perm filename E83.IN[LET,JMC] blob
sn#864706 filedate 1983-10-01 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00417 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00046 00002 ∂03-Jul-83 1621 ME Prancing Pony Bill
C00048 00003 ∂04-Jul-83 2225 ullman@Diablo Re: policy on computer usage
C00050 00004 ∂05-Jul-83 0125 ARK Facilities Charges
C00052 00005 ∂05-Jul-83 0147 ME einit.cmd(18)
C00053 00006 ∂05-Jul-83 0845 LIBRARY@SU-SCORE.ARPA Crytography
C00054 00007 ∂05-Jul-83 0853 DFH TEX Users Course
C00055 00008 ∂05-Jul-83 0930 ullman@Diablo Re: computer usage suggestion
C00057 00009 ∂05-Jul-83 1249 ullman@Diablo supercomputers
C00060 00010 ∂05-Jul-83 1300 JMC*
C00061 00011 ∂05-Jul-83 1314 ME EXECUTE in E
C00062 00012 ∂05-Jul-83 1459 ullman@Diablo Mike, I'm sure, was probing, but he apparently did have a
C00065 00013 ∂05-Jul-83 1622 PACK@SU-SCORE.ARPA Work
C00067 00014 ∂05-Jul-83 1637 DFH tape
C00068 00015 ∂05-Jul-83 1725 MRC@SU-SCORE.ARPA
C00069 00016 ∂05-Jul-83 1804 LLW@S1-A Tomorrow
C00071 00017 ∂05-Jul-83 2321 HST inivitation
C00073 00018 ∂06-Jul-83 0302 RPG
C00076 00019 ∂06-Jul-83 1040 CLT
C00077 00020 ∂06-Jul-83 1200 GENESERETH@SUMEX-AIM discouraging words
C00079 00021 ∂06-Jul-83 1422 GARDNER@SU-SCORE.ARPA Ph.D. oral
C00080 00022 ∂06-Jul-83 1547 MAYR@SU-SCORE.ARPA Supercomputer seminar
C00082 00023 ∂07-Jul-83 0938 GROSZ@SRI-AI.ARPA [Kjb at SRI-AI: Meeting]
C00085 00024 ∂07-Jul-83 1003 ARK Facilities Committee Mtg
C00086 00025 ∂07-Jul-83 1122 DFH
C00087 00026 ∂07-Jul-83 1621 JJW@S1-A
C00088 00027 ∂07-Jul-83 2057 RPG Good news
C00089 00028 ∂07-Jul-83 2120 GROSZ@SRI-AI.ARPA [Kjb at SRI-AI: r.s.v.p.]
C00091 00029 ∂08-Jul-83 0949 DFH Facilities Committee Meeting
C00095 00030 ∂08-Jul-83 1156 DFH
C00096 00031 ∂08-Jul-83 1235 PACK@SU-SCORE.ARPA Meeting
C00097 00032 ∂08-Jul-83 1323 NILSSON@SRI-AI.ARPA no meeting
C00098 00033 ∂08-Jul-83 1816 GOLUB@SU-SCORE.ARPA Re: director of departmental computer facilities
C00109 00034 ∂08-Jul-83 1816 GOLUB@SU-SCORE.ARPA [Charles Hedrick <HEDRICK@RUTGERS.ARPA>: director of departmental computer facilities]
C00111 00035 ∂11-Jul-83 1039 GOLUB@SU-SCORE.ARPA Tandem
C00112 00036 ∂11-Jul-83 1214 MAYR@SU-SCORE.ARPA Supercomputer seminar
C00114 00037 ∂11-Jul-83 1425 RPG Wieneke
C00115 00038 ∂11-Jul-83 1504 SHARON@SU-SCORE.ARPA phone message
C00116 00039 ∂11-Jul-83 1514 Hewitt%MIT-OZ@MIT-MC thanks for your reprint
C00118 00040 ∂11-Jul-83 1833 DEK guibas
C00119 00041 ∂11-Jul-83 2227 ME gone till end of July
C00120 00042 ∂11-Jul-83 2341 cheriton@Diablo Tandem
C00121 00043 ∂12-Jul-83 0832 DFH Final technical report
C00122 00044 ∂12-Jul-83 0925 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #13
C00135 00045 ∂12-Jul-83 1002 GARDNER@SU-SCORE.ARPA Erratum
C00136 00046 ∂12-Jul-83 1030 KJB@SRI-AI.ARPA Thursday
C00138 00047 ∂12-Jul-83 1033 PACK@SU-SCORE.ARPA meeting, natural kinds
C00143 00048 ∂12-Jul-83 1100 JMC*
C00144 00049 ∂12-Jul-83 1127 KJB@SRI-AI.ARPA Update
C00147 00050 ∂13-Jul-83 1408 RPG
C00160 00051 ∂13-Jul-83 1707 DFH
C00163 00052 ∂13-Jul-83 2222 ullman%SU-HNV.ARPA@SU-SCORE.ARPA meeting
C00164 00053 ∂14-Jul-83 0533 @MIT-MC:Burns.boulder@RAND-RELAY
C00166 00054 I think you sent this to me by mistake.
C00168 00055 ∂15-Jul-83 1042 NILSSON@SRI-AI.ARPA interview
C00170 00056 ∂15-Jul-83 1118 RINDFLEISCH@SUMEX-AIM Workstation Purchase Dilemma
C00175 00057 ∂15-Jul-83 1127 DFH phone message
C00176 00058 ∂15-Jul-83 1130 DFH TEX class
C00177 00059 ∂15-Jul-83 1133 BUCHANAN@SUMEX-AIM Re: Workstation Purchase Dilemma
C00179 00060 ∂15-Jul-83 1144 KJB@SRI-AI.ARPA CSLI
C00183 00061 ∂15-Jul-83 1255 RPG Lisp Machines
C00186 00062 ∂15-Jul-83 1552 minker.umcp-cs@UDel-Relay IJCAI Panel on Non-Monotonic Reasoning
C00190 00063 ∂15-Jul-83 1627 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract
C00192 00064 ∂15-Jul-83 1641 KJB@SRI-AI.ARPA Announcing CSLI
C00194 00065 ∂15-Jul-83 1738 ullman@Diablo Re: Equipment Contract
C00195 00066 ∂15-Jul-83 1810 BRESNAN.PA@PARC-MAXC.ARPA Kelly Roach
C00203 00067 ∂15-Jul-83 1814 BRESNAN.PA@PARC-MAXC.ARPA Kelly Roach's resume
C00210 00068 ∂15-Jul-83 2332 burns.boulder@Rand-Relay
C00211 00069 ∂15-Jul-83 2333 burns.boulder@Rand-Relay
C00212 00070 ∂16-Jul-83 1401 KJB@SRI-AI.ARPA csli
C00215 00071 ∂16-Jul-83 1751 LLW@S1-A Festschrift Shackleton
C00217 00072 ∂16-Jul-83 2013 LLW@S1-A Deadline
C00219 00073 ∂17-Jul-83 1041 CLANCEY@SUMEX-AIM Re: Workstation Purchase Dilemma
C00222 00074 ∂17-Jul-83 2007 GENESERETH@SUMEX-AIM
C00223 00075 ∂18-Jul-83 0727 jlh@Shasta Re: Equipment Contract
C00225 00076 ∂18-Jul-83 0950 JMC*
C00226 00077 ∂18-Jul-83 1009 SHORTLIFFE@SUMEX-AIM Re: Workstation Purchase Dilemma
C00228 00078 ∂18-Jul-83 1331 wallace@Whitney AI and space station
C00230 00079 ∂18-Jul-83 1446 GARDNER@SU-SCORE.ARPA CBCL
C00231 00080 ∂18-Jul-83 1758 LENAT@SU-SCORE.ARPA Re: Workstation Purchase Dilemma
C00233 00081 ∂18-Jul-83 2155 wallace@Whitney HAL and spaceborne AI
C00236 00082 ∂19-Jul-83 0759 OP unsupported students
C00237 00083 ∂19-Jul-83 1038 RAC (1) research topic, (2) James P. Hogan
C00248 00084 ∂19-Jul-83 1212 RPG Closures
C00252 00085 ∂19-Jul-83 1706 BOSACK@SU-SCORE.ARPA Fixing your modem
C00253 00086 ∂19-Jul-83 1721 RAC Hogan and Research
C00254 00087 ∂20-Jul-83 1425 DFH
C00256 00088 ∂20-Jul-83 1841 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #14
C00263 00089 ∂20-Jul-83 2317 GENESERETH@SUMEX-AIM meeting
C00264 00090 ∂21-Jul-83 1141 JMC*
C00265 00091 ∂21-Jul-83 1141 JMC*
C00266 00092 ∂21-Jul-83 1214 BOSACK@SU-SCORE.ARPA
C00267 00093 ∂21-Jul-83 1230 RESTIVO@SU-SCORE.ARPA Re: name change
C00268 00094 ∂21-Jul-83 2106 RINDFLEISCH@SUMEX-AIM DEC Changes
C00270 00095 ∂21-Jul-83 2138 LLW@S1-A No Time For Travel
C00271 00096 ∂21-Jul-83 2233 GENESERETH@SUMEX-AIM schedule
C00273 00097 ∂22-Jul-83 0034 FRIEDLAND@SUMEX-AIM Re: DEC Changes
C00275 00098 ∂22-Jul-83 1000 jlh@Shasta Re: DEC Changes
C00277 00099 ∂22-Jul-83 1001 BOSACK@SU-SCORE.ARPA Line to SAIL
C00278 00100 ∂22-Jul-83 1013 GOLUB@SU-SCORE.ARPA Re: DEC Changes
C00279 00101 ∂22-Jul-83 1120 wallace@Whitney AI on space station
C00281 00102 ∂22-Jul-83 1331 BOSACK@SU-SCORE.ARPA Re: It needs work.
C00282 00103 ∂22-Jul-83 1351 Janet.Asbury@CMU-CS-A Re: two documents
C00284 00104 ∂22-Jul-83 1421 Raj.Reddy@CMU-CS-A Re: two documents
C00285 00105 ∂22-Jul-83 1607 LB it seems to work now
C00286 00106 ∂22-Jul-83 1612 DFH
C00288 00107 ∂22-Jul-83 1710 minker.umcp-cs@UDel-Relay Non-Monotonic Panel Session
C00291 00108 ∂22-Jul-83 2340 POURNE@MIT-MC big meeting
C00292 00109 ∂23-Jul-83 1043 CLT Diana's reservations
C00293 00110 ∂23-Jul-83 1143 BOSACK@SU-SCORE.ARPA
C00294 00111 ∂24-Jul-83 2031 LLW@S1-A Pournelle Meeting Revisited
C00296 00112 ∂24-Jul-83 2142 KJB@SRI-AI.ARPA test
C00297 00113 ∂24-Jul-83 2143 LEP traffic lights
C00298 00114 ∂24-Jul-83 2151 LLW@S1-A Big Meeting
C00300 00115 ∂24-Jul-83 2208 KJB@SRI-AI.ARPA Update
C00305 00116 ∂25-Jul-83 0135 POURNE@MIT-MC Big Meeting
C00306 00117 ∂25-Jul-83 0152 POURNE@MIT-MC meeting
C00308 00118 ∂25-Jul-83 0700 Janet.Asbury@CMU-CS-A Centre Mondial
C00310 00119 ∂25-Jul-83 0941 DFH
C00311 00120 ∂25-Jul-83 1517 DFH
C00312 00121 ∂26-Jul-83 1013 DFH phone message
C00313 00122 ∂26-Jul-83 1244 PKANERVA@SUMEX-AIM About my dissertation
C00316 00123 ∂26-Jul-83 1633 minker.umcp-cs@UDel-Relay AAAI and Lerner
C00319 00124 ∂26-Jul-83 1758 reid@Glacier Kelly Roach
C00320 00125 ∂26-Jul-83 1813 @SU-SCORE.ARPA:gardner@Glacier CMOS10 run and wafer prober
C00322 00126 ∂26-Jul-83 1858 GROSZ@SRI-AI.ARPA Kelly Roach
C00323 00127 ∂26-Jul-83 2055 sap%Shasta.#Pup@SU-SCORE.ARPA prober
C00324 00128 ∂26-Jul-83 2317 WALKER@SRI-AI.ARPA IJCAI Executive Committee Meeting
C00327 00129 ∂27-Jul-83 0104 LLW@S1-A In Case You Missed This Earlier. . .
C00329 00130 ∂27-Jul-83 1123 @USC-ISI,OHLANDER@USC-ISI Task Proposal
C00333 00131 ∂27-Jul-83 1321 DFH Kelly Roach
C00335 00132 ∂27-Jul-83 1409 reid@Glacier Re: Kelly Roach
C00336 00133 ∂27-Jul-83 1601 DFH Flight/motel arrangements Roach
C00338 00134 ∂28-Jul-83 0901 BLUM@SUMEX-AIM Vibrio Cholerae
C00340 00135 ∂28-Jul-83 0901 SHORTLIFFE@SUMEX-AIM Re: Vibrio Cholerae
C00344 00136 ∂28-Jul-83 0915 KJB@SRI-AI.ARPA Assistant Director of CSLI, etc.
C00348 00137 ∂28-Jul-83 0942 LLW@S1-A Pournelle Metting Travel Arrangements
C00351 00138 ∂28-Jul-83 1557 minker.umcp-cs@UDel-Relay Non-Monotonic Logic
C00400 00139 ∂28-Jul-83 1603 DFH Roach
C00401 00140 ∂28-Jul-83 1936 KJB@SRI-AI.ARPA My hours
C00403 00141 ∂28-Jul-83 2001 LLW@S1-A Travel Arrangements
C00406 00142 ∂28-Jul-83 2006 KJB@SRI-AI.ARPA Student support for this year
C00409 00143 ∂28-Jul-83 2013 CLT x masda
C00410 00144 ∂29-Jul-83 1032 ULLMAN@SU-SCORE.ARPA meeting
C00411 00145 ∂29-Jul-83 1035 @SU-SCORE.ARPA:DELAGI@SUMEX-AIM.ARPA Please get back to me with additions and amendments.../bruce
C00419 00146 ∂29-Jul-83 1403 ULLMAN@SU-SCORE.ARPA next meeting
C00420 00147 ∂29-Jul-83 1417 KJB@SRI-AI.ARPA Rooms
C00423 00148 ∂29-Jul-83 1649 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA Agenda
C00429 00149 ∂31-Jul-83 1213 RPG Password
C00430 00150 ∂31-Jul-83 1642 ME E substitution speed
C00432 00151 ∂01-Aug-83 0959 CLT plumber
C00433 00152 ∂01-Aug-83 1049 BSCOTT@SU-SCORE.ARPA NSF Grant MCS7905998
C00435 00153 ∂01-Aug-83 1318 EENGELMORE@SUMEX-AIM Technology Magazine
C00436 00154 ∂01-Aug-83 1337 NILSSON@SRI-AI.ARPA net addresses
C00438 00155 ∂01-Aug-83 1359 @SU-SCORE.ARPA:LERMAN@SRI-KL.ARPA AAAI publications comm. meeting
C00441 00156 ∂01-Aug-83 1408 BSCOTT@SU-SCORE.ARPA [OHLANDER at USC-ISI: TAsk Proposals]
C00444 00157 ∂01-Aug-83 1436 SCHMIDT@SUMEX-AIM Re: Workstation Purchase Dilemma
C00454 00158 ∂01-Aug-83 1551 SCHMIDT@SUMEX-AIM Re: Workstation Purchase Dilemma
C00457 00159 ∂01-Aug-83 1605 ULLMAN@SU-SCORE.ARPA Spectrum Supercomputer Survey
C00459 00160 ∂01-Aug-83 1609 ULLMAN@SU-SCORE.ARPA Three messages
C00481 00161 ∂01-Aug-83 1615 DFH
C00482 00162 ∂01-Aug-83 1617 DFH Victor Kuo
C00483 00163 ∂01-Aug-83 1741 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Center for Parallel Computation
C00485 00164 ∂01-Aug-83 1743 ullman%SU-HNV.ARPA@SU-SCORE.ARPA
C00489 00165 ∂01-Aug-83 2107 YOM Back ...
C00490 00166 ∂02-Aug-83 1146 ULLMAN@SU-SCORE.ARPA proposal
C00491 00167 ∂02-Aug-83 1712 BSCOTT@SU-SCORE.ARPA ARPA Equipment Contract
C00493 00168 ∂03-Aug-83 0747 Raj.Reddy@CMU-CS-A
C00495 00169 ∂03-Aug-83 0823 RINDFLEISCH@SUMEX-AIM Symbolics Machines
C00497 00170 ∂03-Aug-83 1003 CLT
C00498 00171 ∂03-Aug-83 1033 DFH travel expenses - ARPA account
C00499 00172 ∂03-Aug-83 1143 ullman%SU-HNV.ARPA@SU-SCORE.ARPA spectrum survey
C00500 00173 ∂03-Aug-83 1149 ullman%SU-HNV.ARPA@SU-SCORE.ARPA
C00516 00174 ∂03-Aug-83 1155 DFH travel arrangements
C00518 00175 ∂03-Aug-83 1156 DFH phone msg.- Pournelle
C00519 00176 ∂03-Aug-83 1206 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: spectrum survey
C00520 00177 ∂03-Aug-83 1306 DFH kuo
C00521 00178 ∂03-Aug-83 1348 BSCOTT@SU-SCORE.ARPA No-Cost Extension, Current ARPA contract
C00523 00179 ∂03-Aug-83 1435 ULLMAN@SU-SCORE.ARPA [Jack Schwartz <SCHWARTZ.CMCL1@NYU.ARPA>: The Challenge is Technological]
C00530 00180 ∂03-Aug-83 1515 BRACEWELL@SU-SIERRA.ARPA NASA Reports
C00531 00181 ∂03-Aug-83 1614 jlh%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: spectrum survey
C00532 00182 ∂03-Aug-83 2026 ZZZ
C00533 00183 ∂03-Aug-83 2027 ZZZ
C00534 00184 ∂04-Aug-83 0707 Raj.Reddy@CMU-CS-A Gunkel-Fredkin-Reddy meeting
C00536 00185 ∂04-Aug-83 0833 STAN@SRI-AI.ARPA Some info on IJCAI panel
C00539 00186 ∂04-Aug-83 0931 KJB@SRI-AI.ARPA Thanks
C00540 00187 ∂04-Aug-83 0932 KJB@SRI-AI.ARPA program
C00542 00188 ∂04-Aug-83 1323 STAN@SRI-AI.ARPA Panel questions
C00546 00189 ∂04-Aug-83 1536 AAAI-OFFICE%SUMEX-AIM.ARPA@SU-SCORE.ARPA AAAI Election Results
C00549 00190 ∂04-Aug-83 1746 JJW@S1-A 206
C00550 00191 ∂04-Aug-83 1750 JJW@S1-A
C00551 00192 ∂04-Aug-83 2144 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Andy Freeman
C00552 00193 ∂04-Aug-83 2217 ullman%SU-HNV.ARPA@SU-SCORE.ARPA survey
C00554 00194 ∂04-Aug-83 2300 MAYR@SU-SCORE.ARPA supercomputer theory seminar
C00556 00195 ∂05-Aug-83 0011 PMF@S1-A
C00557 00196 ∂05-Aug-83 0850 jlh%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: Andy Freeman
C00558 00197 ∂05-Aug-83 0914 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: Andy Freeman
C00560 00198 ∂05-Aug-83 0924 ULLMAN@SU-SCORE.ARPA virtual multiprocessors
C00561 00199 ∂05-Aug-83 0948 BSCOTT@SU-SCORE.ARPA Research Associate Salaries
C00563 00200 ∂05-Aug-83 0946 DFH references -- psycho.4
C00565 00201 ∂05-Aug-83 1023 KJB@SRI-AI.ARPA mail at sail
C00567 00202 ∂05-Aug-83 1038 DFH Lifschitz called
C00568 00203 ∂05-Aug-83 1315 lantz%SU-HNV.ARPA@SU-SCORE.ARPA FYI
C00573 00204 ∂05-Aug-83 1402 JJW@S1-A Tape recorder
C00574 00205 ∂05-Aug-83 1410 BSCOTT@SU-SCORE.ARPA
C00575 00206 ∂05-Aug-83 1430 reid@Glacier File server meeting
C00577 00207 ∂05-Aug-83 1536 GOTELLI@SU-SCORE.ARPA Meeting
C00578 00208 ∂05-Aug-83 1632 DFH key/phone message
C00579 00209 ∂05-Aug-83 2108 GOLUB@SU-SCORE.ARPA Vote for faculty position
C00580 00210 ∂05-Aug-83 2110 WEENING@SU-SCORE.ARPA Tape recorder
C00581 00211 ∂05-Aug-83 2134 ME Prancing Pony Bill
C00583 00212 ∂06-Aug-83 0006 JJW Tape recorder
C00584 00213 ∂06-Aug-83 1945 ZZZ REPORT
C00588 00214 ∂07-Aug-83 0953 BSCOTT@SU-SCORE.ARPA
C00589 00215 ∂07-Aug-83 1022 reid@Glacier Re: Meeting
C00590 00216 ∂08-Aug-83 1725 KJB@SRI-AI.ARPA Sept 1
C00593 00217 ∂08-Aug-83 2243 KJB@SRI-AI.ARPA New CSLI Program
C00608 00218 ∂09-Aug-83 0253 ARK Chris Tucci's Comp Project
C00611 00219 ∂09-Aug-83 0314 ARK Facilities Commitee
C00614 00220 ∂09-Aug-83 1051 AAAI-OFFICE%SUMEX-AIM.ARPA@SU-SCORE.ARPA Halbrecht's suggestion
C00631 00221 ∂09-Aug-83 1538 GOLUB@SU-SCORE.ARPA ballots
C00632 00222 ∂09-Aug-83 2214 PW Common Lisp Proposal
C00633 00223 ∂10-Aug-83 0754 KJB@SRI-AI.ARPA Mailing list for CSLI
C00635 00224 ∂10-Aug-83 1138 ATKINSON@SU-SCORE.ARPA 1983-84 Student Support
C00637 00225 ∂10-Aug-83 1608 KJB@SRI-AI.ARPA Sept 1
C00640 00226 ∂10-Aug-83 1625 KJB@SRI-AI.ARPA Dr. Betsey Macken
C00641 00227 ∂10-Aug-83 1635 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: Center for Parallel Computation
C00643 00228 ∂10-Aug-83 2113 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Re: Center for Parallel Computation
C00644 00229 ∂11-Aug-83 0306 corky@Navajo Summer Research
C00648 00230 ∂11-Aug-83 0308 corky@Navajo Papers
C00649 00231 ∂11-Aug-83 1029 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Freeman
C00651 00232 ∂11-Aug-83 1114 ULLMAN@SU-SCORE.ARPA meeting
C00652 00233 ∂11-Aug-83 1231 RAC Party /Impending departure
C00654 00234 ∂11-Aug-83 1339 @MIT-MC:VVAWH@MIT-MC
C00656 00235 ∂12-Aug-83 1015 @SU-SCORE.ARPA:ATP.BLEDSOE@UTEXAS-20.ARPA My Operation
C00660 00236 ∂14-Aug-83 2218 ULLMAN@SU-SCORE.ARPA bibliography
C00661 00237 ∂15-Aug-83 0849 @USC-ISI,OHLANDER@USC-ISI
C00663 00238 ∂15-Aug-83 0850 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: bibliography
C00667 00239 ∂15-Aug-83 0855 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Meeting of Sept. 15
C00669 00240 ∂15-Aug-83 1352 GOTELLI@SU-SCORE.ARPA Meeting
C00670 00241 ∂15-Aug-83 1524 Bulnes.PA@PARC-MAXC.ARPA Party on Thursday
C00672 00242 ∂16-Aug-83 0826 RINDFLEISCH@SUMEX-AIM Lisp Machines
C00673 00243 ∂16-Aug-83 1816 RAC
C00676 00244 ∂17-Aug-83 1048 CHAPPELL@SRI-AI.ARPA Message ID
C00677 00245 ∂17-Aug-83 1048 CHAPPELL@SRI-AI.ARPA Message ID
C00678 00246 ∂17-Aug-83 1638 ULLMAN@SU-SCORE.ARPA [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
C00681 00247 ∂17-Aug-83 1642 DFH
C00683 00248 ∂17-Aug-83 1731 cheriton%SU-HNV.ARPA@SU-SCORE.ARPA Re: [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
C00685 00249 ∂17-Aug-83 2208 GENESERETH@SUMEX-AIM debate
C00687 00250 ∂18-Aug-83 1009 RINDFLEISCH@SUMEX-AIM Symbolics Deal
C00689 00251 ∂18-Aug-83 1016 DFH Susan called
C00690 00252 ∂18-Aug-83 1117 GOLUB@SU-SCORE.ARPA Vote on Guibas
C00692 00253 ∂18-Aug-83 1205 WUNDERMAN@SRI-AI.ARPA Your Address Memo
C00693 00254 ∂18-Aug-83 1653 YOM CS206
C00694 00255 ∂18-Aug-83 1731 TRATTNIG@SU-SCORE.ARPA Special Meeting
C00696 00256 ∂18-Aug-83 1829 GOLUB@SU-SCORE.ARPA amended vote
C00698 00257 ∂19-Aug-83 0839 BSCOTT@SU-SCORE.ARPA Personnel
C00699 00258 ∂19-Aug-83 1606 SHARON@SU-SCORE.ARPA box
C00700 00259 ∂19-Aug-83 1732 RESTIVO@SU-SCORE.ARPA Phillipe Roussel
C00702 00260 ∂19-Aug-83 1821 RESTIVO@SU-SCORE.ARPA
C00703 00261 ∂20-Aug-83 1000 JMC*
C00704 00262 ∂20-Aug-83 1032 ATP.BLEDSOE@UTEXAS-20.ARPA
C00706 00263 ∂20-Aug-83 1205 KJB@SRI-AI.ARPA Progress!
C00709 00264 ∂20-Aug-83 1551 @MIT-MC:PHW%MIT-OZ@MIT-MC Your AI book
C00711 00265 ∂22-Aug-83 1034 BMACKEN@SRI-AI.ARPA Meeting
C00714 00266 ∂22-Aug-83 1109 ULLMAN@SU-SCORE.ARPA confusion
C00715 00267 ∂22-Aug-83 1450 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Invited Speakers
C00718 00268 ∂22-Aug-83 2159 Guy.Steele@CMU-CS-A Re: Hacker's Dictionary
C00719 00269 ∂23-Aug-83 0521 TARNLUND@SRI-AI.ARPA A draft of your book
C00720 00270 ∂23-Aug-83 1026 CLT kronos
C00721 00271 ∂23-Aug-83 1339 @SU-SCORE.ARPA:DFH@SU-AI
C00725 00272 ∂23-Aug-83 1342 BMACKEN@SRI-AI.ARPA Important Events
C00727 00273 ∂23-Aug-83 1651 ULLMAN@SU-SCORE.ARPA schedule
C00729 00274 ∂23-Aug-83 1740 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA wallich@csnet-sh : 4th question
C00743 00275 ∂24-Aug-83 0837 CL.BOYER@UTEXAS-20.ARPA Re: more on Stoyan
C00745 00276 ∂24-Aug-83 1325 DFH
C00746 00277 ∂24-Aug-83 1357 ARK New CF hire?
C00747 00278 ∂24-Aug-83 1640 ATKINSON@SU-SCORE.ARPA 83-84 Student Support
C00748 00279 ∂24-Aug-83 2025 CLT opera
C00749 00280 ∂25-Aug-83 0155 ARK for your information
C00752 00281 ∂25-Aug-83 0204 ARK Query about rates
C00754 00282 ∂25-Aug-83 1637 ULLMAN@SU-SCORE.ARPA CIS panel session
C00756 00283 ∂25-Aug-83 1645 ULLMAN@SU-SCORE.ARPA seminar course
C00758 00284 ∂25-Aug-83 1722 STUCKY@SRI-AI.ARPA Potluck
C00760 00285 ∂25-Aug-83 1737 RESTIVO@SU-SCORE.ARPA Phillipe Roussel
C00762 00286 ∂25-Aug-83 1803 GOLUB@SU-SCORE.ARPA Computer Science Library Technical Reports List
C00767 00287 ∂26-Aug-83 1604 DFH
C00771 00288 ∂27-Aug-83 1138 @MIT-MC:TK@MIT-MC
C00773 00289 ∂28-Aug-83 0415 POURNE@MIT-MC Marconi
C00774 00290 ∂28-Aug-83 1152 EF@MIT-MC
C00775 00291 ∂28-Aug-83 1159 EF@MIT-MC
C00796 00292 ∂28-Aug-83 1837 RESTIVO@SU-SCORE.ARPA
C00797 00293 ∂28-Aug-83 1934 RESTIVO@SU-SCORE.ARPA
C00798 00294 ∂29-Aug-83 0006 ME XGP
C00799 00295 ∂29-Aug-83 0043 FEIGENBAUM@SUMEX-AIM.ARPA Re: CIS panel session
C00801 00296 ∂29-Aug-83 0933 KJB@SRI-AI.ARPA Sept 1 colloquium
C00802 00297 ∂29-Aug-83 0944 DFH A draft of your book
C00811 00298 ∂29-Aug-83 1529 CHURD@SU-SCORE.ARPA Score
C00812 00299 ∂29-Aug-83 1631 PACK@SU-SCORE.ARPA Re: discussing traffic lights
C00814 00300 ∂30-Aug-83 1345 JMC*
C00815 00301 ∂30-Aug-83 2251 minker%umcp-cs@UDel-Relay Mints
C00817 00302 ∂31-Aug-83 1012 TARNLUND@SRI-AI.ARPA
C00818 00303 ∂31-Aug-83 1018 RINDFLEISCH@SUMEX-AIM [Tom Binford <TOB@SU-AI>: LISP machines ]
C00822 00304 ∂31-Aug-83 1100 DFH
C00823 00305 ∂31-Aug-83 1124 RPG Minor points
C00825 00306 ∂31-Aug-83 1428 Bonnie%UPenn%UPenn@UDel-Relay That resort near Poughkeepsie you mentioned to Ray Reiter & me
C00827 00307 ∂31-Aug-83 1432 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Burton Smith visit - Reschedule Request
C00829 00308 ∂31-Aug-83 2236 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
C00845 00309 ∂01-Sep-83 1000 JMC*
C00846 00310 ∂01-Sep-83 1024 DELAGI%SUMEX-AIM.ARPA@SU-SCORE.ARPA PPA Non-Disclosure
C00848 00311 ∂01-Sep-83 1024 DFH Keith Clark
C00849 00312 ∂01-Sep-83 1031 DFH travel arrangements
C00850 00313 ∂01-Sep-83 1126 Bonnie%UPenn%UPenn@UDel-Relay
C00851 00314 ∂01-Sep-83 1708 DFH Common[e83,jmc]
C00852 00315 ∂01-Sep-83 1834 ME Prancing Pony Bill
C00854 00316 ∂02-Sep-83 1015 JMC*
C00855 00317 ∂02-Sep-83 1019 EF@MIT-MC
C00856 00318 ∂02-Sep-83 1027 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA AAMSI
C00861 00319 ∂02-Sep-83 1038 TRATTNIG@SU-SCORE.ARPA Keith Clark's visit
C00862 00320 ∂02-Sep-83 1107 GOTELLI@SU-SCORE.ARPA Re: directory
C00863 00321 ∂02-Sep-83 1125 NILSSON@SRI-AI.ARPA AAMSI
C00868 00322 ∂02-Sep-83 1253 TUCCI@SU-SCORE.ARPA project
C00869 00323 ∂02-Sep-83 1305 SHARON@SU-SCORE.ARPA dina bola travel
C00870 00324 ∂02-Sep-83 1500 JMC*
C00871 00325 ∂04-Sep-83 0150 ARK Happy birthday!
C00872 00326 ∂04-Sep-83 1331 JK
C00873 00327 ∂04-Sep-83 1940 RPG Congratulations
C00874 00328 ∂04-Sep-83 2204 SHORTLIFFE%SUMEX-AIM.ARPA@SU-SCORE.ARPA forwarded messages regarding AAMSI and AAAI
C00881 00329 ∂05-Sep-83 2008 ULLMAN@SU-SCORE.ARPA
C00882 00330 ∂06-Sep-83 1042 BMACKEN@SRI-AI.ARPA Associations
C00886 00331 ∂06-Sep-83 1151 ullman@Diablo Re: [Tom Binford <TOB@SU-AI>: LISP machines ]
C00888 00332 ∂06-Sep-83 1527 ULLMAN@SU-SCORE.ARPA meeting place
C00889 00333 ∂07-Sep-83 1049 BSCOTT@SU-SCORE.ARPA N00039-82-C-0250
C00890 00334 ∂07-Sep-83 1600 AAAI-OFFICE@SUMEX-AIM [John McCarthy <JMC@SU-AI>:]
C00909 00335 ∂07-Sep-83 1705 KJB@SRI-AI.ARPA Description of research projects
C00911 00336 ∂07-Sep-83 2144 POURNE@MIT-MC HELP
C00912 00337 ∂07-Sep-83 2157 GENESERETH@SUMEX-AIM ai course
C00914 00338 ∂08-Sep-83 0010 HST
C00915 00339 ∂08-Sep-83 0400 @LBL-CSAM.ARPA:uw-beaver!ubc-vision!reiter@LBL-CSAM n.m.workshop
C00917 00340 ∂08-Sep-83 1145 NILSSON@SRI-AI.ARPA Fredkin Meeting
C00919 00341 ∂08-Sep-83 1544 JRP@SRI-AI.ARPA VISITING SCHOLAR CARD
C00921 00342 ∂08-Sep-83 1831 Stefik.PA@PARC-MAXC.ARPA Re: [John McCarthy <JMC@SU-AI>:]
C00925 00343 ∂09-Sep-83 0348 Tenenbaum@SRI-KL.ARPA Re: [John McCarthy <JMC@SU-AI>:]
C00928 00344 ∂09-Sep-83 1035 @SU-SCORE.ARPA:TRATTNIG@SU-SIERRA.ARPA guest speakers
C00930 00345 ∂09-Sep-83 1106 bobrow.PA@PARC-MAXC.ARPA Re: [John McCarthy <JMC@SU-AI>:]
C00934 00346 ∂09-Sep-83 1107 ULLMAN@SU-SCORE.ARPA seminar progress report
C00937 00347 ∂09-Sep-83 1322 LIBRARY@SU-SCORE.ARPA AI and Lisp Books
C00939 00348 ∂09-Sep-83 1419 KJB@SRI-AI.ARPA EMACS on IBM-PC?
C00940 00349 ∂09-Sep-83 1441 WUNDERMAN@SRI-AI.ARPA "Outer" Colloquium Series
C00943 00350 ∂09-Sep-83 1624 WUNDERMAN@SRI-AI.ARPA ANNOUNCEMENT
C00945 00351 ∂09-Sep-83 1739 BMACKEN@SRI-AI.ARPA associations
C00948 00352 ∂09-Sep-83 1809 KJB@SRI-AI.ARPA p.s. to Announcment
C00949 00353 ∂09-Sep-83 2223 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: seminar progress report
C00951 00354 ∂09-Sep-83 2229 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: seminar progress report
C00953 00355 ∂12-Sep-83 1110 DOYLE@CMU-CS-C.ARPA Long-term AI research
C00959 00356 ∂12-Sep-83 1250 LAWS@SRI-AI.ARPA Re: Long-term AI research
C00962 00357 ∂12-Sep-83 1545 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Thursday Meeting (9/15, 3pm)
C00963 00358 ∂12-Sep-83 1808 MRC@SU-SCORE.ARPA removal of pre-1983 Score mailboxes
C00967 00359 ∂13-Sep-83 0638 Bossu.GIA@MIT-MULTICS.ARPA
C00968 00360 ∂13-Sep-83 1312 @SRI-AI.ARPA:BrianSmith.PA@PARC-MAXC.ARPA Area C Meeting
C00973 00361 ∂13-Sep-83 1614 HANS@SRI-AI.ARPA CSLI mailing lists, directory, etc.
C00978 00362 ∂14-Sep-83 0843 BMACKEN@SRI-AI.ARPA Advisory Panal
C00980 00363 ∂14-Sep-83 1044 AAAI-OFFICE@SUMEX-AIM 1984 Conference Facilities
C00985 00364 ∂14-Sep-83 1101 KJB@SRI-AI.ARPA Good news
C00986 00365 ∂14-Sep-83 1226 DOYLE@CMU-CS-C.ARPA Re: Long-term AI research
C00989 00366 ∂15-Sep-83 1121 Winograd.PA@PARC-MAXC.ARPA Re: Advisory Panal
C00991 00367 ∂15-Sep-83 1334 ELYSE@SU-SCORE.ARPA fac. int.
C00992 00368 ∂15-Sep-83 1455 @SU-SCORE.ARPA:TRATTNIG@SU-SIERRA.ARPA Special Supercomputer Seminar
C00994 00369 ∂15-Sep-83 1651 BUCHANAN@SUMEX-AIM Comtex Scientific
C00996 00370 ∂15-Sep-83 1714 @SRI-AI.ARPA:PULLUM%HP-HULK.HP-Labs@Rand-Relay West Coast Conference on Formal Linguistics, 1984
C00999 00371 ∂16-Sep-83 0806 KJB@SRI-AI.ARPA CSLI annoncements
C01002 00372 ∂16-Sep-83 0854 KJB@SRI-AI.ARPA More on csli announcements
C01004 00373 ∂16-Sep-83 0937 KJB@SRI-AI.ARPA Computer language people
C01006 00374 ∂16-Sep-83 1024 AAAI-OFFICE@SUMEX-AIM Presidential address
C01007 00375 ∂16-Sep-83 1052 JRP@SRI-AI.ARPA fall educational program
C01009 00376 ∂16-Sep-83 1132 @SRI-AI.ARPA:BRESNAN.PA@PARC-MAXC.ARPA syntactic constraints on discourse
C01012 00377 ∂16-Sep-83 2044 Brachman@SRI-KL.ARPA AAAI-84 Technical Program
C01024 00378 ∂18-Sep-83 1226 DIETTERICH@SUMEX-AIM AAAI Post-doctoral fellowships
C01027 00379 ∂19-Sep-83 0004 POURNE@MIT-MC Great White Cabin
C01029 00380 ∂19-Sep-83 0907 AAAI-OFFICE@SUMEX-AIM post-doctoral fellowships
C01030 00381 ∂19-Sep-83 1025 RPH BS
C01031 00382 ∂19-Sep-83 1128 WALKER@SRI-AI.ARPA [Derek Sleeman <SLEEMAN@SUMEX-AIM.ARPA>: AI Research Notes series]
C01038 00383 ∂19-Sep-83 1146 DFH Re: space
C01040 00384 ∂19-Sep-83 1501 @MIT-MC:DAVIS%MIT-OZ@MIT-MC Re: [John McCarthy <JMC@SU-AI>:]
C01044 00385 ∂20-Sep-83 0707 PETTY@RUTGERS.ARPA 1983 abstract mailing
C01050 00386 ∂20-Sep-83 1153 DKANERVA@SRI-AI.ARPA Prospective papers, tech reports, monographs
C01052 00387 ∂20-Sep-83 1212 @MIT-MC:Pool@MIT-MULTICS Jackson paper
C01070 00388 ∂21-Sep-83 0116 ARK Chris Tucci's Comprehensive Project
C01071 00389 ∂21-Sep-83 0950 DFH
C01072 00390 ∂21-Sep-83 1006 REG
C01074 00391 ∂21-Sep-83 1421 KJB@SRI-AI.ARPA postdocs
C01077 00392 ∂21-Sep-83 2133 TRATTNIG@SU-SCORE.ARPA Tomorrow's Seminar
C01079 00393 ∂21-Sep-83 2140 @SRI-AI.ARPA:BrianSmith.PA@PARC-MAXC.ARPA No Computation Seminar Meeting
C01080 00394 ∂22-Sep-83 1107 @SRI-AI.ARPA:TW@SU-AI New seminar in CS Dept.
C01083 00395 ∂23-Sep-83 1030 @SRI-AI.ARPA:Barwise.PA@PARC-MAXC.ARPA Working group on Programming Language Semantics
C01086 00396 ∂23-Sep-83 1738 KJB@SRI-AI.ARPA Secretaries at CSLI
C01089 00397 ∂23-Sep-83 1755 GROSZ@SRI-AI.ARPA Seminar Announcement--Context
C01095 00398 ∂24-Sep-83 1538 KJB@SRI-AI.ARPA Newsletter
C01121 00399 ∂26-Sep-83 1007 BMACKEN@SRI-AI.ARPA Grand Opening
C01125 00400 ∂26-Sep-83 1400 rod@Whitney CS200 departmental lectures
C01127 00401 ∂26-Sep-83 1522 @SRI-AI.ARPA:TW@SU-AI Semantics of interrogatives and imperatives
C01128 00402 ∂26-Sep-83 1610 @SRI-AI.ARPA:Kaplan.pa@PARC-MAXC.ARPA Project A1: Phonology, Morphology, Synax
C01130 00403 ∂26-Sep-83 1613 @SRI-AI.ARPA:Kaplan.pa@PARC-MAXC.ARPA Project A3: Strategies and tactics in the processing of
C01132 00404 ∂27-Sep-83 0001 JMC Expired plan
C01133 00405 ∂27-Sep-83 0934 BMACKEN@SRI-AI.ARPA associations
C01136 00406 ∂27-Sep-83 1017 @SRI-AI.ARPA:withgott.pa@PARC-MAXC.ARPA Re: Project A1: Phonology, Morphology, Synax
C01138 00407 ∂27-Sep-83 1031 @SRI-AI.ARPA:withgott.pa@PARC-MAXC.ARPA Re: Project A1: Phonology, Morphology, Synax
C01140 00408 ∂27-Sep-83 1049 WUNDERMAN@SRI-AI.ARPA Announcement of Colloquium, Thurs. Sept. 29
C01143 00409 ∂27-Sep-83 1059 @SRI-AI.ARPA:withgott.pa@PARC-MAXC.ARPA Re: Project A2 : Phonology, Syntax, and Discourse S.
C01145 00410 ∂27-Sep-83 1106 @SRI-AI.ARPA:halvorsen.pa@PARC-MAXC.ARPA Research Seminar, Natural Language, Thursday Sept. 29
C01147 00411 ∂27-Sep-83 1411 KJB@SRI-AI.ARPA newsletter.two
C01148 00412 ∂27-Sep-83 1451 TUCCI@SU-SCORE.ARPA Form signed
C01149 00413 ∂27-Sep-83 1504 ARK ∂27-Sep-83 1451 TUCCI@SU-SCORE.ARPA Form signed
C01150 00414 ∂27-Sep-83 1554 DFH Facilities Committee Meeting
C01151 00415 ∂27-Sep-83 1612 DFH Travel arrangements
C01153 00416 .27-Sep-83 1653 DFH
C01157 00417 ∂27-Sep-83 1704 AAAI-OFFICE@SUMEX-AIM Agenda
C01160 ENDMK
C⊗;
∂03-Jul-83 1621 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 3 July 1983
Previous Balance 0.97
Monthly Interest at 1.5% 0.01
-------
TOTAL AMOUNT DUE 0.98
Please deliver payments to Diana Hall, room 358, Jacks Hall.
Make checks payable to: STANFORD UNIVERSITY.
To ensure proper crediting, please include your Pony account name on your check.
Bills are payable upon presentation. Interest of 1.5% per month will be
charged on balances remaining unpaid 25 days after bill date above.
You haven't paid your Pony bill since 12/82.
Accounts with balances remaining unpaid for more than 55 days are
considered delinquent and are subject to reduction of credit limit.
Please pay your bill and keep your account current.
∂04-Jul-83 2225 ullman@Diablo Re: policy on computer usage
Received: from SU-HNV by SU-AI with PUP; 04-Jul-83 22:25 PDT
Date: Mon, 4 Jul 83 22:27 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Re: policy on computer usage
To: JEF@Sail, JMC@Sail
No, it most definitely is a proposed change; in particular, the
idea of ALTO+(SAIL or SCORE) as a birthright would be replaced by
a SCORE account only; students would have to justify the need for
SAIl or the ALTO's (or anything else). Moreover, the account would
be limited, and would make certain activities (which are suspect
are contrary to departmental policy anyway, like courseork) too
expensive for students with no sponsor.
By the way, Gene is the one who is going to decide the issue,
so he's the one who should receive your comments.
∂05-Jul-83 0125 ARK Facilities Charges
Now that we have an active facilities committee meeting regularly, does
that mean that all important computer facilities decisions will be cleared
or approved by the committee? In particular, will the computer facilities
committee get to clear/approve the charges starting September? I certainly
hope so, as I think that the SAIL/SCORE charges need to be reassessed and
this should include review by non-CSD-CF people (i.e., the payers) before
it becomes effective. You can consider this a request to add an agenda item.
Arthur
∂05-Jul-83 0147 ME einit.cmd(18)
∂05-Jul-83 0143 JMC (on TTY20) exe einit.cmd/18p
To: BUG-e
I would like to be able to use a few macros from einit.cmd by putting
the above command in my einit file. I get a message saying
"illegal file specification", and indeed exe doesn't work if the
specification designates a page. Can this be fixed?
