like a reactor
melt-down. The kinetic stream is literally a megaton-like
explosive, but only a fraction of the downward stream will
hit the earth. The lowest part produces a big crater at the
ground site, and the rest of that fraction leaves a huge
linear trail to the East. Most of the ascending stream will
escape forever into Space, thank heaven (which may not thank
us in return). But the rather large mass in the uppermost
sections -- which is, sadly, most of the mass, will hang
around in eccentric orbits producing a two-dimensional
Saturn ring of great nuisance value -- until it is cleaned
up. What fun.
∂01-Apr-80 1408 LES retransmission
I don't think I heard from you about this.
∂24-Mar-80 1031 LES Account request
To: JMC
Opinion?
∂23-Mar-80 1832 IRWIN.MARIN at CMU-10A Obtaining authorization for an account at SAIL
Date: 23 Mar 1980 2128-EST
From: IRWIN.MARIN at CMU-10A
Subject: Obtaining authorization for an account at SAIL
To: les at SU-AI
Dear Mr. Earnest:
I would like to,have permission to use the system at SU-AI
so to be able to access and run KRL and FOL in connection with
my research in the representation of knowledge.
The project which I require this access for is entitled:
Generative Epistomology (cf: SIGART #70 p17 for details).
Its goals are the development of mathematical dialects for representing
different forms of knowledge.
For more details of some prliminary results see (encyclopedia of Computer
Science and Technology vol 9 pp1-27)
I expect the language delelopment phase of the project to last 1-2 years.
In the event I am given permission to use the system for my research
I would like to have the account designation to be "1,ICM". I have checked
these initials out for uniqueness with the finger program and found them OK.
If any further information is required mail can be sent to my network
address : MARIN@CMUA .
Thank you for considering my request and your anticpated prompt response.
ICM
Well, everything has to paid for somehow. We could ask him to buy a slice,
we could trade it to CMU for something, perhaps just courtesy or we could
ask Richard whether he wants to support him out of his slice. Perhaps you
could find his reference or ask him to say where we can FTP it from.
∂01-Apr-80 2140 LES
∂01-Apr-80 2043 JMC
Well, everything has to paid for somehow. We could ask him to buy a slice,
we could trade it to CMU for something, perhaps just courtesy or we could
ask Richard whether he wants to support him out of his slice. Perhaps you
could find his reference or ask him to say where we can FTP it from.
-------
This must be the answer to something. What was the question?
It concerned a request for an account from IRWIN.MARIN%CMU.