ME - What you're supposed to say is ⊗XEXECUTE EINIT.CMD(18), with
the page number in parens, not after a slash and followed by a "p".
∂05-Jul-83 0845 LIBRARY@SU-SCORE.ARPA Crytography
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Jul 83 08:45:39 PDT
Date: Tue 5 Jul 83 08:47:43-PDT
From: C.S./Math Library <LIBRARY@SU-SCORE.ARPA>
Subject: Crytography
To: mccarthy@SU-AI.ARPA
John,
Cryptography:a primer by Alan Konheim Z103.K66 is in the library. (It came
in during the time of our computer problems and evidently got processed
without cards in the catalog) We have also recently received Meyer and Matyas
Crytography New Dimension in computer data security Z103.M55.
Harry
~
-------
∂05-Jul-83 0853 DFH TEX Users Course
I am signed up to go to a TEX82 course July 11 and 12
(next Mon. and Tues) which will take until about 4 pm
each day. Several people seemed to think it a good
idea that I go. Please let me know if there is any
time conflict as far as you are concerned.
No problem with TEX82 course.
∂05-Jul-83 0930 ullman@Diablo Re: computer usage suggestion
Received: from SU-HNV by SU-AI with PUP; 05-Jul-83 09:29 PDT
Date: Tue, 5 Jul 83 09:31 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Re: computer usage suggestion
To: JMC@Sail
The idea is not bad. I think if we were careful, we could reduce the
load on the department, but it would involve making each person
doing such a project "systems staff", so we could figure it as a
cost of running the system, rather than use of the system.
There are a few points that could cause trouble though:
(1) the students might not be willing to run the project. Stu reges
would pobably like some help, but not THAT much.
(2) We probably should guarantee some free support to everyone, regardless
of whether they worked for it. Perhaps your "workfare" plan should
apply only to those wanting large slices of the pie.
(3) There may be some problems with new students who need the computer
for mail, e.g., but aren't yet capable of making a contribution.
∂05-Jul-83 1249 ullman@Diablo supercomputers
Received: from SU-HNV by SU-AI with PUP; 05-Jul-83 12:48 PDT
Date: Tue, 5 Jul 83 11:40 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: supercomputers
To: jmc@sail
Mike Harrison mentioned to Ernst Mayr that I had a supercomputer proposal
that is "content free." Tracing the remark leads back to something
you said to Mike at IBM. If you really feel that I have no ideas,
you might like to tell me about that. Apparently my proposal has been
better received in other circles.
I read your supercomputer proposal, and I supposed that it more concerned
the organization of collaboration on super computers than a specific
idea for designing one. At least that was what I got out of it.
As to your own ideas of what a supercomputer should be, I didn't
get one from the paper. I would not conclude from that fact that
you don't have ideas. If I carelessly said something that gave that
impression, I apologize. However, I do think that the political
situation vis a vis Japan and the U.S. Government has given rise to
a scientific situation that has some unhealthy aspects. Namely, the
willingness to finance "supercomputers" has outrun ideas for designing
supercomputers or even specific ideas about how the computers should
be super. I'll read your proposal again, and perhaps I'll owe you
a further apology for missing something important when I read it
the first time. I am also surprised that Mike Harrison passed on
such casual remarks as representing his own opinion if he did.
You will discover on Thursday that my own ideas on supercomputers
are alas rather slight.
∂05-Jul-83 1300 JMC*
have tape transcribed
∂05-Jul-83 1314 ME EXECUTE in E
∂05-Jul-83 0151 JMC (on TTY20 0151)
Thanks. It doesn't agree with usage in ε<filename>/page, however.
ME - But it does agree with DO FOO(n) and COPY FOO(n).
∂05-Jul-83 1459 ullman@Diablo Mike, I'm sure, was probing, but he apparently did have a
Received: from SU-HNV by SU-AI with PUP; 05-Jul-83 14:59 PDT
Date: Tue, 5 Jul 83 15:01 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Mike, I'm sure, was probing, but he apparently did have a
To: JMC@Sail
quote to back up the probe. I'm not about to tout my own ideas;
I do think I have a perspective on the problem that may lead somewhere,
maybe not. What really disturbs me is not that you may have said
something that Harrison could exploit to my detriment, but that Ed
is getting so much free publicity and boosting his own project
without the solid ideas to back them up. In turn, he will exploit
his publicity to get the hardware he wants and keep it from others
on campus who are equally, perhaps better, qualified to conduct
supercomputer research. You may think otherwise, and if you
really think Ed has the right ideas, I wouldn't think of asking
you to keep that a secret. However, Mike H. is apparently quite
influential in Washington circles now, and things we say to him
could have substantial impact.
It was not my intention to have any impact at all, because to have
any opinion on who should get what Government money, taking all factors
into account, would require a major effort on my part.
∂05-Jul-83 1622 PACK@SU-SCORE.ARPA Work
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Jul 83 16:22:09 PDT
Date: Tue 5 Jul 83 16:23:33-PDT
From: Leslie E. Pack <PACK@SU-SCORE.ARPA>
Subject: Work
To: jmc@SU-AI.ARPA
I am very interested in working in the area of epistemology of AI.
Don't know if I have enough logic background, though. I have stopped
by to talk to you a couple of times, but keep missing you. Is there a
particular time I'm more likely to find you?
-Leslie
-------
If you get this in time to come in this afternoon before 6pm that will
be fine. Otherwise, Friday would be good or try any afternoon next week.
∂05-Jul-83 1637 DFH tape
The notes I transcribed are in travel[e83,jmc].
I went as far as where you said "end of tape,"
but note that there does appear to be more on
the tape. There were several places where I
couldn't understand what was said.
∂05-Jul-83 1725 MRC@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Jul 83 17:25:05 PDT
Date: Tue 5 Jul 83 17:27:01-PDT
From: Mark Crispin <MRC@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Tue 5 Jul 83 17:23:00-PDT
Postal-Address: 725 Mariposa Ave. #103; Mountain View, CA 94041
Phone: (415) 497-1407 (Stanford); (415) 968-1052 (residence)
On Score, FINGER <last name> gets people with that name.
-------
∂05-Jul-83 1804 LLW@S1-A Tomorrow
Received: from S1-A by SU-AI with TCP/SMTP; 5 Jul 83 18:04:06 PDT
Date: 05 Jul 83 1756 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Tomorrow
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
John, I've rather unexpectedly been dragged off to Phoenix tonite,
and will be there all day tomorrow. I may be back here on Thursday,
or I may be stuck in San Diego. Sorry for the change in plans.
Rod will be here at the Lab both days. Lowell
I'm loaded with things to do here, so maybe I'd better wait till you're
available, but often I can come on one day notice. Weekends are ok too,
and the probability of making it on no notice is higher. You might
look at the file FESTCH[SHA,JMC] at S1 to see how far I got last week.
∂05-Jul-83 2321 HST inivitation
that's ok now.we fix it for the days after ijcai.you did not say
anything about hotel etc.is a bad in our flat enough?
moreover:it may be interesting for you that i gave a very philosophi-
cal lecture on "problems of ai" last winter here.my conclusion was
that your statement(what the philosophers said is uninteresting be←
cause they did not come to a common standpoint after 2000 years)is
not a good advice because the philosophical differences are not in-
teresting for us at all.the rest of their ideas is important.
(i hope you translate the "bad" above to "bed".
A bed in the flat will be fine. I haven't understood whether your
lecture attributes to me the view that what the philosophers said
is uninteresting or the view that the rest of their ideas is
important. Neither exactly is my view. I think that some of their
current work may be helpful for AI, suitably understood and translated.
But we will have time to discuss this in person.
∂06-Jul-83 0302 RPG
∂05-Jul-83 2316 JMC Common Lisp on 68000
I am surprised at your statement that CL is infeasible on a 68000.
Do you have in mind a configuration with a specific limit of memory
and disk or do you think that the architecture itself makes Common
Lisp infeasible?
To test types (for runtime-certified code or for generic arithmetic)
requires several instructions to extract the type bits. And to
use a pointer (such as a CONS cell pointer) as an address means you
have to flush type bits (hard trap otherwise). So, to do a CAR you
move to a data register, flush the type bits, move to an address register,
and then indirect through the register. The distinction between data and
address registers requires a lot of moves. Common Lisp code requires more
type-certification than most other Lisps. Also, the hairy array types require
a similar high degree of shuffling between registers.
My note said that a full Common Lisp is not practical, and I meant that
Lisp programmers on 68000-based computers would not want to use the full
Common Lisp if they were concerned about performance. Another consideration
is that if every Common Lisp function name were in an obarray, that obarray
would be huge.
I think, though, that a Lisp that played fast and loose with type-certification
would run great on, say, a SUN. If we could get a modified 68000 that,
when it moved a pointer into the registers, moved the address field into the
address registers somewhere and the type field into the data registers,
we'd be in fine shape.
Therefore, it is the architecture that I worry about more than some
specific limited configuration.
-rpg-
∂06-Jul-83 1040 CLT
i have a form that needs your signature before i can
register. when will you be in??
∂06-Jul-83 1200 GENESERETH@SUMEX-AIM discouraging words
Received: from SUMEX-AIM by SU-AI with PUP; 06-Jul-83 12:00 PDT
Date: Wed 6 Jul 83 12:00:26-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
Subject: discouraging words
To: jmc@SAIL
John,
Last week I got the word that Kennedy's schedule is full at the time
of the conference, and so he won't be able to make it. My feeling is that
if we can't get a high profile person like Kennedy, then we shouldn't ask
a politician. Therefore, the possibilities are either Cooper/Kahn or
Frank Press. I am intrigued with the idea of having Press as a speaker.
Do you know him? Would you be willing to invite him?
mrg
-------
I don't know him, but I'm willing to telephone. I'll
call Minsky, who might know him, since Press was at M.I.T. and could
call in his role as a past President of AAAI if this seems warranted.
Minsky will be back in an hour.
334-2000, inman
∂06-Jul-83 1422 GARDNER@SU-SCORE.ARPA Ph.D. oral
Received: from SU-SCORE by SU-AI with TCP/SMTP; 6 Jul 83 14:22:30 PDT
Date: Wed 6 Jul 83 14:16:23-PDT
From: Anne Gardner <GARDNER@SU-SCORE.ARPA>
Subject: Ph.D. oral
To: jmc@SU-AI.ARPA
Would you be interested in being a member of my orals committee?
The date has already been arranged for Friday afternoon, July 15.
I apologize for the lateness of the invitation. There was some confusion
about how many members of the reading committee would be present, and
now it turns out we're short one person.
--Anne
-------
∂06-Jul-83 1547 MAYR@SU-SCORE.ARPA Supercomputer seminar
Received: from SU-SCORE by SU-AI with TCP/SMTP; 6 Jul 83 15:47:21 PDT
Date: Wed 6 Jul 83 15:44:27-PDT
From: Ernst W. Mayr <MAYR@SU-SCORE.ARPA>
Subject: Supercomputer seminar
To: su-bboards@SU-SCORE.ARPA
cc: super@SU-SCORE.ARPA
Due to popular demand, I am thinking of setting up a reading seminar on
"Theoretical aspects of massively parallel computation".
The purpose of this seminar is to discuss research papers about
-parallel machine models,
-parallel algorithms and lower bounds,
-parallel architectures,
-parallel programming languages,
-parallel whatelsecanyouthinkof.
This seminar could start in the fall quarter or even next week. In order for
me to find out how big the popular demand really is, let me know (by the
beginning of next week) whether you would take part in this seminar during the
summer or in the fall, and give me some idea about possible times. Due to
other restrictions only Monday or Friday seem to be possible.
-Ernst
-------
∂07-Jul-83 0938 GROSZ@SRI-AI.ARPA [Kjb at SRI-AI: Meeting]
Received: from SRI-AI by SU-AI with TCP/SMTP; 7 Jul 83 09:38:49 PDT
Date: Thu 7 Jul 83 09:39:33-PDT
From: GROSZ@SRI-AI.ARPA
Subject: [Kjb at SRI-AI: Meeting]
To: jmc@SU-AI.ARPA
John--
You inadvertantly got left off the list of addresses.
Barbara
---------------
Mail-From: KJB created at 6-Jul-83 11:49:49
Date: 6 Jul 1983 1149-PDT
From: Kjb at SRI-AI
Subject: Meeting
To: kjb at SRI-AI, BMoore at SRI-AI, Bresnan at PARC,
To: Briansmith at PARC, grosz at SRI-AI, JRP at SRI-AI,
To: Kaplan at PARC, peters at UTEXAS,
To: sag%SU-psych at SUMEX-AIM, Wasow%SU-psych at SUMEX-AIM,
To: winograd at PARC, rosenschein at SRI-AI, lauri at SRI-AI
We need to have a strategy session MOnday$$λ$λ$h$honday morining to prepare
for Thursday's meeting. We will meet here at Ventura at
9 a.m. Those of you who tend to be late should come to an
8:30 a.m. meeting.
As I understand it, Thursday's schedule has us giving presentations
in the morning designed to convince the board
a) that we are serious about the integration,
b) that we have the management issues sorted out.
At 12:30- the board will go off to lunch with Kennedy and people
from SRI and PARC. Then they will meet by themselves. So it is
the morning we need to plan, with (a) and (b) in mind. John, Rich
Dutcher and I are thinking about (b). You should all think about (a)
and come with good ideas.
Jon
-------
-------
∂07-Jul-83 1003 ARK Facilities Committee Mtg
Could you please try to schedule a meeting for Monday or Tuesday July 11-12.
Thanks.
Arthur
∂07-Jul-83 1122 DFH
To: "@FACCOM.[1,DFH]"@SU-AI
To: Facilities Committee
From: Diana Hall (dfh@su-ai)
Re: Facilities Committee Meeting
JMC would like to schedule a facilities committee meeting on Monday or Tuesday,
July 11 or 12. Please let me know what times you could attend on these days.
Thank you.
∂07-Jul-83 1621 JJW@S1-A
Received: from S1-A by SU-AI with TCP/SMTP; 7 Jul 83 16:20:52 PDT
Date: 07 Jul 83 1619 PDT
From: Joe Weening <JJW@S1-A>
To: jmc@SU-AI
I'm planning to come in to Stanford tomorrow instead of Livermore.
Do you want to spend some time discussing operating systems? The
afternoon would be best for me.
Joe
After 2pm I expect to be in.
∂07-Jul-83 2057 RPG Good news
To: JMC@SU-AI, pjb@S1-A
NSF has decided to pay $700 of my trip to IJCAI!
-rpg-
∂07-Jul-83 2120 GROSZ@SRI-AI.ARPA [Kjb at SRI-AI: r.s.v.p.]
Received: from SRI-AI by SU-AI with TCP/SMTP; 7 Jul 83 21:20:49 PDT
Date: Thu 7 Jul 83 18:33:02-PDT
From: GROSZ@SRI-AI.ARPA
Subject: [Kjb at SRI-AI: r.s.v.p.]
To: rperrault@SRI-AI.ARPA, jmc@SU-AI.ARPA
Please reply to Jon.
---------------
Mail-From: KJB created at 7-Jul-83 16:47:35
Date: 7 Jul 1983 1647-PDT
From: Kjb at SRI-AI
Subject: r.s.v.p.
To: kjb at SRI-AI, BMoore at SRI-AI, Bresnan at PARC,
To: Briansmith at PARC, grosz at SRI-AI, JRP at SRI-AI,
To: Kaplan at PARC, peters at UTEXAS, sag%SU-psych at SUMEX-AIM,
To: Wasow%SU-psych at SUMEX-AIM, winograd at PARC,
To: rosenschein at SRI-AI, lauri at SRI-AI
Would you r.s.v.p. to the previous message about
the meeting on monday, just so I will be sure that
everyone got it. Ivan is out of town but Tom is
here, so that should be o.k. Stan is coming.
Jon
-------
-------
∂08-Jul-83 0949 DFH Facilities Committee Meeting
About the best time I can come up with is 1 pm
on Monday, July 11. Gorin is out of town both
days. Of the people who responded, this would
leave out Hennessy and Cheriton. You would have
ARK, Mogul, Bosack, Rindfleisch, Golub, and
probably Mayr. Before sending out this time I
wanted to make sure you would be finished with
what you have scheduled for Monday morning by
1 p.m.
∂08-Jul-83 1156 DFH
To: "@FACCOM.[1,DFH]"@SU-AI
To: Facilities Committee
From: Diana Hall (dfh@su-ai)
Re: Facilities Committee Meeting
Based on the responses I've received to this point, the best time I can come up
with for the meeting is Monday, July 11 at 1 p.m. I have reserved MJH 252.
Apologies to those whose schedules I could not accommodate.
∂08-Jul-83 1235 PACK@SU-SCORE.ARPA Meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 8 Jul 83 12:34:48 PDT
Date: Fri 8 Jul 83 12:36:43-PDT
From: Leslie E. Pack <PACK@SU-SCORE.ARPA>
Subject: Meeting
To: jmc@SU-AI.ARPA
You asked me to come in and talk to you today, but fingering you, I
see you're at home. Will you be in the department this afternoon, or
would it be easier for us to meet next week?
Leslie
-------
I'm in, and if you have some ideas I'll be glad to listen, but I am
more likely to have suggestions if we meet next week.
∂08-Jul-83 1323 NILSSON@SRI-AI.ARPA no meeting
Received: from SRI-AI by SU-AI with TCP/SMTP; 8 Jul 83 13:22:52 PDT
Date: 8 Jul 1983 1319-PDT
From: Nilsson at SRI-AI
Subject: no meeting
To: jmc at SU-AI
cc: nilsson
John, Too many of the "right people" will be away during the next
few weeks for us to schedule you to give a talk here at SRI on
CBCL. Let's postpone it until sometime in September. I'll try to
arrange something after returning from AAAI. -Nils
-------
In that case, the CBCL talk should be in the first two weeks of October.
∂08-Jul-83 1816 GOLUB@SU-SCORE.ARPA Re: director of departmental computer facilities
Received: from SU-SCORE by SU-AI with TCP/SMTP; 8 Jul 83 18:16:21 PDT
Date: Fri 8 Jul 83 18:17:26-PDT
From: Gene Golub <GOLUB@SU-SCORE.ARPA>
Subject: Re: director of departmental computer facilities
To: HEDRICK@RUTGERS.ARPA
In-Reply-To: Message from "Charles Hedrick <HEDRICK@RUTGERS.ARPA>" of Fri 8 Jul 83 19:12:15-PDT
This is to be an honest search. John McCarthy is head of the search
committee and has no fixed candidate.
Yes, you can send your resume by ARPA mail.
GENE GOLUB
-------
∂08-Jul-83 1816 GOLUB@SU-SCORE.ARPA [Charles Hedrick <HEDRICK@RUTGERS.ARPA>: director of departmental computer facilities]
Received: from SU-SCORE by SU-AI with TCP/SMTP; 8 Jul 83 18:16:45 PDT
Date: Fri 8 Jul 83 18:17:46-PDT
From: Gene Golub <GOLUB@SU-SCORE.ARPA>
Subject: [Charles Hedrick <HEDRICK@RUTGERS.ARPA>: director of departmental computer facilities]
To: jmc@SU-AI.ARPA
Return-Path: <HEDRICK@RUTGERS.ARPA>
Received: from RUTGERS.ARPA by SU-SCORE.ARPA with TCP; Fri 8 Jul 83 18:01:24-PDT
Date: 8 Jul 83 19:12:15 EDT
From: Charles Hedrick <HEDRICK@RUTGERS.ARPA>
Subject: director of departmental computer facilities
To: golub@SU-SCORE.ARPA
I noticed your announcement for this position. I think I am interested.
How serious is the search? I had not looked into this because I assumed
that Len was going to get the position. For various reasons I would
prefer not to go out to California to talk if it turned out that they
really intended to hire Len and were just going through the motions. (I
know well what strange things Affirmative Action can make you do, having
done a certain amount of hiring myself.)
Would you accept my resume via Arpanet mail?
-------
-------
∂11-Jul-83 1039 GOLUB@SU-SCORE.ARPA Tandem
Received: from SU-SCORE by SU-AI with TCP/SMTP; 11 Jul 83 10:39:32 PDT
Date: Mon 11 Jul 83 10:43:30-PDT
From: Gene Golub <GOLUB@SU-SCORE.ARPA>
Subject: Tandem
To: jmc@SU-AI.ARPA
cc: gorin@SU-SCORE.ARPA, cheriton@SU-SCORE.ARPA
Tandem has sent me a letter asking if we are interested in any of their
I received a notice that Tandem is willing to give some equipment away.
Are we interested?
GENE
-------
∂11-Jul-83 1214 MAYR@SU-SCORE.ARPA Supercomputer seminar
Received: from SU-SCORE by SU-AI with TCP/SMTP; 11 Jul 83 12:14:33 PDT
Date: Mon 11 Jul 83 11:32:39-PDT
From: Ernst W. Mayr <MAYR@SU-SCORE.ARPA>
Subject: Supercomputer seminar
To: su-bboards@SU-SCORE.ARPA, super@SU-SCORE.ARPA
It seems to me there is enough interest in it over the summer, so we
shall start the (reading) seminar on
"Theoretical aspects of massively parallel computation"
right away. It will meet on a regular basis on Fridays, at 1:15pm, in
MJ352. The constituting meeting is coming Friday, July 15.
The purpose of the seminar is getting us theoretically ready for
supercomputers (the sixth generation + or - one), presenting research
papers from the literature and/or results of our own in anything which
has some relation to the title of the seminar.
P.S.: I'll be maintaining a mailing list for interested parties. So if you
haven't responded yet but are nonetheless interested (which is a slight
possibility) send me a message. I shall ask the systems management in time
to set up an alias for our group.
I am accepting voluntary applications for giving presentations!
-Ernst
-------
∂11-Jul-83 1425 RPG Wieneke
To: DFH
CC: JMC
The new Common Lisp contract from ARPA (that's to pay for Wieneke) has money
for travel. I want Wieneke to go to AAAI, since Common Lisp things will be brewing.
Diana, can you book him a flight, arriving the night before and leaving the evening
of the last day? I think the days are August 22 through August 26. He has a room
booked already, but he'll need registration paid for,
-rpg-
∂11-Jul-83 1504 SHARON@SU-SCORE.ARPA phone message
Received: from SU-SCORE by SU-AI with TCP/SMTP; 11 Jul 83 15:04:35 PDT
Date: Mon 11 Jul 83 15:02:17-PDT
From: Sharon Bergman <SHARON@SU-SCORE.ARPA>
Subject: phone message
To: mccarthy@SU-SCORE.ARPA
Your daughter, Sarah, called. Please call her back.
Sharon
-------
∂11-Jul-83 1514 Hewitt%MIT-OZ@MIT-MC thanks for your reprint
Received: from SU-DSN by SU-AI with PUP; 11-Jul-83 15:08 PDT
Received: From MIT-MC by SU-DSN.ARPA; Mon Jul 11 15:09:14 1983
Date: Monday, 11 July 1983, 18:01-EDT
From: Carl Hewitt <Hewitt%MIT-OZ@MIT-MC>
Subject: thanks for your reprint
To: JMC%su-ai@su-dsn
Cc: Hewitt%MIT-OZ@MIT-MC, dejong%MIT-OZ@MIT-MC
John,
Thank you very much for sending us a reprint
of your paper "The Common Business Communication Language".
It addresses some of the very same problems we are attacking.
I like the examples which you present; they are a good challenge.
We are using the description system Omega (see the references
in our open systems paper for further info) which allows a partial
description to be used wherever an identifier is allowed. Thus
our communication language already satisfies the important principle
which you attribute to Chomsky in your paper.
Cheers,
Carl
P.S. What is the proper citation for your paper? There is a lot
of German on the cover which you sent us which I cannot read.
I can't read it either. I suggest you cite it, after verifying meaning,
and remark that reprints are available from the author. It was given
at an IBM Symposium in Germany.
∂11-Jul-83 1833 DEK guibas
Have you had a chance to look at his papers yet?
I'm gung ho to keep the ball rolling on this before it's too late.
The papers are in Elyse's office, and I think I have a key to it if you
need access after hours.
∂11-Jul-83 2227 ME gone till end of July
∂11-Jul-83 1736 JMC July or August?
Your note in the log says you`ll be gone til July 31. Your PLAN says
August 31.
ME - Thanks, it's 31 July. I fixed the plan file (I was going to say
1 Aug....)
∂11-Jul-83 2341 cheriton@Diablo Tandem
Received: from SU-HNV by SU-AI with PUP; 11-Jul-83 23:40 PDT
Date: Mon, 11 Jul 83 23:44 PDT
From: David Cheriton <cheriton@Diablo>
Subject: Tandem
To: golub@score
Cc: gorin@score, jmc@sail
I am not very enthused about Tandem equipment because it is 16-bit
and thus rather limited and archaic for work here. (One of my major
battles with DEC was getting 32-bit rather than 16-bit equipment.)
Since it is primarily oriented towards database applications,
maybe Gio would be interested?
∂12-Jul-83 0832 DFH Final technical report
To: JMC
CC: RPG
Marlie got a call yesterday from ONR asking about this
again. This is for the old ARPA contract, MDA903-80-C-0102,
which Binford and Lantz were also on. Betty had asked for
this while you were away, and RPG had started to look into
it. In any event, it still needs to be done.
∂12-Jul-83 0925 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #13
Received: from SU-SCORE by SU-AI with TCP/SMTP; 12 Jul 83 09:25:50 PDT
Date: Sunday, July 10, 1983 5:36PM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #13
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Tuesday, 12 Jul 1983 Volume 1 : Issue 13
Today's Topics:
Representation - Declaring Predicates Transitive,
Implementations - Performance In Programs,
Prolog Availability,
LP Library - Trace Routine
----------------------------------------------------------------------
Date: Thu 7 Jul 83 17:58:09-PDT
From: Vivek Sarkar <JLH.Vivek@SU-SIERRA>
Subject: More On Declaring Predicates Transitive
Having only read the initial issues of the Digest a few days ago, I
hope that I may be forgiven for bringing up an old problem:
The problem was to come up with a predicate, transitive(R) which
checks if R is a transitive (binary) relation.
This is what my solution looks like:
transitive( R ) :- not( nontransitive(R) ).
nontransitive( R ) :- R(A, B), R(B, C), not( R(A, C) ) .
The reason that this kind of problem arises is because a Prolog goal
with uninstantiated variables is like a predicate with an
*existential* ("there exists ...") quantifier. I.e. p(X) is true if
there exists some value a, such that p(a) is true (this is only in the
case that X is uninstantiated to start with). So when we approach
this transitivity problem we may think of
transitive( R ) :- R(A, B), R(B, C), R(A, C).
Then we pause because what we really want to say is "for all possible
A, B & C". But what this says is that R is transitive if we can find
any one set of values for A,B & C such that the above RHS holds. We
want to use a *universal* ("for all ...") quantifier for a goal, but
Prolog goals only lend themselves to existential quantifiers. This is
where the "not" predicate comes into play. From predicate calculus we
know that a universal quantifier can be written as the negation of an
existential quantifier, and so we can expect the "not" predicate to
help us. It turns out to be straightforward to write a predicate for
non-transitivity, since it's an existential assertion. The predicate
for transitivity then follows by a simple negation.
This issue of existential vs. universal quantifiers is a bit of an
anomaly in Prolog. The semantics of variables which occur in the LHS
(or head) of a clause is that they are bound by universal quantifiers.
I.e.
p(X) :- q(X), r(X).
is a statement about *all* X. But if the variable occurs only in the
RHS, then it is effectively bound by an quantifier. I.e.
p(X) :- q(X), r(X, Y).
is a statement which holds for *all* X and *some* Y.
A final note on the solution:
Since most Prolog implementations do not allow a variable functor in a
goal (I.e. R(A, B), where R is a variable), =.. & call have to be used
to actually get it to work:
transitive( R ) :- nott( nontransitive(R) ) .
nontransitive( R ) :- X =.. [R, A, B], call( X ),
Y =.. [R, B, C], call( Y ),
Z =.. [R, A, C], NZ =.. [nott, Z], call( NZ ) .
nott( P ) :- call( P ), !, fail. nott( ← ) .
I have tried this on the DEC-10 Prolog interpreter, and found that it
works. You may notice that I have defined my on negation predicate
"nott"; for some reason, the standard "not" didn't work correctly.
( Even '' ?- not(fail). '' was not satisfied! )
-- Vivek Sarkar
------------------------------
Date: Thu 7 Jul 83 19:37:44-EDT
From: STEVE@COLUMBIA-20
Subject: Performance In Prolog Programs
I would like to do some statistical analysis on large Prolog programs.
I am particularly interested in AI programs in the following areas:
1) Expert Systems
2) Data Bases,
3) Planning or Robotics,
4) NLP
Can anyone provide sample programs that I can use? They should be
large programs that run on Edinburgh Prolog 3.47 (Dec-20) or C-Prolog
1.2 (Unix 4.1/Vax). I would like to collect a good variety, so any
programs will be useful. I would also appreciate a sample journal of
a session with the program so that it can be exercised quickly and
effectively.
Many Thanks,
--Stephen Taylor
------------------------------
Date: 2 Jul 83 13:11:36 EDT (Sat)
From: Bruce T. Smith <BTS.UNC@UDel-Relay>
Subject: How Many Prologs Are There ?
Here's Randy Harr's latest list of Prolog systems. He's away
from CWRU for the summer, and he asked me to keep up the list for him.
Since there have been several requests for information on finding a
Prolog lately, I've recently submitted it to net.lang.prolog. The
info on MU-Prolog is the only thing I've added this summer, from a
recent mailing from the U. of Melbourne. (Now, if I could only find
$100, I would like to try it...)
--Bruce T. Smith, UNC-CH
duke!unc!bts (USENET)
bts.unc@udel-relay (lesser NETworks)
list compiled by: Randolph E. Harr
Case Western Reserve University
decvax!cwruecmp!harr
harr.Case@UDEL-RELAY
{ the list can be FTP'd as [SU-SCORE]PS:<PROLOG>Prolog.Availability.
SU-SCORE observes Anonymous Login convention. If you cannot FTP,
I have a limited number of hard copies I could mail. -ed }
------------------------------
Date: 8 July 1983 1344-PDT (Friday)
From: Foonberg at AEROSPACE (Alan Foonberg)
Subject: Trace Routine
Trace - July 8, 1983 By Russ Abbott and Allen Foonberg
Included is a trace routine which may be useful to users debugging
Prolog programs. If you have any questions, please contact Alan
Foonberg (Foonberg@Aerospace) or Russ Abbott (Abbott@Aerospace).
This is a trace program similar to that provided by the Prolog system.
A term's execution is traced step-by-step by typing "trace(term).",
where term is that term to be traced. At each step of the trace, you
have an option to enter a different trace mode. The list of modes can
be obtained by answering this question with a '?'. A typical trace
line looks like this:
2. (3) enter b(Var, constant)
The number before the period indicates the count, the number in
parenthesis is the depth, the following phrase indicates the port, and
the rest is the current term being traced. To obtain the full trace
as a parameter, type "trace(term, Full←Trace)." Full←Trace will
contain a list of the trace lines, which can be printed out formatted
by typing "trace$←write←term(trace←list(Full←Trace)." A 'true trace'
is also provided by adding a third parameter to the trace call, and
this can be printed in a manner similar to that used to print the full
trace. The true trace contains only those calls that succeeded which
led to the satsification of the main goal. If you have any questions
or suggestions, please send network mail to Foonberg@Aerospace or
Abbott@Aerospace.
[ Trace can be FTP'd as {SU-SCORE}PS:<PROLOG>Trace.Pl or
{SRI-AI}PS:<PROLOG>Trace.Pl
Note: The syntax used is only valid in C-Prolog, not in
DEC-10/20 Prolog or other available Prologs. The problem
is that C-Prolog accepts $ as an alphabetic character. -ed ]
------------------------------
End of PROLOG Digest
********************
∂12-Jul-83 1002 GARDNER@SU-SCORE.ARPA Erratum
Received: from SU-SCORE by SU-AI with TCP/SMTP; 12 Jul 83 10:00:56 PDT
Date: Tue 12 Jul 83 10:04:59-PDT
From: Anne Gardner <GARDNER@SU-SCORE.ARPA>
Subject: Erratum
To: Winograd@PARC-MAXC.ARPA, Buchanan@SUMEX-AIM.ARPA, Lenat@SU-SCORE.ARPA,
Golub@SU-SCORE.ARPA, jmc@SU-AI.ARPA
cc: gardner@SU-SCORE.ARPA
I just found a confusing mistake in the draft I gave you yesterday.
On page 110, the second full paragraph should begin,
The program finds that, if the SECOND rule is followed . . .
--Anne
-------
∂12-Jul-83 1030 KJB@SRI-AI.ARPA Thursday
Received: from SRI-AI by SU-AI with TCP/SMTP; 12 Jul 83 10:30:39 PDT
Date: 12 Jul 1983 1004-PDT
From: Kjb at SRI-AI
Subject: Thursday
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI,
To: Kaplan at PARC, kjb at SRI-AI, lauri at SRI-AI,
To: peters at UTEXAS, rperrault at SRI-AI,
To: sag%SU-psych at SUMEX-AIM, stan at SRI-AI,
To: Wasow%SU-psych at SUMEX-AIM, winograd at PARC
cc: kjb
Dear all,
Thanks for being so helpful at the meeting yesterday. I hope
all our meetings can go so well. And thanks for being on time.
The sherry hour will be in the President Dining room, and
adjacent patio, at the faculty club. It will run from 11:30-12:30.
This is the same room where the lunch will take place, so be alert to
when you should leave, presumably at 12:30. You should come a little early
to find your name tag. Please wear it, so the board members will know
who they are talking to.
Finally, please read over the proposal before Thursday. It
would be a disaster for some of us to appear not to know what is in it.
Thanks for your help. Jon
-------
∂12-Jul-83 1033 PACK@SU-SCORE.ARPA meeting, natural kinds
Received: from SU-SCORE by SU-AI with TCP/SMTP; 12 Jul 83 10:33:13 PDT
Date: Tue 12 Jul 83 10:37:38-PDT
From: Leslie E. Pack <PACK@SU-SCORE.ARPA>
Subject: meeting, natural kinds
To: jmc@SU-AI.ARPA
I stopped by your office around 5 yesterday, but I guess you had
already gone home. When is a good time to reach you?
I have been thinking more about natural kinds. I kind of like the
idea of paradigms for separating natural kinds. Every object we see
(or hear of, etc) can be represented as a point in multi-dimensional
object space. As we develop, clusters can be seen to be developing in
this space. We find the 'center of mass' of each of these clusters
and call that the paradigm for that particular kind. (The paradigm
may well not exist). This is not immune to all of the arguments about
where kinds intersect, etc., but seems more intuitive than the idea of
lists of necessary characteristics. I guess the representation you
might have is the equation of the 'sphere' defining the kind -- i.e.
size value between a and b; color value between a1 and b1; etc for
each dimension in the space.
Leslie
-------
The conceptual utility of natural kinds depends on the fact that
they are natural. It works best to consider children. A kind,
e.g. horse, is identified by those properties that are observed
the first time the kind is encountered, whether it is directly
perceived or merely talked about. The key thing is that the kind
has additional, yet to be discovered, properties. Everyone, but
a child most clearly, relies on this fact. For example, a child
sees his first horse, develops a subconscious visual ability to
recognize horses based on shape features that he has already
learned are more stable than markings. He is then prepared
to learn other facts about horses, e.g. that they eat hay and
grain, that they live in pastures and stables, and that they
can be ridden or used to pull carts. None of these are in any
way consequences of the appearance of horses. When the kind is not
discrete, e.g. a pond as compared to a lake or a puddle, a child
has more difficulty.
As for five yesterday, I think I had just gone to the
coffee shop or something like that. I'll be around all day
today except for such expeditions.
It would be useful if you would MAIL me a bibliography
of three or four items characteristic of the recent philosophical
interest in natural kinds. I believe that the considerations
I mentioned above are the most important ones for the AI interest
in natural kinds, but I'm not sure they are the ones the philosophers
have most emphasized.
∂12-Jul-83 1100 JMC*
Guibas papers.
∂12-Jul-83 1127 KJB@SRI-AI.ARPA Update
Received: from SRI-AI by SU-AI with TCP/SMTP; 12 Jul 83 11:26:35 PDT
Date: 12 Jul 1983 1115-PDT
From: Kjb at SRI-AI
Subject: Update
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI, Kaplan at PARC,
To: kjb at SRI-AI, lauri at SRI-AI, peters at UTEXAS,
To: rperrault at SRI-AI, sag%SU-psych at SUMEX-AIM,
To: stan at SRI-AI, Wasow%SU-psych at SUMEX-AIM,
To: winograd at PARC
More on Thursday:
1) We have just learned that the SDF lawyers need to talk to the
board of SDF in the morning, which is going to move our schedule
back somewhat. The sherry hour probably will not start till 11:45
or 12. Also, after that, there is now a lunch at the faculty
club scheduled for other sdf personnel and all of you. Please try
to stay for this, and talk with the sdf people at the lunch.
2) We understand that the board will perceive that the following
dangers exist:
a) that they will be seen as supporting one side of a
factional dispute among linguists
b) that CSLI will fail to adequately control quality
c) that CSLI will fail to develop momentum
d) that instead of being synergistic, CSLI will bring
all participants down to a mediocre level
e) that the ideas might not lead to fruitful research
3) The issue the board wants to get clear about are:
a)quality of vision
b) quality of personnel
c) adequacy of leadership (sigh)
d) commitment of personnel
e) commitment of institutions
f) proper appreciation of scope of project
e) credibility of management structure
Anything you can do to reassure them about any of this will be
appreciated.
Jon
-------
∂13-Jul-83 1408 RPG
To: ullman@SU-SHASTA
CC: JMC@SU-AI
Multiprocessing Lisp
Jeff,
About your comment on n↑2 messages: The situation I envision is that you
want to search a space for some best attribute, perhaps for a shortest
path. We do not want to have too much message traffic at the start
of the search process, but when a good idea of the space is obtained,
message traffic to stifle long, possibly unprofitable searchs is tolerable.
To accomplish this, we will spawn a total of n+1 tasks, where n is the
number that would be spawned naturally for the problem, and one is a
controller task.
We want to search happily along until some fraction of the
processes finish with their own searches. They report their result to the
controller. Before each process that has completed commits suicide via
THROW, it updates a global variable containing the number of currently
active processes. This variable is incremented when a new process is
spawned. Each process has code that checks whether its best expected result
is better than a globally posted result, but this code is not enabled
during the initial part of the search.
When a suitable fraction of the processes have completed their part of the
search, the controller process posts the best result thus far and each
process can check to see how it is doing, killing itself early if necessary.
I think John McCarthy does not like message-passing as a control-structure.
There are, then, two ways to implement this search strategy in Lisp.
If we allow interrupts and message-passing, then when each process
finishes it sends a message to the controller. When enough processes die
off, the controller sends a message to the remainder to check with the
currently-posted best result. If some process does not expect to better
that result, it THROWs to suicide. As each process returns a new result to
the controller, the controller can then send messages to the remainder to
check once more. After they are all finished, or after a time limit has
been reached, the controller kills itself (and the remainder), and a final
result is determined.
We can use the code:
((QLAMBDA QLET ( () V1...Nn) ...) (CATCH 'FOO (CONTROLLER ..)) ...)
to activate the controller and to ignore the value (the "()" does this).
If we do not allow interrupts or message-passing, then there can be two
variables: one says whether each process should compare its best expected
result to the best so far, and the other is the best so far. The
controller again watches the count of active processes to decide when to
set the first to T. We might wish to disallow the check early in the
search because it is expensive to execute. Later it might be too expensive
to continue long searches.
If we take the QLET variable seriously, then there is a problem with
interrupts and message-passing. Not all logical processes have corresponding
actual processes. If we have a QLAMBDA, then perhaps the arguments are
being evaluated sequentially and there is no process that corresponds to
the nth argument to which a message can be sent. In this case it isn't
obvious what the controller's role is except to `kill itself' immediately.
The code for each `process' becomes a bit unpleasant if it must take into
account the possibility that QLET is () at its activation time. In essence,
the controller code is spread amongst the individual search `processes.' The
difficulty of understanding code written this way can be minimized either
by adopting a continuation-passing style of programming or by using
a closure to implement the controller. In the latter case, we write a piece
of code which has local storage, which is passed as an argument to
each process, and which executes the controller's activities when applied.
Each process can apply this closure periodically.
As I sit here it occurs to me that this might be the beginning of a good solution
to several problems. The syntax for the creation of a closure is:
(lambda (x) <body>)
So if you say:
(defun create (x)
(lambda (y)(+ x y)))
then (create 7) returns a function (closure) that when given a numeric argument,
adds 7 to it.
Suppose we extend the closure-creating syntax as with QLAMBDA:
(lambda qlet (x) ...)
which creates a simple closure if QLET is (), and which creates a closure
that exists as a separate process if QLET is not (). Consider the code:
(defun addemup (l)
(let ((result 0))
(flet ((adder
(lambda qlet (mess)
(cond ((null mess)
result)
(t (setq result (+ mess result)))))))
(do ((l l (cdr l)))
((null l) (adder ()))
(adder (car l))))))
LET is a macro that becoms a lambda. (LET ((A B)(C D)) ...) is just
((LAMBDA (A c) ...) B D). FLET is a lambda-binder that binds functions to
variables. DO is a loop mechanism. The DO above binds a new variable, L,
to the value of the old L, it then CDRs down the list until the new L is
(). After each failing test of (NULL L), it does (ADDER (CAR L)). When
(NULL L) is true, it does (ADDER ()).
Now, suppose that QLET is (). Then the function that is bound to ADDER is
a simple closure. The expression (ADDER (CAR L)) just adds (CAR L) to RESULT.
(ADDER ()) returns RESULT (in a hackish way).
The behavior of (ADDER (CAR L)) when QLET was not () is that the
application of the closure sends the values of the arguments to the
process that is running the closure. The closure is then run. If the value
of the application is thrown away by the code applying the closure, the
process running the applying code does not wait for the value. If the code
applying the closure binds or assigns the value returned, it waits.
The code:
(let ((n (adder x))) ...)
adds X to RESULT, and since ADDER returns the current value of RESULT, N is bound
to this value. The process running this code (the (let ((n (adder x))) ...))
waits for ADDER to return. If you write:
(progn (adder x) (baz x))
X is sent to ADDER, and control immediately passes to (baz x).
In the ADDEMUP example, we send the next value to ADDER at (ADDER (CAR L)),
and then proceed through the loop while the addition is going on. The
hope is that ADDER finishes about the same time that the loop code is
ready to send the next value to add.
In the search algorithm above, instead of writing:
((QLAMBDA QLET ( () ...) ...) (CATCH 'FOO (CONTROLLER)) ...)
with horrible message-passing and conditional code all over the place that
worries about the value of QLET, you write:
(FLET ((CONTROLLER (LAMBDA QLET ...)))
((QLAMBDA QLET (V1 ... Vn) ...)
(CATCH 'SEARCH (SEARCH A1 CONTROLLER)) ...))
In each SEARCH process there is code that looks like closure-applications, and
the code works whether QLET is T or ().
The upshot is that I think we can use the following primitives for a
simple multiprocessing capability in Lisp:
QLET as a global variable that allows or disallows true multiprocessing
ala JMC.
(QLAMBDA v ...) that spawns parallel processes or queues sequential
processes depending on the value v.
(LAMBDA v ...) that creates closures in separate processes or creates
a simple closure depending on v. We may need a name other LAMBDA for this,
perhaps.
Closure application is either function application in the simple case
or message-passing in the multiprocessing case.
Closure application is asynchronous in the value-ignoring context and
is a synchronizer in the value-binding context.
CATCH/THROW is a non-local exit in the simple case and is a process
killing, non-local exit in the multiprocessing case.
I think this extends JMC's ideas a bit in the Hewitt direction without
committing entirely to a true multiprocessed situation in all cases. That is,
if we decide to run a program with no multiprocessing, the same code works,
the code looks good, and there isn't a lot of user-written control hair that
worries about whether multiprocessing is going on.
-rpg-
∂13-Jul-83 1707 DFH
Re: Travel Arrangements
Both TWA and Pan Am have similar schedules to Frankfurt. You would leave SF on
the morning of Aug. 6 and arrive in Frankfurt the morning of Aug. 7. The return
would get you into Boston late afternoon-
early evening on Tues. Aug. 16. On Aug 17, TWA has a Boston/SFO flight leaving
at 7:10 pm, arriving SFO 10:11 pm. The SFO/Frankfurt and Frankfurt/Boston
portions all change planes in NY.
The regular coach fare is sold out on Pan Am, but available on TWA. Only
business class is available on Pan Am, which will cost about $300 more.
The University travel rules appear to allow reimbursement for anything
less than first class, though I am double-checking this.
It appears to me from reading over the Pan Am World Pass rules that you
be able to get an upgrade to Clipper Class from coach on your London/Paris
trip. This is based on your flying 10,000 miles SFO/Frankfurt. Alternatively,
you could use the upgrade on your Brazil trip or on a domestic trip. If you
wait until after the Newcastle/Paris trip, you could get a 25% discount off
any domestic fare. (There are probably some variations on this).
In view of all this--should I try to use Pan Am, or should I take the
cheaper fare on TWA?
∂13-Jul-83 2222 ullman%SU-HNV.ARPA@SU-SCORE.ARPA meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 13 Jul 83 22:22:07 PDT
Received: from Diablo by Score with Pup; Wed 13 Jul 83 22:26:23-PDT
Date: Wed, 13 Jul 83 22:26 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: meeting
To: super@score
We meet as usual Thursday 3PM in the air conditioned CIS conference
room.
∂14-Jul-83 0533 @MIT-MC:Burns.boulder@RAND-RELAY
Received: from MIT-MC by SU-AI with TCP/SMTP; 14 Jul 83 05:32:31 PDT
Date: 14 Jul 83 1:58:26-MDT (Thu)
From: Burns.boulder@Rand-Relay
Return-Path: <Burns%boulder.boulder@Rand-Relay>
Received: from rand-relay.ARPA by udel-relay.ARPA ; 14 Jul 83 07:30:47 EDT (Thu)
To: McCarthy.MIT-Mc@Udel-Relay
Via: boulder; 14 Jul 83 4:08-PDT
did you ever write the 'advice taker'?
No. Various people have written programs along that line starting with
Fischer Black in 1964, but the epistemology has never been in good enough
shape for me to want to emphasize the program, especially since I'm better
at epistemology than at programming.
Incidentally, I'm at SU-AI. It is purely accidental that MC forwarded
the message.
I think you sent this to me by mistake.
∂14-Jul-83 1328 jouppi@Shasta MIPS & TV
Received: from SU-SHASTA by SU-AI with PUP; 14-Jul-83 13:26 PDT
Date: 14 July 1983 12:42:21-PDT (Thursday)
From: jouppi@Shasta
Subject: MIPS & TV
To: at , MCCARTHY, SIERRA
Cc: jouppi
I am currently verifying my results by comparisons with SPICE
and fab'ed performance of entire chips (e.g., Geometry Engine).
TV is oriented towards circuits with thousands of transitors,
hence the best test structures are VLSI chips themselves.
TV's models are calibrated by using SPICE parameters. If I
could get the SPICE parameters for the IC Lab run, as well as
the circuit for the ring oscillator (including all parasitics),
it should be enough.
Norm
∂15-Jul-83 1042 NILSSON@SRI-AI.ARPA interview
Received: from SRI-AI by SU-AI with TCP/SMTP; 15 Jul 83 10:34:07 PDT
Date: 15 Jul 1983 0913-PDT
From: Nilsson at SRI-AI
Subject: interview
To: jmc at SU-AI
cc: nilsson
John, A Mr. Alexander from Fortune Magazine is visiting me now.
He wants to write a book on AI based on his previous articles in
Fortune about AI. As you remember, they did emphasize the
Shank/Minsky point of view about AI, and I am telling him that he
ought to at least acknowledge in his book that there is a point
of view based more on the use of formal methods. He would like to
talk to you about all of this. Any chance that you could see him?
Call me at 859-2311 if you read this today. (He will be talking to
me until about 10:15 or so today.) -Nils
-------
∂15-Jul-83 1118 RINDFLEISCH@SUMEX-AIM Workstation Purchase Dilemma
Received: from SUMEX-AIM by SU-AI with PUP; 15-Jul-83 11:18 PDT
Date: Fri 15 Jul 83 11:18:21-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Workstation Purchase Dilemma
To: HPP-Exec@SUMEX-AIM, Yeager@SUMEX-AIM, Schmidt@SUMEX-AIM, JMC@SAIL, RPG@SAIL
cc: Rindfleisch@SUMEX-AIM
This msg is to solicit updated comments and strategy recommendations
for our Lisp Machine acquisition plans. I'm getting increasingly
nervous about the utility and cost-effectiveness of the current
incarnations of these systems.
Recent Events:
1) Prices are going up. Xerox welched on a discount deal so that the
effective price of Dandelions has jumped from $26.7K to $29.3K (no
sales tax). The effective price of Dorados goes from $120.4K up to
$123.0K.
I also just heard that Symbolics is reducing their discount schedule
so that our price for 3600's will go up from $68.8K to $80.2K. We
may be able to hold on to our current price for a while by
preordering a number of machines before the middle of August.
2) Performance problems. Our experience with Dolphins is that they
are very slow for most of our projects -- Ted Shortliffe has not
"hit the wall" yet. Much of this is due to memory limitations and
swapping overhead. Since the Dandelion is packaged so that it can
hold only 1.5 Mbytes, even though the machine is 25% faster, the
net effect will likely be very unsatisfactory with no relief path.
The 3600 is also not performing up to expectations. Current
benchmarks show it to be little faster than the LM-2 (as opposed
to the expected factor of 4 or so). This apparently comes from some
microcode problems (e.g., CONS is not microcoded as was true in the
early Dolphin releases). It is not clear how fast Symbolics will
fix this.
Thus, we have the following relative situation currently:
Dandelion:3600:Dorado
Price 1:2.7:4.2
Performance 1:2(?):6-8(?)
Some questions/problems are:
1) Is the Dandelion useful to anyone at this point with its apparent
dead-end memory limitations => should we buy any?
2) What are the implications of making more commitment to the 3600
(InterLisp conversion, faith in performance improvements, etc.)?
3) Even though the Dorado appears to be the most cost-effective per
cycle, it is expensive in terms of unit price and no serious
production facility exists at Xerox.
4) Our needs for more cycles are desperate.
5) We are faced with having to spend about $700K soon for bureaucratic
reasons.
I welcome any comments or proposed strategies.
Tom R.
-------
∂15-Jul-83 1127 DFH phone message
Lynn from Dr. Kim's office , Univ. of S. Florida
called again regarding the Korean Computer Science
Society conference. Could you phone Dr. Kim,
(813) 974-4184 until 3 pm our time, after that his
home phone is (813)985-5597. (Lynn's no. is (813)
974-4232).
∂15-Jul-83 1130 DFH TEX class
ARK is giving a beginning TEX tutorial this
afternoon 1 - 2:15 or so. I am planning to
go, unless you need me to do something at that
time.
∂15-Jul-83 1133 BUCHANAN@SUMEX-AIM Re: Workstation Purchase Dilemma
Received: from SUMEX-AIM by SU-AI with PUP; 15-Jul-83 11:32 PDT
Date: Fri 15 Jul 83 11:33:50-PDT
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM>
Subject: Re: Workstation Purchase Dilemma
To: Rindfleisch@SUMEX-AIM
cc: HPP-Exec@SUMEX-AIM, Yeager@SUMEX-AIM, Schmidt@SUMEX-AIM, JMC@SAIL, RPG@SAIL
In-Reply-To: Your message of Fri 15 Jul 83 11:19:09-PDT
I like the model of having a Dorado with satellite Dandelions.
It depends on the utility of small machines for development of
small parts of programs and depends on swift loading of many
files onto one machine or another. The former requires changing
work habits of programmers (not impossible if the price is not
too high). The latter is a technical question on which the
Dolphin currently would fail -- would the Dandelions & Dorado also?
I think we make better use of hardware if we do not limit ourselves
to single-(or dual)-user machines.
bgb
-------
∂15-Jul-83 1144 KJB@SRI-AI.ARPA CSLI
Received: from SRI-AI by SU-AI with TCP/SMTP; 15 Jul 83 11:44:10 PDT
Date: 15 Jul 1983 1125-PDT
From: Kjb at SRI-AI
Subject: CSLI
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI, Kaplan at PARC,
To: kjb at SRI-AI, lauri at SRI-AI, peters at UTEXAS,
To: rperrault at SRI-AI, sag%SU-psych at SUMEX-AIM,
To: stan at SRI-AI, Wasow%SU-psych at SUMEX-AIM,
To: winograd at PARC
Dear all, As you have all heard by now, things went well yesterday.
We have received funding of 15 million for 4 years. There will be
a review after year 1 to see that we are up to what we claimed and
have our act together. There will be another review at the end
of year 3 to determine what kind of funding will be provided after
year 4. This should all us to really put CSLI on the map and make
the vision we have had begin to become a reality.
Yesterday was very important, and many people helped make
it a success, including the secretaries here, Henry Organ, Dean
Chase, the provost, the president, William Miller from SRI and
Georege Pake. Oh, yes, and all of us principals. But important
as it was, in retrospect it was just the successful completion of
an enormous task, most of which was carried out by John Perry,
Barbara Grosz and Brian Smith. We all owe them an enormous amount
for all the work they have done over the past months.
NOw for the other side of things. Our orginal budget request
for the first 4 years was 19 million. I am meeting with Charlie
tomorrow to find out the rational for the new figures. I simply
have no information, only conjectures at present.
I will be back in Madison next week, but still in touch via
computer. I am very enthusiastic about what we can do here and want
to thank all of you for your support. It let me be very confident
with the board yesterday about my future job, which is part of
what was needed. I'm very sorry you all had to stand in the
sun for an hour, but it was important for you to be there, even if
it didn't feel like it.
Great things ahead. Jon
-------
∂15-Jul-83 1255 RPG Lisp Machines
To: hpp-exec@SUMEX-AIM, yeager@SUMEX-AIM, schmidt@SUMEX-AIM, JMC@SU-AI,
rindfleisch@SUMEX-AIM
Folks:
As some of you know, I have been benchmarking the various Lisps around, and the
message that Tom sent out about buying Lisp machines touches on performance.
I know a little about the comparative performances of the machines in question.
The problem is that in order to have gotten to the position of knowing these
facts, I have committed myself to non-disclosure agreements with Symbolics
and Xerox. Therefore I am certainly not willing to state my facts over netmail,
and I'm not sure to what extent I can talk about hard data to you in any forum
at the moment. Since I'm benchmarking under DARPA contract, then it may
be the case that the non-disclosure agreements are actually between Stanford
as a whole and these companies. In this case we can use my data to make an
informed decision.
Tom's comparative performance figures are not accurate. I think that if we are
to make any judgement based on performance that the degree of inaccuracy is
significant. I do not, on the other hand, believe performance to be a conclusive
consideration by any means.
A further note is that Symbolics has been distributing, locally, new CONS
microcode, and preliminary results show a good improvement.
I would be happy to discuss these issues in private individually or in a group.
-rpg-
∂15-Jul-83 1552 minker.umcp-cs@UDel-Relay IJCAI Panel on Non-Monotonic Reasoning
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 15 Jul 83 15:52:39 PDT
Date: 15 Jul 83 12:09:30 EDT (Fri)
From: JACK MINKER <minker.umcp-cs@UDel-Relay>
Return-Path: <minker.umcp-cs@UDel-Relay>
Subject: IJCAI Panel on Non-Monotonic Reasoning
To: "uw-beaver!ubc-vision!Reiter"@LBL-UNIX, Moore@SRI-AI, Mcdermott@YALE,
JMC@SU-AI
Cc: BUNDY@RUTGERS, Amarel@RUTGERS
Via: UMCP-CS; 15 Jul 83 18:12-EDT
I have discussed our IJCAI session on non-monotonic logic
with Ray Reiter who visited Maryland recently. We thought that the session might best focus on issues in non-monotonic logic. Each panelist would be given approximately 10 minutes to disuss the various issues. A first approach to the
overall organization of the panel is:
Minker -10 minutes on Lerner accomplishments
Minker - 2 minutes on non-monotonic logic
McCarthy -10 minutes
McDermott -10 minutes
Minker -10 minutes
Moore -10 minutes
Reiter -10 minutes
Inter Panel Discussion-10 minutes
Audience Participation-10 minutes
Summary -5 minutes (one minute per panelist)
The topics that might be discussed are:
1. Is logic an appropriate vehicle for non-monotonic reasoning?
2. What are some open technical issues?
What are reasonable approaches to them?
3. What classes of AI problems are inherently non-monotonic?
Knowledge Based Systems?
Presuppositions in Natural Language?
Reasoning About Prototypes?
4. What are the computational problems in dealing with non-monotonic
reasoning?
5. Does non-monotonic reasoning play a central role in AI?
I would appreciate your comments on the organization of the panel
and on the issues to be discussed.I look forward to hearing from you shortly
and to seeing you at IJCAI
Best regards,
Jack Minker
∂15-Jul-83 1627 RINDFLEISCH@SUMEX-AIM Re: Equipment Contract
Received: from SUMEX-AIM by SU-AI with PUP; 15-Jul-83 16:27 PDT
Date: Fri 15 Jul 83 16:22:34-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Re: Equipment Contract
To: Feigenbaum@SUMEX-AIM, JMC@SAIL, CSL.JLH@Score, Ullman@Diablo, RPG@SAIL,
REG@SAIL
cc: CSD.BScott@Score, Rindfleisch@SUMEX-AIM
In-Reply-To: Your message of Mon 23 May 83 17:25:49-PDT
The DARPA Equipment Contract is fully signed now and has been received
at Stanford. We will need to set up the accounts for managing the
money. Are there any objections to the functional subdivision plan I
mentioned before, i.e., separate accts for HPP, Formal Reasoning, CSL,
and CSD-CF?
Tom R.
-------
∂15-Jul-83 1641 KJB@SRI-AI.ARPA Announcing CSLI
Received: from SRI-AI by SU-AI with TCP/SMTP; 15 Jul 83 16:33:34 PDT
Date: 15 Jul 1983 1616-PDT
From: Kjb at SRI-AI
Subject: Announcing CSLI
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI,
To: Kaplan at PARC, kjb at SRI-AI, lauri at SRI-AI,
To: peters at UTEXAS, rperrault at SRI-AI,
To: sag%SU-psych at SUMEX-AIM, stan at SRI-AI,
To: Wasow%SU-psych at SUMEX-AIM, winograd at PARC
We will be making up an announcement of CSLI and its relation to
SDF next week for distribution. It would be better not to send
the existing proposal around, especially with the budget and
such in it. (The latter in response to a couple of questions.)
-------
∂15-Jul-83 1738 ullman@Diablo Re: Equipment Contract
Received: from SU-HNV by SU-AI with PUP; 15-Jul-83 17:37 PDT
Date: Fri, 15 Jul 83 17:37 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Re: Equipment Contract
To: CSL.JLH@SU-Score, Feigenbaum@Sumex-Aim, JMC@Sail, REG@Sail,
RPG@Sail, Rindfleisch@Sumex-Aim
Cc: CSD.BScott@SU-Score
we never resolved the issue of how many SUN's are for ERL, how many for
MJH. I think everything else is clear, no?
∂15-Jul-83 1810 BRESNAN.PA@PARC-MAXC.ARPA Kelly Roach
Received: from PARC-MAXC by SU-AI with TCP/SMTP; 15 Jul 83 18:10:15 PDT
Date: Fri, 15 Jul 83 18:09 PDT
Sender: BRESNAN.PA@PARC-MAXC.ARPA
Subject: Kelly Roach
To: jmc@sail.ARPA
From: Bresnan.PA@PARC-MAXC.ARPA, Kaplan.PA@PARC-MAXC.ARPA
cc: Bresnan.PA@PARC-MAXC.ARPA, Kaplan.PA@PARC-MAXC.ARPA
Dear John,
This message is to follow up on our conversation about Kelly Roach
yesterday. As we indicated yesterday, we would like to have him
transfer from MIT to Stanford, if it is possible, by this September.
Here, briefly, is his background.
Kelly came to the MIT computer science dept in 1980 from Cal Tech, where
he graduated with honors with three simultaneous degrees--a B.S. in
mathematics and a B.S. and M.S. in computer science. He had worked with
Fred Thompson at Cal Tech, and went to MIT to work with Bill Martin on
natural language. Bill Martin died of liver cancer in Kelly's first
year, but Kelly continued working on language at MIT, with Pat Winston
serving as his supervisor. In his second year Kelly took several
graduate courses in the MIT linguistics dept and built a parser for our
lexical-functional grammar formalism. He was one of 4 top students in
EECS (Electrical Engineering & Computer Science), Linguistics and
Philosophy, and Psychology to be awarded a MIT Cognitive Science Center
Fellowship. The summer after his second year, he came to Xerox PARC as
a Summer Intern, and he stayed on to work in the Xerox PARC environment
during the academic year (this year), taking a leave of absence from
MIT. During this year he accomplished the following: (1) he designed
and brought up a grammar induction system incorporating truth
maintenance and applying Bayesian inference to the theory of context
free languages; (2) he worked for John Seely Brown on an algebra problem
solver for children; (3) he worked on the display-based LFG parser for
the Xerox 1100 series personal LISP machines, implementing a chart
window facility; (4) he developed some remarkable results in the
mathematical theory of LFGs, showing (i) that the indexed languages are
properly contained in LFG languages, and (ii) that every LFG language is
weakly equivalent to a left-linear LFG language (i.e. an LFG with a
left-linear grammar base); (5) right now he is doing a floppy operating
system for the Dandelion.
In all of these projects Kelly has demonstrated a number of
extraordinary qualities. Not only is he is an excellent programmer,
but he is capable of serious and intense work at a high theoretical
level. The theorems that he has proved, for example, provide an elegant
mathematical understanding of information flow in LFG grammars,
indicating what formal devices are responsible for its descriptive
power. Recently a number of people have worked on these problems (e.g.
Pereira, Friedman, Berwick, and ourselves), but none have obtained such
a comprehensive and powerful set of results as Kelly has His
combination of programming skills and mathematical sophistication is
extremely rare.
We think Kelly would be an ideal student to have at Stanford, and he
will undoubtedly receive fellowship support from the CSLI. As we
mentioned, he wants to be a computer scientist, and intends to get his
degree in that. Much as we want to have Kelly here, neither of us can
admit him or officially supervise him in Computer Science, and that's
why we are turning to you.
We hope you can help in bringing Kelly here. Joan spoke about Kelly to
David Kennedy and to Norm Wessels after the application deadline had
passed, and they indicated that there are some things Deans CAN do to
help. But the real problem is to find the best connection with Computer
Science.
In general, we are a little concerned about the problem of connecting
the CSLI with good Stanford CS students who want to work on theoretical
and computational problems in language. At MIT, AI students are
encouraged to do serious work in the knowledge domain they are
specializing in--computer vision people work with the brain people in
psychology, computer language people work with the linguists (e.g. Mitch
Marcus, Ken Church, Bob Berwick, Kelly Roach), etc. It is extremely
important for us to find faculty cooperation in CS to encourage this.
Thanks for your help. Please let us know what further steps we need to
take to make this happen. (We are sending a copy of his resume in a
separate message.)
Joan and Ron
------------------------------------------------------------
∂15-Jul-83 1814 BRESNAN.PA@PARC-MAXC.ARPA Kelly Roach's resume
Received: from PARC-MAXC by SU-AI with TCP/SMTP; 15 Jul 83 18:14:03 PDT
Date: Fri, 15 Jul 83 18:13 PDT
Sender: BRESNAN.PA@PARC-MAXC.ARPA
Subject: Kelly Roach's resume
To: jmc@sail.ARPA
From: Bresnan.PA@PARC-MAXC.ARPA, Kaplan.PA@PARC-MAXC.ARPA
cc: Bresnan.PA@PARC-MAXC.ARPA, Kaplan.PA@PARC-MAXC.ARPA
Dear John,
Here is Kelly's resume, as promised in a separate message.
RESUME
Name: Kelly Roach
Home Address: 958 Addison
Palo Alto, CA 94301
Home Phone: (415) 321-5132
Work Address: Xerox PARC
3333 Coyote Hill Road
Palo Alto, CA 94304
Work Phone: (415) 494-4372
Social Security Number: 536-72-8199
EDUCATION
1976-1980: Undergraduate at California Institute of Technology
B.S. in Mathematics (1980)
B.S. in Computer Science (1980)
M.S. in Computer Science (1980)
GPA 3.95
Graduated with Honors
1980-1982: Graduate Student at Massachusetts Institute of
Technology
2 years towards Ph.D. in Computer Science (1985)
EXPERIENCE
6/83-9/83
Employer: Xerox PARC
Manager: Dr. Beau Sheil
Salary: $1526/mo
Position: Parttime Summer
Description: Member of Xerox PARC's LISPSUPPORT group and
implementing software to support Shugart floppy disk drives and
connections to existing filing system in Interlisp.
11/82-6/83
Employer: Xerox PARC
Manager: Dr. John Brown
Salary: $1526/mo
Position: Research Intern
Description: Developed menu based symbolic children's algebra
problem solving system using pattern matching techniques. Worked in
Interlisp.
6/82-9/82
Employer: Xerox PARC
Manager: Dr. Ronald Kaplan
Salary: $2200/mo
Position: Summer Intern
Description: Research into Lexical Functional Grammar, a formal
system for representation of human grammatical knowledge of natural
languages. Worked in Interlisp.
9/81-6/82
Employer: MIT Artificail Intelligence Laboratory
Manager: Dr. Patrick Winston
Salary: $1383/mo non-taxable
Position: Research Assistant
Description: Implemented a natural language parser based on
Lexical Functional Grammar. Worked in Maclisp.
6/81-9/81
Employer: IBM Data Systems Division/Poughkeepsie, NY 12602
Manager: William Benson
Salary: $1650/mo
Position: Computer Programmer
Description: Worked on and had complete responsibility over the
B5 EC Projection System. This system used a database of previous
product histories to project Engineering Changes of a new product over
time. Worked in APL.
9/80-6/81
Employer: MIT Laboratory for Computer Science/Cambridge, MA 02139
Manager: Dr William Martin
Salary: $1199/mo non-taxable
Position: Research Assistant
Description: Worked on EQSP natural language system interfaced
with the naval ONRODA Data Base created by the Office of Naval
Research and various other data bases. Worked in MACLISP.
6/80-9/80
Employer: IBM Data Systems Division/Poughkeepsie, NY 12602
Manager: Doug Wolfe
Salary: $1565/mo
Position: Computer Programmer
Description: Coresponsible for the design and portions of the
implementation of the Alpine Cable Data Base System, a data base for
computer cables. Worked in PL/S and IBM Assembler.
9/79-6/80
Manager: Dr. Fred Thompson (Caltech)
Salary: None
Position: Student
Description: Helped create the POL System, a rapidly expansible
relational database system with conversation english as one front end.
Worked in Pascal.
MACHINES USED:
Xerox Dolphin, Xerox Dorado, IBM370, PDP10, VAX, personal computers,
and microprocessors
LANGUAGES USED:
High Level: Interlisp, Maclisp, APL, Pascal, Fortran, Basic, PL/S,
PL/I, Simula, C, Snobol
Low Level: Z80, 6502, IBM Assembler
Joan and Ron
∂15-Jul-83 2332 burns.boulder@Rand-Relay
Received: from RAND-RELAY by SU-AI with TCP/SMTP; 15 Jul 83 23:32:39 PDT
Date: 15 Jul 1983 23:38:14-PDT
From: burns.boulder@Rand-Relay
Return-Path: <burns.boulder@Rand-Relay>
To: JMC@SU-AI
Cc: `λλ\`λλZ.boulder@Rand-Relay
Via: boulder; 15 Jul 83 23:21-PDT
Thank you, sir. What conclusions have you reached to date?
∂15-Jul-83 2333 burns.boulder@Rand-Relay
Received: from RAND-RELAY by SU-AI with TCP/SMTP; 15 Jul 83 23:33:02 PDT
Date: 16 Jul 1983 00:01:07-PDT
From: burns.boulder@Rand-Relay
Return-Path: <burns.boulder@Rand-Relay>
To: JMC@SU-AI
Via: boulder; 15 Jul 83 23:22-PDT
Thank you, sir. What are you working now?
∂16-Jul-83 1401 KJB@SRI-AI.ARPA csli
Received: from SRI-AI by SU-AI with TCP/SMTP; 16 Jul 83 14:01:15 PDT
Date: Sat 16 Jul 83 13:56:21-PDT
From: KJB@SRI-AI.ARPA
Subject: csli
To: BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA, Briansmith@PARC-MAXC.ARPA,
grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA, JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA,
kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA, peters@UTEXAS-20.ARPA,
rperrault@SRI-AI.ARPA, sag%SU-psych@SUMEX-AIM.ARPA, stan@SRI-AI.ARPA,
Wasow%SU-psych@SUMEX-AIM.ARPA, winograd@PARC-MAXC.ARPA
Dear all,
I have just returned from a long talk with c.s. My impression
is that we are in good shape, if we pull together. I will make a long
memo summarizing where we are after I have sorted it all out in my mind.
But he insists on having a budget that is tied to the program we have
outlined, and that has to get done first thing. That is to be used to
manage the project, and to judge us down the road. I have put Ducher
to work on it, but he may well need help from all of you in the next few
days to make sense of it.
The building budget is not in the 15,000,000, and there is still
a lot to be done on that score.
We also have to write proposals to Xerox and DEC for equipment
very soon. I hope that we can get a lot of help on that. If I have to
do it, I might order 16 apples or something.
Jon
-------
∂16-Jul-83 1751 LLW@S1-A Festschrift Shackleton
Received: from S1-A by SU-AI with TCP/SMTP; 16 Jul 83 17:50:59 PDT
Date: 16 Jul 83 1749 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Festschrift Shackleton
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
∂15-Jul-83 0114 JMC@SU-AI Shackleton for festschrift
Received: from SU-AI by S1-A with TCP/SMTP; 15 Jul 83 01:14:21 PDT
Date: 15 Jul 83 0113 PDT
From: John McCarthy <JMC@SU-AI>
Subject: Shackleton for festschrift
To: llw@S1-A, rah@S1-A
Have you looked at FESTSC[SHA,JMC] or have you written anything?
[John: I've read it (and consider it to be an excellent framework),
and have discussed it with Rod. I'm frantically preparing for
fund-raising seances this coming week and probably won't have
a chance to write on it this weekend, but will certainly encourage
Rod to do so. Lowell]
∂16-Jul-83 2013 LLW@S1-A Deadline
Received: from S1-A by SU-AI with TCP/SMTP; 16 Jul 83 20:13:30 PDT
Date: 16 Jul 83 2011 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Deadline
To: jmc@SU-AI
CC: LLW@S1-A
∂16-Jul-83 1920 JMC@SU-AI
Received: from SU-AI by S1-A with TCP/SMTP; 16 Jul 83 19:19:53 PDT
Date: 16 Jul 83 1920 PDT
From: John McCarthy <JMC@SU-AI>
To: llw@S1-A, rah@S1-A
What is the real deadline for Festschrift Shackleton?
[John: Whenever Hans tells me he's going to drop the manuscript
collection on the publisher he's selected. Lowell]
∂17-Jul-83 1041 CLANCEY@SUMEX-AIM Re: Workstation Purchase Dilemma
Received: from SUMEX-AIM by SU-AI with PUP; 17-Jul-83 10:35 PDT
Date: Sun 17 Jul 83 10:36:18-PDT
From: William J. Clancey <CLANCEY@SUMEX-AIM>
Subject: Re: Workstation Purchase Dilemma
To: Rindfleisch@SUMEX-AIM
cc: HPP-Exec@SUMEX-AIM, Yeager@SUMEX-AIM, Schmidt@SUMEX-AIM, JMC@SAIL, RPG@SAIL
In-Reply-To: Your message of Fri 15 Jul 83 11:19:13-PDT
Tom,
Given that we have the money and must spend it, get at least SOME
Dandelions. We need places for programmers to sit today. I see the
crunch already beginning with Wilkins using the machine in Doug's office
and Mark Richer and myself vying for time on the "Novak" machine. Diane
Warner will be fulltime in September, and we might get another Master's
student. This mix will require another machine. A Dandelion might be
less than ideal, but it definitely can be used for development and is better
than nothing. As I understand it, the fully configured Dandelion is
better than the Dolphins we are using today, and they are adequate.
If you can, consider cutting the Dandelion order to a bare minimum
and buy another Dorado. While individual use of a machine expands to fit the
resources available, 3 people might prefer sharing a Dorado to having
Dandelions for themselves, even though this would require working nights and
weekends (but why not? that's how I did my thesis work on Sumex). What is
Xerox experience in this regard?
Bill
-------
∂17-Jul-83 2007 GENESERETH@SUMEX-AIM
Received: from SUMEX-AIM by SU-AI with PUP; 17-Jul-83 20:07 PDT
Date: Sun 17 Jul 83 20:09:06-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
To: JMC@SAIL
In-Reply-To: Your message of Sat 16 Jul 83 22:41:00-PDT
John,
My home phone is 322-4707, but I am at the office at the moment and will
be for a while. The number here is 497-0324. I'll also be in early
tomorrow morning.
mrg
-------
∂18-Jul-83 0727 jlh@Shasta Re: Equipment Contract
Received: from SU-SHASTA by SU-AI with PUP; 18-Jul-83 07:27 PDT
Date: Monday, 18 Jul 1983 07:36-PDT
To: T. C. Rindfleisch <Rindfleisch at SUMEX-AIM>
Cc: Feigenbaum at SUMEX-AIM, JMC at SAIL, CSL.JLH at Score,
Ullman at Diablo, RPG at SAIL, REG at SAIL, CSD.BScott at Score
Subject: Re: Equipment Contract
In-reply-to: Your message of Fri 15 Jul 83 16:22:34-PDT.
From: John Hennessy <jlh@Shasta>
Subdivisions are ok with me. Do you want actual subaccounts for record
keeping? I'd like to avoid a long process to make small changes in what
we purchase.
The systems people can handle the distribution of systems suns (the
bulk of which are currently needed in MJH). I presume any CSD-CF sun's
will go into MJH.
∂18-Jul-83 0950 JMC*
Phone Len.
∂18-Jul-83 1009 SHORTLIFFE@SUMEX-AIM Re: Workstation Purchase Dilemma
Received: from SUMEX-AIM by SU-AI with PUP; 18-Jul-83 10:08 PDT
Date: Mon 18 Jul 83 09:09:23-PDT
From: Ted Shortliffe <Shortliffe@SUMEX-AIM>
Subject: Re: Workstation Purchase Dilemma
To: Rindfleisch@SUMEX-AIM
cc: HPP-Exec@SUMEX-AIM, Yeager@SUMEX-AIM, Schmidt@SUMEX-AIM, JMC@SAIL, RPG@SAIL
Office: Room TC-117, Stanford Med Center; Phone: (415) 497-6979
In-Reply-To: Your message of Fri 15 Jul 83 11:19:40-PDT
Tom,
Although our group has not yet "hit the wall", we haven't yet
connected the Reasoner and Interview on the Dandelion so I don't know
how serious an impact that will have on performance. With the
separate screen processor, I am hopeful that performance will be adequate.
From our point of view, given our deadlines and commitment to
Interlisp, there is really no alternative to Dandelions. Given the
immediate HPP need for cycles, and the amount of Dolphin-based work in
Interlisp-D that could move to Dandelions, I'd agree with Bill Clancey
that some Dandelions in the short term seem mandatory, despite concerns
about performance and the recent changes in pricing structure.
Ted
-------
∂18-Jul-83 1331 wallace@Whitney AI and space station
Received: from SU-WHITNEY by SU-AI with PUP; 18-Jul-83 13:27 PDT
Date: 18 July 1983 13:27:47-PDT (Monday)
From: wallace@Whitney
Subject: AI and space station
To: jmc at su-ai
I am with the NASA Summer Study looking into automation and
the proposed space station. (I am also a graduate student in
computer science at CMU). I am interested in the space station
as an intelligent system. Would you like to communicate with me
about this?
Best,
Richard Wallace
I have no special ideas about "the space station as an intelligent
system", though I'm all for space stations. What are your ideas?
∂18-Jul-83 1446 GARDNER@SU-SCORE.ARPA CBCL
Received: from SU-SCORE by SU-AI with TCP/SMTP; 18 Jul 83 14:46:24 PDT
Date: Mon 18 Jul 83 14:46:29-PDT
From: Anne Gardner <GARDNER@SU-SCORE.ARPA>
Subject: CBCL
To: jmc@SU-AI.ARPA
I read the paper, and I'd definitely like to read more. The representation
problems you raise are all going to come up in doing more offer-and-
acceptance cases, too.
--Anne
-------
∂18-Jul-83 1758 LENAT@SU-SCORE.ARPA Re: Workstation Purchase Dilemma
Received: from SU-SCORE by SU-AI with TCP/SMTP; 18 Jul 83 17:58:27 PDT
Date: Mon 18 Jul 83 17:57:38-PDT
From: Doug Lenat <LENAT@SU-SCORE.ARPA>
Subject: Re: Workstation Purchase Dilemma
To: Shortliffe@SUMEX-AIM.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA, HPP-Exec@SUMEX-AIM.ARPA, Yeager@SUMEX-AIM.ARPA,
Schmidt@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA, RPG@SU-AI.ARPA
In-Reply-To: Message from "Ted Shortliffe <Shortliffe@SUMEX-AIM>" of Mon 18 Jul 83 09:09:50-PDT
I agree with Bill and Ted: grumble and get some Dandelions,
perhaps stretch and get one extra Dorado.
Doug
-------
∂18-Jul-83 2155 wallace@Whitney HAL and spaceborne AI
Received: from SU-WHITNEY by SU-AI with PUP; 18-Jul-83 21:54 PDT
Date: 18 July 1983 21:55:24-PDT (Monday)
From: wallace@Whitney
Subject: HAL and spaceborne AI
To: jmc at su-ai
Once I wrote an article comparing HAL (from 2001) with
contemporary state-of-the-art AI systems. We have computers
onboard spacecraft now (programmed in a language called HAL-S).
We have computers that can beat most humans at chess, as HAL did.
We have computers that can do limited speech understanding
(in 2001, HAL listens to the voices of only 2 people; they typically
speak in short utterances and very clearly with little
background noise). NASA has contracts
out to private companies to develop expert systems for fault
diagnosis of spacecraft hardware. Remember HAL's line "I have
detected a fault in the AE-35 unit..." HAL is able to
recognize line drawings of the crew members as well as to
detect when a crewmember passes before one of its cameras.
It's not much more sophisticated than we could expect from
a limited-domain vision system today. HAL's speech (text) generation
ability is better than what most machines I know about can do;
but overall it's reasonable to assert that most of what HAL did in
2001 is doable by taking the union of a lot of today's AI systems.
The analogy breaks down where HAL did, since we wouldn't expect a
current system to exhibit willfully deceitful behavior. In any case
my vision of an intelligent spacecraft controller is based upon HAL.
Do you think this vision is realistic?
I can't imagine basing a spacecraft controller on HAL. There is to
big a gap between movie scenarios and a concrete list of functions
and means of achieving them.
∂19-Jul-83 0759 OP unsupported students
do you want to meet today? (actually, I'm in no hurry, I just don't want to
forget)
--Oren
∂19-Jul-83 1038 RAC (1) research topic, (2) James P. Hogan
(1) Could we sit down together sometime soon to talk about a choice
of a possible research topic for me. If you have any ideas of how to
combine learning with the things we did in the winter quarter course
you taught that would certainly be interesting. Basically, what I am
interested in is working on representations to improve learning.
(2) James P. Hogan (one of the Sci Fi authors on Marvin Minsky's
Sci Fi panel at the last AAAI) is a friend of mine. He is comming to
town this Thursday evening and will be here through Saturday afternoon.
Would you like to join bull sessions on space, AI, and such on any of
(a) Thu dinner, (b) Fri dinner, (c) Sat afternoon picnic, (d) you pick
a time?
Rodger Cliff
(1) It should be this week, because I'll be away next week, and after
that I'll be doing a lot of traveling.
(2) Thursday or Friday dinner with Hogan would be interesting with
a slight preference for Thursday.
∂19-Jul-83 1212 RPG Closures
To: ullman@SU-HNV
CC: JMC@SU-AI
`Passing global values' is simple. Since I assume a shared address
space implemented as shared physical memory, each of the processes can
obtain the values trivial. I doubt that is what you meant. The problems
come up when you want to take hold of a value, muck with it awhile,
possibly alter it, and then write it out. Here's how to do that:
Suppose we have a variable, *BAZ* that has as its value a record of
3 slots:
(defstruct (vt
(conc-name vt-))
(x 0.0)(y 0.0)(z 0.0))
(setq *baz* (make-vt x 9.0 y 6.0)) ;default z to 0.0
Now we can do one of two things, depending on the style we wish to adopt.
First we can make a closure that takes a message and some data, and hacks *baz*:
(defun hack-baz (mess data)
(caseq mess
(x (vt-x *baz*))
(y (vt-y *baz*))
(z (vt-z *baz*))
(set-x (setf (vt-x *baz*) data))
(set-y (setf (vt-y *baz*) data))
(set-z (setf (vt-z *baz*) data))
(update-x (setf (vt-x *baz*) (update data (vt-x *baz*))))
...
(t "Unknown message type for *baz*)))
Second we can make a closure that takes a closure and applies that closure
to *baz*:
(defun hack-baz (closure)
(closure *baz*))
and then set-x is implemented as:
...
(let ((data ..)) ;whatever DATA should be
(hack-baz (lambda (z)
(setf (vt-x z) data))))
...
Since simple closure-invocation is a message-passing operation, not an interruption,
we get critical regions this way.
So I imagine a `typical' algorithm would look like:
...
(let ((x (vt-x *baz*)))
(cond ((and (= x 99.) ;looks like x should be incremented by 27
;based on its current value
(not
(eq 'foobar
(hack-baz
(lambda (z)
(cond ((= (vt-x z)
99.) ;still true!
(incf (vt-x z) 27.))
(t 'foobar))))))))
...
Notice that this code looks at the global value of *baz*, which costs
cache coherence at worst, decides to update, and then looks again in a
critical region.
I would like to discuss this further with the committee, but I will be out of
town thursday.
-rpg-
∂19-Jul-83 1706 BOSACK@SU-SCORE.ARPA Fixing your modem
Received: from SU-SCORE by SU-AI with TCP/SMTP; 19 Jul 83 17:06:31 PDT
Date: Tue 19 Jul 83 17:06:31-PDT
From: Len Bosack <BOSACK@SU-SCORE.ARPA>
Subject: Fixing your modem
To: JMC@SU-AI.ARPA
If tomorrow afternoon is OK, I will take Joe Chang (a summer student from
EE) over and work on the modem.
Len
-------
That will be fine. Do you have the combination?
∂19-Jul-83 1721 RAC Hogan and Research
Jim Hogan is supposed to meet me at my apartment at about 5pm on Thu.
My suggestion is that you come here at 4pm. We can talk while we wait
for him.
Address: 535 Everett St. Palo Alto., Apt 404
Phone: 327-1730
The apartment is in the 6-story building on Everett (parallel to Univ. Av.)
between Cowper and Webster.
Rodger Cliff
∂20-Jul-83 1425 DFH
To: CLT, JMC
September Travel Arrangements (Newcastle, IFIP)
I got some tentative flight information from Dina Bolla. Please look this over
and let me know what you would like to do.
9/4 SFO/London Pan Am 5:30pm/11:30am next day
9/5 London/Newcastle 3:00pm/4:00pm
(There are a variety of daily flights between London and Newcastle.
You could overnight in London, and take a 7:30 am flight the
following morning if you want)
9/9 evening
or
9/10 a.m. Newcastle/London
9/10 London/Venice 9:40am/12:45pm
or 2:35pm/5:35pm
9/18 Venice/Paris, 1:45pm/3:20 pm
or 5:50pm/7:25 pm
9/19 - 9/23 IFIP - Paris
9/23 Paris/SFO 11:30am/3:30pm (Change planes in London)
or (Pan Am flight) 1:00pm/7:22 pm (Change planes in NYC)
9/24 5:00pm/7:35 pm (to Los Angeles -- then
fly up from LA)
Most of the flights mentioned operate daily.
∂20-Jul-83 1841 RESTIVO@SU-SCORE.ARPA PROLOG Digest V1 #14
Received: from SU-SCORE by SU-AI with TCP/SMTP; 20 Jul 83 18:41:06 PDT
Date: Wednesday, July 20, 1983 7:31AM
From: Chuck Restivo (The Moderator) <PROLOG-REQUEST@SU-SCORE.ARPA>
Reply-to: PROLOG@SU-SCORE.ARPA
US-Mail: P.O. Box 4584 Stanford University, Stanford CA 94305
Phone: (415) 326-5550
Subject: PROLOG Digest V1 #14
To: PROLOG@SU-SCORE.ARPA
PROLOG Digest Thursday, 21 Jul 1983 Volume 1 : Issue 14
Today's Topics:
Applications - Definitions,
Implementations - DEC-20 Environment & CProlog Bug
----------------------------------------------------------------------
Date: Mon 18 Jul 83 08:47:20-PDT
From: SHardy@SRI-KL
Subject: Equal On Terms
An earlier message asked how to define a predicate EQUAL on two terms
that didn't mind if variables were renamed but otherwise wanted the
terms to be identical, thus:
equal(A + B, C + D), equal(A + A, C + C)
are true, but not:
equal(A + A, B + C)
The proposed ``quick and dirty'' solution was something like:
equal(X, Y) :- assert(temp(1, X)), assert(temp(2, Y)),
retract(temp(1, A)), retract(temp(2, B)),
numbervars(A, 1, N), numbervars(B, 1, N),
A = B.
A ``better'' (I.e. faster) solution is:
equal(X, Y) :- \+notequal(X, Y).
notequal(X, Y) :- \+notnotequal(X, Y).
notnotequal(X, Y) :- numbervars(X, 1, N),
numbervars(Y, 1, N), X = Y.
This is ``better'' because it doesn't use ASSERT. It's best to avoid
Assert because:
* It is slow - it has to copy the structure being asserted.
* On the DEC-20, it consumes memory until the procedure doing
the ASSERT is backtracked over - not just until the item is
RETRACTed. This might be for the ENTIRE computation.
A useful heuristic is that predicates that side-effect their arguments
in undesirable ways (eg by using NUMBERVARS) can sometimes be rendered
safe by enclosing them in a double negation.
-- Steve Hardy,
Teknowledge
------------------------------
Date: Mon 18 Jul 83 11:55:26-PDT
From: SHardy@SRI-KL
Subject: Environment On DEC-20
I have some queries about DEC-20 Prolog:
(1) Does anyone know a way of avoiding having to copy library files
into one's own directory?
(2) Is there some secret way of running inferior forks, such as Emacs.
(3) If not, is it possible for a Prolog program to exit to the monitor
and then, later when Prolog is CONtinued, resume execution to, for
example, REConsult a file.
Also, some queries about CProlog:
(4) Is it distributed with source code?
(5) Is it possible to define new predicated using C?
(6) Would it be feasible to run it on a 16-bit machine?
-- Steve Hardy,
Teknowledge
------------------------------
Date: Sun Jul 10 20:54:08 1983
From: OgcVax!Tombl@Berkeley
Subject: CProlog Bug
There is a bug in the CProlog grammar rules. Translation of
expressions of the following form may be affected:
(expr1 ; expr2)
The current rules produce code which may incorrectly bind the return
que for some instances of user code. For example, the expression
( funct(arg1,...,argn) ; {expr1} )
is translated to produce
funct(arg1,...,argn,Input,Output)
for the left member, which is correct. Translation of the right member
returns expr1, also correct, but has the side effect of binding its
input que to its output que. These happen also to be the input and
output ques for the expression as a whole, so the left member
translation effectively becomes
funct(arg1,...,argn,Input,Input)
which is incorrect.
The bug fix is straightforward, and should be applied to
prolog/pl/grammar:
30,32c30,32 < $t←rp((T;R),S,SR,(Tt;Rt)) :- !, /* distributed code */ <
$t←rp(T,S,SR,Tt), < $t←rp(R,S,SR,Rt).
--- > $t←rp((T;R),S,SR,(Tt,SR=SR1;Rt,SR=SR2)) :- !, /* modified code
*/ > $t←rp(T,S,SR1,Tt), > $t←rp(R,S,SR2,Rt).
I'd be interested in seeing a better solution, if someone should come
up with one.
-- Tom Blenko
decvax!tektronix!ogcvax!metheus!tombl
ucbvax!tektronix!ogcvax!metheus!tombl
------------------------------
End of PROLOG Digest
********************
∂20-Jul-83 2317 GENESERETH@SUMEX-AIM meeting
Received: from SUMEX-AIM by SU-AI with PUP; 20-Jul-83 23:17 PDT
Date: Wed 20 Jul 83 23:18:25-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
Subject: meeting
To: jmc@SAIL
John,
Something has come up and it would be helpful to me if we could postpone
our meeting until early afternoon, say 1:30. If that's possible send me
a note. Otherwise I will assume we are still on. Sorry for the late notice.
mrg
-------
∂21-Jul-83 1141 JMC*
wpps
∂21-Jul-83 1141 JMC*
Philip Holmes 607 256-4335
Michel Herman,
1965 arnold ams trans ser. 2 vol. 46, 213-284
small denominators
Herman, thesis, math. pub. IHES
1976
Spring vol. 597
p.271-293, Mesure de
geometry and topology
John Guckenheimer, Math. Santa Cruz
Non linear osc. of dynamical systems and bifurcations of vector fields
Springer applied math. no. 42
Leon Glass, McGill, cardiac arhythmias,
∂21-Jul-83 1214 BOSACK@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 21 Jul 83 12:13:56 PDT
Date: Thu 21 Jul 83 12:14:31-PDT
From: Len Bosack <BOSACK@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Thu 21 Jul 83 01:14:00-PDT
The problem is at the Telephone company. They will be around sometime this
afternoon. I'll lead them as appropriate...
Len
-------
∂21-Jul-83 1230 RESTIVO@SU-SCORE.ARPA Re: name change
Received: from SU-SCORE by SU-AI with TCP/SMTP; 21 Jul 83 12:30:17 PDT
Date: Thu 21 Jul 83 09:15:52-PDT
From: Chuck Restivo <RESTIVO@SU-SCORE.ARPA>
Subject: Re: name change
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Wed 20 Jul 83 21:46:00-PDT
Done!
-------
∂21-Jul-83 2106 RINDFLEISCH@SUMEX-AIM DEC Changes
Received: from SUMEX-AIM by SU-AI with PUP; 21-Jul-83 21:06 PDT
Date: Thu 21 Jul 83 21:07:13-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: DEC Changes
To: HPP-Exec@SUMEX-AIM
cc: Rindfleisch@SUMEX-AIM, Ullman@Diablo, Golub@Score, JMC@SAIL, RPG@SAIL,
REG@SAIL, CSL.JLH@Score, Cheriton@Diablo, Lantz@Diablo
A rumor, afoot for some time, has been or is about to be confirmed that
Gordon Bell, VP of Engineering, is leaving DEC. Further unsubstantiated
rumor says he is moving to PRIME Computer. I don't have any information
about what changes this bodes for DEC. Gordon has had some serious
health problems recently but the fact that he may be joining another
computer manufacturer (rather than an academic place like CMU or
Stanford) suggests that other changes are occurring within DEC beyond
the recent freezing of the 36-bit line. I don't know what this means
for the DEC collaboration we had been planning on for the architecture
project.
If anyone has more information, please let me know. Tom R.
-------
∂21-Jul-83 2138 LLW@S1-A No Time For Travel
Received: from S1-A by SU-AI with TCP/SMTP; 21 Jul 83 21:38:18 PDT
Date: 21 Jul 83 2138 PDT
From: Lowell Wood <LLW@S1-A>
Subject: No Time For Travel
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
∂20-Jul-83 2326 JMC@SU-AI
Received: from SU-AI by S1-A with TCP/SMTP; 20 Jul 83 23:26:42 PDT
Date: 20 Jul 83 2323 PDT
From: John McCarthy <JMC@SU-AI>
To: LLW@SU-AI, RAH@SU-AI
Are you planning to go to Pournelle's meeting?
[John: I'm afraid I'm not going to be able to make it. Lowell]
∂21-Jul-83 2233 GENESERETH@SUMEX-AIM schedule
Received: from SUMEX-AIM by SU-AI with PUP; 21-Jul-83 22:33 PDT
Date: Thu 21 Jul 83 22:33:59-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
Subject: schedule
To: jmc@SAIL
John,
Here's th schedule we discussed.
mrg
Course schedule
Monday Tuesday Wednesday Thursday Friday
------------------------------------------------------------------------
AI Intro Formal Reas Disj Search NonMon Reas Meta-Level
Lisp Unification Conj Search
JMC MRG MRG JMC MRG
------------------------------------------------------------------------
ES Intro BC and FC Game-Playing Simulation Common Sense
Pred Calc Resolution Planning Future Directions
Semantics
MRG MRG JMC MRG JMC
------------------------------------------------------------------------
1. The last three sessions can be juggled around subject to the constraint
that planning come before meta-level.
2. Reception Tuesday afternoon at 5. We need to finish early.
3. There will probably be need for discussion of exercises each morning.
-------
∂22-Jul-83 0034 FRIEDLAND@SUMEX-AIM Re: DEC Changes
Received: from SUMEX-AIM by SU-AI with PUP; 22-Jul-83 00:34 PDT
Date: Fri 22 Jul 83 00:34:54-PDT
From: Peter Friedland <FRIEDLAND@SUMEX-AIM>
Subject: Re: DEC Changes
To: Rindfleisch@SUMEX-AIM
cc: HPP-Exec@SUMEX-AIM, Ullman@Diablo, Golub@Score, JMC@SAIL, RPG@SAIL, REG@SAIL,
CSL.JLH@Score, Cheriton@Diablo, Lantz@Diablo
In-Reply-To: Your message of Thu 21 Jul 83 21:07:18-PDT
Gordon is now with ENCORE, a new company formed by former PRIME executives.
There is an article about ENCORE in Thursday's Wall Street Journal on page
2. Their mission is to buy up small high-tech companies (mainly in computers)
and use their marketing expertise to sell the small company products. Good
question about how that changes things with DEC, Tom.
Peter
-------
∂22-Jul-83 1000 jlh@Shasta Re: DEC Changes
Received: from SU-SHASTA by SU-AI with PUP; 22-Jul-83 10:00 PDT
Date: Friday, 22 Jul 1983 08:46-PDT
To: T. C. Rindfleisch <Rindfleisch at SUMEX-AIM>
Cc: HPP-Exec at SUMEX-AIM, Ullman at Diablo, Golub at Score, JMC at SAIL,
RPG at SAIL, REG at SAIL, CSL.JLH at Score, Cheriton at Diablo,
Lantz at Diablo
Subject: Re: DEC Changes
In-reply-to: Your message of Thu 21 Jul 83 21:07:13-PDT.
From: John Hennessy <jlh@Shasta>
Bell has left DEC and is joining a start-up ONCOR that the ex-president of
Prime (Fisher) is involved in. Apparently, Gordon felt a strong to
desire to come to either Stanford or CMU, but his wife's involvement in
the DEC computer museum (she runs it) made that difficult.
See yesterday's Wall Street Journal.
John
∂22-Jul-83 1001 BOSACK@SU-SCORE.ARPA Line to SAIL
Received: from SU-SCORE by SU-AI with TCP/SMTP; 22 Jul 83 10:01:21 PDT
Date: Fri 22 Jul 83 09:51:20-PDT
From: Len Bosack <BOSACK@SU-SCORE.ARPA>
Subject: Line to SAIL
To: JMC@SU-AI.ARPA
After much discussion with Ma Bell, we think your line is working again.
Len
After a very short experiment, I think it's working pretty well, but it
sems to be dropping occasiona charac and skipping around. e errors
in tis message are not mine. Sometimes a carriage return causes it
to go up rather than down. I haven't seen this kind of malfunction
before.
∂22-Jul-83 1013 GOLUB@SU-SCORE.ARPA Re: DEC Changes
Received: from SU-SCORE by SU-AI with TCP/SMTP; 22 Jul 83 10:13:06 PDT
Date: Fri 22 Jul 83 09:09:00-PDT
From: Gene Golub <GOLUB@SU-SCORE.ARPA>
Subject: Re: DEC Changes
To: FRIEDLAND@SUMEX-AIM.ARPA, Rindfleisch@SUMEX-AIM.ARPA
cc: HPP-Exec@SUMEX-AIM.ARPA, Ullman@SU-HNV.ARPA, JMC@SU-AI.ARPA, RPG@SU-AI.ARPA,
REG@SU-AI.ARPA, CSL.JLH@SU-SCORE.ARPA, Cheriton@SU-HNV.ARPA,
Lantz@SU-HNV.ARPA
In-Reply-To: Message from "Peter Friedland <FRIEDLAND@SUMEX-AIM>" of Fri 22 Jul 83 00:35:05-PDT
There is a similiar story in Thursday's NYT. GENE
-------
∂22-Jul-83 1120 wallace@Whitney AI on space station
Received: from SU-WHITNEY by SU-AI with PUP; 22-Jul-83 11:19 PDT
Received: by Whitney with TCP/SMTP; 22 Jul 83 11:20:26 PDT (Fri)
Date: Fri, 22 Jul 83 11:20 PDT
From: Richard Wallace <wallace@Whitney>
Subject: AI on space station
To: jmc@su-ai
The problem with giving a concrete list of space station functions
and means of achieving them is that:
The space station has 13 major subsystems. The Environmental
Control and Life Support System is one of these. It has six major
sub-subsystems. One of these is Water Restoration. It has seven
sub-sub-subsystems. Giving the list is easy, giving the means
is not easy since
I have 5 weeks to say something meaningful to NASA about how AI
can be used on the space station. I gave the HAL simile because it
is a picturesque way of saying what AI might look like on a space
station (e.g. fault diagnosis, game-playing with crew, vision and
speech understanding). Exactly how AI will be applied to the
various subsystems is a question for the engineers who actually
design these subsystems, wouldn't you agree?
∂22-Jul-83 1331 BOSACK@SU-SCORE.ARPA Re: It needs work.
Received: from SU-SCORE by SU-AI with TCP/SMTP; 22 Jul 83 13:31:23 PDT
Date: Fri 22 Jul 83 13:24:14-PDT
From: Len Bosack <BOSACK@SU-SCORE.ARPA>
Subject: Re: It needs work.
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Fri 22 Jul 83 11:52:00-PDT
We have more of Ma Bell's finest working on the problem.
Len
-------
∂22-Jul-83 1351 Janet.Asbury@CMU-CS-A Re: two documents
Received: from CMU-CS-A by SU-AI with TCP/SMTP; 22 Jul 83 13:51:22 PDT
Received: from [128.2.254.192] by CMU-CS-PT with CMUFTP; 22 Jul 83 16:47:36 EDT
Date: 22 July 1983 1648-EDT (Friday)
From: Jan.Asbury <Janet.Asbury@CMU-CS-A>
To: John McCarthy <JMC@SU-AI>
Subject: Re: two documents
CC: Janet.Asbury@CMU-CS-A
In-Reply-To: "John McCarthy's message of 22 Jul 83 12:37-EST"
Message-Id: <22Jul83.164822.JA61@CMU-CS-A>
Dear Dr. McCarthy,
Professor Reddy did receive the document by U.S. mail. I would imag
The last line of your message was cut off. What I received ends
"I would imag". I also received an acknowledgement from him. Also
please MAIL me a complete U.S. mailing address for the Centre Mondial.
∂22-Jul-83 1421 Raj.Reddy@CMU-CS-A Re: two documents
Received: from CMU-CS-A by SU-AI with TCP/SMTP; 22 Jul 83 14:20:55 PDT
Received: from [128.2.254.192] by CMU-CS-PT with CMUFTP; 22 Jul 83 17:17:39 EDT
Date: 22 July 1983 1718-EDT (Friday)
From: Raj.Reddy@CMU-CS-A
To: John McCarthy <JMC@SU-AI>
Subject: Re: two documents
In-Reply-To: "John McCarthy's message of 22 Jul 83 12:37-EST"
Message-Id: <22Jul83.171816.RR29@CMU-CS-A>
John,
Yes. I have received the messages.
I will be in Paris July 28 and 29.
Will be in touch from there.
Raj
∂22-Jul-83 1607 LB it seems to work now
The phone company adjusted one of the repeaters and now it seems to work well.
This message was typed at your terminal and (other than my own typing errors)
everything seems to work without error.
Let me know how it works for you.
Len
∂22-Jul-83 1612 DFH
IJCAI Travel Arrangements
Aug 6 SFO/NYC/Paris PA201 lv 8:30am ar.NY 4:45pm
PA114 lv NY 6:35 pm ar. 7:25am on Aug 7
Aug 8 Paris/Frankfurt Lufthansa 117 lv.6:55pm ar. 8:05pm
will request transfer from Frankfurt to Karlsruhe
Hilton Hotel in Karlsruhe through night of Aug. 12
Aug 13 To Erlangen. I have made no arrangements.
Aug 16 Frankfurt/NYC/Boston PA67 lv. 11:00am ar. NY 1:20pm
PA532 lv.NY 3:15pm ar. Boston 4:17 pm
Aug 17 Boston/SFO TWA 61 lv. 7:10pm ar. 10:11pm
1. Let me know if this schedule looks OK to you. Note that if you want
to maximize use of Pan Am, for the Boston/SFO leg you could probably
use them by going through NY and changing planes. (Incidentally, you
do need to complete the form and pay them--partially completed form
in your in-box).
2. I am assuming that part of the air fare will be charged to III (SFO/
Boston one-way) and the Karlsruhe portion to ARPA. Is Centre
Mondial covering a portion of it also?
∂22-Jul-83 1710 minker.umcp-cs@UDel-Relay Non-Monotonic Panel Session
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 22 Jul 83 17:10:36 PDT
Date: 22 Jul 83 17:32:34 EDT (Fri)
From: JACK MINKER <minker.umcp-cs@UDel-Relay>
Return-Path: <minker.umcp-cs@UDel-Relay>
Subject: Non-Monotonic Panel Session
To: Moore@SRI-AI
Cc: "uw-beaver!ubc-vision!Reiter"@LBL-UNIX, Mcdermott@YALE, JMC@SU-AI
Via: UMCP-CS; 22 Jul 83 18:23-EDT
Dear Bob,
Knowledge Based systems which invoke the Closed World Assumption
(assuming the Knowledge Base consists of Horn clauses) are non-monotonic.
In addition, when the clauses are non-Horn, a sort of negation by failure
exists as I have shown in a paper that I am sending to you next week. I
refer to this as the Generalized Closed World Assumption.
There are also Knowledge Based systems that do not comply with
either the CWA or the GCWA in the sense that we are willing to accept
negation for all things that we cannot prove except for some items that
we are unsure of. We are , for example, not sure that we ever knew our
friend's address. This also leads to non-monotonicity. It is also not
covered by McCarthy's circumscription. In some recent work with Don Perlis,
we have generalized McCarthy's concept of circumscription to account for
this kind of a situation. If all goes well, I will send you a DRAFT of the
paper over the net.
I hope that this clarifies the terse statement in my outline.
I look forward to receiving your paper.
Best regards,
Jack
I would also like to get copies of the generalization of circumscription
and also the GCWA.
∂22-Jul-83 2340 POURNE@MIT-MC big meeting
Received: from MIT-MC by SU-AI with TCP/SMTP; 22 Jul 83 23:40:10 PDT
Date: 23 July 1983 01:32 EDT
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: big meeting
To: POURNELLE @ MIT-MC, llw @ SU-AI, jmc @ SU-AI
can you perhaps call me relative to big meeting next week. Gen
grahm and others similar importance coming. hyde said llw can
come but have not heard for while.
∂23-Jul-83 1043 CLT Diana's reservations
I need to call Alice before settling the London/Venice/Paris times.
I'll do it this week.
Also, I thought you wanted to stay a week after IFIP?
Thats later than the 24th I assume?
∂23-Jul-83 1143 BOSACK@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 Jul 83 11:43:23 PDT
Date: Sat 23 Jul 83 11:24:13-PDT
From: Len Bosack <BOSACK@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Sat 23 Jul 83 02:26:00-PDT
Yes, PUPFTP should work. Note that most LOTS users can't get off of the
LOTS subnet. Anyone who has an account elsewhere on campus can be
blessed just by asking (there may be a form....).
Len
-------
∂24-Jul-83 2031 LLW@S1-A Pournelle Meeting Revisited
Received: from S1-A by SU-AI with TCP/SMTP; 24 Jul 83 20:31:10 PDT
Date: 24 Jul 83 2030 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Pournelle Meeting Revisited
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
∂21-Jul-83 2138 LLW No Time For Travel
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
∂20-Jul-83 2326 JMC@SU-AI
Received: from SU-AI by S1-A with TCP/SMTP; 20 Jul 83 23:26:42 PDT
Date: 20 Jul 83 2323 PDT
From: John McCarthy <JMC@SU-AI>
To: LLW@SU-AI, RAH@SU-AI
Are you planning to go to Pournelle's meeting?
[John: I'm afraid I'm not going to be able to make it. Lowell]
[[John: Rod has convinced me that we have to go, in spite of the general
wretchedness of the schedule situation. Would you care to ride down and
back with the two of us, so that the three of us could try to get the
Shackleton paper dictated/written in the process? Lowell]]
∂24-Jul-83 2142 KJB@SRI-AI.ARPA test
Received: from SRI-AI by SU-AI with TCP/SMTP; 24 Jul 83 21:42:03 PDT
Date: 24 Jul 1983 2122-PDT
From: Kjb at SRI-AI
Subject: test
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI,
To: Kaplan at PARC, kjb at SRI-AI, lauri at SRI-AI,
To: peters at UTEXAS, rperrault at SRI-AI,
To: sag%SU-psych at SUMEX-AIM, stan at SRI-AI,
To: Wasow%SU-psych at SUMEX-AIM, winograd at PARC
I am trying to use a new <different> terminal without any obvious
backspace so forgive the problems in what follows.
-------
∂24-Jul-83 2143 LEP traffic lights
I have a fairly simple traffic light program running in prolog now. I'm having
trouble with the frame problem, though, and would like to talk to you about it.
When would be a good time for such a talk?
-Leslie
I'll be here only this week. Please come by when you can.
∂24-Jul-83 2151 LLW@S1-A Big Meeting
Received: from S1-A by SU-AI with TCP/SMTP; 24 Jul 83 21:51:17 PDT
Date: 24 Jul 83 2150 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Big Meeting
To: pourne@MIT-MC
CC: LLW@S1-A, RAH@S1-A, jmc@SU-AI
∂22-Jul-83 2340 @SU-AI:POURNE@MIT-MC big meeting
Received: from SU-AI by S1-A with TCP/SMTP; 22 Jul 83 23:40:00 PDT
Received: from MIT-MC by SU-AI with TCP/SMTP; 22 Jul 83 23:40:10 PDT
Date: 23 July 1983 01:32 EDT
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: big meeting
To: POURNELLE @ MIT-MC, llw @ SU-AI, jmc @ SU-AI
can you perhaps call me relative to big meeting next week. Gen
grahm and others similar importance coming. hyde said llw can
come but have not heard for while.
[Jerry: Rod and I are planning to come down Friday PM and stay through
Sunday PM. Have you invited leading people from the `Third Camp'--the
space chemical laser battle station advocates--such as Maxwell Hunter of
Lockheed Missiles & Space Systems? If not, I'd strongly recommend it--it
really seems that, if everyone is to sing in chorus hereafter, we should
all get the same music passed out to them at the same time.
When-and-where is a good number-and-time to call you? Lowell]
∂24-Jul-83 2208 KJB@SRI-AI.ARPA Update
Received: from SRI-AI by SU-AI with TCP/SMTP; 24 Jul 83 22:08:48 PDT
Date: 24 Jul 1983 2156-PDT
From: Kjb at SRI-AI
Subject: Update
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI, Kaplan at PARC,
To: kjb at SRI-AI, lauri at SRI-AI, peters at UTEXAS,
To: rperrault at SRI-AI, sag%SU-psych at SUMEX-AIM,
To: stan at SRI-AI, Wasow%SU-psych at SUMEX-AIM,
To: winograd at PARC
Dear all,
Here is a progress report on the SDF proposal.
CS says that money can start flowing as of July 15, but only
after we get the budget ironed out. There are many problems, but none
that should cause major upheavel.
Brian Smith is now acting computer officer. Terry and Daniel
S have put in some good work getting proposals to Xerox and DEC
started, but we need one person to represent the Center in negotiations
and orchestration until such time as we hire a permanent computer officer.
Hopefully this will happen reasonably soon. Please give Brian all
the help you can. In particular, he has to get a list of what we need
to CS who will be at DEC Tuesday p.m.
CS is insisting that the budexx budget be tied to the program
much more closely thatxx than in the current proposal. For example,
the way the proposal is structure d now, of the 5 parts, (3.2, 3.3,
3.4, 3.5 , 3.6) alxx most effort is going into 3.2 and later into
3.5). We do not have to rewrite the proposal, but we do have to
reorganize the projects under headings to make budget allocations
tied to the projects in recognizable hunks. He wants 3.2 split in
two, 3.3 deleted with projects placed elsewhere, and reorgaization
of 3.4 - 3.5λλλλxx 3.4 - 3.6. I will try to do this in time to get
his reaction next weekend when he returns. I think none of us will
be entirely happy with the results, since it gives us less flexibility
to do what we wnat, but that is the way hx it has to be.
As this implies, things are still up in the air. He still
bx makes noises about deleting this or that project that either
does not make sense to him or got bad press in the referees reports.
This would take the form of explicit workdingxx wording in the letter
of award that they are not supporting work in a given direction.
I hope we can forstall this, but it is a possibility.
We are now aiming for August 15th for getting the official
letter of funding. I assure you that no one is more frustrated than
I am bux by the extra dely (well, maybe some of the associates who
would like to get paid are), but there is nothing to be done, since
the computational enviromnment is still up in the air and needs
reactions from people at PARC. Also, we have to get architect's
estimates on building costs. Also, CS will be ayxx away this next week.
I will undoubtedly have to ax call on many of you for help
in the days to comesxxx come. Thanks for your help -- in advance.
Jon
-------
∂25-Jul-83 0135 POURNE@MIT-MC Big Meeting
Received: from MIT-MC by SU-AI with TCP/SMTP; 25 Jul 83 01:35:02 PDT
Date: 25 July 1983 04:34 EDT
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: Big Meeting
To: LLW @ S1-A
cc: jmc @ SU-AI, RAH @ S1-A
In-reply-to: Msg of 24 Jul 83 2150 PDT from Lowell Wood <LLW at S1-A>
Max will be here. Also Stu Meyer from Huntspatch. Anyone else
from Huntr Camp I shold know of?
I'm home daytimes after noon 213-762-2256
Glad you'll be here. Perhaps something good comes of
all this...
Best,
JEP
∂25-Jul-83 0152 POURNE@MIT-MC meeting
Received: from MIT-MC by SU-AI with TCP/SMTP; 25 Jul 83 01:52:49 PDT
Date: 25 July 1983 04:51 EDT
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: meeting
To: JMC @ SU-AI
In-reply-to: Msg of 23 Jul 83 0033 PDT from John McCarthy <JMC at SU-AI>
Rod & Lowell will be here. Sorry you won't be but I can
understand; next meeting may be different. Howefver, one goal
is to integrate the civil and military programs--given that we
have such and such a military program happening, what does it
NOW make sense to do in civil side. Also Keyworth talking about
national goals. Will miss you.
JEP
∂25-Jul-83 0700 Janet.Asbury@CMU-CS-A Centre Mondial
Received: from CMU-CS-A by SU-AI with TCP/SMTP; 25 Jul 83 07:00:14 PDT
Received: from [128.2.254.192] by CMU-CS-PT with CMUFTP; 25 Jul 83 09:57:18 EDT
Date: 25 July 1983 0958-EDT (Monday)
From: Jan.Asbury <Janet.Asbury@CMU-CS-A>
To: John McCarthy <JMC@SU-AI>
Subject: Centre Mondial
CC: Janet.Asbury@CMU-CS-A
In-Reply-To: "John McCarthy's message of 22 Jul 83 16:24-EST"
Message-Id: <25Jul83.095812.JA61@CMU-CS-A>
I am sorry about the garbled message.
Here is the address you requested:
Centre Mondial
Informatique and Resource Humaine
22, Avenue Matignon
75008 Paris
Phone: 011-33-1-268-1100
Jan
∂25-Jul-83 0941 DFH
∂23-Jul-83 1652 JMC reference
I have mislaid my copy of Daniel Dennett's Herbert Spence lecture,
and I need to refer to it. Please find it or the reference. The
last resort is to telephone the Philosophy Department at Tufts
University and speak to his secretary.
------------------------
I found your copy of the paper. It is: True Believers: The Intentional
Strategy and Why it Works, from SCIENTIFIC EXPLANATION, The Herbert Spencer
Lectures, ed. A. Heath, Oxford University Press, 1981.
Do you want me to mail a copy of the paper to you in Santa Cruz?
--Diana
∂25-Jul-83 1517 DFH
Doug Ferguson of the Library returned your call.
His office no. is 497-9724, or if you can only
call in the evening, his home phone is (415) 526-4081
(Berkeley).
∂26-Jul-83 1013 DFH phone message
1. Ron Ohlander, (202) 694-5051. Wants to know if you could talk to
Defense Science Board here at Stanford Aug. 16 or 17 (I said as
things stood now you had a conflict).
Also needs to talk about proposal issues.
2. Judith Jennings, Univ. of Illinois. Wants to confirm your visit this
fall. (217) 333-6138.
∂26-Jul-83 1244 PKANERVA@SUMEX-AIM About my dissertation
Received: from SUMEX-AIM by SU-AI with PUP; 26-Jul-83 12:44 PDT
Date: Tue 26 Jul 83 12:44:54-PDT
From: Pentti Kanerva <PKANERVA@SUMEX-AIM>
Subject: About my dissertation
To: McCarthy@SAIL
This is to bring you up to date on what is happening with my
dissertation. Pat Suppes (my advisor) and Jaakko Hintikka have read it.
Suppes said that he liked it and that he was going to read it a second
time and then give me detailed comments but that he was not going to
ask for major revisions. Hintikka said that he was ready to sign it
as is but wanted to see first any revisions based on your and Suppes'
comments.
This quarter's deadline for turning in the dissertation is August
11. I still hope to make it, although it will be close. To complicate
matters, Hintikka is leaving for Finland on the 31st of this month, so
getting the final version to him and then the signature page back will
take over a week. I would like to know what rewriting to expect based
on your comments. Also, are you going to be here in August?
I have a related question to you. I did my thesis work all alone
and have no idea of who might be interested in research of this nature.
I am interested in continuing the work, but I could not do it in
isolation. Can you suggest people or organizations I should approach?
- Pentti
-------
∂26-Jul-83 1633 minker.umcp-cs@UDel-Relay AAAI and Lerner
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 26 Jul 83 16:33:44 PDT
Date: 26 Jul 83 13:05:49 EDT (Tue)
From: JACK MINKER <minker.umcp-cs@UDel-Relay>
Return-Path: <minker.umcp-cs@UDel-Relay>
Subject: AAAI and Lerner
To: JMC@SU-AI
Via: UMCP-CS; 26 Jul 83 18:11-EDT
Dear John,
The AAAI has given approval to a session in honor of Alexander
Lerner to be held at 7:00PM August 23 (Tuesday). Would you agree to be on
another panel with me? I would like the panel to be different from the
IJCAI panel. The topic I thought that might be of interest is,
Future Directions in AI
Nils Nilsson has agreed to be on a panel (with an unspecifed title). If you
agree to be on the panel and with the title, I will try to obtain some of the
following individuals:
Marv Minsky
Herb Simon
Alan Robinson.
I would greatly value your thoughts on the topic, participants, and any other
thoughts that you may have. Since time is short, considering that we have
to go to IJCAI, I would appreciate a response from you at your earliest convenience.
Best regards,
Jack
I agree to be on the AAAI panel, but having two such sessions in one
year may be regarded by some as our using Lerner to get unrefereed
papers into the meeting. Therefore, if we do it at AAAI this year,
I would not be agreeable to doing it next year unless there was some
strong new reason.
∂26-Jul-83 1758 reid@Glacier Kelly Roach
Received: from SU-GLACIER by SU-AI with TCP/SMTP; 26 Jul 83 17:58:06 PDT
Date: Tuesday, 26 July 1983 17:57:52-PDT
To: JMC@Sail
Cc: Grosz@SRI-AI
Subject: Kelly Roach
From: Brian Reid <reid@Glacier>
John,
I agree with you that we should be very accommodating about Kelly
Roach. I will send a message to the admissions committee about it. I do
not think that we can act without telling the admissions commitee about
it.
Brian
∂26-Jul-83 1813 @SU-SCORE.ARPA:gardner@Glacier CMOS10 run and wafer prober
Received: from SU-SCORE by SU-AI with TCP/SMTP; 26 Jul 83 18:13:16 PDT
Received: from Glacier by SU-SCORE.ARPA with TCP; Tue 26 Jul 83 18:13:46-PDT
Date: Tuesday, 26 July 1983 18:12:52-PDT
To: shott@Glacier
Cc: gardner@Glacier, sap@Helens, mccarthy@Score
Subject: CMOS10 run and wafer prober
From: Don Gardner <gardner@Glacier>
John et al.,
I was wondering if you would be interested in doing some of
the cmos wafers(contacts)using the Sigma-80? Jim says "It will probably work.
The reservations are that it hasn't been tested for devices and step
coverage", but if we use Amdahl metal and if I can get some time on the
wafer prober, these two issues would be resolved. This brings me to
the second question. I need to use the wafer prober to make those
measurements and others for my paper for the IEDM conference. The
deadline was yesterday, but Krishna says I have until Thursday or
Friday. When can I use it?
cheers,
Don
∂26-Jul-83 1858 GROSZ@SRI-AI.ARPA Kelly Roach
Received: from SRI-AI by SU-AI with TCP/SMTP; 26 Jul 83 18:57:16 PDT
Date: Tue 26 Jul 83 18:53:50-PDT
From: GROSZ@SRI-AI.ARPA
Subject: Kelly Roach
To: jmc@SU-AI.ARPA
John--
I would like to see his folder. Need to do it within the next week
since I leave soon for IJCAI etc. What's the best time for me to
get it and from whom? thanks
Barbara
-------
∂26-Jul-83 2055 sap%Shasta.#Pup@SU-SCORE.ARPA prober
Received: from SU-SCORE by SU-AI with TCP/SMTP; 26 Jul 83 20:55:35 PDT
Received: from Shasta by Score with Pup; Tue 26 Jul 83 20:56:09-PDT
Date: Tuesday, 26 Jul 1983 20:56-PDT
To: : gardner at Glacier
Cc: sap at Shasta, mccarthy at Score, shott at Glacier
Subject: prober
From: Steve Przybylski <sap@Shasta>
As far as I'm concerned, you can use it wednesday.
Steven
McCarthy@SCORE is me. Perhaps you meant MCCARTHY@SIERRA?
∂26-Jul-83 2317 WALKER@SRI-AI.ARPA IJCAI Executive Committee Meeting
Received: from SRI-AI by SU-AI with TCP/SMTP; 26 Jul 83 23:17:30 PDT
Date: Tue 26 Jul 83 23:12:07-PDT
From: Don Walker <WALKER@SRI-AI.ARPA>
Subject: IJCAI Executive Committee Meeting
To: IJCAI-83-Executive-Committee: ;
PLEASE RESPOND IMMEDIATELY IF YOU GET THIS MESSAGE; LET ME KNOW IF YOU CAN COME
You are cordially invited to participate as a member of the Executive
Committee at IJCAI-83. The functions of the Committee are to select
the site of the next conference, elect the Conference and Program
Chairs for that conference, elect a trustee to the IJCAI Board of
Trustees for a three conference term, and discuss other issues that
may arise. Executive Committee members are selected to be broadly
representative of the artificial intelligence community.
The site being proposed for IJCAI-85 is UCLA; a group from the Los Angeles
area will present that bid. The dates tentatively set aside for the
conference are 18-24 August 1985.
The Executive Committee will meet from 12 noon to 1:30 pm on Wednesday,
10 August, at the NANCY Restaurant.
Cordially,
Don Walker, Secretary-Treasurer
International Joint Conferences on Artificial Intelligence, Inc.
-------
I will be at IJCAI and will be glad to attend the Executive Committee
meeting.
∂27-Jul-83 0104 LLW@S1-A In Case You Missed This Earlier. . .
Received: from S1-A by SU-AI with TCP/SMTP; 27 Jul 83 01:04:40 PDT
Date: 27 Jul 83 0105 PDT
From: Lowell Wood <LLW@S1-A>
Subject: In Case You Missed This Earlier. . .
To: jmc@SU-AI
CC: LLW@S1-A
∂24-Jul-83 2030 LLW Pournelle Meeting Revisited
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
∂21-Jul-83 2138 LLW No Time For Travel
To: jmc@SU-AI
CC: LLW@S1-A, RAH@S1-A
∂20-Jul-83 2326 JMC@SU-AI
Received: from SU-AI by S1-A with TCP/SMTP; 20 Jul 83 23:26:42 PDT
Date: 20 Jul 83 2323 PDT
From: John McCarthy <JMC@SU-AI>
To: LLW@SU-AI, RAH@SU-AI
Are you planning to go to Pournelle's meeting?
[John: I'm afraid I'm not going to be able to make it. Lowell]
[[John: Rod has convinced me that we have to go, in spite of the general
wretchedness of the schedule situation. Would you care to ride down and
back with the two of us, so that the three of us could try to get the
Shackleton paper dictated/written in the process? Lowell]]
∂27-Jul-83 1123 @USC-ISI,OHLANDER@USC-ISI Task Proposal
Received: from USC-ISI by SU-AI with TCP/SMTP; 27 Jul 83 11:22:58 PDT
Date: 27 Jul 1983 1121-PDT
Sender: OHLANDER at USC-ISI
Subject: Task Proposal
From: OHLANDER at USC-ISI
To: RPG at SU-AI
Cc: JMC at SU-AI, BSCOTT at SU-SCORE
Message-ID: <[USC-ISI]27-Jul-83 11:21:42.OHLANDER>
Dick,
I have talked to John and Betty Scott about the funding
levels for the new task effort under the umbrella proposal. What
has been sent to me is generally acceptable with the exception of
three items. First of all, the available funding will come in
the following increments: $525K for FY84, $630K for FY85, and
$500K for FY86. You don't have to propose it specifically at
these exact levels for each year, as long as your proposed figures
total out the same. The second issue is that I need a statement
of work for the task. This should consist of short statements of
what work will be accomplished and when. Include milestones such
as demonstrations. The third change that is required is that
some text be put in about benchmark testing of the S-1 running
lisp. This should also be put in the statement of work. As you
should recall, I talked to you about a benchmark task previously.
I will leave it up to you to decide what appropriate benchmarks
are, but we should consider some fairly substantial Lisp
programs. After the changes are made, get the revised task
proposal to Betty Scott for forwarding to us.
Ron P.S.
You may have heard by now that Martin Griss is
remaining at HP, at least for a while. This changes the
complexion of what will be done at Utah in regards to Common Lisp
work. I suggest that you get to me a draft proposal on
development of a portable Common Lisp compiler as soon as you
can. John says he is willing to act as PI as long as you have an
interest in the work. We still might have some informal
collaboration with Griss but that is yet to be worked out if it
happens at all.
∂27-Jul-83 1321 DFH Kelly Roach
To: reid@SU-GLACIER
CC: JMC@SU-AI
Barbara Grosz wants to see Kelly Roach's folder. When
JMC phoned me this morning, he said he did not have it,
and I can find nothing in his office of that nature. Rita
has a message from you indicating that JMC does have the
folder, and I am wondering what this information is based
on. In fact, I am wondering if such a folder actually
exists, as Rita never made one. Grosz needs to see the
folder this week, so I would appreciate a response with
any information you might have. If there is no folder,
should I forward her the recommendation and resume that
I mailed you this morning?
(I believe I mentioned in the previous message that JMC was
in Santa Cruz for the week).
--Diana
∂27-Jul-83 1409 reid@Glacier Re: Kelly Roach
Received: from SU-GLACIER by SU-AI with TCP/SMTP; 27 Jul 83 14:07:01 PDT
Date: Wednesday, 27 July 1983 14:06:50-PDT
To: Diana Hall <DFH@SU-AI>
Cc: reid@SU-GLACIER, JMC@SU-AI
Subject: Re: Kelly Roach
In-Reply-To: Your message of 27 Jul 83 1321 PDT.
From: Brian Reid <reid@Glacier>
JMC left a message on my telephone answering machine which led me to
believe that he has the application. He never actually stated that he
had it, but I drew that inference. Rita would know for sure whether or
not a formal application actually exists.
Brian
∂27-Jul-83 1601 DFH Flight/motel arrangements; Roach
I am sending this in case you have Carolyn read you messages from home.
Have reserved a seat for you on PSA 306 lv. San Jose 6:40pm ar. Burbank 7:34pm
on Friday, July 29. Next earlier PSA flight was at 3:21 pm.
Reserved a room at The Chalet Lodge, 19170 Ventura Blvd., Tarzana, phone
(213)-345-9410. This was the motel name I got from (I guess) Pournelle's wife.
There does not appear to be an actual application or admissions folder for
Roach. I am passing on his resume and pertinent part of recommendation
letter to Barbara Grosz. I am also giving a copy of the information to Rita
because until today she had never heard of him.
∂28-Jul-83 0901 BLUM@SUMEX-AIM Vibrio Cholerae
Received: from SUMEX-AIM by SU-AI with PUP; 28-Jul-83 08:56 PDT
Date: Wed 27 Jul 83 19:06:41-PDT
From: Robert Blum <BLUM@SUMEX-AIM>
Subject: Vibrio Cholerae
To: jmc@SAIL
cc: shortliffe@SUMEX-AIM, blum@SUMEX-AIM
John,
I'm going to forward your msg to Ted Shortliffe. I don't currently
have access to MYCIN, but perhaps someone in Ted's shop can run your
Cholera case thru MYCIN and explain the source of its therapeutic
recommedation.
I would guess that it doesn't have any. rules explicitly abt.
cholera and simply pulled the tetracycline recommendation from an
organism/drug sensitivity table.
Bob (Reply to Blum@SUMEX)
-------
∂28-Jul-83 0901 SHORTLIFFE@SUMEX-AIM Re: Vibrio Cholerae
Received: from SUMEX-AIM by SU-AI with PUP; 28-Jul-83 08:57 PDT
Date: Wed 27 Jul 83 23:13:43-PDT
From: Ted Shortliffe <Shortliffe@SUMEX-AIM>
Subject: Re: Vibrio Cholerae
To: BLUM@SUMEX-AIM
cc: jmc@SAIL
Office: Room TC-117, Stanford Med Center; Phone: (415) 497-6979
In-Reply-To: Your message of Wed 27 Jul 83 19:06:43-PDT
Bob (and John),
One of the reasons we stopped working on MYCIN in the late '70s
was the overwhelming task remaining if we were to broaden its knowledge
of infectious diseases beyond bacteremia and meningitis. The rule set
was tuned only for those two categories of disease, and it was evaluated
only for its performance in treating septicemia (bacteremia) and
meningitis.
Cholera is, of course, a gi/diarrheal infection that does not
present primarily as either bacteremia or meningitis. As a result, there
are no specific rules in the system about clinical cholera and it is
not surprising that the program failed to handle a case of cholera well.
Apparently the organism (vibrio-cholerae) was known because someone
methodically filled in the entire organism table, but I wouldn't be
surprised if no one has every previously tried to run a cholera case
through the system. I don't know why the program chose tetracycline --
perhaps MYCIN has that on the list of drugs that should be used to treat
vibrio meningitis, but since I don't know if such a disease even exists,
I have no idea how that particular drug was selected. We could look into
it, but the answer would not be particularly illuminating. The best
bottom line answer is that MYCIN doesn't know anything much about cholera
or vibrio cholerae, and that its poor performance on such a case is
therefore to be expected.
In testing the system, I'd suggest that you always use meningitis
or bactermia cases. It'll make dumb errors on most other disease
entities. Incidentally, Bob, everyone on SUMEX has access to MYCIN.
Just type @MYCIN or @<MYCIN>MYCIN.
Cheers,
Ted
-------
∂28-Jul-83 0915 KJB@SRI-AI.ARPA Assistant Director of CSLI, etc.
Received: from SRI-AI by SU-AI with TCP/SMTP; 28 Jul 83 09:13:38 PDT
Date: 28 Jul 1983 0907-PDT
From: Kjb at SRI-AI
Subject: Assistant Director of CSLI, etc.
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI,
To: Kaplan at PARC, kjb at SRI-AI, lauri at SRI-AI,
To: peters at UTEXAS, rperrault at SRI-AI,
To: sag%SU-psych at SUMEX-AIM, stan at SRI-AI,
To: Wasow%SU-psych at SUMEX-AIM, winograd at PARC
Dear all,
I have decided to hire Dr. Betsy Maken to be the assitant
director. She is now Vice President for Curriculum and Publications
at CCC (Suppes company) where she has managed many large research/
development projects. She has degrees in mathematics, statistics
and mathematical education. She has written 15 papers, some in the
area of syntax and semantics. She has also worked extensively with
computers and people using computers. Pat Suppes is reluctant to
lose her but thinks she and this project are made for each other.
We are in the proecess of getting cost figures out of the
university on building plans. This involved Joyce and I coming up
with two shopping lists of space needs, one what we want, a second
of what we could get by with. Believe it or not, we are already very
short of space. Ventura is full, or at least those parts of it that
are for CSLI are. We will not be able to make any space allocations
in Casita until we have some idea as to the extent to which it will
be replaced or remodeled. On a brighter note, it turns out that
there is are a couple of rarely used classrooms near Ventura that
will help with our larger semanars and colloquia.
Brian is making excellent progress on the computing front.
The projects are being reorganized as of now under 4
headings.
A. Evaluate and further develop linguistic theories in light of
constraints from computation, semanics and language use
B. Develop theories of language use
C. Understanding theories and practice of computing within
theories of language use
D. Foundations
Except in C, where Brian and I are negotiating, no projects have
been cut, but this has not been past Charlie, who is out of town.
My messages to Tom and Ivan are returned as undeliverable.
If any of you see them, or have ways of forwarding mssages to them,
please keep them posted.
Jon
-------
∂28-Jul-83 0942 LLW@S1-A Pournelle Metting Travel Arrangements
Received: from S1-A by SU-AI with TCP/SMTP; 28 Jul 83 09:40:34 PDT
Date: 28 Jul 83 0236 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Pournelle Metting Travel Arrangements
To: dfh@SU-AI
CC: LLW@S1-A, RAH@S1-A, jmc@SU-AI
∂27-Jul-83 1345 DFH@SU-AI Pournelle meeting
Received: from SU-AI by S1-A with TCP/SMTP; 27 Jul 83 13:44:57 PDT
Date: 27 Jul 83 1345 PDT
From: Diana Hall <DFH@SU-AI>
Subject: Pournelle meeting
To: llw@S1-A
CC: rah@S1-A
JMC is teaching a class at UCSC this week. In response
to your message about a ride to the above meeting, he
will have to fly down because of schedule in Santa Cruz,
but would be happy to ride back up with you.
--Diana Hall, secy. to JMC.
[Diana: We can shift around departure times considerably to accommodate
John's needs, and could even pick him up somewhere on the San
Jose-Watsonville leg of 101 on the way down, if this would work out.
Perhaps it would be best if he could give me a ring so that we could to to
mesh travel arrangements--(415) 422-7286 is my office number. If he just
can't possibly arrange to ride down with us (and get our paper half-done),
we'll certainly be glad to have him come back with us Sunday evening.
Lowell]
∂28-Jul-83 1557 minker.umcp-cs@UDel-Relay Non-Monotonic Logic
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 28 Jul 83 15:56:07 PDT
Date: 28 Jul 83 16:09:17 EDT (Thu)
From: JACK MINKER <minker.umcp-cs@UDel-Relay>
Return-Path: <minker.umcp-cs@UDel-Relay>
Subject: Non-Monotonic Logic
To: "uw-beaver!ubc-vision!Reiter"@LBL-UNIX, Moore@SRI-AI, Mcdermott@YALE,
JMC@SU-AI
Via: UMCP-CS; 28 Jul 83 18:15-EDT
Below is a PRELIMINARY DRAFT of a paper thar Don Perlis and I
are writing. The paper is not yet completed. We are sending you the DRAFT
since it is relevant to our panel session at IJCAI. We would welcome any
comments that you may have on the material. Ilook forward to our panel session.
(D R A F T)
On the Semantics of Circumscription
J. Minker and D. Perlis
Computer Science Department
University of Maryland
College Park, MD 20742
July 1983
I. Introduction
We investigate the model theory of the notion of
circumscription (McCarthy [1980]), and find a completeness theorem that
provides both a generalization and its converse, to a result of
McCarthy. In particular, we extend the idea of circumscription to
allow prescription of what objects are or are not to be included in the
circumscription process, broadening the applicability of the technique.
Then we show that the circumscriptive theorems are precisely the truths
of the minimal models, in the case of 'finitary' theories.
As an example, consider the following problem: Someone asks
you whether you have ever known the phone number of a movie star. You
pause only a split second before answering 'No.' Later, on being asked
whether you have ever known the phone number of your uncle in
Chattanooga, you hesitate, frown, and end up saying that you are not
sure.
We apparently circumscribe on a movie star's number, but not on
an uncle's. That is, if our data base does not contain the fact that
we have ever known a star's number, we conclude we have not ever so
known; but we are not so hasty in other cases. Some things we realize
may well be true (we may have once known Uncle's number after all) even
if we cannot get our hands on them. So we would like a kind of
built-in 'protection' for the uncertainty of certain things while still
circumscribing over others. Indeed, it would seem that this is the
more common condition, for rarely outside of artificial problems like
the missionary-cannibal one of McCarthy, is it wise to assume we know
all that is relevant about a given matter. Of course, as McCarthy
points out, circumscription may be useful for a first pass at a problem
and later discarded or revised. But at some point the kind of
protected circumscription we urge here becomes relevant.
As a further motivation, suppose a database DB is given, and
that as is usual the queries Q that are answerable affirmatively are
the ones that are true with respect to an intended real-world model,
with the exception of certain queries regarding items that we know have
not been specified completely yet in DB. E.g., we may know data is
still being gathered on these items, such as, say, incomes of
middle-level management in a large company, while all the other entries
in DB may be complete. We may wish to reason about DB assuming that all
data is known (closed world assumption) except for these incomes.
II. The idea of circumscription
We review briefly the idea of circumscription. Given a
predicate symbol P and a formula A[P] containing P, the circumscription
of P by A[P] can be thought of as saying that the P-things consist of
certain ones as needed to satisfy A[P] and no more, in the sense that
any P-things Z satisfying A[Z] already include ALL P-things:
P
C [Z]: [A[Z] & (x)(Z(x)-->P(x))] --> (x)(P(x) --> Z(x)),
A
or alternatively
[A[Z] & (x)(Z(x)-->P(x))] --> (x)(-Z(x) --> -P(x)),
which emphasizes the use of circumscription to prove something is NOT a
P-thing.
Reiter, in unpublished work, has shown that the following
schema is equivalent to McCarthy's:
[A[P&Z] --> (x)(P(x)-->Z(x)),
which can be interpreted intuitively as saying that any strengthening
of P, such as P&Z, which conforms to the axioms A under consideration
for defining P, is no real strengthening: P already entails Z. This
is a very satisfying form for it incorporates the puzzling second
antecedent of McCarthy's schema in a natural version of the first.
A key example, emphasized by McCarthy, is the following: let
A[P] be P(a) v P(b). Let Z1(x) be x=a and Z2(x) be x=b. Then from
P(a) v P(b) we get that either Z1 or Z2 can serve for circumscription,
i.e., either P(x) --> Z1(x) or P(x) --> Z2(x). Thus either a is the
only P-thing, or b is: (E!x)(P(x) & (y)(y=/=x --> -P(y)). We will use
this idea later.
We point out the following lemma, that is instructive and is
related to later results:
Lemma 1: Let A[P] be -P(a) v Q. Then -P(a) is a result of
circumscription of P by A[P]. So -P(a) v Q is redundant,
since -P(a) is already a result of circumscription from the
empty set.
Proof: Let Z(x) be x=/=a & P(x). Then it is easy to see that A[Z]
holds. Furthermore, Z(x) --> P(x) trivially. Thus by
circumscription we have P(x) --> Z(x) and so we get P(x) -->
x=/=a, i.e., -P(a). <>
Note that in Reiter's form mentioned above, this is especially
suggestive, in that we simply let Z(x) be x=/=a and check that A[P&Z]
holds.
Now, an obvious reply to the above problem is that if we are
thinking of circumscribing the predicate symbol P in an axiom A, except
that we wish to protect the uncertainty of certain items, say those
with property S, then we should really be circumscribing the predicate
P(x) & -S(x), and then simply follow McCarthy. However, McCarthy's
technique requires that a single predicate symbol be used, not a
non-atomic formula. If we try to accomodate to this by introducing a
new symbol, say Q, with the additional axiom Q(x) <--> P(x) & -S(x),
then we will have immediate difficulties, since it is no longer clear
what axioms A to circumscribe Q in: if we include the defining formula
for Q then any circumscribing predicate substitution Z for Q simply
leads to the need to prove what we want to conclude, i.e., Z <--> Q.
And if we leave it out, then we have nothing to work on. Now of course
we can carefully analyze whatever axioms A had already been chosen for
circumscribing P in the first place, and re-tailor them for the new
situation protecting S. But this is more work, and it is not clear
what a general method for this would be.
III. Circumscription with protected terms
Here we show that a simple syntactic device will yield the
desired result in such cases. We suggest that once A has been selected
as appropriate for circumscribing P, and if (perhaps later) it is
desired to protect S-things from this process so that circumscription
will not be used to show S-things are not P-things, we can keep the
same criteria A, but alter the form of the schema itself. Starting
with P(x) & -S(x) as above, which we write P/S(x) (and more generally
T/U(x) for T(x) & -U(x)), we alter the circumscription schema to read
as follows:
P/S
C [Z]: [A[Z] & (x)(Z/S(x)-->P(x))] --> (x)(-Z/S(x) --> -P(x))
A
for all predicate symbols Z. (Here -Z/S means (-Z)/S and not -(Z/S).)
Intuitively, we are saying that conclusions are drawn only about
non-S-things, as far as ruling out possible P-things goes. Notice that
the conclusion is equivalent to (x)(P/S(x) --> Z(x)). We refer to this
schema as protected circumscription; unless so indicated,
circumscription will refer to McCarthy's schema. We write C[Z] when
context makes clear what the A, P, and S (if protected) are.
We consider some examples. Suppose HKP(x) says, "have known
the phone number of x." We would like to stipulate as a first
approximation that if we cannot determine that we have known x's phone
number, then in fact we have not so known. This might be accomplished
by stating that HKP is circumscribed by the formula A[HKP]:
HKP(a)&HKP(b)&HKP(c)&...&HKP(n) where a,b,...,n are particular
individuals whose numbers we do recall having known. Now suppose we
want to alter this so that it applies only to famous people (F(x)) or
lovers (L(x))--not uncles, etc. That is, we expect to recall if we
have ever known the number of a movie star just as we would recall if
we have not learned the number of our lover, but in other cases we are
often unsure. We then set up the protected circumpscription schema:
[A[Z] & (x)(Z/S(x)-->HKP(x)] --> (x)(-Z/S(x) --> -HKP(x))
where S(x) <--> -F(x) & -L(x).
Consider how this works now for a movie star s. Assume in fact
that HKP(s) is not in our database. Then we circumscribe on the
predicate Z(x): (x = a)v(x = b)v...v(x = n). Since A[Z] is easily
verified in the database, and since in this case in fact Z(x) -->
HKP(x), we obtain (x)(-Z/S(x) --> -HKP(x). Hence, instantiating s and
assuming F(s) to be in the database, we end up with -Z(s) --> -HKP(s).
But since -Z(s) (recall s is not any of a,...,n), we conclude that we
have never known s's phone number. (Admittedly this is not the
split-second reasoning of our opening motivational
statement--circumscription seems to share, with other non-monotonic
modes of inference, a rather substantial computational burden.
However, we will see later that in a precise sense the only instances
of Z that are of significance are those as above: disjunctions of the
form x=a v ... v x=n, providing a guide of sorts to the calculations.
Consider a variation on this. Suppose we recall having once
known the phone number of a movie star, but cannot recall who. Thus
(Ex)(HKP(x)&F(x)). We circumscribe HKP by this formula as well as the
above HKP(a)&...&HKP(n), i.e., A is now the conjunction of these two
formulas. Again, we want to protect those who are not famous or
lovers, so that we will not conclude we have not known x's phone number
merely because we cannot recall so, if x is not 'noteworthy'. Let s be
a Skolem constant for (Ex)(HKP(x) & F(x)). This time, we use for Z(x)
the formula (x = s)v(x = a)v...v(x = n). Then A[Z] holds, for
it is
(Ex)((x=s) v (x=a) v ... v (x=n) & F(x))
&
(a=y) v (a=a) v ... v (a=n)
&
.
.
.
&
(n=y) v (n=a) v ... v (n=n)
where we know x=s satisfies the first conjunct, and the rest are
tautologies.
Now we know HKP(s)&F(s); from this we can conclude (x)(Z/S(x)-->HKP(x))
since the latter really amounts to the conjunction of
F(s)vL(s) --> HKP(s)
F(a)vL(a) --> HKP(a)
.
.
.
F(n)vL(n) --> HKP(n)
where we already know all the right-hand sides. Then by protected
circumscription we get (x)(-Z/S(x) --> -HKP(x)). We existentially
quantify this, to get
(Ey)(x)([x=/=y & x=/=a &...& x=/=n & (F(x)vL(x))] --> -HKP(x)),
ie, we have known the phone number of at most one famous person or
lover different from a,...,n, and in particular then of at most one
famous person different from a,...,n.
IV. Models
We must re-define minimal model in an manner appropriate to the
new version of circumscription. Let M and N be models of A[P]. We say
M <P/S N if the atomic truths of M are contained in those of N, if
those atomic truths of M not using P are precisely those of N, and if
the extension of P&S in M is also that in N. Then M is a P/S-minimal
model of A[P] if M is a model of A[P] minimal with respect to the
relation <P/S.
As an example, suppose P(a)&P(b)&P(c)&-P(d)&Q(d) is the
sentence A[P], and we wish to protect the constant c: S(c). Then the
only model is {P(a) P(b) P(c) Q(d) S(c)}. (Here we indicate a model by
writing the positive ground clauses that hold in it.) This model is
the only minimal model. In this case protection is superfluous since
P(c) is required to hold.
Now consider the sentence P(a)&P(b) where c is still a
protected constant--S(c)--and d is an unprotected constant. Here we
obtain four models:
M1={P(a) P(b) P(c) P(d) S(c)}
M2={P(a) P(b) P(c) S(c)}
M3={P(a) P(b) P(d) S(c)}
M4={P(a) P(b) S(c)}.
Of these only M2 and M4 are minimal, M2 being a P/S-minimal submodel of
M1, and M4 of M3.
Finally, consider P(a) v P(b) v P(c) with S(a) and S(c). Then
the models are
M1={P(a) P(b) P(c) S(a) S(c)}
M2={P(a) P(b) S(a) S(c)}
M3={P(a) P(c) S(a) S(c)}
M4={P(b) P(c) S(a) S(c)}
M5={P(a) S(a) S(c)}
M6={P(b) S(a) S(c)}
M7={P(c) S(a) S(c)}.
The minimal ones are M3, M5, M6, M7.
V. Completeness results
McCarthy [1980] provides the soundness half of a model theory
for circumscription, but not the completeness part. As noted by Davis
[1980] (see below) the fully general completeness result would be
false. Nonetheless, Minker [1982] has a soundness and completeness
result for the highly analogous situation of generalized closed world
assumptions. Here we look for a parallel to Minker's work in
restricted cases, and obtain both soundness and completeness there for
circumscription and protected circumscription.
We begin by demonstrating that our schema C[Z] does provide an
extension of McCarthy's circumpscription for protected terms. First,
we prove the following lemma:
Lemma 2: A[P] |P/S== C[Z], where the P/S-double-turnstile means the
consequent holds in any P/S-minimal model of the antecedent,
and Z and S are any formulas not involving P. (When we intend
ordinary circumscription we will write A[P] |c== B to mean B
holds in any minimal model of A[P], and A[P] |c-- B to mean B
is a circumscriptive theorem of A[P].)
Proof: We follow McCarthy, giving further details to make clear where
there is a difference with our approach.
Let Z be a formula not involving P. Then we must show that the
schema C[Z] holds in an arbitrary P/S-minimal model M. So
assume A[Z] & (x)(Z/S(x) --> P(x)) holds in M but (x)(-Z/S(x)
--> -P(x)) fails in M. Define M' to be the submodel of M where
all atomic formulas P(x) for which -Z/S(x) holds in M are
removed.
We wish to show M' is a model for A[P], contradicting M being a
P/S-minimal model for A[P]. For clearly M' <P/S M, and M' =/=
M. But A[Z] holds in M, so that it also holds in M'. Thus any
logical consequences of A[Z] also hold in M'. Now if A[P]
fails in M' then some consequence P(a)v...vP(n) must fail,
whereas Z(a)v...vZ(n) will hold in M' as a consequence of A[Z].
Moreover, none of a,...,n are in the extension of S (in M or
M') since only elements not in S are removed from the extension
of P in M'. Thus some Z(k) holds in M' while P(k) does not,
nor does S(k). This violates the condition that Z/S(x) -->
P(x) in M. <>
Theorem 1: A[P] |P/S-- B implies A[P] |P/S== B for any formula B.
Proof: Assume the hypothesis. Then any P/S-minimal model M of A[P] is
a model not only of A[P] but also of all the C[Z] in the
schema. But A[P] + {C[Z]} |-- B, so by the usual
completeness theorem of first-order logic, we are done. <>
Unfortunately the converse does not hold, as shown by Davis
[1980]. Let A[=] be Peano arithmetic (with or without the induction
postulate; here '=' is a binary predicate symbol). Then the minimal
models are isomorphic to the natural numbers; but no recursive
first-order theory, including one of the form A[=] + {C[Z]}, has as its
theorems precisely those sentences true of the natural numbers.
However, we shall show that a restricted converse does hold,
which seems more useful than one given by Davis. Namely, we show that
if a theory A[P] has a finite Herbrand universe (no function letters),
then the converse holds: if B holds in all (P/S-)minimal models of
A[P] then B is a theorem by circumscription of P in A[P]. This we then
generalize to certain cases with function letters. In the sequel we
shall assume A[P] to have a finite Herbrand universe unless otherwise
specified; let us call this assumption FHU (finite Herbrand universe),
which we will append to 'Lemma' or 'Theorem' to indicate such an
hypothesis. We first prove our results for ordinary (unprotected)
circumscription, and then observe that the proofs carry over for our
version as well.
We will proceed by induction on the number of quantifiers and
connectives in B. It will turn out that negation is the only difficult
case, and then only when applied to a single atom P(a). We will
attempt to use Z(x) as x=/=a in the complementary form x=c1 v x=c2 v
... v x=ck where the cj are all the Herbrand constants other than a.
This is where our assumption of finiteness comes in. In fact the
actual situation will be slightly more complicated, but suggested by
these comments.
Lemma 3: (FHU) If A[P] |c== P(a) then A[P] |-- P(a)
Proof: If P(a) is true in all minimal models of A[P] then since any
model M of A[P] has a minimal submodel N (recall the finite
Herbrand universe assumption) we have N |== P(a) and thus M |==
P(a): N can differ from M only in having possibly fewer P's,
never new ones. Finally, by completeness of first order logic,
A[P] |-- P(a).
Lemma 4: (FHU) If A[P] |c== (x)P(x) then A[P] |-- (x)P(x)
Proof: Same as for Lemma 3.
Lemma 5: (FHU) If A[P] |c== -P(a) then A[P] |c-- -P(a)
Proof: We will define a variation on disjunctive normal form: we say
F is a P-DNF formula if it is a (possibly universally
quantified) disjunction whose disjuncts are conjuncts of either
literals P(t) or -P(t) or subformulas Q not containing P.
Clearly any formula is equivalent to a P-DNF formula: we
Skolemize existential quantifiers in the prenex form and then
change CNF to DNF in the standard manner.
Let F be a P-DNF formula equivalent to A[P]. (Note for later
use that at least one of the Q's must hold if A[P] does.) Let
F' be the formula resulting from F by eliminating any disjuncts
that subsume other disjuncts, i.e., if D1 and D2 are disjuncts
of F and D1 --> D2, then remove D1. Note that F' is equivalent
to F and to A[P].
We show first that each disjunct Dc of F', say
P(c1)&...&P(ck)&-P(c'1)&...&-P(c'j)&Qc, determines a minimal
model of A[P]. For let M be the structure for A[P] with true
atoms P(c1),...,P(ck) as well as those entailed by Qc. Then M
is a model of A[P], and M is minimal since if any P(ci) could
be removed from M and still leave a model, then the disjunct Dc
would subsume another: each model must satisfy some disjunct.
(This shows that conversely every minimal model corresponds to
a disjunct. As a consequence, there are only a finite number
of minimal submodels of any model M of A[P]--viewed as sets of
true atoms--though possibly an infinite number of models M.)
Now if P(a) appears as a conjunct in any disjunct of F' then
P(a) will be true in some minimal model by the above argument.
So we conclude that P(a) does not so appear, assuming the
hypothesis of our Lemma, ie that -P(a) holds in each minimal
model of A[P].
To finish the proof we will show that P(a) cannot hold in any
model of A[P] + {C(Z)}. We will find formulas which will serve
for use in circumscribing P and will lead to the conclusion
-P(a). Again consider F' as before. For each disjunct Dc, ie,
P(c1)&...P(ck)&-P(c'1)&...&-P(c'j)&Qc
form
Zc(x): x=c1 v ... v x=ck .
The formulas Zc(x) will be our candidates for circumscription
of P in A[P]. Let us call the set of terms ci that appear in
any of the Zc(x) the P-requirement of A[P], or P-REQ for short.
Note that given Dc, the term a is provably distinct from
the terms c1 to ck since ci=/=a holds in each minimal model
of A[P] (otherwise since P(ci), then P(a) would hold there) and
thus also ci=/=a in each model of A[P]. (Actually, if P(x) is
the predicate x=/=a, then this argument is inadequate, but then
P(ci) gives ci=/=a directly.)
We first observe that each disjunct Dc(x) of F entails that for
the corresponding disjunct Zc(x), Zc(x) --> P(x). This is
because for all the terms ci in Zc(x), P(ci) is already a
conjunct of Dc(x): that is how we got Zc(x) in the first
place. So then we have that F (and therefore A[P]) entails the
disjunction of the formulas
Zc(x) --> P(x),
and specifically Dc entails Zc(x) --> P(x).
Now we observe that at least one of the A[Zc] holds given A[P].
Specifically, each disjunst Dc of A[P] entails A[Zc]. To show
this, first note that the P-DNF formula F, when all occurrences
of P are replaced by Zc, yields a result equivalent to A[Zc]
and also each disjunct Dc in F, i.e., P(c1)&...&P(ck) &
-P(c'1)&...&-P(c'j) & Qc, is thereby replaced by
[c1=c1 v ... v c1=ck] & ... & [ck=c1 v ... v ck=ck]
&
[c'1=/=c1 & ...& c'1=/=ck] & ... & [c'j=\=c1 & ... & c'j=\=ck]
&
Qc. (Note that each ci is in P-REQ.)
Now for any true disjunct Dc of F, Qc must be true. But then
the altered form of this disjunct as we have written just above
follows from Dc since none of the c'i's is in P-REQ (assuming
A[P] consistent), and Qc was already true. This shows that
A[Zc] holds if Dc does.
Now we have the ingredients for performing the circumscription
of P in A[P]. We know that at least one of the conditionals
Zc(x) --> P(x)
holds, as well as one of A[Zc], and in fact any Dc entails both
of these. Since one of the Dc holds (F is their disjunction)
then for some c we get both A[Zc] and Zc(x) --> P(x). This is
all we need for circumscription, and the conclusion is that
P(x) --> Zc(x) for some c, i.e., in each model M of A[P] there
will be a true Dc, and also P(x)-->Zc(x) holds there. Further,
a=/=ci in such M, since Dc entails P(ci) for each i, yet since
-P(a) holds in a minimal submodel N of M, then also -P(a) in M.
Since a is not among c1,...,ck then Zc(x) clearly forces x to
be distinct from a:
[x=c1 v ... v x=ck] & [x=/=c'1 & ... & x=/=c'j] & Qc.
Therefore P(x) forces x=/=a also, and the proof is complete. <>
Now we can establish our first completeness theorem:
Theorem 2: If A[P] has a finite Herbrand base, then for all
sentences B, A[P] |c== B iff A[P] |c-- B.
Proof: The left-to-right order of entailment we have already shown in
Theorem 1. Let us proceed to the converse.
By Lemma 4 above, the converse is true when B is an atom P(a)
or the formula (x)P(x). If B is instead -P(a) then by Lemma 5
again the converse is true. If B is some other formula having
one or fewer connectives, then it must not contain the letter
P. So again the converse is true since B will then hold not
only in minimal models but also in all models.
This shows that any sentence B with one or fewer connectives
and quantifiers satisfies the theorem. We proceed by induction
on the number of connectives and quantifiers in B. Assume the
theorem holds for all sentences with n or fewer connectives and
quantifiers. Let B have n+1 connectives and quantifiers. Then
B is of one of the forms
1) C v D
2) C & D
3) C --> D
4) -C
5) (x)C(x)
Now we may assume that v and - are being used as primitive
connectives which serve to define & and -->. Thus we work only
with v, -, and (x). Furthermore, we may assume B is written as
a P-DNF formula so that the case of -C arises only when C is
either P(a) or does not contain P.
Let M be a model of A[P]+{C(Z)}, and N a minimal submodel of M
(guaranteed to exist by the finite Herbrand base). We will
show that if B holds in N then it also holds in M, thereby
establishing the theorem.
If then B is of the form C v D, since C and D will have fewer
than n connectives we may apply the inductive assumption to
conclude that either C or D holds in M, since either C or D
will hold in N.
If B is of the form -C, then either C is of the form P(a) or
does not contain P, and in each case we already know -C holds
in M.
Finally, if B is of the form (x)C(x), then since M and N have
the same domain, we have C(d) holds in N for each d in M, and
therefore also C(d) holds in M by inductive assumption.
With this we are done. <>
This proof provides the justification for our earlier claim
that the only cases of significance for circumscription C[Z] are those
in which Z is a disjunction x=c1 v ... v x=ck. That is, no further
circumscribing will lead to new information. As a consequence we will
have new characterizations of circumscription over theories with finite
Herbrand universes, and, we anticipate, even infinite universes in
certain cases. Reiter [1982] addresses the problem of selecting
appropriate formulas Z, in his work relating circumscription and the
closed world assumption using ideas in Clark [1978]. Here we have
delimited precisely the class of Z's that are useful in general in the
FHU case. McCarthy [1980] uses two examples to illustrate
circumscription and chooses judiciously a disjunction in one and
multiple cases for Z in the other; we have used exactly these
situations to show that all circumscriptive theorems are obtainable
this way, indicating that no further ingenuity is needed in
constructing Z's in such cases. The P-DNF version of A[P] determines
the minimal models and hence the theorems, in analogy with the work of
Reiter and Clark but for arbitrary formulas (not just Horne clauses)
and allowing for possible protected data.
Let us note finally that the above result also holds for the
protected form of circumscription, giving us the following:
Theorem 3: If A[P] has a finite Herbrand base, then for all
sentences B, A[P] |P/S== B iff A[P] |P/S-- B.
Proof: We simply use the same proofs as in the above Lemmas and Theorem
in which it is easily seen that the addition of the 'protection
clause' S(x) in no way affects the arguments, except possibly
in the very last line of the proof of Theorem 2, namely, that
'P(x) forces x=/=a'. We should have here 'P(x) & -S(x) forces
x=/=a,' and then the only new case to handle is that in which
a itself is protected: S(a). But then P/S-minimal models N of M
will have to agree with M on P(a), so automatically we get P(a)
to fail in all models. <>
It is instructive to consider the following example: Let A[P]
consist of the data P(a), -P(b) v -P(c). Then there are three models
of A[P]:
1. {P(a)}
2. {P(a), P(b)}
3. {P(a), P(c)}
Of these, only 1 is minimal, and so the formulas true in 1 are the
circumscriptive theorems of A[P], for all choices of Z at once! Notice
that the theory A'[P] having ONLY P(a) as axiom also has these three
models as well as: 4. {P(a), P(b), P(c)} which still is not minimal.
So A and A' have the same minimal models and hence the same
circumscriptive theorems. In fact in both theories we have the
theorems -P(b) and -P(c), so that the axiom -P(b) v -P(c) in A is
circumscriptively redundant.
Now suppose we wish to protect b and c in A so that ALL we know
about P(b) and P(c) is that they are not jointly true, i.e., -P(b) v
-P(c) represents real uncertainty. Then we find that 1, 2, and 3 are
the only models and all are minimal. Furthermore, although -P(b) v
-P(c) holds in each, neither -P(b) nor -P(c) does, so that the
protection has really worked. But now if we pass to A' and protect b
and c, we find still all four models as before and all are minimal, so
that not even -P(b) v -P(c) holds.
Although the completeness result has shown us what the theorems
of these four theories are, we see from this example that negative data
(-P(b) v -P(c)) can have a non-redundant effect when there are
protected constants. This shows a strong distinction with the situation
in Lemma 1 for ordinary circumscription.
←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←←
We think that these results may generalize to certain cases of
infinite Herbrand bases, and we have tentative conjectures on what such
cases might be.
------------------------------------------------------------------------------
Acknowledgements
Our work obviously depends greatly on that of John McCarthy. We have
also been influenced by work of and discussions with Ray Reiter.
This paper was written with support from the following grants:
AFOSR-82-0303, for J. Minker and D. Perlis
NSFD MCS 79 19418, for J. Minker
Summer Research Award, Univ. of Md. General Research Board, for D. Perlis
Bibliography
Clark, K. Negation as failure, in Logic and Databases, Gallaire and
Minker (eds) (pp 293-322). Plenum Press, NY 1978.
Davis, M. The mathematics of non-monotonic reasoning. Artificial
Intelligence 13, (pp 73-80), 1980.
McCarthy, J. Circumscription--A form of non-monotonic
reasoning. Artificial Intelligence 13 (pp 27-39), 1980.
Minker, J. On theories of definite and indefinite databases.
Tech. Report. Univ. of Maryland
Minker, J. On indefinite databases and the closed-world assumption.
Springer-Verlag Lecture Notes in Computer Science, v.138, pp
292-308. Sixth Conference on Automated Deduction. New
York, NY. 1982.
Reiter, R. A logic for default reasoning. Artificial Intelligence
13 (pp 81-132), 1980.
Reiter, R. On closed world databases. In Logic and Data Bases.
(pp55-76) Gallaire and Minker (eds.) Plenum, 1978.
Reiter, R. Circumscription implies predicate completion
(sometimes). Proceedings of AAAI-82. (pp418-420)
∂28-Jul-83 1603 DFH Roach
I contacted him at Xerox PARC, where he is working this summer.
I requested the graduate admissions office to send application
materials there.
∂28-Jul-83 1936 KJB@SRI-AI.ARPA My hours
Received: from SRI-AI by SU-AI with TCP/SMTP; 28 Jul 83 19:36:44 PDT
Date: 28 Jul 1983 1656-PDT
From: Kjb at SRI-AI
Subject: My hours
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI,
To: Kaplan at PARC, kjb at SRI-AI, lauri at SRI-AI,
To: peters at UTEXAS, rperrault at SRI-AI,
To: sag%SU-psych at SUMEX-AIM, stan at SRI-AI,
To: Wasow%SU-psych at SUMEX-AIM, winograd at PARC
cc: wasow at PARC
Just for your information, I plan to restrict my administrative
duties to the afternoons, so will generally be hard to get hold
of before 1:30 or 2 p.m.
Jon
-------
∂28-Jul-83 2001 LLW@S1-A Travel Arrangements
Received: from S1-A by SU-AI with TCP/SMTP; 28 Jul 83 20:01:21 PDT
Date: 28 Jul 83 1800 PDT
From: Lowell Wood <LLW@S1-A>
Subject: Travel Arrangements
To: clt@SU-AI
CC: LLW@S1-A, RAH@S1-A, jmc@SU-AI
∂28-Jul-83 0959 CLT@SU-AI pournelle mtg
Received: from SU-AI by S1-A with TCP/SMTP; 28 Jul 83 09:58:48 PDT
Date: 28 Jul 83 0956 PDT
From: Carolyn Talcott <CLT@SU-AI>
Subject: pournelle mtg
To: llw@S1-A, rah@S1-A
John would like to ride down with you on Friday, if possible.
He is in Santa Cruz and can leave there at noon.
ETA Livermore about 2pm. He is sans badge so will call from the gate.
He is also basically inaccessible, so could you reply to
me whether or not this is ok? He will call me this evening.
Perhaps you should send also a telephone number where he can
reach one of you this evening or tomorrw am.
- Carolyn -
(CLT@SU-AI)
[Carolyn: I got a net note with a time-stamp of a few minutes ago from
Diana, John's secretary, telling me that he couldn't get away from Santa
Cruz until 1630 tomorrow. I hope that your data is more correct. Meeting
John here at Livermore anytime after 1400 tomorrow (until 1700 or so)
would be fine, as it's about a 5.5 hour drive down to LA; driving down
with him as well as coming back would be great, as I estimate that we'll
need all 10 hours together to get the paper dictated that we're overdue in
finishing. I'll be at my office (415-422-7286; try 422-0758 if no answer)
until after midnight this evening, and most of the day tomorrow, and would
be delighted to get a call from John. Lowell]
∂28-Jul-83 2006 KJB@SRI-AI.ARPA Student support for this year
Received: from SRI-AI by SU-AI with TCP/SMTP; 28 Jul 83 20:06:37 PDT
Date: 28 Jul 1983 1644-PDT
From: Kjb at SRI-AI
Subject: Student support for this year
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI,
To: Kaplan at PARC, kjb at SRI-AI, lauri at SRI-AI,
To: peters at UTEXAS, rperrault at SRI-AI,
To: sag%SU-psych at SUMEX-AIM, stan at SRI-AI,
To: Wasow%SU-psych at SUMEX-AIM, winograd at PARC
cc: wasow at PARC
Barbara has asked when we can start making student support decisions
for this next year. The answer is, basically, now, though it will
not be official until the budget is, which should be Aug 15 if all
goes well. Here are some points:
We do not want to pick up weak students, who have run out of spport
becuase it has taken them so long to finish. We want to have only
very good students.
We want to be fair, so we need to make the availablity of this support
widely known in the departments involved, though we will not be
short.
We have not come to any mechanism for getting applicatins and
recommendatins yet, of course, so I propose that we run it on
an ad hoc basis THIS YEAR only. That is, get students to send
letters to
Graduate Support Committee
CSLI
Ventura Hall
Stanford Univeristy
and we will meet to review them periodically as we get them.
(I realize this latter "we" is ambiguous. It is meant to be.)
BE sure that you send a letter of support for the students you
support, not a computer mssage, since at present we cannot get
the messages printed. The dover is broken.
Jon
-------
∂28-Jul-83 2013 CLT x masda
Has Sarah disposed of her x masda and cancelled the insurance?
∂29-Jul-83 1032 ULLMAN@SU-SCORE.ARPA meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 29 Jul 83 10:31:30 PDT
Date: Wed 27 Jul 83 17:45:58-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: meeting
To: super@SU-SCORE.ARPA
We are going to meet tomorrow at 3PM, CIS conference room, as usual.
The agenda will include a discussion of hardware and plans for
future talks.
-------
∂29-Jul-83 1035 @SU-SCORE.ARPA:DELAGI@SUMEX-AIM.ARPA Please get back to me with additions and amendments.../bruce
Received: from SU-SCORE by SU-AI with TCP/SMTP; 29 Jul 83 10:35:36 PDT
Received: from SUMEX-AIM.ARPA by SU-SCORE.ARPA with TCP; Wed 27 Jul 83 19:32:26-PDT
Date: Wed 27 Jul 83 19:32:49-PDT
From: Bruce Delagi <DELAGI@SUMEX-AIM.ARPA>
Subject: Please get back to me with additions and amendments.../bruce
To: super@SU-SCORE.ARPA
27-Jul-83 18:23:39-PDT,4559;000000000000
Received: from Shasta by SUMEX-AIM with Pup; Wed 27 Jul 83 18:23:08-PDT
Received: from decwrl by Shasta with UUCP; Wed, 27 Jul 83 18:23 PDT
Mail-from: DEC-Enet site RHEA rcvd at 27 Jul 1983 18:15:04-PDT (Wednesday)
Date: Wednesday, 27 Jul 1983 18:15:04-PDT
From: decwrl!RHEA::DELAGI@Shasta
To: alien::duffy, omega::dchen, rdvax::gannon, @tome
Subject: QUESTIONS ON THE /784 AND PPA (FOR THE FOLKS AT STANFORD)
In talking with Jeff Ullman and Ernst Meyr here at Stanford, I've
gathered some questions on the 784 and the PPA that I hope you can
answer. Perhaps we can discuss these when we meet next week.
OS, Diagnostic, and Language Support
1. What is the diagnostic support? How much of it is stand-alone
only? Is error-logging supported? What faults are flagged? Are
there degraded operation modes providing soft failure? Are OS control
structure inconsistencies error-logged or otherwise reported for
analysis?
2. What is the language support for process fork and join, for
interlocked access, for semaphores, ..... Are there language
constructs to do the primative concurrency and synchronization
operations? Is this run-time-library stuff? What is the time to
perform such operations - as seen from the perspective of a
non-assembly language program. (Do any languages available from DEC
support in-line escape to assembly language? What are the
consequences of such escapes with respect to compiler optimization?)
3. How does one represent and aggregate separate processes as part of
the same program - e.g. to the loader?
4. What is the mechanism for manipulating the runable process queues?
What is the plan with respect to distinguishing processors as
scheduling targets for given processes? Is priority globally relevant
or may it be particular to each processor for a given process?
5. For experimentation, it may be necessary to supplant some of the
VMS scheduling services. What is the best way in which that can be
done?
6. What are the time requirements for process scheduling? Attempts
to do intra-process forking are probably attempts to re-invent
threads. There are probably good reasons why threads are planned for
the PPA but not for the 784. What are they?
7. To what extent are the /784 system concepts (and understandings)
scalable to the PPA?
8. What measurement results are available? How were they taken?
What flexibility is there in instrumenting the system?
9. What operations and styles of operation are cheap/optimal and
expensive/pessimal in the /784 and PPA.
Hardware Nitties-Gritty (Assume the MA780-D cache invalidation map)
1. How do VAX interlocked instructions deal with the cache? Is the
cache bypassed for reads as well as writes when the instruction that
caused them is of the interlocked persuasion?
2. What is the (detailed) sequence of bus, cache, processor, and
memory operations for mixed interlocked and non-interlocked
instructions.
3. If each of the four processors writes to the same location, what
is the precise sequence of operations that transpire. Will the target
location wind up with one of the four values that were being written?
4. What are the best and worst case timings for simultaneous writes,
simultaneous reads, and mixed read-modify-writes.
5. How are processor conflicts resolved? Is there some "fairness"
criteria on access? What are the timing implications in conflict
resolution? Are there programmer-accessible ways to tune low-level
scheduling behavior?
6. What instruction sequence is recommended to do (counted and simple
flagged) semaphore operations? What exactly happens in the caches, on
the busses, in the shared memory, and in the processors during this
sequence? How long does it take? What are the best and worst case
conditions?
7. What is the granularity of cache invalidation? Are there times
when reads yield results that would be surprising to the typical OS
programmer? (Please share some of the "training" dialog). What are
the cache control rules? How are they different from a standard 780?
8. What role does the CI (as opposed to the MA780) play in the /784?
and in the PPA?
Documentation
1. Assume the background of a Unix C programmer. What sequence of
documents do you recommend in order to make effective use of the
parallel operation of the /784 and PPA.
2. How does one get them?
Look forward to meeting with you (all) next week.
/bruce
-------
∂29-Jul-83 1403 ULLMAN@SU-SCORE.ARPA next meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 29 Jul 83 14:03:35 PDT
Date: Fri 29 Jul 83 13:55:04-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: next meeting
To: super@SU-SCORE.ARPA
We meet as usual, in the CIS conference room, 3PM next Thursday.
Werner T. will be the speaker.
-------
∂29-Jul-83 1417 KJB@SRI-AI.ARPA Rooms
Received: from SRI-AI by SU-AI with TCP/SMTP; 29 Jul 83 14:17:23 PDT
Date: 29 Jul 1983 1408-PDT
From: Kjb at SRI-AI
Subject: Rooms
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI,
To: Kaplan at PARC, kjb at SRI-AI, lauri at SRI-AI,
To: peters at UTEXAS, rperrault at SRI-AI,
To: sag%SU-psych at SUMEX-AIM, stan at SRI-AI, wasow at PARC,
To: Wasow%SU-psych at SUMEX-AIM, winograd at PARC
I have been getting requests for offices from everyone and their
brother the last couple of days. There are the following points
to be made:
1. The space given us in Ventura+Casita back in January
had a much smaller proposal in mind that has come to pass.
2. We will be remodelling or rebuilding or expanding
Casita this year, so it may be largely unusable. We won't know for
a while. Also, the 2nd floor, which we will have eventually, is
not ours yet, due to the delays caused by the SDF boards postponing
things back in June. We should have it by second quarter, however.
3. Lots of "promises" have been made, by implicit understandings,
by Pat Suppes, thinking it would fit, by JRP, thinking it would fit, by
the Dean, perhaps, thinking it would fit, and maybe even by me, in my
naivite. There is simply not room for everyone that thinks they have
a promise, without a lot of sharing and thought, and cooperation. One
of Dr. Maken's first jobs will be to sort this all out with Pat Suppes
in the way that is best for the project as a whole. If she can do that,
it will be a miracle. But it is just the sort of thing that is going
to be her responsibility. She will start here on August 15, so nothing
will be settled on that front till then, except for the offices that
are already assigned.
4. So. patience, please. Thanks, Jon
-------
∂29-Jul-83 1649 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA Agenda
Received: from SU-SCORE by SU-AI with TCP/SMTP; 29 Jul 83 16:49:23 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Fri 29 Jul 83 16:49:50-PDT
Date: 29 Jul 1983 1621-PDT
From: Nilsson at SRI-AI
Subject: Agenda
To: Amarel at RUTGERS, bobrow at PARC, buchanan at SUMEX,
To: BEngelmore at SRI-KL, LErman at SRI-KL,
To: csd.genesereth at SCORE, grosz at SRI-AI, hart at SRI-KL,
To: csd.mccarthy at SCORE, mcdermot at YALE, minsky at MIT-MC,
To: nilsson at SRI-AI, reddy at CMU-10A, rich at MIT-MC,
To: aaai-office at SUMEX-AIM, stan at SRI-AI, gjs at MIT-MC,
To: tenenbaum at SRI-KL, walker at SRI-AI, dwaltz at BBNG,
To: bonnie.upenn at UDEL, carbonell at CMUA,
To: rhayes-roth at SRI-KL, bledsoe at UTEXAS, lenat at SU-SCORE,
To: lourobinson
The following is the agenda (as of this date) for our AAAI executive
meeting in WDC. Claudia, could you please make sure that Chuck Rieger
gets a copy by U.S. Mail? If anyone notes that there is someone who
ought to attend who is not on the list of people at the end of agenda,
please coordinate with Claudia Mazzetti who can arrange to invite
him/her. Also, as soon as the election results for the new Councilors
are definite, could you invite those people, Claudia? I will be
"unavailable" until I see you all at IJCAI or AAAI--suggestions for
changes can go to Claudia (AAAI@sri-ai). -Nils
AAAI EXCOM AGENDA
Wednesday Evening 7:00 pm, August 24, 1983
Buffet Dinner Meeting
Thoroughbred Room, Concourse Level, Washington Hilton Hotel
. Approval of Minutes
Staff Reports and Discussion -- Claudia Mazzetti
Financial -- Raj Reddy
Conference Local Arrangements
Corporate Participants Program
AAAI Office Procedures and Capabilities
. Committee Reports and Discussion
Financial (see above)
Personnel -- Nils Nilsson
Elections -- Marvin Minsky
Program -- Mike Genesereth
Publications -- Lee Erman and Bob Engelmore
Membership -- Bruce Buchanan
Tutorials -- Charles Rich
Relations with IJCAI -- Saul Amarel
Future Conference Sites -- Marty Tenenbaum
Future Program Chairpeople
Possible Expansion of AAAI Interests and Activities
Industrial and Business -- Rick Hayes-Roth
Educational -- Claudia Mazzetti
On-Line Abstracts and Reports -- Michael Genesereth
Other Business
Installation of New President and Officers
Balance of Meeting to be Chaired by John McCarthy
Officers and Councilors:
Saul Amarel, Dan Bobrow, Mike Genesereth, Barbara Grosz, Peter Hart,
John McCarthy, Drew McDermott, Claudia Mazzetti (Executive Director),
Marvin Minsky (Past President), Nils Nilsson, Raj Reddy, Stan
Rosenschein, Gerald Sussman, Don Walker, David Waltz, Bonnie Lynn
Webber, Charles Rieger III.
Committee Chair People:
Bruce Buchanan (membership), Charles Rich (tutorial), Marty Tenenbaum
(conference), Lee Erman (publications)
Others:
Jaime Carbonell (representing SIGART), Rick Hayes-Roth (on agenda),
Woody Bledsoe (possible President-Elect), Doug Lenat, Bob Engelmore
(on agenda), Newly elected Councilors, Lou Robinson
-------
∂31-Jul-83 1213 RPG Password
To: JMC, JK, DFH
Since Jussi will be hacking ARP83 and WORK83 on [PRO,JMC], I suggest that
a password we all know ought to be installed on that area.
-rpg-
∂31-Jul-83 1642 ME E substitution speed
∂18-Jul-83 1712 JMC slow substitution in E
In order to count the spaces is a page, and thus to count words,
I asked E to substitute space for space, i.e αf α\ αβ∞. It took
41 seconds according to the WHO line to do 2114 substitutions, i.e.
about 20 milliseconds per substitution. Why is that?
ME - Substitution isn't terribly efficient, as it copies the whole line
each time it is changed (2114 times here). Why didn't you just search for
spaces instead of substituting for them. To get a better count of words,
you should also count lines (easily done) and subtract the number of
periods that have two spaces after them (or just the number of occurrences
of two consecutive spaces).
∂01-Aug-83 0959 CLT plumber
i called dahl, some one will come between 13:00 and 15:30,
they will call me first. if the car were here near the
office i could go meet the plumber when he calls.
if you are going to be home then, thats ok too.
∂01-Aug-83 1049 BSCOTT@SU-SCORE.ARPA NSF Grant MCS7905998
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Aug 83 10:49:30 PDT
Date: Mon 1 Aug 83 10:49:53-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: NSF Grant MCS7905998
To: RWW@SU-AI.ARPA
cc: Rindfleisch@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA, JK@SU-AI.ARPA,
BScott@SU-SCORE.ARPA
Richard,
A final report on the referenced grant is past due. I have been informed
by Dr. Chien's office that the second year funding for Jussi Ketonen's
grant will be delayed until your final report has been received at NSF.
Can you please take care of this right away, and please give the report
to me for processing.
Thanks,
Betty
-------
∂01-Aug-83 1318 EENGELMORE@SUMEX-AIM Technology Magazine
Received: from SUMEX-AIM by SU-AI with PUP; 01-Aug-83 13:17 PDT
Date: Mon 1 Aug 83 13:02:53-PDT
From: Ellie Engelmore <EENGELMORE@SUMEX-AIM>
Subject: Technology Magazine
To: Genesereth@SUMEX-AIM, Lenat@Score, JMC@SAIL, Clancey@SUMEX-AIM
cc: EENGELMORE@SUMEX-AIM
Sue Onions, London 439-9606, is the Technology Magazine write on AI.
She will be attending a portion of IJCAI and would like to talk with
you there if possible.
She will try to get in touch with you in Karlsruhe.
-------
∂01-Aug-83 1337 NILSSON@SRI-AI.ARPA net addresses
Received: from SRI-AI by SU-AI with TCP/SMTP; 1 Aug 83 13:37:14 PDT
Date: 1 Aug 1983 1334-PDT
From: Nilsson at SRI-AI
Subject: net addresses
To: jmc at SU-AI
John, Here is a list of net addresses of the people to whom
I sent the AAAI agenda (in case you want to send them anything).
(Chuck Rieger is, as far as I know, not net-addressable.) -Nils
-------
Your message didn't include the net addresses. However, I had
already copied and edited the list associated with your previous
message about the agenda, and I assume that's complete. I'll remember
about poor Chuck. I hope to get the supplementary agenda out this
week. Tentatively, the two items will be proposals for AAAI
sponsored workshops in particular topics and Mike Genesereth's ideas
on on-line reports. I'm waiting for Mike to agree. He isn't already
on the agenda, is he? My goof. I see he is. OK, one item.
∂01-Aug-83 1359 @SU-SCORE.ARPA:LERMAN@SRI-KL.ARPA AAAI publications comm. meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Aug 83 13:59:24 PDT
Received: from SRI-KL.ARPA by SU-SCORE.ARPA with TCP; Mon 1 Aug 83 13:59:29-PDT
Date: Mon 1 Aug 83 14:55:07-PDT
From: Lee Erman <LErman@SRI-KL.ARPA>
Subject: AAAI publications comm. meeting
To: ATP.Bledsoe@UTEXAS-20.ARPA, BEngelmore@SRI-KL.ARPA,
Feigenbaum@SUMEX-AIM.ARPA, McCarthy@SU-SCORE.ARPA,
Minsky%MIT-OZ@MIT-MC.ARPA, Nilsson@SRI-AI.ARPA, Price@USC-ECLC.ARPA,
Reddy@CMU-CS-B.ARPA, Walker@SRI-AI.ARPA, AAAI-office@SRI-AI.ARPA
The AAAI publications committee will meet on Thursday, Aug. 25, 7:30-9:00 AM
in the Hemisphere Room. Breakfast will be provided.
In my absence, Bob Engelmore has agreed to chair the meeting. Please
let Bob and Claudia (AAAI@sri-ai) know if you are coming. If you have any
agenda items, please send them to Bob and to me.
--Lee
-------
Agenda item for AAAI publication committee
I would like to discuss the possibility of semi-annual journal, "Critical
reviews in AI". Each issue would contain two to four critical reviews
of pieces of recent AI research. The object is to raise the standards
of evaluation of work. Of course, there would have to be an editor and
people willing to write such reviews.
∂01-Aug-83 1408 BSCOTT@SU-SCORE.ARPA [OHLANDER at USC-ISI: TAsk Proposals]
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Aug 83 14:06:04 PDT
Date: Mon 1 Aug 83 14:06:16-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: [OHLANDER at USC-ISI: TAsk Proposals]
To: DCL@SU-AI.ARPA, TOB@SU-AI.ARPA, JMC@SU-AI.ARPA, ZM@SU-AI.ARPA
I am assuming that each of you (except Zohar) talked with Ron Ohlander
about the revisions to be made in the proposals.
Ron Ohlander will be out of the office until about mid-August, but I
hope to get the revised task proposals (and budgets) in the mail by
the first of next week.
Betty
---------------
Return-Path: <@USC-ISI:OHLANDER@USC-ISI>
Received: from USC-ISI.ARPA by SU-SCORE.ARPA with TCP; Wed 27 Jul 83 11:32:51-PDT
Date: 27 Jul 1983 1131-PDT
Sender: OHLANDER at USC-ISI
Subject: TAsk Proposals
From: OHLANDER at USC-ISI
To: BSCOTT at SU-SCORE
Message-ID: <[USC-ISI]27-Jul-83 11:31:02.OHLANDER>
Betty,
There is a slight change in what I told you by telephone. For the
revised task efforts, I will need a revised proposal in addition to revised
budget statements for some of the tasks. This will be the case for
Binford, Manna, McCarthy, and Luckham. For Wiederhold, all I need
are the revised budget pages. I also need a new forwarding letter with
the "Poor Boy" statement and current date.
Ron
P.S. If you can't get a response from Mana, send the others'
task proposals.
-------
Creary will be leaving my group in December, so I'll need to revise the
parts of the proposal that concern him. He will be replaced, but the
empahsis will change. I also understand that there will be revisions
of Gabriel's activities. Will there be anything else.
∂01-Aug-83 1436 SCHMIDT@SUMEX-AIM Re: Workstation Purchase Dilemma
Received: from SUMEX-AIM by SU-AI with PUP; 01-Aug-83 14:36 PDT
Date: Mon 1 Aug 83 14:36:38-PDT
From: Christopher Schmidt <SCHMIDT@SUMEX-AIM>
Subject: Re: Workstation Purchase Dilemma
To: Rindfleisch@SUMEX-AIM
cc: HPP-Exec@SUMEX-AIM, Yeager@SUMEX-AIM, JMC@SAIL, RPG@SAIL, Lane@SUMEX-AIM,
Spurgeon@SUMEX-AIM
I would like to volunteer my [colored] view of the abilities of
Symbolics and Xerox to address the problem of providing adequate computing
resources for our future needs.
There seems to be a consensus that none of the machines available today
is adequate for A.I. research beyond the feasability study level.
If this is correct (and I believe it is for some projects and will be
for others) then better machines will
ultimately be produced and marketed. I can't prove the validity of this
statement, but since most of the projects concerned are going forward on this
assumption anyway, I won't attempt to prove or disprove it.
Given the above, it is quite likely that either Xerox or Symbolics or
both will produce such machines. Other companies may enter the market, but
I won't speculate on them because (1) I'm unqualified, and (2) they would, of
necessity, be "first generation" lisp machines, and would almost certainly
not address our problems.
Now, on to a comparison. We may be fortunate in that both
manufacturers have produced more than one generation of machine and we can
examine how each effort fared.
(Roughly: Xerox produced the Alto, D-Lisp for the Alto, D-Lisp for
the Dolphin, Interlisp-D for the Dolphin, Interlisp-D for the Dorado, and
Interlisp-D for the Dandelion. MIT produced the CONS, the CADR, and
Zetalisp, Symbolics copied the CADR, producing the LM-2, and enhancing
Zetalisp in small ways, and then finally Symbolics produced the LM-3600, and
a new implementation of Zetalisp.)
Both groups have experience in bringing up new systems--if you
count the MIT effort as "part" of Symbolics.
You can take an Interlisp-D
sysout to any of the three machines with no changes. This means that if
feature X is debugged once, it will be available now, everywhere.
You can't take a world
load from one machine to another. (If you could, their chaosnet implementation
is so poor that it would take 40 minutes to transfer the smallest world load.)
Can you take compiled code from one LM to another? Nope. Why is this? I will
suggest that it is because Symbolics has produced not two machines, but
one--the 3600. Since we are considering buying not from MIT, but from
Symbolics, I believe that this is the only machine we can think of as theirs.
I have heard Xerox criticized regularly that if we ordered a Dorado (which
they don't exactly stress as a mainstream product) that we probably couldn't
have it delivered next week. Well, what about the 3600, Symbolics's only
product? We were supposed to have one here by Nov. 1 1982. Who can't
deliver his product? What happened to the IFU, touted as "within a couple
of months of completion" at last AAAI? What happened to the "Symbolics
phone net" that I was assured would be added to all LM2's at last AAAI?
Why can't they make a left control key that lasts more than six months? :-)
Here is a company, which, when adding a NEW feature can't resist putting it
on a different combination of select/hyper/system keys on each machine.
They've never seen a menu before?
I have heard Xerox criticized because they won't learn from
suggestions from the outside.
Let's look at Symbolic's notion of a menu:
A function breaks. Does a menu pop up? No. Instead what looks like a
list of options is printed at the end the TTYoutput. Each selection is
preceded by a printed representation of some combination of hyper/top/greek
keys, which when typed in the proper way will cause that option to be
selected. Do you think Symbolics' egos would let them walk over to the
next booth at AAAI and see how it is done right? No way in hell.
Let's talk scrolling. I don't think anyone has wanted a terminal to
"scroll" by simply sending the cursor to the "home" position since 1965.
When someone suggested this to Allan Wechsler on his visit here, his reply
was that "people don't really want windows to scroll." That's like your
mechanic telling you "you don't really want your transmission fixed."
Let's talk editors. Face it, ZMACS is EMACS. The editor that
comes with the VIC20 is better tuned for BASIC that EMACS is for LISP.
Their last big editor change was to coerce ZMACS windows on the 3600 to be
the same width as those on the LM-2, so that lines wouldn't wrap around if
you moved a file from one machine to another. This is engineering?
Let's talk network protocols. They call themselves the world
purveyor of Choasnet, but we've never had a completely working set of services.
Their "documentation" is a photocopy of an old ARPA document which they don't
even implement correctly. Their host tables are hard coded into the C source,
and they completely ignore the Chaosnet version compatability protocol. And
we expect these guys to bring up IP/TCP ???
Let's talk about their file system salvager
For months we couldn't salvage our system because to do so had
the undesirable side-effect of trashing it. This is engineering?
Only within the past month has Symbolics released a garbage collector
for the 3600. This is a production system?
In other words, when better machines are made, I rather suspect that
there is a better chance that Xerox will make them.
Don't get me wrong. I think Symbolics made a lot of design decisions
better than those Xerox made. A lot. For example, I think the decision to
separate the front-end from the lisp engine was a good idea. The only catch
is, I now want to use a Dandelion as my front end. :-)
--Christopher
P.S. If this seems choppy and antecedents don't match up, it's because I wrote
it over 5 days in odd minutes.
-------
∂01-Aug-83 1551 SCHMIDT@SUMEX-AIM Re: Workstation Purchase Dilemma
Received: from SUMEX-AIM by SU-AI with PUP; 01-Aug-83 15:51 PDT
Date: Mon 1 Aug 83 15:51:34-PDT
From: Christopher Schmidt <SCHMIDT@SUMEX-AIM>
Subject: Re: Workstation Purchase Dilemma
To: Rindfleisch@SUMEX-AIM
cc: HPP-Exec@SUMEX-AIM, Yeager@SUMEX-AIM, JMC@SAIL, RPG@SAIL, Lane@SUMEX-AIM,
Spurgeon@SUMEX-AIM
I would here like to react to the LM-3600 fever that was the
response to the partial revelation of Dick Gabriel's numbers.
What we have to keep in mind is that the benchmarks were not made
on either of the machines we've been considering purchase of. A Dolphin is
not a Dandelion, and a 3600 with 6 memory boards is not a configuration that
anyone (within my earshot) has suggested buying.
That a Dolphin is not a Dandelion I needn't go into.
That a Dandelion is not a Dandelion I won't even speculate on...
I think it would be safe to assert that few of our systems fit
entirely into core and rely a good deal on the paging mechanism. Dick even
suggested that paging performance might be critical for us. Since I have
never heard anyone quote a price for the 3600 other than $80,000 I conclude
that we are envisioning 3600's with only one memory board. Does anyone
believe that the 6 board numbers are applicable? Try this
gedankenexperiment: Imagine a Dolphin with only one memory board instead of
the usual 6. Meaningful? Of course not.
What would a 3600 with 6 boards cost? Guess about $80,000 plus
5 times $14,000; or about $150,000. Any implications I leave to the reader
so as to absolve myself of any responsibility.
--Christopher
-------
∂01-Aug-83 1605 ULLMAN@SU-SCORE.ARPA Spectrum Supercomputer Survey
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Aug 83 16:05:33 PDT
Date: Mon 1 Aug 83 16:05:34-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: Spectrum Supercomputer Survey
To: super@SU-SCORE.ARPA
I have gotten involved in a project to develop a consensus (?)
among a variety of people regarding the "new generation" of
computers. The vehicle is a month-long discussion via CSNET.
Apparently the discussion is completely public, so I am going to
forward questions and replies to the list.
If anyone has any thoughts they'd like to interject into the
discussion, forward them to me and I'll give them the
proper attribution.
The next three messages are: (1) the introduction from the "moderators,"
(2) the first question, and (3) a reply already received.
-------
∂01-Aug-83 1609 ULLMAN@SU-SCORE.ARPA Three messages
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Aug 83 16:08:59 PDT
Date: Mon 1 Aug 83 16:07:32-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: Three messages
To: super@SU-SCORE.ARPA
29-Jul-83 13:02:46-PDT,4376;000000000001
Return-Path: <SHORT.HP-HULK@Rand-Relay>
Received: from csnet-sh by SU-SCORE.ARPA with TCP; Fri 29 Jul 83 13:02:32-PDT
Date: 28 Jul 1983 1004-PDT
From: SHORT.HP-HULK@Rand-Relay
Return-Path: <SHORT%HP-HULK.HP-Labs@Rand-Relay>
Subject: TELECONFERENCE ON NEW GENERATION COMPUTERS
Received: by HP-VENUS via CHAOSNET; 28 Jul 1983 10:06:06-PDT
To: NEW-GEN@CSNET-SH
Cc: RAPHAEL.HP-HULK@Rand-Relay
Message-Id: <428259967.23117.hplabs@HP-VENUS>
Via: HP-Labs; 28 Jul 83 10:23-PDT
The November issue of IEEE Spectrum will be devoted to the issues raised by
the Japanese "Fifth Generation" challenge. Paul Wallich has already
discussed briefly with you some of the topics of concern, along with the
proposal to conduct a month-long computer teleconference.
This message launches the teleconference. We anticipate about forty
participants, including Lynn Conway, Gordon Bell, Erich Bloch, Ivan and
Bert Sutherland, and Ed Feigenbaum. Our purpose today is to outline the
goals, provide some starting framework, and ask for some specific starting
information. Our first question of the teleconference will follow shortly,
along with procedures of conduct that we expect to follow. While the
teleconference will be a short-term project, we are hopeful that a longer-
term dialogue will continue after the initial needs are met.
Our goals are to produce informal commentary on what the issues are, how
they are being addressed, what prospects are most promising, and what
changes in direction or approach may be appropriate to recommend. This
obviously is served by each of you responding to each question we pose,
but more importantly, by responding to each others' inputs. It will be
Paul's challenge to distill our messages into a meaningful and coherent
story.
Rather than concentrating on the Japanese "Fifth Generation", we would
prefer to focus upon American "New Generation" issues. Our preliminary
thoughts revolve around four basic ideas:
1) What is the definition of New Generation? What goals, structures,
architectures, and developments are critical to its achievement?
2) What is the current status of New Generation research and
development in America? Where are the issues being confronted,
and which alternatives? What are the commercialization options and
plans? This is germane not only in technological and capital
investment arenas, but in societal impact as well.
3) What major new disciplines will be needed and may evolve as a result
of New Generation needs? (e.g. Knowledge Engineering, Advanced
Architecture, Advanced VLSI, Network Interconnection). How will
the structure of disciplines as separate entities be changed by New
Generation technology? (Much of computer science is impacting most
engineering disciplines.)
4) What new infrastructures or new uses of old infrastructures will be
needed to make New Generation work possible? What kinds of networks,
servers, facilities, etc. will be needed? What kind of funding will
be needed for the new infrastructures (not just how much, but where
will it come from, and how is it disbursed and evaluated)?
The moderator will pose a question every few days, for which your answers
should be addressed to NEW-GEN@CSNET-SH, which will automatically forward
them to the whole group. (You will also be able to engage in private side
conversations by addressing messages directly to the sender of any message.
We will provide a full mailing list later.)
We'd like to pose two questions now. Please send your answer only to the
moderators, by addressing it to NG-MOD@CSNET-SH.
1) What favorite issue or two do you feel is most important for our
discussion agenda?
2) Please send us a short (one paragraph) biosketch so that we can
prepare a roster of the conference by August 5.
Thanks for your involvement. Your response will serve to acknowledge that
your network address is correct.
The Moderators:
Alan Bell, Xerox PARC
Chuck House, Hewlett Packard
Bert Raphael, Hewlett Packard
Paul Wallich, IEEE
z
-------
29-Jul-83 13:02:46-PDT,4376;000000000001
Return-Path: <SHORT.HP-HULK@Rand-Relay>
Received: from csnet-sh by SU-SCORE.ARPA with TCP; Fri 29 Jul 83 13:02:32-PDT
Date: 28 Jul 1983 1004-PDT
From: SHORT.HP-HULK@Rand-Relay
Return-Path: <SHORT%HP-HULK.HP-Labs@Rand-Relay>
Subject: TELECONFERENCE ON NEW GENERATION COMPUTERS
Received: by HP-VENUS via CHAOSNET; 28 Jul 1983 10:06:06-PDT
To: NEW-GEN@CSNET-SH
Cc: RAPHAEL.HP-HULK@Rand-Relay
Message-Id: <428259967.23117.hplabs@HP-VENUS>
Via: HP-Labs; 28 Jul 83 10:23-PDT
The November issue of IEEE Spectrum will be devoted to the issues raised by
the Japanese "Fifth Generation" challenge. Paul Wallich has already
discussed briefly with you some of the topics of concern, along with the
proposal to conduct a month-long computer teleconference.
This message launches the teleconference. We anticipate about forty
participants, including Lynn Conway, Gordon Bell, Erich Bloch, Ivan and
Bert Sutherland, and Ed Feigenbaum. Our purpose today is to outline the
goals, provide some starting framework, and ask for some specific starting
information. Our first question of the teleconference will follow shortly,
along with procedures of conduct that we expect to follow. While the
teleconference will be a short-term project, we are hopeful that a longer-
term dialogue will continue after the initial needs are met.
Our goals are to produce informal commentary on what the issues are, how
they are being addressed, what prospects are most promising, and what
changes in direction or approach may be appropriate to recommend. This
obviously is served by each of you responding to each question we pose,
but more importantly, by responding to each others' inputs. It will be
Paul's challenge to distill our messages into a meaningful and coherent
story.
Rather than concentrating on the Japanese "Fifth Generation", we would
prefer to focus upon American "New Generation" issues. Our preliminary
thoughts revolve around four basic ideas:
1) What is the definition of New Generation? What goals, structures,
architectures, and developments are critical to its achievement?
2) What is the current status of New Generation research and
development in America? Where are the issues being confronted,
and which alternatives? What are the commercialization options and
plans? This is germane not only in technological and capital
investment arenas, but in societal impact as well.
3) What major new disciplines will be needed and may evolve as a result
of New Generation needs? (e.g. Knowledge Engineering, Advanced
Architecture, Advanced VLSI, Network Interconnection). How will
the structure of disciplines as separate entities be changed by New
Generation technology? (Much of computer science is impacting most
engineering disciplines.)
4) What new infrastructures or new uses of old infrastructures will be
needed to make New Generation work possible? What kinds of networks,
servers, facilities, etc. will be needed? What kind of funding will
be needed for the new infrastructures (not just how much, but where
will it come from, and how is it disbursed and evaluated)?
The moderator will pose a question every few days, for which your answers
should be addressed to NEW-GEN@CSNET-SH, which will automatically forward
them to the whole group. (You will also be able to engage in private side
conversations by addressing messages directly to the sender of any message.
We will provide a full mailing list later.)
We'd like to pose two questions now. Please send your answer only to the
moderators, by addressing it to NG-MOD@CSNET-SH.
1) What favorite issue or two do you feel is most important for our
discussion agenda?
2) Please send us a short (one paragraph) biosketch so that we can
prepare a roster of the conference by August 5.
Thanks for your involvement. Your response will serve to acknowledge that
your network address is correct.
The Moderators:
Alan Bell, Xerox PARC
Chuck House, Hewlett Packard
Bert Raphael, Hewlett Packard
Paul Wallich, IEEE
z
-------
1-Aug-83 08:20:03-PDT,1175;000000000001
Return-Path: <ABell.PA@PARC-MAXC.ARPA>
Received: from csnet-sh by SU-SCORE.ARPA with TCP; Mon 1 Aug 83 08:19:59-PDT
Date: Mon, 1 Aug 83 08:07 PDT
From: ABell.PA@PARC-MAXC.ARPA
Subject: Opening Question
To: NEW-GEN@CSNET-SH.ARPA
cc: ABell.PA@PARC-MAXC.ARPA
Reply-to: NEW-GEN@CSNET-SH.ARPA
This message contains the first question to be discussed
on the NEW-GEN net conference. Please answer the
question by sending a response to the whole distribution
list at NEW-GEN@CSNET-SH. If you have any
procedural comments or questions for the moderators,
please send them to NG-MOD@CSNET-SH. A second
question will be sent out in a few days.
First Question:
What does "fifth" or "new" or "supercomputer" generation
computing mean? What should the goals of these programs be?
What will be the form of their structures, architectures,
developments, etc. What will the results of this research look
like after the technology has matured? Will there be several
different efforts going to occur which are heading toward
different goals with similar names (superscientific, fifth gen, etc.)
How can each of them be characterized and distinguished?
1-Aug-83 13:32:48-PDT,3118;000000000001
Return-Path: <blb@lanl>
Received: from csnet-sh by SU-SCORE.ARPA with TCP; Mon 1 Aug 83 13:32:33-PDT
Date: 1 Aug 1983 14:11:35-MDT
From: Bill Buzbee C-DO <blb@lanl>
Reply-to: blb@lanl
To: NEW-GEN@CSNET-SH
Subject: Terminology, etc
I'll begin with a short discussion of terminology and then offer some comments
on the Japanese project.
1. Terminlogy. Seems to me this is an "area of opportunity" for this
teleconference. For example,
a. "5th generation"--at first blush it is easier to define "5th gen"
than "New Gen" because (presumably) 5th gen is things built out of VLSI.
However, clearly the Japanese have a lot more in mind than component
technology. My problem with "new" is that what is "new" today is "old"
tommorrow.
b. logic processing. Part of what is new in 5th gen is development
and support of applications involving "non-numeric computation."
Since the phrase "symbolic manipulation" already has an established
definition, we at Los Alamos have been considering the use of
of the phrases "numeric processing" and "logic processing" to
distinguish between the basic functionalities of computers.
Have others noted a need for such and what terms are you using?
2. The Japanese project. In this country, many people equate this project to
R&D in AI. As I understand it, it is somewhat broader than that, but clearly
AI is a vital component. So I'll review their objectives and offer some
comments on them.
The Japanese objectives are
a. make computers easier to use, i.e. usable by a large fraction of
the population
b. improve productivity in software development and
c. improve reliability and cost-effectiveness of hardware/software
systems.
Comments:
Ease of use. The Japanese technical approach involves natural language
processing, audio, grpahics, AI, etc. This is an area where the US personal
computer industry may be relevant because their market is proportional to
ease-of-use. Some of the spread-sheet packages are conspicous examples of
this point.
Software productivity. I'm sure all of us appreciate the merits of work in
this area. The Japanese technical approach involves programming
environments and ultrahigh-level languages that have a high degree of
verifiiability. To what extent, if any, do the Japanese enjoy an advantage
in this area because of the tremendous amount of software that the US has
in place and must carry into the future?
System reliability. The Japanese seek to further improve the reliability
and performance of their components, as well as to improve fault tolerance
and automatic diagnostics in their systems. This area is their forte.
The best of both worlds. The point of the previous comments is that the
Japanese can make progress toward their objectives without any technological
breakthroughs. This is not to discount the revolutionary aspects of the
the project (aspects that have already captured the imagination of the
world's computing community), but to note that to some extent they can
have their cake and eat it too.
-------
∂01-Aug-83 1615 DFH
∂01-Aug-83 1333 JMC aaai mailing list
aaai.dis[e83,jmc] is a mailing list for the executive committee of aaai.
In addition, Charles Rieger gets announcements by U.S. Mail. Please
get his address from Claudia Mazzetti at AAAI.
---------------
Address (from last year) is:
Charles Rieger
Scion
12310 Pinecrest Rd.
Weston, VA 22091
(703) 476-6100
∂01-Aug-83 1617 DFH Victor Kuo
I spoke to his wife. He is apparently still in
Peking (the last she heard, anyway) and having
some trouble with getting the Chinese to give him
the proper papers so he can leave. She did not
think the trouble was too serious. She thought
perhaps it might help if we wrote a letter inquiring
as to what had happened, not mentioning any contact
with her. Do you want me to write, or do you want
to, or?
Please draft something to Kuo for me to sign.
∂01-Aug-83 1741 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Center for Parallel Computation
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Aug 83 17:40:57 PDT
Received: from Diablo by Score with Pup; Mon 1 Aug 83 17:41:08-PDT
Date: Mon, 1 Aug 83 17:36 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Center for Parallel Computation
To: super@score
Gene G. forwarded to me a letter from Richard Hale of Technology Development
of Calif., the folks who apparently brought you the ILLIAC IV,
concerning the establishment of a center for the sorts of research
we as a group have been contemplating, e.g., architectures, networks,
languages. They would provide centralized machines in a mode similar
to the way we expect the VAX 784/PPA to be used.
The proposal looks not unreasonable, although it might cut into
our operation rather than enhance it; I'm not sure.
At any rate, I'll send a copy to anyone who wants to see it quickly,
and distribute it next Thursday.
P. S.: For a good laugh, the next message is my response to question #1.
∂01-Aug-83 1743 ullman%SU-HNV.ARPA@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Aug 83 17:42:50 PDT
Received: from Diablo by Score with Pup; Mon 1 Aug 83 17:41:23-PDT
Date: Mon, 1 Aug 83 17:36 PDT
From: Jeff Ullman <ullman@Diablo>
To: super@score
Response to Question #1:
The "new generation" of computers arrived this year not because of
the recent Japanese preoccupation with AI, but because it suddenly
became unreasonable to see computer technology progress in terms
of cheaper processors. Once we have a $10 processor chip, I
don't see a big future for trying to invent the 30-cent processor
that can be used to control the ink flow in my 50-cent ball-point pen.
Instead roughly from here on, progress is going to be measured by
how many processors we can bring to bear on a single problem.
The primary research goals, as I see it, concern software and
algorithms as much as they do architecture.
Really, everything has got to fit together: the application, the
algorithm, the language, the communication support, the operating
system, and the hardware, or we won't get useful work out of
more than a few processors.
However, I think we can classify the architectures for supercomputers
in to a few groups, each of which will have its place.
1. High-flux machines, such as ultracomputers (shuffle-exchange machines),
butterfly networks, hypercubes, cube-connected-cycles, etc.
These are characterized by the ability to move massive quantities of
data in parallel from any set of processors to any other, quickly.
I predict that these machines will have their greatest impact in the
AI domain, where, despite our best attempts to narrow search, we shall
find that logical inferences depend substantially on "everything
meeting everything else."
2. Tree machines. These are in a sense opposite to high-flux
machines. They have a massive bottleneck at the root and cannot be
used for pairing rules and data in arbitrary ways efficiently.
However, they may play a role as retrieval devices for massive databases.
3. Arrays of processors, e.g., systolic arrays. These are intermediate
in flux between the ultracomputers and the trees. For example, they
can sort n elements in time sqrt(n), while high-flux machines take
log n and tree machines take time n. They have a crucial role to
play because they are tuned well to the data transmission requirements
of a number of problems, e.g., feature extraction and various numerical
problems.
∂01-Aug-83 2107 YOM Back ...
To: JMC@SU-AI, DFH@SU-AI, CLT@SU-AI, JJW@SU-AI, RSC@SU-AI
from my summer trip. I'm currently working at IBM, but should be around
reading mail etc. every other day or so.
Yoram.
∂02-Aug-83 1146 ULLMAN@SU-SCORE.ARPA proposal
Received: from SU-SCORE by SU-AI with TCP/SMTP; 2 Aug 83 11:45:57 PDT
Date: Tue 2 Aug 83 11:45:40-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: proposal
To: super@SU-SCORE.ARPA
cc: harrison@UCB-VAX.ARPA
The draft of the proposal I have been working on is done.
I'll distribute copies at the Thursday meeting if the Dover ever
starts working. Meanwhile, scp.dvi and scp.press are readable files
in <ullman> if anyone cares to make a copy.
-------
∂02-Aug-83 1712 BSCOTT@SU-SCORE.ARPA ARPA Equipment Contract
Received: from SU-SCORE by SU-AI with TCP/SMTP; 2 Aug 83 17:12:30 PDT
Date: Tue 2 Aug 83 17:12:58-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: ARPA Equipment Contract
To: JMC@SU-AI.ARPA, JLH@Shasta, Feigenbaum@SUMEX-AIM.ARPA
cc: Golub@SU-SCORE.ARPA, Ullman@SU-HNV.ARPA, Rindfleisch@SUMEX-AIM.ARPA,
BScott@SU-SCORE.ARPA
The separate accounts are now established. Pertinent information:
Contract No.: MDA903-83-C-0188
Stanford Fund Number: 187X007
The four separate account numbers and internal reference names (the
Contract Number is the official name), P.I. names, and the amounts
currently allocated to each account are set out below. Note that
$1,615,000 is available for the period 6/1/83 through 12/31/83, and this
amount has been divided among the accounts per instructions from Tom
Rindfleisch.
2-FCZ780 Formal Reasoning $ 224,000
McCarthy
2-FCZ781 CSD Equipment 393,000
McCarthy
2-FCZ782 CSL Equipment 514,000
Hennessy
2-FCZ783 HPP Equipment 484,000
Feigenbaum
I believe this is all the necessary information. If you have questions,
please let me know.
Betty
-------
∂03-Aug-83 0747 Raj.Reddy@CMU-CS-A
Received: from CMU-CS-A by SU-AI with TCP/SMTP; 3 Aug 83 07:46:41 PDT
Received: from [128.2.254.192] by CMU-CS-PT with CMUFTP; 3 Aug 83 10:39:03 EDT
Date: 3 August 1983 1045-EDT (Wednesday)
From: Raj.Reddy@CMU-CS-A
To: JMC@SU-AI
CC: Raj.Reddy@CMU-CS-A
Message-Id: <03Aug83.104518.RR29@CMU-CS-A>
- - - - Begin forwarded message - - - -
Received: ID <A610RR29@CMU-CS-A>; 2 Aug 83 16:35:51 EDT
Date: 2 August 1983 1635-EDT (Tuesday)
From: Raj.Reddy@CMU-CS-A
To: John McCarthy@SU-AI
Subject: Letter from JJSS
CC: Raj.Reddy@CMU-CS-A
Message-Id: <02Aug83.163519.RR29@CMU-CS-A>
Dear John,
A letter from JJSS of the kind that we both talked about is in the mail to
you today and we should talk about how to proceed next. I won't be able
to be in Paris or at IJCAI next week because of some last minute problems.
I will be calling to talk about plans for September, October and November.
Best regards,
Raj
- - - - End forwarded message - - - -
∂03-Aug-83 0823 RINDFLEISCH@SUMEX-AIM Symbolics Machines
Received: from SUMEX-AIM by SU-AI with PUP; 03-Aug-83 08:23 PDT
Date: Wed 3 Aug 83 08:19:41-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Symbolics Machines
To: JMC@SAIL, RPG@SAIL
cc: Rindfleisch@SUMEX-AIM
I'm putting together the HPP Symbolics order. Do you want me to include
your 3 machines as well now or do you want to proceed separately?
Tom R.
-------
I'm afraid I must proceed separately, because I want to have a session first
with RPG and the members of the Formal Reasoning Group. RPG: would you
put together a draft proposal for our order that we can discuss?
∂03-Aug-83 1003 CLT
Could you please tidy up the bedroom before you go today?
e.g. put away clothes, papers, suitcases, etc.
Merci mille fois
∂03-Aug-83 1033 DFH travel expenses - ARPA account
It looks to me that there is a chance that with both
Lew Creary and Paul Wieneke going to AAAI and Corky's
travel expenses that we may go over the amount allotted
for domestic travel, and that something in the neighborhood
of $100 might have to go on your unrestricted. This amount
is probably not worth asking for rebudgeting; however, if there
is any other likely domestic travel that I don't know about yet,
perhaps we should consider such a request.
∂03-Aug-83 1143 ullman%SU-HNV.ARPA@SU-SCORE.ARPA spectrum survey
Received: from SU-SCORE by SU-AI with TCP/SMTP; 3 Aug 83 11:43:32 PDT
Received: from Diablo by Score with Pup; Wed 3 Aug 83 11:37:24-PDT
Date: Wed, 3 Aug 83 11:36 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: spectrum survey
To: super@score
I received two more messages regarding the meaning of new-generation.
I'm going to forward them in the next messages, but I'm wondering if
all this chatter is valuable. Would you prefer to browse the
directory /vlsi/ullman/scsurv@diablo that I set up to hold these
messages, or that I forget the whole thing?
∂03-Aug-83 1149 ullman%SU-HNV.ARPA@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 3 Aug 83 11:46:44 PDT
Received: from Diablo by Score with Pup; Wed 3 Aug 83 11:38:29-PDT
Date: Wed, 3 Aug 83 11:37 PDT
From: Jeff Ullman <ullman@Diablo>
To: super@score
2-Aug-83 19:35:08-PDT,3012;000000000001
Return-Path: <Fahlman@CMU-CS-C.ARPA>
Received: from csnet-sh by SU-SCORE.ARPA with TCP; Tue 2 Aug 83 19:35:01-PDT
Received: ID <FAHLMAN@CMU-CS-C.ARPA>; Tue 2 Aug 83 22:20:55-EDT
Date: Tue, 2 Aug 1983 22:20 EDT
From: Scott E. Fahlman <Fahlman@CMU-CS-C.ARPA>
To: new-gen@UWISC.ARPA
Cc: fahlman@CMU-CS-C.ARPA
Subject: reply to question #1
In-reply-to: Msg of Tue 2 Aug 83 15:50:45-PDT from Jeffrey D. Ullman <ULLMAN at SU-SCORE.ARPA>
I agree with Jeffrey Ullman's comments that the next generation is going
to be largely concerned with parallelism. I also agree with his view
that within the world of parallel architectures the amount and pattern
of connectivity is usually the critical question. Once we start finding
microprocessors in Carcker Jack boxes, the real bottleneck is going to
be communication among the processors, and not in the amount of
available processor power.
Let me propose an orthogonal classification that complements Ullman's
division into high flux machines, low flux tree machines, and
intermediate flux arrays. Instead of looking at the amount of
connectivity, we might look at how it is used. This view gives rise to
the following sorts of categories:
1. General-puropse machines with reconfigurable connection networks.
All of Ullman's "high flux" examples fall into this category: butterfly
nets, N-cubes, etc.
2. Special purpose machines with connectivity tailored to a specific
application. A hard-wired systolic sorting network would be an example.
A 2-D hardware retina or image-transformation processor would be
another. Note that the communication bandwidth can be very high in such
architectures, since you do not have to pay the price for general
interconnection.
3. "Connectionist" machines (Jerry Feldman's term) in which the resident
knowledge is represented by the pattern of connectivity -- the presence
or absence of a wire between two points, perhaps with some variable
weight on the wire -- and not as bits inside some processor's memory.
Most knowledge-base machines are full of pointers; these machines just
take that view a bit more literally than most. These can be viewed as
custom-wired machines like those in class 2, but new connections must be
added from time to time as new knowledge is added.
Of course, the real world is never so clean as our taxonomies would
suggest. An Illiac-IV is sometimes used as a class 2 machine with the
right connectivity to munch square arrays, and sometimes as a class 1
machine with the existing connections used to implement arbitrary
patterns or virtual connectivity. A connectionist machine is a nice
abstraction, but when you want to implement it you might use an N-cube
network of processors, each representing a thousand lower-level nodes.
Despite these "A masquerading as B implementing C" games, however, I
have found this sort of division useful in thinking about the mapping
between architectures and applications.
-- Scott Fahlman, CMU
Received: from SU-SCORE.ARPA by Diablo with TCP; Wed, 3 Aug 83 11:30:11 PDT
Date: Wed 3 Aug 83 11:29:23-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: [David Shaw <DAVID@COLUMBIA-20.ARPA>: New generation supercomputers]
To: ullman@SU-HNV.ARPA
Return-Path: <DAVID@COLUMBIA-20.ARPA>
Received: from csnet-sh by SU-SCORE.ARPA with TCP; Wed 3 Aug 83 11:20:19-PDT
Date: Wed 3 Aug 83 13:01:46-EDT
From: David Shaw <DAVID@COLUMBIA-20.ARPA>
Subject: New generation supercomputers
To: new-gen@UWISC.ARPA
Colleagues,
Let me apologize in advance for what will probably be a fairly long
message. Because of other commitments during the period of our
teleconference, though, it will be a lot easier for me to express my
thoughts in a small number of long contributions than to send quick
responses on an "almost daily" basis.
David
-----
I'm personally quite fond of the term "new generation supercomputers" as a
description of machines that might incorporate
1. A very large number (thousands to millions) of processing elements, and
2. Very high level mechanisms for the description of computational tasks
As Bill Buzbee and Jeff Ullman pointed out, the development of such
machines will depend on closely coupled advances in both the hardware and
software arenas. On the basis of the mutual reinforcement that Backus and
others have observed between old generation machine architectures and old
generation programming languages, I would predict that that these changes
will necessarily be revolutionary, and not evolutionary, in nature. For
several years now, I have also believed that such advances are likely to
draw heavily on techniques and formalisms borrowed from the areas of
artificial intelligence and logic programming, and on such closely related
computational operations as the relational algebraic primitives.
While the taxonomic classification of new generation supercomputers is, as
Scott Fahlman pointed out, often problematic, I'd like to share the scheme
I've been using for three years in teaching our course on Parallel
Architectures and VLSI Systems at Columbia. This scheme assigns each
"homogeneous machine" (I'll explain that term shortly) a position in each
of three interacting dimensions:
1. TOPOLOGY, which I divide into the following three categories:
1.1 The Meshes. In this category, I include the linear array, the two-,
three-, and other constant-dimensional meshes, and such "nearly isomorphic"
interconnection topologies as ring networks, chordal rings and toroidal
meshes.
1.2 The Trees. In practice, these almost always seem to be binary.
1.3 The High-Bandwidth Machines. This category includes machines based on
interconnection networks having a logarithmic number of stages, and
offering high thoughput with log-time latency.
The binary n-cube doesn't fit neatly into any of these slots, but is
closely related to both the first and last categories.
2. GRANULARITY, represented by a spectrum with the following end points:
2.1 Course Granularity Machines. These have a relatively small number of
fast processing elements, each associated with a substantial amount of
local memory.
2.2 Fine Granularity Machines. Such machines are characterized by a very
large number of comparatively slow processing elements having very little
local storage capacity.
3. SYNCHRONY, which is used to distinguish between the following two
categories:
3.1 Synchronous Machines. This category corresponds to Flynn's notion of
single instruction stream, multiple data stream (SIMD) machines, in which
each instruction is broadcast to a large number of processing elements,
which execute it simultaneously on different data elements.
3.2 Asynchronous Machines. These are the multiple instruction stream,
multiple data stream (MIMD) machines, which are capable of executing
independent instruction streams generated by programs stored in each of
their processing elements.
In my view, it is the interaction between choices along these three
dimensions that is of central interest to the architect of a new generation
machine. As an example of the interaction between topology and
granularity, note that the tree-structured and linear array topologies are
"fully scalable" (that is, they use only O(n) silicon area and O(1) pins,
where n is the number of embedded processing elements), and can thus be
used to build extremely fine granularity machines. The high-bandwidth
machines scale very poorly, while the higher-dimensional meshes fall
somewhere in between.
On the other hand, the high-bandwidth machines are able, for example, to
permute data elements from all processing elements much more rapidly than
the trees and linear arrays, with the higher-dimensional meshes again
occupying an intermediate position. We thus observe that very fine
granularity machines tend to employ one of the more scalable topologies,
while coarse granularity machines are free to employ more powerful,
asymptotically high-bandwidth interconnection schemes.
Granularity and synchrony also interact strongly, since the finest
granularity machines, while offering tremendous raw computational power,
have an insufficient amount of local memory to store their own programs
locally, and must "import" them from one or more external "control
processors". Such machines are thus typically synchronous. Coarser
granularity machines, on the other hand, are able to store substantial
programs locally, and thus tend to take advantage of the greater
algorithmic flexibility offered by asynchronous execution.
I should emphasize that this classification scheme applies on to what might
be called "homogeneous machines". In fact, I think there is good reason to
believe that most new generation supercomputers will ultimately be
"heterogeneous" in terms of topology, granularity and synchrony. Such
machines, exemplified by MIT's Connection Machine and the emerging design
for NON-VON 4, incorporate more than one kind of interconnection topology
to take advantage of the benefits of different classes of homogeneous
architectures.
As a final note, I think it's worth remembering Larry Snyder's warning that
"we don't live in Asymptopia". Although I've found the conceptual
framework outlined above to be useful in practice, I wouldn't really be too
surprised if some clever constant-juggler ultimately designed a superior
new generation supercomputer that represented a radical departure from the
"first principles" arguments I've just advanced.
David
-------
-------
∂03-Aug-83 1155 DFH travel arrangements
New flight to Frankfurt is:
Aug 7 PA72 lv. SFO 8:30am ar. Frankfurt 7:40am on Aug 8.
There is one stop, may or may not be an actual change of planes.
AAAI travel
Current reservations are:
Aug 21 TWA64 lv. SF 8:45am ar. Wash DC 4:40pm
Aug 27 TWA 63 lv Wash DC 4:45 pm ar. 7:22pm
Possible alternate:
Aug 21 UA lv. 1:10pm ar. 9:10 pm
Aug 27 UA lv. 9:30am ar. 12 noon.
I am not clear if you wanted to be there the night of Aug 26 or not, though
hotel reservation is made that way.
Let me know if you want to change any of this.
The flight to Frankfurt is ok. I prefer the 1:10pm flight to AAAI,
and I want to return Friday evening. I made the reservation for an
extra night just in case there is a reason to do otherwise.
∂03-Aug-83 1156 DFH phone msg.- Pournelle
Please phone him this afternoon about letter to editor--
needs to decide which signatures to use. He is going out
of town tomorrow. 213-762-2256
∂03-Aug-83 1206 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: spectrum survey
Received: from SU-SCORE by SU-AI with TCP/SMTP; 3 Aug 83 12:06:49 PDT
Received: from Shasta by Score with Pup; Wed 3 Aug 83 12:01:10-PDT
Date: Wednesday, 3 Aug 1983 12:05-PDT
To: Jeff Ullman <ullman at Diablo>
Cc: super at Score
Subject: Re: spectrum survey
In-reply-to: Your message of Wed, 3 Aug 83 11:36 PDT.
From: Steve Lundstrom <slndstrm@Shasta>
For the time being, I appreciate your forwarding the messages.
I have no problem browsing, but will need to setup an account
to gain access to your machine. Steve
∂03-Aug-83 1306 DFH kuo
kuo.2[let,jmc] is a draft of a brief letter to
Kuo. Let me know if it is ok.
∂03-Aug-83 1348 BSCOTT@SU-SCORE.ARPA No-Cost Extension, Current ARPA contract
Received: from SU-SCORE by SU-AI with TCP/SMTP; 3 Aug 83 13:48:03 PDT
Date: Wed 3 Aug 83 13:43:16-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: No-Cost Extension, Current ARPA contract
To: JMC@SU-AI.ARPA, TOB@SU-AI.ARPA, ZM@SU-AI.ARPA, Wiederhold@SUMEX-AIM.ARPA
cc: DFH@SU-AI.ARPA, MAS@SU-AI.ARPA, Yearwood@SU-SCORE.ARPA,
Atkinson@SU-SCORE.ARPA, BScott@SU-SCORE.ARPA
Have just learned that David Luckham has on his own requested a four-month
no-cost extension of your current ARPA contract. If granted, this will
extend the expiration date from 9/30/83 to 1/31/84.
I had intended to message you this week to ask whether you think you need
an extension. But in view of David's action, now all I need to know is
whether any of you anticipates needing longer than four months to expend
your funds. Unless I hear from you to the contrary by the end of next week,
I will assume that no further extension request is necessary, at least at
the present time.
Thanks,
Betty
-------
∂03-Aug-83 1435 ULLMAN@SU-SCORE.ARPA [Jack Schwartz <SCHWARTZ.CMCL1@NYU.ARPA>: The Challenge is Technological]
Received: from SU-SCORE by SU-AI with TCP/SMTP; 3 Aug 83 14:34:28 PDT
Date: Wed 3 Aug 83 14:29:53-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: [Jack Schwartz <SCHWARTZ.CMCL1@NYU.ARPA>: The Challenge is Technological]
To: ullman@SU-HNV.ARPA, super@SU-SCORE.ARPA
Return-Path: <SCHWARTZ.CMCL1@NYU.ARPA>
Received: from csnet-sh by SU-SCORE.ARPA with TCP; Wed 3 Aug 83 12:39:02-PDT
Date: 3 Aug 83 14:59 EDT
From: Jack Schwartz <SCHWARTZ.CMCL1@NYU.ARPA>
To: new-gen@UWISC.ARPA
Subject: The Challenge is Technological
Cc: Malvin Kalos <KALOS.CMCL1@NYU.ARPA>
Message-ID: <215524F7F.00C7001E;1983@CMCL1.NYU.ARPA>
I concur with the opinion, expressed by Ullman and Fahlman, that
the essential new opportunity is technological, viz. to exploit
parallelism on a very large scale, and that a first crucial question is
what form of interconnection to use. I agree with Ullman's
classification of interconnection architectures. Another parameter
entering into design is the size of the individual processing element to
be used. Suggestions have ranged from 'very large, fast as possible'
(leading to S1-like designs incorporating a relatively small number of
100MIP individual processors); thru 'most cost-effective, high
performance one-chip microprocessors' (characteristic for the
'ultracomputer' type of machine with which we are working), to 'largest
possible number of highly simplified processors' (characteristic for
image processors, systolic arrays, and various tree machines). It seems
to me that these two parameters, namely interconnection pattern and
processor size, are the key issues which the fifty-odd university
architectural suggestions now extant have been exploring. Naturally
enough, designs based on more substantial processors tend to use the
powerful anywhere-to-anywhere connections like the shuffle;
many-tiny-processor machines tend to favor connections, like the tree
and nearest neighbor, which lay out well on silicon.
The most important organisational issue now is to get a number of
really powerful parallel machines manufactured and into the hands of
application-oriented groups, so that real experience can begin to
accumulate on a large scale, efficiency questions be understood, etc.
Software and algorithm designers will have to learn how to organise
concurrent computation on a very large scale, develop operating systems,
manage I/O, etc.
I feel strongly that, with the possible exception of designs
specialised for image processing, this first round of machines should
aim at designs which are general rather than highly specialised, since
there is as yet no field in which parallel algorithms are well enough
understood for a large specialised engine to be a wise bet. In
particular, I think it rather unlikely that PROLOG will have major
programming advantages, or even that the best way to move toward an
efficent parallel implementation of PROLOG can be understood without a
great deal of penetrating research that still remains to be done. The
phrase 'Fifth Generation' is of course a catchword which should not be
allowed to blind us to the essentially technological nature of the
opportunity which fast cheap silicon now offers.
Once the technology matures, I believe we will see a continuing
contest to upgrade machines, and to offer better price-performance, much
like what we see now. It seems clear that the US and Japan will both
succeed in producing parallel supercomputers, and this within a very few
years of each other. Hence we are not so much racing for a unique
golden ring as positioning ourselves for the long haul.
Jack Schwartz
-------
-------
∂03-Aug-83 1515 BRACEWELL@SU-SIERRA.ARPA NASA Reports
Received: from SU-SIERRA by SU-AI with TCP/SMTP; 3 Aug 83 15:15:06 PDT
Date: Wed 3 Aug 83 15:16:08-PDT
From: BRACEWELL@SU-SIERRA.ARPA
Subject: NASA Reports
To: mccarthy@SU-AI.ARPA
Do you have or know where to get the following?
NASA NBSIR 82-2479 "Robotics"
82-2505 "Expert systems"
82-2582 "Computer vision"
NASA TM85635
NBSIR 83-2687 "Natural language processing"
NASA TM85636 "Artificial intelligence"
BRACEWELL@STAR
-------
∂03-Aug-83 1614 jlh%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: spectrum survey
Received: from SU-SCORE by SU-AI with TCP/SMTP; 3 Aug 83 16:13:22 PDT
Received: from Shasta by Score with Pup; Wed 3 Aug 83 16:05:55-PDT
Date: Wednesday, 3 Aug 1983 15:59-PDT
To: Jeff Ullman <ullman at Diablo>
Cc: super at Score
Subject: Re: spectrum survey
In-reply-to: Your message of Wed, 3 Aug 83 11:36 PDT.
From: John Hennessy <jlh@Shasta>
prefer to look through files - maybe with notes from anybody in the
group who sees something really worthwhile. John
∂03-Aug-83 2026 ZZZ
I WILL HAVE IT IN A FEW DAYS
∂03-Aug-83 2027 ZZZ
IF ITS MORE URGENT THAN THAT PLEASE LET ME KNOW
∂04-Aug-83 0707 Raj.Reddy@CMU-CS-A Gunkel-Fredkin-Reddy meeting
Received: from CMU-CS-A by SU-AI with TCP/SMTP; 4 Aug 83 07:07:40 PDT
Received: from [128.2.254.192] by CMU-CS-PT with CMUFTP; 4 Aug 83 09:57:25 EDT
Date: 4 August 1983 1001-EDT (Thursday)
From: Raj.Reddy@CMU-CS-A
To: JMC@SU-AI
Subject: Gunkel-Fredkin-Reddy meeting
CC: Raj.Reddy@CMU-CS-A, ef%oz@MIT-MC
Message-Id: <04Aug83.100126.RR29@CMU-CS-A>
Dear John,
The so-called Gunkel-Fredkin-Reddy meeting has nothing of Reddy except as
a possible participant. I don't want to get involved with one more thing at
this point. Happy to help Ed with it but don't want too much to do with
Mr. Gunkel.
Raj
Actually it seems to be the Fredkin meeting now. Frankly, I'm glad you're
not promoting it. We'll listen to Ed, but I think he'll have hard work
convincing us that the meeting would be worthwhile. I even suspect that
the task of producing a one page written plan may prove overwhelming.
∂04-Aug-83 0833 STAN@SRI-AI.ARPA Some info on IJCAI panel
Received: from SRI-AI by SU-AI with TCP/SMTP; 4 Aug 83 08:33:46 PDT
Date: 4 Aug 1983 0832-PDT
From: Stan at SRI-AI
Subject: Some info on IJCAI panel
To: jmc at SAIL
cc: stan
IJCAI Panel session on 5th Generation Project
Tuesday, 9 August 1983, 1:30 pm
Participants:
Chair, Ehud Shapiro
Herve Gallaire
John McCarthy
Stan Rosenschein
Toshio Yokoi (ICOT)
Questions to the panel [transcribed as they were read to me over the
phone by Ehud]
1. What is the 5th generation project really about?
2. Is it a coherent project or just a little something for
everyone?
3. Are its technical choices appropriate, e.g. logic programming,
dataflow?
4. Are the proposals to collaborate w/ the West serious? Can the
West be expected to help Japan leapfrog IBM?
5. Is the 5th generation project really about general-purpose
computing or special purpose for AI?
6. Are these the right questions to ask?
-------
I am dubious about the propriety of discussing question 4; it would lead
to a lot of speculation, veiled attacks on the Japanese motives and
equally veiled defenses. I would also favor greater emphasis on what are
the problems in achieving the expert system, knowledge engineering and AI
goals. My own interest is in common sense which was part of the original
goals but seems to have been de-emphasized. When I visited ICOT in May, I
asked who was working on it and was told "no-one". If common sense has
been dropped as a goal, what is the level of AI ambition in performance
terms. Is it well-defined. Is it suitable for a development project or
does it, as I suspect, require fundamental advances.
∂04-Aug-83 0931 KJB@SRI-AI.ARPA Thanks
Received: from SRI-AI by SU-AI with TCP/SMTP; 4 Aug 83 09:31:46 PDT
Date: 4 Aug 1983 0918-PDT
From: Kjb at SRI-AI
Subject: Thanks
To: jmc at SAIL
John, Thanks for coming to dinner last night. Jon
-------
∂04-Aug-83 0932 KJB@SRI-AI.ARPA program
Received: from SRI-AI by SU-AI with TCP/SMTP; 4 Aug 83 09:31:56 PDT
Date: 4 Aug 1983 0916-PDT
From: Kjb at SRI-AI
Subject: program
To: BMoore at SRI-AI, Bresnan at PARC, Briansmith at PARC,
To: grosz at SRI-AI, jmc at SU-AI, JRP at SRI-AI, Kaplan at PARC,
To: kjb at SRI-AI, lauri at SRI-AI, peters at UTEXAS,
To: rperrault at SRI-AI, sag%SU-psych at SUMEX-AIM,
To: stan at SRI-AI, wasow at PARC, Wasow%SU-psych at SUMEX-AIM,
To: winograd at PARC
Help, please. The new program has 4 parts, as I have said, one of
which is described as follows. "This collection of projects will
evaluate and further develop theories of syntax, phonology and
morphology, i.e. areas of traditional linguistics, in light of our
commitment to computation, semantics and language use." What I need
is a descriptive title for this area quickly. I take it that Charlies
"looking backward" is not what we want.
Jon
-------
∂04-Aug-83 1323 STAN@SRI-AI.ARPA Panel questions
Received: from SRI-AI by SU-AI with TCP/SMTP; 4 Aug 83 13:22:41 PDT
Date: 4 Aug 1983 1318-PDT
From: Stan at SRI-AI
Subject: Panel questions
To: jmc at SAIL
cc: stan
I agree with you about question 4, especially since IJCAI is an
international conference and not a strategy session for the US
computer industry. I'll pass your comments on to Shapiro. I also
agree with you about common sense. I just glanced into the ICOT
Journal (1983, No. 1) which has a piece by Yokoi in which he
acknowledges common sense as one of the two components of knowledge
(the other being expert knowledge), but doesn't say anything about
what they plan to do about it. More fundamentally, however, the sharp
distinction between expert & commonsense knowledge is surely absurd.
Everyone (or almost everyone) acknowledges that real expert knowledge
is seamlessly embedded in broad commonsense knowledge of time, space,
physical objects, events, purposeful action, etc. Expertise, after
all, is about things in the real world; how expert could a system
really be without knowledge of these things?
On the other hand, there is an argument (from Japan's point of view)
against working on commonsense knowledge. After all, applications
programming has been around for a long time, and many "conventional"
applications programs show considerable expertise. One might argue
that "expert systems" represent an improved (?) methodology for
building a class of applications program. Since Japan is in the 5th
generation business to make money and ICOT was set up as part of a
plan for making money, it boils down to evaluating risk & potential
return on investment. Apparently somebody believes that gigalips are
possible in the intermediate term and common sense is not! Well, how
much would YOU personally invest today in common sense reasoning? The
difference between a ten-year horizon and a thirty-year horizon may be
crucial to a business enterprise in ways that it can't be to a
science. Do you see any way of predicting, controlling, or even
measuring the rate of progress on fundamental issues like common sense
reasoning that would make sense to a Japanese manager? Is creative
research in AI really compatible with focused development programs
aimed at very specific goals. If not, when governments decide to
invest large sums in ICOT-style R&D, should scientists applaud,
cringe, warn, try to divert funds to their pet projects, ...?
--Stan
-------
∂04-Aug-83 1536 AAAI-OFFICE%SUMEX-AIM.ARPA@SU-SCORE.ARPA AAAI Election Results
Received: from SU-SCORE by SU-AI with TCP/SMTP; 4 Aug 83 15:35:58 PDT
Received: from SUMEX-AIM by Score with Pup; Thu 4 Aug 83 15:31:50-PDT
Date: Thu 4 Aug 83 15:33:54-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM>
Subject: AAAI Election Results
To: amarel@RUTGERS.ARPA, bobrow@PARC-MAXC.ARPA, buchanan@SUMEX-AIM,
bengelmore@SRI-KL.ARPA, lerman@SRI-KL.ARPA, genesereth@SUMEX-AIM,
grosz@SRI-AI.ARPA, hart@SRI-KL.ARPA, mccarthy@Score, mcdermott@YALE.ARPA,
minsky%mit-oz@MIT-MC.ARPA, nilsson@SRI-AI.ARPA, reddy@CMU-CS-A.ARPA,
rich%mit-oz@MIT-MC.ARPA, stan@SRI-AI.ARPA, gjs%mit-oz@MIT-MC.ARPA,
tenebaum@SRI-KL.ARPA, walker@SRI-AI.ARPA, dwaltz@BBNG.ARPA,
bonnie.upenn@UDEL-RELAY.ARPA
cc: aaai-office@SUMEX-AIM
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
The compilation of the election results have been completed. This office
is pleased to annnounce Woody Bledsoe has been elected President-elect for
1984-85. The councilors for the 1984-86 will be Randall Davis, Stan
Rosenschein,Eugene Charniak, and Mark Stefik.
You can express your personal congradulations to them at the EXCOM meeting, AUg24, in the Thoroughbred Room at 7:00 p.m.
--- Claudia Mazzetti
-------
∂04-Aug-83 1746 JJW@S1-A 206
Received: from S1-A by SU-AI with TCP/SMTP; 4 Aug 83 17:46:12 PDT
Date: 04 Aug 83 1747 PDT
From: Joe Weening <JJW@S1-A>
Subject: 206
To: jmc@SU-AI
Would you like to TA cs206 again?
Well, I'd rather not, to be honest. Have you considered Yoram? The main
reason is that I would like to begin some serious thesis work this fall,
and I'd find it hard to concentrate when something else would be taking up
so much time.
I do intend to provide some support for EKL on LOTS, if needed.
OK. I just wanted to give you first choice if you wanted it.
∂04-Aug-83 1750 JJW@S1-A
Received: from S1-A by SU-AI with TCP/SMTP; 4 Aug 83 17:50:23 PDT
Date: 04 Aug 83 1752 PDT
From: Joe Weening <JJW@S1-A>
To: jmc@SU-AI
It was announced today that Tom McWilliams and Jeff Rubin are leaving the
S-1 project to work for VALID, which develops and markets SCALD. This is
effective in about a month. They'll continue as part-time consultants, and
the goals of the project have not changed.
I would suggest not spreading this news around, though I'm sure in a few
days "everyone" will know about it.
∂04-Aug-83 2144 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Andy Freeman
Received: from SU-SCORE by SU-AI with TCP/SMTP; 4 Aug 83 21:44:21 PDT
Received: from Diablo by Score with Pup; Thu 4 Aug 83 21:39:25-PDT
Date: Thu, 4 Aug 83 21:38 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Andy Freeman
To: super@score
I have a request from Freeman to join the "group".
How do people feel about this?
∂04-Aug-83 2217 ullman%SU-HNV.ARPA@SU-SCORE.ARPA survey
Received: from SU-SCORE by SU-AI with TCP/SMTP; 4 Aug 83 22:17:30 PDT
Received: from Diablo by Score with Pup; Thu 4 Aug 83 22:12:41-PDT
Date: Thu, 4 Aug 83 22:00 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: survey
To: super@score
It is obvious that I cannot mail everything to everyone on the list,
because the fur is beginning to fly pretty fast.
I'm going to keep files in directory <ullman>@score and
/vlsi/ullman/scsurv@diablo. The first file is named quest1.
As it is getting rather big, I'll probably start putting things
in quest1a, quest1b, etc., then quest2, 2a, etc., when the next
questions get discussed.
∂04-Aug-83 2300 MAYR@SU-SCORE.ARPA supercomputer theory seminar
Received: from SU-SCORE by SU-AI with TCP/SMTP; 4 Aug 83 23:00:47 PDT
Date: Thu 4 Aug 83 22:55:30-PDT
From: Ernst W. Mayr <MAYR@SU-SCORE.ARPA>
Subject: supercomputer theory seminar
To: mpc-theory@SU-SCORE.ARPA, super@SU-SCORE.ARPA
A reminder for tomorrow's (Friday) meeting of the seminar on
"Theoretical aspects of massively parallel computation"
at 1:15pm, in MJH352.
On the agenda are:
-solution to last time's puzzle
-how to compute the EXOR in less than log←2 n steps (or not)
-Richard Anderson (about universal circuits, and uniform circuits),
-a drive for volunteer speakers
-Ernst
-------
∂05-Aug-83 0011 PMF@S1-A
Received: from S1-A by SU-AI with TCP/SMTP; 5 Aug 83 00:11:48 PDT
Date: 05 Aug 83 0012 PDT
From: Mike Farmwald <PMF@S1-A>
To: jmc@SU-AI
I'd like to talk sometime about ``Virtual Multiprocessors'' (the
idea of using simultaneously running many processes on a single pipeline.)
I have a lot of new ideas.
Good. Let's talk the next time I come to Livermore which unfortunately
won't be soon, because I'm travelling. Let me also suggest that you
consider talking in Jeff Ullman's supercomputer seminar. That's why
I'm forwarding this to him.
∂05-Aug-83 0850 jlh%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: Andy Freeman
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 08:50:40 PDT
Received: from Shasta by Score with Pup; Fri 5 Aug 83 08:38:05-PDT
Date: Friday, 5 Aug 1983 08:43-PDT
To: Jeff Ullman <ullman at Diablo>
Cc: super at Score
Subject: Re: Andy Freeman
In-reply-to: Your message of Thu, 4 Aug 83 21:38 PDT.
From: John Hennessy <jlh@Shasta>
ok for technical discussions - less appropriate for poltical and
planning ones. John
∂05-Aug-83 0914 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: Andy Freeman
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 09:13:56 PDT
Received: from Shasta by Score with Pup; Fri 5 Aug 83 09:08:41-PDT
Date: Friday, 5 Aug 1983 09:13-PDT
To: Jeff Ullman <ullman at Diablo>
Cc: super at Score
Subject: Re: Andy Freeman
In-reply-to: Your message of Thu, 4 Aug 83 21:38 PDT.
From: Steve Lundstrom <slndstrm@Shasta>
I agree with Hennessey.
In addition, I have had another request. Ken Stevens, who is
responsible for research in high performance numeric processing
at NASA Ames Research Center (and who will be sponsoring some of
my research), is interested in participating in an informal
way with the group. He has considerable experience in this area
having been involved both with architectural and software (languages
and operating systems) issues since the days of Illiac IV, the
development of the Flow Model Processor, and planning for future
NASA needs.
Steve
∂05-Aug-83 0924 ULLMAN@SU-SCORE.ARPA virtual multiprocessors
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 09:24:33 PDT
Date: Fri 5 Aug 83 09:18:40-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: virtual multiprocessors
To: pmf@S1-A.ARPA, jmc@SU-AI.ARPA
We'd like you to very much to speak. I suggest that we make the
talk sometime in the fall, because we are slowly converting to a
public seminar. In the meanwhile, I'm sure the group would be
happy to have you join us for informal discussions on Thursday afternoons.
-------
∂05-Aug-83 0948 BSCOTT@SU-SCORE.ARPA Research Associate Salaries
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 09:48:05 PDT
Date: Fri 5 Aug 83 09:30:19-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Research Associate Salaries
To: JMC@SU-AI.ARPA
cc: BScott@SU-SCORE.ARPA
John, your research associate salary recommendations for next year have
been approved:
Creary 8% $35,100
(even though he will be terminating at the end of November, we can't
hold up a salary increase without documenting that his work is un-
satisfactory; I have been waiting for this salary approval before
writing him about his lay-off benefits--will write this letter very
soon)
Gabriel 20.1% 37,000
Goad 14% 37,050
Ketonen 18.1% 44,400
(haven't forgotten about the promotion to senior research associate;
I'll get the form done soon)
Do you want to tell your people, or would you prefer that I do this
in your behalf?
Betty
-------
You might as well do it.
∂05-Aug-83 0946 DFH references -- psycho.4
1. The Allen Newell item is probably "The Knowledge Level" (rather than
The Logic Level) and it was the Presidential Address at AAAI80, as well
as appearing in the Jan. 1982 Artificial Intelligence and the Summer 1981
AI Magazine. Should I list all three references or which one do you want?
2. There is an article by John Searle in the Sept. 1980 issue of The
Behavioral and Brain Sciences entitled "Minds, Brains, and Programs". Does
this sound like the right one? I checked only the issues I could find in
your office, and there seems to be a gap in 1981, so if you think it could
be another article by him I will go check in the library.
Use "Artificial Intelligence" for the Newell reference, and you've got the
Searle reference correct.
∂05-Aug-83 1023 KJB@SRI-AI.ARPA mail at sail
Received: from SRI-AI by SU-AI with TCP/SMTP; 5 Aug 83 10:22:48 PDT
Date: Fri 5 Aug 83 10:22:38-PDT
From: KJB@SRI-AI.ARPA
Subject: mail at sail
To: jmc@SU-AI.ARPA
John, I can't figure out how to read my mail at sail the way I
now log in there, telneting from sri, on this terminal. In particular
I see I have a message from you that I have not read. Who should I
ask (net address) to get my mail there forwarded here?
You aren't going to steal Ostrom for C.S. are you? If so, how about
splitting him, if he would be interested?
Jon
-------
The message concerned going to dinner the other night. Probably you
should send ME a note asking to have mail to BYY as SAIL forwarded
automatically to KJB at SRI-AI. ME (Martin Frost) can also tell you
how to make your virtual terminal from SRI look like a SAIL terminal.
I plan no action until the beginning of September on Ostrom. Splitting
him might be if we decide to go in that direction.
∂05-Aug-83 1038 DFH Lifschitz called
Would like to see you today if possible.
(408) 446-0586
z=0∧E ∨ z≠1∧F : ∀xz.
∂05-Aug-83 1315 lantz%SU-HNV.ARPA@SU-SCORE.ARPA FYI
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 13:14:54 PDT
Received: from Diablo by Score with Pup; Fri 5 Aug 83 13:09:25-PDT
Date: Friday, 5 Aug 1983 13:09-PDT
To: super at Score, ng at Diablo
Subject: FYI
From: Keith Lantz <lantz@Diablo>
Looks like an interesting bibliography.
If you didn't already receive this message and want to be added to the
(strangely named) mailing list of Vineet's, I presume you can send him
a message.
Keith
------- Forwarded Message
Received: from SUMEX-AIM by Diablo with PUP; Thu, 4 Aug 83 17:41 PDT
Date: Thu 4 Aug 83 17:41:01-PDT
From: Vineet Singh <vsingh@SUMEX-AIM>
Subject: Distributed Problem Solving: An Annotated Bibliography
To: dps:;@SUMEX-AIM
For all of you who expressed interest in the annotated bibliography on
Distributed Problem Solving, here is some important information on how
to ftp a copy if you don't know this already.
The bibliography manuscript file "<vsingh.dps>dpsdis.bib" will be kept
on sumex-aim.arpa. Please login as "anonymous" with password
"sumexguest" (one word).
The file is by no means complete as you can see. It will be
continually updated. You may notice that the file is prepared for
Scribe formatting.
Please mail additional entries/annotations/corrections/suggestions to
me and I will incorporate them in the file as soon as possible. The
turnaround time will be a lot shorter if the new entries are also in
Scribe format. If you know anything about Scribe, please save me a
lot of effort and put your entries in Scribe format.
For those of you that did not see the original message, I have
reproduced it below.
-------------------------------------------------------------------------------
This is to request contributions to an annotated bibliography of
papers in *Distributed Problem-Solving* that I am currently compiling.
My plan is to make the bibliography available to anybody that is
interested in it at any stage in its compilation. Papers will be from
many diverse areas: Artificial Intelligence, Computer Systems
(especially Distributed Systems and Multiprocessors), Analysis of
Algorithms, Economics, Organizational Theory, etc.
Some miscellaneous comments. My definition of distributed
problem-solving is a very general one, namely "the process of many
entities engaged in solving a problem", so feel free to send a
contribution if you are not sure that a paper is suitable for this
bibliography. I also encourage you to make short annotations; more
than 5 sentences is long. All annotations in the bibliography will
carry a reference to the author. If your bibliography entries are in
Scribe format that's great because the entire bibliography will be in
Scribe.
Vineet Singh
(VSINGH@SUMEX-AIM.ARPA)
-------------------------------------------------------------------------------
-------
------- End of Forwarded Message
∂05-Aug-83 1402 JJW@S1-A Tape recorder
Received: from S1-A by SU-AI with TCP/SMTP; 5 Aug 83 14:02:24 PDT
Date: 05 Aug 83 1400 PDT
From: Joe Weening <JJW@S1-A>
Subject: Tape recorder
To: jmc@SU-AI
I have your tape recorder now. It would be easiest for me to bring it in to MJH
at about 7:30 or 8:00 p.m. tonight, or it could be earlier (6:00 to 6:30) if
you need it by then. I don't have a key to Diana's office, but if you're not
there when I bring it, I could put it in my office instead.
∂05-Aug-83 1410 BSCOTT@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 14:10:03 PDT
Date: Fri 5 Aug 83 14:05:21-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
cc: BSCOTT@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Fri 5 Aug 83 10:37:00-PDT
O.K., John, I'll take care of it.
Betty
-------
∂05-Aug-83 1430 reid@Glacier File server meeting
Received: from SU-GLACIER by SU-AI with TCP/SMTP; 5 Aug 83 14:30:01 PDT
Date: Friday, 5 August 1983 14:29:49-PDT
To: JMC@Sail
Cc: Mogul@Shasta, Mogul@Navajo
Subject: File server meeting
From: reid@Glacier
John,
I wrenched my back this afternoon and I am not going to be able to
make it to school for the 3:00 meeting on the file servers. If you can
reach Jeff Mogul, he will certainly represent my interests and will
know more about it than I do.
Brian
It's not a question of interests but of making the best decision on
how to spend $393,000 in a short time. I would really like your
help in making these decisions. However, as it turns out I have a
very bad cold and welcome any excuse to postpone the meeting. Len
Bosack will try to reschedule it for Thursday the 18th or Friday
the 19th. (If my cold lets up, I'll be going to Germany for the
IJCAI and won't be back till then). I hope your back is ok.
∂05-Aug-83 1536 GOTELLI@SU-SCORE.ARPA Meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 15:29:09 PDT
Date: Fri 5 Aug 83 15:23:50-PDT
From: Lynn Gotelli <GOTELLI@SU-SCORE.ARPA>
Subject: Meeting
To: REG@SU-AI.ARPA, Bosack@SU-SCORE.ARPA, Reid@SU-SHASTA.ARPA, JMC@SU-AI.ARPA
cc: gotelli@SU-SCORE.ARPA
The meeting on the expenditure of the ARPA funds has been
cancelled for 3:00 p.m. today and tentatively rescheduled
on August 19th at 3:00 p.m. I will send out a reminder
message about a week before.
Lynn Gotelli
-------
∂05-Aug-83 1632 DFH key/phone message
The key to rm. 353 is in the top right hand drawer.
Richard Schroeppel called again. Day phone is (213) 390-8611
nights (213) 398-9130.
∂05-Aug-83 2108 GOLUB@SU-SCORE.ARPA Vote for faculty position
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 21:08:35 PDT
Date: Fri 5 Aug 83 18:26:00-PDT
From: Gene Golub <GOLUB@SU-SCORE.ARPA>
Subject: Vote for faculty position
To: fullprof@SU-SCORE.ARPA, BUCHANAN@SU-SCORE.ARPA, binford@SU-AI.ARPA
Please indicate your vote for Leo Guibas as a full professor.
Vote
YES
NO
ABSTAIN
---------------------------------------------------------------
Comments
-------
∂05-Aug-83 2110 WEENING@SU-SCORE.ARPA Tape recorder
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Aug 83 21:10:46 PDT
Date: Fri 5 Aug 83 19:51:08-PDT
From: Joe Weening <WEENING@SU-SCORE.ARPA>
Subject: Tape recorder
To: jmc@SU-AI.ARPA
I put the tape recorder in my office, on the bookshelf.
Joe
-------
∂05-Aug-83 2134 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 5 August 1983
Previous Balance 0.98
Monthly Interest at 1.5% 0.01
-------
TOTAL AMOUNT DUE 0.99
Please deliver payments to Diana Hall, room 358, Jacks Hall.
Make checks payable to: STANFORD UNIVERSITY.
To ensure proper crediting, please include your Pony account name on your check.
Bills are payable upon presentation. Interest of 1.5% per month will be
charged on balances remaining unpaid 25 days after bill date above.
You haven't paid your Pony bill since 12/82.
Accounts with balances remaining unpaid for more than 55 days are
considered delinquent and are subject to reduction of credit limit.
Please pay your bill and keep your account current.
∂06-Aug-83 0006 JJW Tape recorder
I saw Carolyn here tonight, so I gave her the recorder.
∂06-Aug-83 1945 ZZZ REPORT
THANKS FOR SIGNATURE. I AM ON VACATION,
BUT WILL RETURN ON TUES AND WILL
GIVE IT TO BETTY WED MORN.
HAVE A NICE TRIP.
RICHARD
∂07-Aug-83 0953 BSCOTT@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 7 Aug 83 09:53:30 PDT
Date: Sun 7 Aug 83 09:49:28-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
cc: BSCOTT@SU-SCORE.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Sat 6 Aug 83 22:37:00-PDT
John, Joe Oliger, Marlie Yearwood and Richard Treitel are dealing with space
assignments/problems. I will forward your message to them.
Betty
-------
∂07-Aug-83 1022 reid@Glacier Re: Meeting
Received: from SU-GLACIER by SU-AI with TCP/SMTP; 7 Aug 83 10:22:37 PDT
Date: Sunday, 7 August 1983 10:22:39-PDT
To: Lynn Gotelli <GOTELLI@SU-SCORE>
Cc: REG@SU-AI, Bosack@SU-SCORE, Reid@SU-SHASTA, JMC@SU-AI
Subject: Re: Meeting
In-Reply-To: Your message of Fri 5 Aug 83 15:23:50-PDT.
From: Brian Reid <reid@Glacier>
I will be out of town August 17th through 20th, and I will be involved
in Dean Kays' summer school session most mornings the following week.
∂08-Aug-83 1725 KJB@SRI-AI.ARPA Sept 1
Received: from SRI-AI by SU-AI with TCP/SMTP; 8 Aug 83 17:25:04 PDT
Date: Mon 8 Aug 83 17:21:11-PDT
From: KJB@SRI-AI.ARPA
Subject: Sept 1
To: csli.people: ;
August 8, 1983
Dear members of CSLI,
I have been so preoccupied with ironing out the details
of the program and budget to be approved by SDF that I have been
remiss in keeping everyone posted on what is going on. To remidy
this, I would like to have a "colloquium" on September 1, where I
will explain what we have gotten ourselves involved in, outline
the program, answer questions, try to get volunteers, and so on.
I thought we might follow this by a gormet potluck dinner.
I will be very busy getting moved for some time, as my family and
furnishings start arriving this week, so I will ask people to
orgainize this.
In general, we should get things going in September.
September 1 is also the first TINLunch to be held at Ventura.
I hope everyone knows that Thursdays are the official
CSLI day, the day that everyone should keep regularly free to
attend things connected with CSLI. Specific projects will arrange
meeting times and places depending on the preferences of the people
involved. I will send out a list of the projects as soon as I learn
enough about emacs to pretty it up. For now, I hope to know enough
about it to get this saved and sent.
More soon,
Jon
-------
∂08-Aug-83 2243 KJB@SRI-AI.ARPA New CSLI Program
Received: from SRI-AI by SU-AI with TCP/SMTP; 8 Aug 83 22:43:30 PDT
Date: Mon 8 Aug 83 22:38:23-PDT
From: KJB@SRI-AI.ARPA
Subject: New CSLI Program
To: BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA, Briansmith@PARC-MAXC.ARPA,
grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA, JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA,
kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA, peters@SRI-AI.ARPA,
rperrault@SRI-AI.ARPA, sag%SU-psych@SUMEX-AIM.ARPA, stan@SRI-AI.ARPA,
wasow@PARC-MAXC.ARPA, Wasow%SU-psych@SUMEX-AIM.ARPA,
winograd@PARC-MAXC.ARPA, appelt@SRI-AI.ARPA, ford@PARC-MAXC.ARPA,
halvorsen@PARC-MAXC.ARPA, hans@SRI-AI.ARPA, hobbs@SRI-AI.ARPA,
kay@PARC-MAXC.ARPA, konolige@SRI-AI.ARPA, pollard.hplabs@UDEL-RELAY.ARPA,
pullum.hplabs@UDEL-RELAY.ARPA, pcohen@SRI-KL.ARPA, pereira@SRI-AI.ARPA,
shieber@SRI-AI.ARPA, stickel@SRI-AI.ARPA, stucky@SRI-AI.ARPA,
withgott@PARC-MAXC.ARPA
Dear CSLI Folks:
I think I have finally learned enough EMACS to send a real note. The
following describes the current (modified) research program. There
are four basic parts to the program, with three to five projects each.
Some of the projects continue work in progress, while some must be
initiated. Some will be phased out over time while some will continue
indefinitely. More specifics will be given in the individual
instances. The projects in parts A, B, and D are taken from the
original proposal, though they have been reorganized and, in some
instances, combined into a single project. The projects in part C
have been rethought and somewhat reduced in scope. For this reason,
they will be described in greater detail than those in Parts A, B, and
D.
As you all know, arriving at an acceptable research program has
involved some rather tortuous negotations with SDF, but I think this
is a version we can all live with. Feel free to send either
compliments or violent objections, but nothing in between.
Research Program for SDF Project
A. Traditional Linguistics and Situated Language
This collection of projects will evaluate and further develop theories
of syntax, phonology and morphology, i.e. areas of traditional linguistics,
in light of our commitment to computation, semantics, and language use.
1. Develop and integrate theories of phonology, morphology, and syntax.
Project manager: Karttunen (Kaplan during fall 1983)
Description: old 3.2.6
Time frame: It is expected that this project will be a
component of the project all four years.
2. Examine the effect of syntactic and phonological structure at the
level of discourse.
Project manager: Bresnan
Description: old 3.2.7
Time frame: This will be phased into B gradually starting in year 2.
3. Study strategies and tactics involved in processing utterances.
Project manager: Kaplan
Description: old 3.3.2
Time frame: This will be phased into Part B in years 3 and 4.
4. Analyze computational properties of alternative parsing algorithms.
Project manager: Peters
Description: old 3.3.1
Time frame: This will continue all for years, contibuting to B7
B. Theories of Situated Human Languages
This collection of projects aims at developing scientific theories of
natural language use consonant with our basic perspective on language
users as finite information processors.
1. Extend semantical theories
Project manager: Peters
Description: old 3.2.1, but incorporating as subprojects old
3.2.5 and 3.2.8.
Time frame: This will be a major project all four years.
2. Develop a semantics for sentences about mental states
Project manager: Perry
Description: old 3.2.2
Time frame: This will be phased into D4 after year 2.
3. Integrate syntactic and semantic accounts of discourse phenomena
Project manager: Sag
Description: old 3.2.3
Time frame: All four years
4. Integrate semantical and computational treatments of discourse
Project manager: Grosz
Description: old 3.2.4
Time frame: This will be phased into B5 during the second year.
5. Theory of communication
Project manager: Perrault
Description: old 3.5.1, with all of its subprojects.
Time frame: This will continue all four years, assuming
greater importance as progress is made in other projects that
feed into it.
C. Theories of Situated Computer languages
These projects aim at developing an understanding of current theories
and practices of computing, and at developing new languages,
architectures and theories, all within the perspective of computation
as linguistic activity. There will be a constant interplay between
these different activities, but we can break the task down into three
projects as follows.
1. Semantics of Computer Languages
Project manager: Barwise (acting, until a principal in this area is
added) Description: The aim is to develop a semantic account of
computation, rich enough to account for current computational
practise, that can lead into the kind of theory developed below. A
substantial body of work has been done on this by Dana Scott and
others working with Scott's domain theory. One of our first aims will
be to connect with this work. In particular, we plan to develop a
semantic account that can encompass both natural and computer
languages. Similarities between the information systems in domain
theory and informational relations in situation semantics point in
this direction. (For more on this, see old 3.4.1)
A number of other aspects of this project are described in the new
Xerox PARC subcontract in some detail. In particular, work will be
done on the semantics of the reflective lambda-calculus, leading into
the semantics of 3-LISP. Work will also be carried out on the
semantics of computational and reflective logics. Work on "syntactic
unification" will also be carried out -- the development of a general
theory of the syntactic and structural operations on which all formal
systems are based (procedure calls, structural unifications, schema
instantiation, etc.).
Time frame: During year 1 we will use one visitor position to bring in
a series of experts in this area to give lecture series. Smith will
hire one person at PARC as of about January 1, 1984. Smith himself
will devote about 30% of his time to this project this year. It is
expected that McCarthy and Feferman will contribute to this work as
well.
During year 2, Barwise plans to spend about 25% of his time on this
project. There will be a regular visitor. The level of activity at
PARC will remain constant over all 4 years, as will the contributions
of McCarthy and Feferman. In year 3 we hope to add a regular
principal at the Center in this area, if not before (such person to
become project manager as soon as possible).
2. Analysis and design of linguistically coherent computer languages
Project manager: Smith
Description: This work will be carried out primarily by Winograd and
an assistant, and by a team at PARC directed by Brian Smith. It will
concentrate on languages with declarative structures (ALEPH, for
example, and 4-LISP), on the relation between description and control,
and on the relation between input/output and communication
(essentially the "semantic translation" routines in a computational
setting). These are described in more detail in old 3.4.2.1 and in
the new subcontract to PARC.
Time frame: Brian Smith will spend about 50% of his time on this the
first year, decreasing to a small amount by the fourth year. Winograd
and his assistant will work on this all four years, it is expected.
In addition, there will be two research associates working with Smith
during the first two years, one the last two.
3. Computational architecture of reasoning.
Project manager: Smith
Description: old 3.5.2.4 and 3.4.3.3, plus McCarthy's work on
circumscription and on reasoning with concepts.
Time frame: McCarthy is expected to spend a fair amount of his time on
this project. Smith will begin with a relatively minor effort on his
own, bring in a research associate in year two, and another in year 3,
to continue work indefinitely.
D. Foundations underlying a unified view of language
These projects aim at developing the philosophical and mathematical
foundations needed to support the theories on situated language
developed here and elsewhere.
1. Computation, information, and logic
Project manager: Barwise
Description: old 3.6.1 and 3.6.4
Time frame: This will begin immediately with the visit of
Stavi. It will become a major project only as results from B
and C permit and personnel become available. We hope it can
begin in earnest in year 3.
2. Theories of reasoning and planning
Project manager: Rosenschein
Description: old 3.5.2 with all its subprojects
Time frame: This will continue current work and continue all
four years.
3. Theories of mind and action
Project manager: Perry
Description: old 3.6.2
Time frame: same as D2
4. Common sense theories of the world
Project manager: Moore
Description: old 3.6.3
Time frame: this will begin immediately, with work increasing as
dictated by need of other projects
-------
∂09-Aug-83 0253 ARK Chris Tucci's Comp Project
To: JMC@SU-AI
CC: Tucci@SU-SCORE, PW@SU-AI, ARK@SU-AI
I have looked at the project moderately carefully. Most of it looks pretty
good. I have sent him some comments on the writeup and on the program,
most of them due to my not understanding things and wanting some clarification.
It may be desirable for him to do some more work on the algorithm described
in figure 2 of his write-up (page 11). Most of the program is clear, but
the procedure GetNextState and the function Legal might merit a little
more attention.
I am leaving for New York this morning to return on the evening of 16th.
If, after looking at the two small things I have specified above, you decide
that it merits passing on to the Comp Committee, feel free to do so.
Otherwise, if you desire that Chris try to improve these sections, then
these are the only ones in my opinion that need looking at, so you could
check them over if you wish. I would be willing to look at them again
upon my return, but I want to ensure that I am not "holding things up"
in my absence.
I will be reading my mail occasionally from New York, so please let me
know if there is anything I can do or clarify about this review. Thanks.
Arthur
∂09-Aug-83 0314 ARK Facilities Commitee
To: JMC
CC: ARK
I'd suggest we meet sometime soon to discuss the rates for SAIL and SCORE
and to decide upon a successor for Ralph. It's fine with me if Len wants
to do it, but we need to decide that and give him the title. I'd suggest
the following:
* A meeting be set up for the afternoon of Aug 17 or Aug 18.
* Ralph and/or Len be instructed to supply the following information:
Expenses estimate for 1983-84 total and by system
Income by month by system and by category (A/B/C connect/CPU, disk)
Usage projections for 1983-84 by system by category per month
for worst, expected, and best case scenarios.
Proposed rates consonant with these figures
Reports requested
* Status of Dover charging
* Status of equipment acquisition
* Who applied for Ralph's job and what are their qualifications
Things to do once Ralph's successor is chosen
* Ensure feedback and communication between CSD-CF operations and CF oversight
committee
* Equipment and space plans
* What to do about SCORE's not-as-useful modems, CANON printer attached to SAIL,
BOISE
* What to do about random LOTS people using Jacks terminals for LOTS access
* Building and equipment security
Thanks. Feel free to recirculate extracts of this message.
Arthur
∂09-Aug-83 1051 AAAI-OFFICE%SUMEX-AIM.ARPA@SU-SCORE.ARPA Halbrecht's suggestion
Received: from SU-SCORE by SU-AI with TCP/SMTP; 9 Aug 83 10:50:44 PDT
Received: from SUMEX-AIM by Score with Pup; Tue 9 Aug 83 10:43:41-PDT
Date: Tue 9 Aug 83 10:49:07-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM>
Subject: Halbrecht's suggestion
To: JMC@Score
cc: aaai-office@SUMEX-AIM
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
John,
Herb Halbrecht, Pres of Halbrecht Associates - headhunters, has approached
me with the suggestion that he sponsor another award for the best published
applied research results. He's willing to sponsor this award for several
consecutive years. Should this be an agenda item for the EXCOM meeting?
Also, here is the financial report we will present at the EXCOM meeting.
Pls note that the figures for the registration and tutorial revenues will
change because of walk-ins.
----Claudia Mazzetti
BALANCE SHEETS for the YEARS
1980, 1981, 1982, and 1983 (projected)
1980 1981 1982 1983
(projected)
---------------------------------------------------------------------------
ASSETS
Current Assets 36,799 69,940 150,850 146,552
Fixed Assets 116 0 3,556 14,160
TOTAL ASSETS 36,915 69,940 154,406 160,712
LIABILITIES
Current Liabilities 0 0 1,821 2,000
Fund Balance 36,915 75,572 152,584 158,712
36,915 ( 5,631) 154,406 160,712
-1-
!
BALANCE SHEET
for 1983
JAN FEB MAR APR MAY JUN
ASSETS
Current Assets 129,239 125,389 102,386 73,069 49,988 93,578
Fixed Assets 4,019 4,210 13,635 13,635 13,635 14,160
TOTAL ASSETS 133,258 129,599 116,021 86,704 63,623 107,739
LIABILITIES
Current Liabilities 2,208 1,948 522 434 820 813
Fixed Liabilities 131,050 127,650 115,499 86,269 62,802 106,925
TOTAL LIABILITIES 133,258 129,599 116,021 86,704 63,623 107,739
2
! MEMBERSHIP COST CENTER
ACTUAL BUDGET YR TO DATE PROJECTED
1982 1983 (as of 6/83) 1983
---------------------------------------------------------------------------
MAINTENANCE
-----------
Dues 31,461 31,441 26,232 35,957
Sale of Mailing
list 960 4,800 2,239 6,739
------- ------- ------- -------
REVENUES 32,421 36,241 28,471 42,696
Refunds ( 912) ( ) ( 90) ( 900)
Postage ( 3,202) ( 4,500) ( 820) ( 2,000)
Printing ( 558) ( 2,500) ( 3,201) ( 6,000)
Misc ( 61) ( ) ( 31) ( 60)
AAAI labor ( 9,484) (13,800) ( 9,530) (14,865)
office expenses ( 5,219) ( 2,898) ( 3,538) ( 4,714)
-------- -------- -------- --------
EXPENSES (19,536) (23,698) (17,210) (28,539)
MAINTENANCE
GROSS MARGIN 12,985 12,543 11,261 14,157
AI MAGAZINE
-----------
Advertising 4,546 12,000 9,503 12,077
------- ------- ------- -------
REVENUES 4,546 12,000 9,503 12,077
L. Robinson
(35% gross) ( 1,410) ( 4,200) ( 4,538) ( 8,090)
Misc ( 1,348) 0 0 ( 50)
Printing (12,939) (20,000) (10,204) (21,000)
Typesetting,
graphics ( 5,250) (10,000) ( 3,808) ( 7,808)
Mailing ( 3,724) ( 8,000) ( 5,719) ( 7,719)
AAAI Labor ( 5,419) (11,000) ( 5,446) ( 8,506)
Office exp ( 2,982) ( 1,320) ( 2,022) ( 2,694)
-------- -------- -------- --------
EXPENSES (33,072) (54,520) (31,737) (55,867)
AI MAGAZINE
GROSS MARGIN (28,526) (42,520) (22,234) (39,252)
MEMBERSHIP
----------
TOTAL REVENUES 36,967 48,241 37,974 54,773
TOTAL EXPENSES (52,608) (78,218) (48,947) (84,406)
-------- -------- -------- --------
GROSS MARGIN (15,641) (29,977) (10,973) (29,633)
3
! AAAI-83 CONFERENCE COST CENTER
ACTUAL BUDGET YR TO DATE PROJECTED
1982 1983 (as of 6/83) (1983)
---------------------------------------------------------------------------
Registration 126,247 135,000 10,159 135,000
(1982 includes banquet)
Banquet 0 25,000 5,810 18,750
T-Shirts 2,500 3,450 0 3,450
-------- -------- -------- --------
TOTAL REVENUES 128,247 163,450 15,969 157,200
Facilities (14,532) 0 0 0
Refunds ( 5,489) 0 ( 663) ( 2,500)
Catering* (23,330) (23,148) 0 (25,000)
Program Comm. ( 6,172) (10,000) ( 5,516) (10,000)
Advertising 0 (10,000) ( 2,704) ( 2,704)
Postage ( 4,900) ( 6,000) (12,317) (12,317)
Conf Packets ( 3,974) ( 3,000) ( 2,998) ( 3,500)
Travel ( 3,676) ( 5,000) ( 884) ( 5,000)
A/V ( 1,238) ( 5,000) ( ) ( 5,000)
Printing ( 8,267) (16,229) (21,903) (21,903)
Design, Type-
setting ( 1,810) ( 4,024) ( 5,589) ( 6,000)
Banquet (13,875) (25,000) ( 1,059) (18,750)
AAAI Labor (21,677) (30,800) (21,783) (30,000)
Office exp (11,929) (14,784) ( 8,087) (10,775)
--------- --------- --------- ---------
TOTAL EXPENSES (120,869) (146,735) (83,473) (151,939)
CONFERENCE,
GROSS MARGIN 7,378 10,465 67,504 5,261
* Late '82 expenses
4
! AAAI-83 TUTORIAL COST CENTER
ACTUAL BUDGET YR TO DATE PROJECTED
1982 1983 (as of 6/83) 1983
-------------------------------------------------------------------------
Tutorial Fees 169,042 162,000 17,191 150,000
TecH Transfer
Symposium 0 52,500 11,375 52,500
(300@$175)*
-------- -------- ------- --------
TOTAL REVENUES 169,042 214,500 28,566 202,500
Lecturers'Fees (32,040) (31,000) ( ) (37,500)
(tutorials)
Lecturers' Fees* 0 ( 6,000) ( ) (19,125)
(tech transfer)
L. Robinson
(10% gross) (16,146) (13,500) ( ) (20,250)
Printing of
Syllabuses (14,588) (11,740) ( ) (11,740)
Refunds ( 160) ( 2,000)
Supplies ( 168) ( 200) ( 173) ( 173)
AAAI Labor ( 1,806) ( 5,000) ( 1,815) ( 2,835)
Office Exp ( 994) ( 200) ( 674) ( 2,039)
-------- -------- -------- --------
TOTAL EXPENSES (65,742) (80,765) ( 2,822) (95,489)
TUTORIAL,
GROSS MARGIN 103,300 146,860 25,744 106,838
* New entry
5
! AAAI-83 EXHIBIT COST CENTER
ACTUAL BUDGET YR TO DATE PROJECTED
1982 1983 (as of 6/83) 1983
-----------------------------------------------------------------------
Booth Fees 20,440 25,000 13,000 36,640
------- ------- ------- -------
TOTAL REVENUES 20,440 25,000 13,000 36,640
Space Rental ( 559) ( 2,000) ( ) ( 3,000)
Security 0 ( 3,000) ( ) ( 3,000)
Exhibit PR
materials ( 918) 0 ( 666) ( 666)
Mailing ( 38) 0 ( 655) ( 655)
L. Robinson
(25% gross) ( 6,769) ( 5,208) ( ) ( 7,925)
AAAI labor ( 4,516) ( 8,100) ( 4,538) ( 7,038)
Office Expenses( 2,485) ( 810) ( 1,685) ( 2,245)
--------- -------- -------- --------
TOTAL EXPENSES (15,285) (19,118) (17,544) (24,529)
EXHIBIT, GROSS
MARGIN 5,155 5,882 5,456 12,111
6
! PUBLICATION COST CENTER
ACTUAL BUDGET YR TO DATE PROJECTED
1982 1983 (as of 6/83) 1983
-----------------------------------------------------------------------
AI Magazine 1,000 2,700 983 1,500
Tutorial
Syllabuses 700 2,400 1,402 2,500
Proceedings 13,644 14,500 15,203 23,000
Page Overruns 3,600 3,600 6,900 7,100
------ ------ ------ ------
TOTAL REVENUES 18,944 23,200 24,448 34,100
W. Kaufmann
Printing ( 8,612) ( 7,000) ( ) ( 9,000)
Postage ( 1,162) ( ) ( ) ( )
Proceedings
Reprinting ( 3,491) ( 9,918) ( 9,918) ( 9,918)
ICJAI Reim-
bursement* 0 0 ( 8,241) ( 8,240)
AAAI Labor ( 2,258) ( 2,300) ( 2,269) ( 3,540)
Office Exp. ( 1,243) ( 115) ( 842) ( 1,122)
------ ------ ------ ------
TOTAL EXPENSES (16,766) (19,333) (21,270) (31,824)
PUBLICATION,
GROSS MARGIN 2,178 7,167 13,136 2,276
* New entry (for Proceedings Sales)
7
!
-------
-------
-------
Claudia:
Please put the Halbrecht proposal for awards on the agenda and arrange for
someone to present the proposal - yourself if you like.
Note that I am JMC@SAIL, not SCORE.
∂09-Aug-83 1538 GOLUB@SU-SCORE.ARPA ballots
Received: from SU-SCORE by SU-AI with TCP/SMTP; 9 Aug 83 15:37:50 PDT
Date: Tue 9 Aug 83 15:27:32-PDT
From: Gene Golub <GOLUB@SU-SCORE.ARPA>
Subject: ballots
To: full-professors@SU-SCORE.ARPA
Please sign your ballot which you turn in on the Guibas vote.
I have one unsigned ballot ; I need to know who cast it.
GENE
-------
∂09-Aug-83 2214 PW Common Lisp Proposal
To: JMC, RPG, DFH
PCL.TEX[PRO,JMC] contains the latest version of the common lisp proposal
to ARPA.
∂10-Aug-83 0754 KJB@SRI-AI.ARPA Mailing list for CSLI
Received: from SRI-AI by SU-AI with TCP/SMTP; 10 Aug 83 07:54:20 PDT
Date: Wed 10 Aug 83 07:49:01-PDT
From: KJB@SRI-AI.ARPA
Subject: Mailing list for CSLI
To: csli.people: ;
cc: wunderman@SRI-AI.ARPA, macken@SRI-AI.ARPA
Pat Wunderman is compiling a (hard-copy) mailing list for
CSLI. Please send her (wunderman@sri-ai) your address, your
net address, and the same for anyone you feel should be included
that might not be yet -- i.e., not a principal or associate.
Thanks, Jon
-------
My U.S. mailing address is:
John McCarthy
Computer Science Department
Stanford, CA 94305
tel: 497-4430
arpanet: JMC@SU-AI
∂10-Aug-83 1138 ATKINSON@SU-SCORE.ARPA 1983-84 Student Support
Received: from SU-SCORE by SU-AI with TCP/SMTP; 10 Aug 83 11:38:37 PDT
Date: Wed 10 Aug 83 11:31:24-PDT
From: Amy Atkinson <Atkinson@SU-SCORE.ARPA>
Subject: 1983-84 Student Support
To: jmc@SU-AI.ARPA
cc: dfh@SU-AI.ARPA
Stanford-Phone: (415) 497-9354
Though Autumn quarter still seems to be far off, it is again time for
you to let me know which students you intend to support for the
1983-84 academic year. I understand that things are apt to change at
some point during that time, but we would like to set up appointments
for the entire year so that we encounter less confusion than usual,
and less changes need to be made each quarter. I will, of course,
check with you each quarter to be sure that you wish to continue the
current support.
I ask that you please give me the names of incoming students you intend
to support as well as returning students. We must have all of these
appointments done by September 1st, which means that we need the
information from you by August 24th.
Thank you,
Amy
-------
∂10-Aug-83 1608 KJB@SRI-AI.ARPA Sept 1
Received: from SRI-AI by SU-AI with TCP/SMTP; 10 Aug 83 16:07:58 PDT
Date: Wed 10 Aug 83 15:54:41-PDT
From: KJB@SRI-AI.ARPA
Subject: Sept 1
To: csli.people: ;
cc: bmacken@SRI-AI.ARPA
The colloquium will be early on September 1, from 3 to around 5.
(There is a lot to say, even if a lot of it is boring. It still
needs to be known.) Then we will have our feast.
At the July 11 meeting of the principals, a vote was taken to
have a colloquium only every other week. However, I have an idea I would like to float, to get your reaction. It is based on something done
in the math department at MAdison which is a great success.
Namely, they have both an "inner" and an "outer" series. The
inner series is your typical math colloquium, but the outer
series consists of people from all over the university with
interesting things to say, no matter how loosely related to
math. Examples: the physics of the violin, mathematics in
James Joyces's Ulysses, the geometry of oriental rugs, group
theory and contra dances, ...
So the idea would be to do something similar here, with an
outer series where we would invite people from around the
area interested in aspects of language and information that
would be of a more recreational nature (I don't mean this in
any pejorative sense) than our regular series. IT might
be fun as well as give us a higher visibility around the area,
for whatever that is worth. What do you think of this?
Straw vote:
Great idea
OK idea
terrible idea
no opinion
_
-------
∂10-Aug-83 1625 KJB@SRI-AI.ARPA Dr. Betsey Macken
Received: from SRI-AI by SU-AI with TCP/SMTP; 10 Aug 83 16:25:05 PDT
Date: Wed 10 Aug 83 16:20:06-PDT
From: KJB@SRI-AI.ARPA
Subject: Dr. Betsey Macken
To: csli.people: ;
cc: bmacken@SRI-AI.ARPA
Dr. Betsey Macken, who will be assistant director of CSLI, and
project manager for the SDF grant, will begin work on Monday.
(Thank heavens.) Her net address at sri-ai is bmacken. Go
easy on the questions for a while, while she finds out what state
I have gotten things to (or is it in?).
Jon
-------
∂10-Aug-83 1635 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: Center for Parallel Computation
Received: from SU-SCORE by SU-AI with TCP/SMTP; 10 Aug 83 16:35:31 PDT
Delivery-Notice: While sending this message to SU-AI.ARPA, the
SU-SCORE.ARPA mailer was obliged to send this message in 50-byte
individually Pushed segments because normal TCP stream transmission
timed out. This probably indicates a problem with the receiving TCP
or SMTP server. See your site's software support if you have any questions.
Received: from Shasta by Score with Pup; Wed 10 Aug 83 16:25:12-PDT
Date: Wednesday, 10 Aug 1983 16:29-PDT
To: Jeff Ullman <ullman at Diablo>
Cc: super at Score
Subject: Re: Center for Parallel Computation
In-reply-to: Your message of Mon, 1 Aug 83 17:36 PDT.
From: Steve Lundstrom <slndstrm@Shasta>
Jeff: I've been maintaining contact with the Technology Development
people for some time. Do you want me to invite them to talk to the
group sometime? (They did not, of course, develop Illiac IV, they
were the operators of Illiac IV at IAC for many years. I can provide
more history if you are interested.) Steve
∂10-Aug-83 2113 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Re: Center for Parallel Computation
Received: from SU-SCORE by SU-AI with TCP/SMTP; 10 Aug 83 21:13:27 PDT
Received: from Diablo by Score with Pup; Wed 10 Aug 83 21:09:19-PDT
Date: Wed, 10 Aug 83 21:08 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Re: Center for Parallel Computation
To: slndstrm@Shasta
Cc: super@SU-Score
Let's talk about it at the Thursday meeting.
∂11-Aug-83 0306 corky@Navajo Summer Research
Received: from SU-NAVAJO by SU-AI with PUP; 11-Aug-83 03:06 PDT
Date: 11 Aug 1983 2:39-PDT
From: Robert.S..Cartwright@Navajo
Subject: Summer Research
To: JMC@SAIL
Message-Id: <83/08/11 0239.133@Navajo>
Adios until next visit (POPL?). I hope that you are recovering rapidly
from your cold and enjoying IJCAI. I talked to Diana and Carolyn
about the possibility of issuing some of the papers that I worked
on this summer under your sponsorship as Stanford Tech Reports. If you
are interested, I'm willing to have any or all of the papers
published in that form. Although they will all be published
elsewhere (conferences, journals), I would welcome your issuing them
as technical reports funded by the Formal Reasoning research program.
On the other hand, if you judge that it is not worth the expense and effort,
I will not be offended. So do what you think best.
During my visit, I worked on 4 papers and 2 summaries that will
shortly become papers. They are:
.np
Recursive Programs as Definitions in First Order Logic
(to be published imminently in SICOMP)
Non-standard Fixed Points in First Order Logic
(to be published in the Proceedings of the NSF/ONR
1983 Workshop on Logics of Programs)
Constructive Data Type Definitions (under
revision for TOPLAS)
The Semantics of Lazy and Industrious Evaluation
(with J. Donahue; to appear as XEROX PARC report;
submitted to SICOMP)
Computational Models for Programming Logics (summary;
to be submitted to POPL)
Lambda: the Ultimate Combinator
(joint with J. McCarthy; summary to be submitted
to POPL)
The final summary above, is out pending paper on
on extensional forms; by the time you read this
you should have received a draft from me in the
mail. If you object to either the title or the
content of the paper, please contact me
immediately by electronic mail or the phone.
Although we didn't interact as much I had hoped,
I had a very productive summer. For one of the
few times in my life, I was left free to work without
interruption. Thank you for your financial and
moral support.
Corky
∂11-Aug-83 0308 corky@Navajo Papers
Received: from SU-NAVAJO by SU-AI with PUP; 11-Aug-83 03:08 PDT
Date: 11 Aug 1983 3:06-PDT
From: Robert.S..Cartwright@Navajo
Subject: Papers
To: JMC@SAIL
Message-Id: <83/08/11 0306.550@Navajo>
I put copies of two of the papers in your box; I will
send you the others as soon as I create output suitable
for reproduction.
Corky
∂11-Aug-83 1029 ullman%SU-HNV.ARPA@SU-SCORE.ARPA Freeman
Received: from SU-SCORE by SU-AI with TCP/SMTP; 11 Aug 83 10:29:27 PDT
Received: from Diablo by Score with Pup; Thu 11 Aug 83 10:24:57-PDT
Date: Thu, 11 Aug 83 10:24 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Freeman
To: super@score
I few people felt (as I do) that it is fine for Andy to attend
the technical talks, but not the "political" side of the meeting.
Andy understands the constraints.
It seems that to be even-handed, we should see if there are othe
students, like Tom Blank, who have done something in parallel
computation and have the same sort of claim to join the group.
∂11-Aug-83 1114 ULLMAN@SU-SCORE.ARPA meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 11 Aug 83 11:13:55 PDT
Date: Thu 11 Aug 83 11:09:32-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: meeting
To: super@SU-SCORE.ARPA
There is a meeting today at 3PM;
the speaker is Fouad Tobagi.
-------
∂11-Aug-83 1231 RAC Party /Impending departure
To: OK@SU-AI, HHB@SU-AI, TOB@SU-AI, MAS@SU-AI, NAN@SU-AI,
ejh@SU-SCORE, cheeseman@SRI-AI, brown@SRI-KL,
park@SRI-AI, bach@SUMEX-AIM, ARG@SU-AI, ALL@SU-AI,
sally@SU-WHITNEY, PB@SU-AI, JMC@SU-AI, buchanan@SUMEX-AIM,
nilsson@SRI-AI
As most of you probably know, my time at Stanford is drawing to a close.
As some of you know, last weekend I got married in Maryland.
My wife, Lynn, is comming here at the end of the month and we are going to
drive back to Maryland together. A party is being planned for the dual purpose
of commemorating our marriage and my departure. The party will be on Sunday
afternoon the 28th of August at Bill Park's house. Stay tuned for details.
Rodger
∂11-Aug-83 1339 @MIT-MC:VVAWH@MIT-MC
Received: from MIT-MC by SU-AI with TCP/SMTP; 11 Aug 83 13:39:28 PDT
Date: 11 August 1983 11:28 EDT
From: Anthony W. Hackenberg <VVAWH @ MIT-MC>
To: JMC @ MIT-MC
RE: Trouble with IBM Asychronous Communication package to MACSYMA
I am having trouble with this communication package when logging onto
the MC machine at MIT. The line feed character is somehow being interpreted as a linefeed plus a CARRIAGE RETURN. This makes Macsyma's printout of integral
signs look like garbage.
Would appreciate any info you may have on this problem.
Dr. Edward C. Bittner of Lewis Research Center suggested that I contact you.
You can respond via MC ITS's mail commands or telephone me at (216) 433-4000 Ext. 478.
I cannot imagine why anyone would suggest contacting me about difficulties
with MC. I'm a very occasional user at Stanford.
∂12-Aug-83 1015 @SU-SCORE.ARPA:ATP.BLEDSOE@UTEXAS-20.ARPA My Operation
Received: from SU-SCORE by SU-AI with TCP/SMTP; 12 Aug 83 10:15:40 PDT
Received: from UTEXAS-20.ARPA by SU-SCORE.ARPA with TCP; Fri 12 Aug 83 08:33:47-PDT
Date: Fri 12 Aug 83 10:35:45-CDT
From: Woody Bledsoe <ATP.Bledsoe@UTEXAS-20.ARPA>
Subject: My Operation
To: Amarel@RUTGERS.ARPA, Bobrow@PARC-MAXC.ARPA, Buchanan@SUMEX-AIM.ARPA,
BEngelmore@SRI-KL.ARPA, LErman@SRI-KL.ARPA, csd.genesereth@SU-SCORE.ARPA,
Grosz@SRI-AI.ARPA, Hart@SRI-KL.ARPA, csd.mccarthy@SU-SCORE.ARPA,
mcdermot@YALE.ARPA, Minsky@MIT-MC.ARPA, Nilsson@SRI-AI.ARPA,
Reddy@CMU-CS-A.ARPA, Rich@MIT-MC.ARPA, aaai-office@SUMEX-AIM.ARPA,
stan@SRI-AI.ARPA, gjs@MIT-MC.ARPA, Tenenbaum@SRI-KL.ARPA,
Walker@SRI-AI.ARPA, DWaltz@BBNG.ARPA, Bonnie.upenn@UDEL-RELAY.ARPA,
Carbonell@CMU-CS-A.ARPA, Rhayes-roth@SRI-KL.ARPA, Bledsoe@UTEXAS-20.ARPA,
Lenat@SU-SCORE.ARPA, Rich@UTEXAS-20.ARPA, bundy@RUTGERS.ARPA,
wrightson@RUTGERS.ARPA
cc: ATP.Bledsoe@UTEXAS-20.ARPA
Dear Friends and Colleagues:
Let me tell you about my operation!! Not really. I am now
home doing nicely after gaul bladder removal. The bandages are all
off and I am doing really well (and enjoying some enforced leisure).
My surgeon won't let me drive my car until August 19, and no airplane
trips before September. But I will be able to start my classes as
planned on August 29.
So I not only missed IJCAI-83 (and all my friends there) but also
will miss AAAI-83 in Washington. Sorry that I will be unable to be
there for the AAAI Committee Meeting August 24, but Elaine Rich will
be there, and she can discuss AAAI-84, etc.
Since I have a terminal at home I can at least communicate that
way with the outside world. I Will give Bob and J their PV prize
checks when J returns from Germany. I hope all went will with the
prizes and with IJCAI in general.
Woody Bledsoe
-------
Glad you're better. IJCAIs seem to be unhealthy for you. We'll keep
you informed.
Woody, I just ran into your message again, and I must warn you not
to have anything to do with Frenchmen now that you are without a
gaul bladder.
∂14-Aug-83 2218 ULLMAN@SU-SCORE.ARPA bibliography
Received: from SU-SCORE by SU-AI with TCP/SMTP; 14 Aug 83 22:18:47 PDT
Date: Sun 14 Aug 83 22:19:25-PDT
From: ULLMAN@SU-SCORE.ARPA
Subject: bibliography
To: super@SU-SCORE.ARPA
The most recent question in the supercomputer survey asked for
a bibliography of relevant publications. Can you guys give me citations
for anything you have done in the supercomputer area (broadly interpreted)?
-------
∂15-Aug-83 0849 @USC-ISI,OHLANDER@USC-ISI
Received: from USC-ISI by SU-AI with TCP/SMTP; 15 Aug 83 08:49:25 PDT
Date: 15 Aug 1983 08:44-PDT
Sender: OHLANDER@USC-ISI
From: OHLANDER@USC-ISI
To: JMC@SU-AI
Message-ID: <[USC-ISI]15-Aug-83 08:44:33.OHLANDER>
In-Reply-To: The message of 01 Aug 83 1415 PDT from John McCarthy <JMC@SU-AI>
John,
It is important to get any changes in the task as soon as
possible. I am not sure what will be worked out with Dick in the
long run. For now, we have agreed to do some of the common Lisp
work and some benchmarking of S-1. Eventually, I hope to come to
some agreement on a portable Common Lisp compiler but we should
ignore that work for purposes of this task proposal. It can be
easily incorporated at a later date.
Ron
∂15-Aug-83 0850 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: bibliography
Received: from SU-SCORE by SU-AI with TCP/SMTP; 15 Aug 83 08:50:43 PDT
Received: from Shasta by Score with Pup; Mon 15 Aug 83 08:51:19-PDT
Date: Monday, 15 Aug 1983 08:51-PDT
To: ULLMAN at SU-SCORE.ARPA
Cc: super at SU-SCORE.ARPA
Subject: Re: bibliography
In-reply-to: Your message of Sun 14 Aug 83 22:19:25-PDT.
From: Steve Lundstrom <slndstrm@Shasta>
Jeff: See the following:
S.F.Lundstrom and G.H.Barnes, "A Controllable MIMD Architecture,"
PROCEEDINGS OF THE 1980 INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING,
IEEE Computer Society, 1980, pp 19-27.
(reprinted in:
TUTORIAL ON PARALLEL PROCESSING, by Robert H. Kuhn and David Padua,
IEEE Computer Society - The Computer Society Press, pp 165-173.)
G.H.Barnes and S.F.Lundstrom, "Design and Validation of a Connection
Network for Man-Processor Multiprocessor Systems," COMPUTER Magazine,
Vol. 14, No. 12, Dec 1981, pp 31-41.
Numerical Aerodynamic Simulation Facility, Preliminary Study, Burroughs
Corporation, October 1977 (NASA CR-152060 & CR-152061 & CR-152062).
Numerical Aerodynamic Simulation Facility, Preliminary Study Extension,
Burroughs Corporation, February 1978 (NASA CR-152106 & CR-152107).
Numerical Aerodynamic Simulation Facility, Feasibility Study, Burroughs
Corporation, March 1979 (NASA CR-152284 & CR-152285).
There were also reports on the NASF (Numerical Aerodynamic Simulation
Facility) by CDC which I suspect that they will cite. If they do not,
I can look them up. Note that the most important work (the system
design studies for the Numerical Aerodynamic Simulator Processing
System, NASPS) was not published and was generally considered to be
proprietary. Burroughs has given me permission to publish the work
we did, and I am in the process of preparing a number of papers on
the Burroughs Flow Model Processor (the computational portion of the
NASPS).
If the general area of data flow and data driven machines is to be
cited, Al Davis (now at Fairchild R&D Labs) whould be invited to
provide his list of citations. I can also provide some of those, but
they are likely not to be complete.
Steve Lundstrom
∂15-Aug-83 0855 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Meeting of Sept. 15
Received: from SU-SCORE by SU-AI with TCP/SMTP; 15 Aug 83 08:55:01 PDT
Received: from Shasta by Score with Pup; Mon 15 Aug 83 08:55:34-PDT
Date: Monday, 15 Aug 1983 08:55-PDT
To: super at Score
Cc: slndstrm at Shasta
Subject: Meeting of Sept. 15
From: Steve Lundstrom <slndstrm@Shasta>
Dick Wilson or Harry Heard of Technology Development Corporation
are planning to attend the 9-15-83 meeting with us (tentatively)
if that schedule fits everyone. Does the group have any specific
agenda you would like me to suggest to them?
Steve Lundstrom
∂15-Aug-83 1352 GOTELLI@SU-SCORE.ARPA Meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 15 Aug 83 13:52:07 PDT
Date: Mon 15 Aug 83 13:52:29-PDT
From: Lynn Gotelli <GOTELLI@SU-SCORE.ARPA>
Subject: Meeting
To: REG@SU-AI.ARPA, JMC@SU-AI.ARPA, Bosack@SU-SCORE.ARPA
The meeting on the expenditure of the ARPA funds has been
scheduled for this Friday 8/19 at 3:00 p.m. in Ralph's
or Len's office. John, if you cannot attend please let me
know.
Lynn
-------
∂15-Aug-83 1524 Bulnes.PA@PARC-MAXC.ARPA Party on Thursday
Received: from PARC-MAXC by SU-AI with TCP/SMTP; 15 Aug 83 15:24:35 PDT
Date: 15 Aug 83 15:25:13 PDT (Monday)
From: Bulnes.PA@PARC-MAXC.ARPA
Subject: Party on Thursday
To: JMC@su-ai.ARPA
cc: , Bulnes.PA@PARC-MAXC.ARPA
John,
Stefania and myself would like to invite you and your wife to an
informal party on Thursday, August 18, from 5.30 pm until dusk, in the
gardens of Kingscote Gardens (i.e. at the side of the Stanford Faculty
Club towards the Lagunita, i.e. in fornt of the International Center).
Sorry for the short notice, but we only decided on this party yesterday.
Hope you can come.
Juan
∂16-Aug-83 0826 RINDFLEISCH@SUMEX-AIM Lisp Machines
Received: from SUMEX-AIM by SU-AI with PUP; 16-Aug-83 08:26 PDT
Date: Tue 16 Aug 83 08:28:41-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Lisp Machines
To: JMC@SAIL, RPG@SAIL
cc: Rindfleisch@SUMEX-AIM
When can we meet so I can find out what is on your minds? I have all
sorts of machinery at Stanford and Symbolics stopped and hanging fire.
Tom R.
-------
∂16-Aug-83 1816 RAC
To: OK@SU-AI, HHB@SU-AI, TOB@SU-AI, MAS@SU-AI, NAN@SU-AI, ARG@SU-AI,
PB@SU-AI, park@SRI-AI, brown@SRI-KL, cheeseman@SRI-AI,
nilsson@SRI-AI, buchanan@SUMEX-AIM, bach@SUMEX-AIM,
sally@SU-WHITNEY, feigenbaum@SUMEX-AIM, GFS@SU-AI, JMC@SU-AI,
ejh@SU-SCORE, rtxenm@AMES-TSS
Goodbye party
Below is the map and details. Hope you can make it. Lynn, my wife of just
over one week now, will be flying in on Sat and will be there too.
| | |O<--Bill Park's house (3rd on right) | SUNDAY
| | |O 1430 Mills Court, Menlo Park | AUG 28
| + |O (415) 326-6369 | Noon-6:00pm
E |--|--+---+-------------------------------------| M
l | | |M |L Glenwood Ave | i POT LUCK
| +S |i |a | d GOODBYE PARTY
C | |P |l |u | d FOR
a | + |l |r | l RODGER CLIFF
m | | |s |e | e
i | + | |l Oak Grove Ave | f BRING A MAIN
n |--|--+---+-------------------------------------| i COURSE &/OR
o | +t |S | e LIQUIDS.
| |r |t | l
R | +a |r | d BBQ COALS,
e | |c |e | SALAD, DIP,
a | +k |e | R MUNCHIES
l | |s |t Ravenswood Ave | d PROVIDED
|--|------+-------------------------------------|
| | | ** SRI ** |
-------
------- End undelivered message -------
------- End undelivered message -------
∂17-Aug-83 1048 CHAPPELL@SRI-AI.ARPA Message ID
Received: from SRI-AI by SU-AI with TCP/SMTP; 17 Aug 83 10:47:42 PDT
Date: 17 Aug 1983 1016-PDT
From: Chappell at SRI-AI
Subject: Message ID
To: jmc at SU-AI, jmc at SAIL
cc: chappell
I am in the process of making a list of all Principals and Associates
in the CSLI. I need to know your correct mail address to send
TINLUNCH announcement. When I sent a msg. to JMC-list@sail it came
back as undeliverable.
Thanks
--Elsie Chappell
chappell@sri-ai
-------
∂17-Aug-83 1048 CHAPPELL@SRI-AI.ARPA Message ID
Received: from SRI-AI by SU-AI with TCP/SMTP; 17 Aug 83 10:48:11 PDT
Date: 17 Aug 1983 1016-PDT
From: Chappell at SRI-AI
Subject: Message ID
To: jmc at SU-AI, jmc at SAIL
cc: chappell
I am in the process of making a list of all Principals and Associates
in the CSLI. I need to know your correct mail address to send
TINLUNCH announcement. When I sent a msg. to JMC-list@sail it came
back as undeliverable.
Thanks
--Elsie Chappell
chappell@sri-ai
-------
That should have been jmc-lists@sail.
∂17-Aug-83 1638 ULLMAN@SU-SCORE.ARPA [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
Received: from SU-SCORE by SU-AI with TCP/SMTP; 17 Aug 83 16:38:10 PDT
Date: Wed 17 Aug 83 16:38:29-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
To: super@SU-SCORE.ARPA
Any thoughts??
---------------
Return-Path: <wallich@Uwisc>
Received: from csnet-sh by SU-SCORE.ARPA with TCP; Wed 17 Aug 83 15:33:44-PDT
Date: 17 Aug 1983 17:22:00-CDT
From: wallich@csnet-sh.ARPA (Paul Wallich)
Reply-to: wallich@csnet-sh.ARPA
To: new-gen@Uwisc.ARPA
Subject: Fourth question
This conference has been taking place by means of a distribution
list on the CSNET service host, which is sponsored by the NSF.
Most of the participants are on the ARPAnet, which is sponsored
by DARPA. Many people have reached those nets via intracompany or
intrauniversity nets such as those at HP, PARC, IBM, CMU, and
MIT. [There have been some adventures in trying to find people on
these nets, which points to an interesting question for future
co-operation.]
If computer networks are going to be an important part of multi-site
cooperation for next-generation research, who is going to sponsor
them, how will they be connected, and who will have access to them? If
other facilities are to spring up for such cooperation, such as
hardware and software prototyping and simulation services, who will
pay for them? In short who is going to sponsor next-generation
research and how will the resources get where they should go? How will
work sponsored by different agencies be co-ordinated?
Paul Wallich Alan Bell
Chuck House
Bert Raphael
-------
∂17-Aug-83 1642 DFH
Messages
1. Lewis called on Aug 8 to invite you to something, which you weren't around
for. He'll be in touch later.
2. Wendy Rauch-Hindin called, wants to set up a dinner appointment with you
at AAAI. Her no. is 516-667-7278. I didn't take this message, so don't
know anything further about who she is.
3. Wendell Mead of Spectra Research Systems came by. He was here just for the
day with someone from the Army. He will phone you Thurs. or Fri. to
try to set up some time to talk with you, perhaps at AAAI in Washington.
4. A former student of yours, Bulnes, called, inviting you to a party Thurs
night, Aug. 18. He said he would send you a separate message about
this, however.
5. Beverly Parker of ILEX called, wants to talk with you about bridge for
CPM capabilities. 415-652-3900.
6. ARK wants to schedule a facilities committee meeting sometime prior to
September 1. It looks to me like the only three days that would be
possible for you would be Aug 29, 30,31, unless you want to try
for this Thurs or Fri.
∂17-Aug-83 1731 cheriton%SU-HNV.ARPA@SU-SCORE.ARPA Re: [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
Received: from SU-SCORE by SU-AI with TCP/SMTP; 17 Aug 83 17:31:10 PDT
Received: from Diablo by Score with Pup; Wed 17 Aug 83 17:31:38-PDT
Date: Wed, 17 Aug 83 17:21 PDT
From: David Cheriton <cheriton@Diablo>
Subject: Re: [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
To: ULLMAN@SU-Score, super@SU-Score
Seems like a dumb question to me. Up to epsilon, money will come from where
it has always come from, resources will go to those with money, and
work will be no better coordinated than it has ever been. Coordinated
{research (beyond fairly harmless categories) produces poor research
results in my view.
Finally, everyone who is anyone will be (is?) connected to some network
which is connected to all other networks, although perhaps by arbitrarily
ugly means (including if networking standards have any effect).
Next question?
∂17-Aug-83 2208 GENESERETH@SUMEX-AIM debate
Received: from SUMEX-AIM by SU-AI with PUP; 17-Aug-83 22:08 PDT
Date: Wed 17 Aug 83 22:10:18-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
Subject: debate
To: jmc@SAIL
John,
Don't forget you are scheduled to spar with Schank in the Great
Debate next Thursday afternoon. I've asked Cordell to moderate,
and I hope he has had a chance to reach you by now. If not
let me remind you that it is a formal dbate (15 minutes of
position statement, 5 minutes of rebuttal, and optional rejoinder);
and the topic is (loosely) "Methodologies for AI Research".
You'll recall that the reaon for the vague topic is that Roger
didn't want to be billed as being AGAINST anything, e.g. logic.
However, he intends to present his position accurately
during the debate. Good luck!
mrg
-------
∂18-Aug-83 1009 RINDFLEISCH@SUMEX-AIM Symbolics Deal
Received: from SUMEX-AIM by SU-AI with PUP; 18-Aug-83 10:09 PDT
Date: Thu 18 Aug 83 09:59:35-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: Symbolics Deal
To: JMC@SAIL
cc: Rindfleisch@SUMEX-AIM, Feigenbaum@SUMEX-AIM, RPG@SAIL
John, I need to talk to you asap about the Symbolics deal. I believe
I've worked out all the administrative and legal details and Symbolics
is ready to deliver 2 machines in September. HPP (Genesereth) badly
needs at least one of those machines right now but this is all wrapped
up in the LM-2 trade for 5 3600's. So I have to know why you changed
your mind and what is hanging you up.
Tom R.
-------
∂18-Aug-83 1016 DFH Susan called
Your daughter Susan called. Please phone her
at home.
∂18-Aug-83 1117 GOLUB@SU-SCORE.ARPA Vote on Guibas
Received: from SU-SCORE by SU-AI with TCP/SMTP; 18 Aug 83 11:17:33 PDT
Date: Thu 18 Aug 83 11:08:48-PDT
From: Gene Golub <GOLUB@SU-SCORE.ARPA>
Subject: Vote on Guibas
To: fullprof@SU-SCORE.ARPA
!Germund Dahlquist!
!George Dantzig! YES
!Ed Feigenbaum! ABSTAIN
!Bob Floyd! YES
!Gene Golub! YES
!Don Knuth! YES
!Zohar Manna! YES
!John McCarthy! ABSTAIN
!Ed McCluskey! YES
!W.F. Miller!
!Christos Papadimitriou!YES
!Vaughn Pratt! YES
!James Wilkinson!
!Jeff Ullman! YES
Based on this vote, we will go ahead with the appointment. Don is in
the process of completing the paper work.
I think it should be understood that this appointment is being made
with the understanding that Guibas has no other institutional appointment.
GENE
-------
∂18-Aug-83 1205 WUNDERMAN@SRI-AI.ARPA Your Address Memo
Received: from SRI-AI by SU-AI with TCP/SMTP; 18 Aug 83 12:05:28 PDT
Date: 18 Aug 1983 0950-PDT
From: Wunderman at SRI-AI
Subject: Your Address Memo
To: JMC at SU-AI
I did receive the information--thanks, John.
Pat Wunderman at CSLI
-------
∂18-Aug-83 1653 YOM CS206
The reading list has been orderred the way you wished. The bookstore
will be distributing buth the Revised MacLisp Man. and the Lisp
Programming and Proving book. Pitman at MIT has reserved 75 copies
for our use.
-yom
∂18-Aug-83 1731 TRATTNIG@SU-SCORE.ARPA Special Meeting
Received: from SU-SCORE by SU-AI with TCP/SMTP; 18 Aug 83 17:30:57 PDT
Date: Thu 18 Aug 83 17:31:30-PDT
From: Werner Trattnig <TRATTNIG@SU-SCORE.ARPA>
Subject: Special Meeting
To: super@SU-SCORE.ARPA
cc: Pratt@SU-NAVAJO.ARPA
Jack Dennis from M.I.T. will visit the Computer Systems Lab on Friday,
September 16. He would like to give a short presentation of his
current research in data flow computer architecture and programming
language to our supercomputer group followed by a discussion. Thus we
have decided in our meeting today to have a special seminar on
Friday, September 16
3 pm (s.t.)
CIS Conference Room (ERL)
W-
-------
∂18-Aug-83 1829 GOLUB@SU-SCORE.ARPA amended vote
Received: from SU-SCORE by SU-AI with TCP/SMTP; 18 Aug 83 18:29:42 PDT
Date: Thu 18 Aug 83 18:30:05-PDT
From: Gene Golub <GOLUB@SU-SCORE.ARPA>
Subject: amended vote
To: Full-Professors: ;
Apparently the full professors distribution list had not been properly
updated. Here is the vote for Leo.
Vote on Guibas
!Tom Binford! YES
!Bruce Buchanan! YES
!Germund Dahlquist!
!George Dantzig! YES
!Ed Feigenbaum! ABSTAIN
!Bob Floyd! YES
!Gene Golub! YES
!Don Knuth! YES
!Zohar Manna! YES
!John McCarthy! ABSTAIN
!Ed McCluskey! YES
!W.F. Miller!
!Christos Papadimitriou!YES
!Vaughn Pratt! YES
!James Wilkinson!
!Jeff Ullman! YES
!Andrew Yao! YES
Based on this vote, we will go ahead with the appointment. Don has
completed the first draft of the paper work.
GENE
-------
-------
∂19-Aug-83 0839 BSCOTT@SU-SCORE.ARPA Personnel
Received: from SU-SCORE by SU-AI with TCP/SMTP; 19 Aug 83 08:38:58 PDT
Date: Fri 19 Aug 83 08:39:39-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: Personnel
To: JMC@SU-AI.ARPA
You're quite right. Personnel is plural--we MUST get the form changed!
Betty
-------
∂19-Aug-83 1606 SHARON@SU-SCORE.ARPA box
Received: from SU-SCORE by SU-AI with TCP/SMTP; 19 Aug 83 16:06:10 PDT
Date: Fri 19 Aug 83 16:06:11-PDT
From: Sharon Bergman <SHARON@SU-SCORE.ARPA>
Subject: box
To: jmc@SU-AI.ARPA
cc: DFH@SU-AI.ARPA
There is a box that just arrived for you via UPS.
Sharon
-------
∂19-Aug-83 1732 RESTIVO@SU-SCORE.ARPA Phillipe Roussel
Received: from SU-SCORE by SU-AI with TCP/SMTP; 19 Aug 83 17:32:29 PDT
Date: Fri 19 Aug 83 17:19:11-PDT
From: Chuck Restivo <RESTIVO@SU-SCORE.ARPA>
Subject: Phillipe Roussel
To: jmc@SU-AI.ARPA, restivo@SU-SCORE.ARPA
A friend named Phillipe Roussel is going to be visiting the area
from Caracas from August 30th thru September 1st or second. He
would like to make an appointment with you during this time.
Phillipe mentioned that he was involved with a group in Marseille
that you had visited or worked with. If getting together is
feasible during this time, might you let me know ? Thank you.
Chuck Restivo
-------
I would be glad to see Phillippe Roussel and will be at Stanford then.
∂19-Aug-83 1821 RESTIVO@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 19 Aug 83 18:21:20 PDT
Date: Fri 19 Aug 83 18:22:11-PDT
From: Chuck Restivo <RESTIVO@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Fri 19 Aug 83 17:34:00-PDT
I will tell him the good news this Monday. I hope to have the opportunity
to meet you myself.
-------
∂20-Aug-83 1000 JMC*
super@score
∂20-Aug-83 1032 ATP.BLEDSOE@UTEXAS-20.ARPA
Received: from UTEXAS-20 by SU-AI with TCP/SMTP; 20 Aug 83 09:11:46 PDT
Date: Sat 20 Aug 83 11:12:30-CDT
From: Woody Bledsoe <ATP.Bledsoe@UTEXAS-20.ARPA>
To: JMC@SU-AI.ARPA
cc: ATP.Bledsoe@UTEXAS-20.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Thu 18 Aug 83 09:26:00-CDT
Thanks John. I am doing pretty well now. But the doctor advised me not to
come to Washington.
I'd still like to work out the arrangement for you to come to Austin
to discuss the Chair (To advise us if nothing else, but hopefully more).
I believe you can pretty well pick you dates. We are having the UT
centenial program Oct 20-21, so that isn't the best time. But since
you will be here then anyway, we could probably work that out for you,
on say Oct 19-20, If that would help. (I know how busy you are).
Woody
-------
∂20-Aug-83 1205 KJB@SRI-AI.ARPA Progress!
Received: from SRI-AI by SU-AI with TCP/SMTP; 20 Aug 83 12:05:25 PDT
Date: Sat 20 Aug 83 12:01:40-PDT
From: KJB@SRI-AI.ARPA
Subject: Progress!
To: BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA, Briansmith@PARC-MAXC.ARPA,
grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA, JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA,
kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA, peters@SRI-AI.ARPA,
rperrault@SRI-AI.ARPA, sag%SU-psych@SUMEX-AIM.ARPA, stan@SRI-AI.ARPA,
wasow@PARC-MAXC.ARPA, Wasow%SU-psych@SUMEX-AIM.ARPA,
winograd@PARC-MAXC.ARPA
Dear all,
Charlie returned late Thursday after being away two weeks.
Brian, Rich Dutcher, Betsy Macken and I met with him yesterday
for two hours and presented him with piles of stuff. Things seemed
to go very well (by the standard of other meetings with Charlie)
and he now hopes to be able to "pull all the pieces together and
issue the grant letter next week".
I will be busy with the SID meeting Mon-Thursday, but
Betsy Macken is going to start the wheels in motion to get the
PARC subcontract and SRI gift awarded as soon after the grant
becomes official as possible. Whether this will be by Sept 1 or
not is unclear. We will put all the pressure to bear that we can
to make this happen.
There is a lot more to say, but I will save it for the
family gathering on Sept 1.
Jon
-------
∂20-Aug-83 1551 @MIT-MC:PHW%MIT-OZ@MIT-MC Your AI book
Received: from MIT-MC by SU-AI with TCP/SMTP; 20 Aug 83 07:19:33 PDT
Date: Sat, 20 Aug 1983 10:19 EDT
Message-ID: <[MIT-OZ].PHW.20-Aug-83 10:19:13>
From: Patrick H. Winston <PHW%MIT-OZ@MIT-MC>
To: John McCarthy <JMC@SU-AI>
Cc: phw%oz@MIT-MC
Subject: Your AI book
In-reply-to: Msg of 18 Aug 1983 19:15-EDT from John McCarthy <JMC at SU-AI>
Let me reply thankfully. The word "thankfully" is misused 6 times in
the version of the second edition you have. I have replaced all by
"fortunately."
Thank you for the remark about patterns as well. I'll try to do
something about that.
Patrick
∂22-Aug-83 1034 BMACKEN@SRI-AI.ARPA Meeting
Received: from SRI-AI by SU-AI with TCP/SMTP; 22 Aug 83 10:34:15 PDT
Date: Mon 22 Aug 83 10:32:48-PDT
From: BMACKEN@SRI-AI.ARPA
Subject: Meeting
To: BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA, Briansmith@PARC-MAXC.ARPA,
grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA, JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA,
kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA, peters@SRI-AI.ARPA,
rperrault@SRI-AI.ARPA, sag%SU-psych@SUMEX-AIM.ARPA, stan@SRI-AI.ARPA,
wasow@PARC-MAXC.ARPA, Wasow%SU-psych@SUMEX-AIM.ARPA,
winograd@PARC-MAXC.ARPA, Bmacken@SRI-AI.ARPA
I am very excited to be part of this far-reaching program and anxious
to meet all of you, gain a better understanding of your projects, and hear
your views about how they fit into the long range goals of the Center.
Specific topics we need to discuss are: the function of the Advisory
Board (issues related to the evaluation of the SDF program) and potential
members, centers of related research and mechanisms for creating ties
to their work, other projects at Stanford that CSLI might sponsor, and
the inner and outer colloquium series. Would you call Pat W. (497-0930)
to plan a time this week to meet me in my office in the Ventura Garret
(room 24)?
B.
-------
∂22-Aug-83 1109 ULLMAN@SU-SCORE.ARPA confusion
Received: from SU-SCORE by SU-AI with TCP/SMTP; 22 Aug 83 11:08:57 PDT
Date: Mon 22 Aug 83 11:13:11-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: confusion
To: super@SU-SCORE.ARPA
I'm getting confused as to who is speaking when.
COuld those who have invited speakers send me a message (and maybe a
message to the group) announcing that fact; I'll keep a schedule so
we don't get signals crossed.
By the way, we have no agenda for this Thursday, although most of
the subsequent weeks are taken (I think).
-------
∂22-Aug-83 1450 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Invited Speakers
Received: from SU-SCORE by SU-AI with TCP/SMTP; 22 Aug 83 14:49:58 PDT
Received: from Shasta by Score with Pup; Mon 22 Aug 83 14:53:28-PDT
Date: Monday, 22 Aug 1983 14:49-PDT
To: super at Score
Subject: Invited Speakers
From: Steve Lundstrom <slndstrm@Shasta>
I have made the following invitations (on behalf of our group, I hope).
Sept. 15 - Either Harry Heard or Dick Wilson, from Technology Development
Corporation
topic: TDC plans in advanced parallel architectures and
modeling
October 6 - Burton Smith - Denelcor Corporation
topic: details on the HEP, recent performance results,
development status, answer questions, ...
NOTE: Burton has been invited to give the EE380 video
seminar on Wed. 10/5. He will have time available
during the day on Thrus. 10/6 to visit individual
faculty. Please let me know if you want to spend any
time with him (and your time constraints). I will
set up his schedule for the day.
In addition, Joe Oliger, Ken Stevens, and I are planning to summarize
our visit to the Frontiers of Supercomputing meeting at Los Alamos
at some future meeting. Ken Stevens will be able to report on the
Parallel Processing Conference at the same time. The date for this
report has not yet been firmed up.
Steve Lundstrom
∂22-Aug-83 2159 Guy.Steele@CMU-CS-A Re: Hacker's Dictionary
Received: from CMU-CS-A by SU-AI with TCP/SMTP; 22 Aug 83 21:59:22 PDT
Received: from [128.2.254.192] by CMU-CS-PT with CMUFTP; 23 Aug 83 00:52:10 EDT
Date: 23 Aug 83 0059 EDT (Tuesday)
From: Guy.Steele@CMU-CS-A
To: John McCarthy <JMC@SU-AI>
Subject: Re: Hacker's Dictionary
In-Reply-To: "John McCarthy's message of 18 Aug 83 15:22-EST"
Hope you enjoy reviewing the dictionary. I *think* I caught that
"advisors" typo; we'll see when the final book comes out.
--Guy
∂23-Aug-83 0521 TARNLUND@SRI-AI.ARPA A draft of your book
Received: from SRI-AI by SU-AI with TCP/SMTP; 23 Aug 83 05:21:44 PDT
Date: Tue 23 Aug 83 05:23:10-PDT
From: TARNLUND@SRI-AI.ARPA
Subject: A draft of your book
To: McCarthy@SU-AI.ARPA
May I send you a little reminder of our agreement in
Karlsruhe that you send me a draft of your book and
I send you a dollar per student that is using it.
-------
We sent him that didn't we?
∂23-Aug-83 1026 CLT kronos
Fri. 16 Sep 20:00 Kronos
Fri. 21 Oct 20:00 Kronos
Fri. 18 Nov 20:00 Kronos
Fri. 20 Jan 20:00 Kronos
Fri. 24 Feb 20:00 Kronos
Fri. 16 Mar 20:00 Kronos
∂23-Aug-83 1339 @SU-SCORE.ARPA:DFH@SU-AI
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 Aug 83 13:38:53 PDT
Received: from SU-AI.ARPA by SU-SCORE.ARPA with TCP; Tue 23 Aug 83 13:42:31-PDT
Date: 23 Aug 83 1337 PDT
From: Diana Hall <DFH@SU-AI>
To: super@SU-SCORE
Parlog Seminar
Keith Clark will give a seminar on Parlog Thursday, Sept. 1 at 3 p.m in
Room 252 MJH.
PARLOG: A PARALLEL LOGIC PROGRAMMING LANGUAGE
Keith L. Clark
ABSTRACT
PARLOG is a logic programming language in the sense that nearly every
definition and query can be read as a sentence of predicate logic. It differs
from PROLOG in incorporating parallel modes of evaluation. For reasons of
efficient implementation, it distinguishes and separates and-parallel and
or-parallel evaluation.
PARLOG relations are divided into two types: and-relations and
or-relations. A sequence of and-relation calls can be evaluated in parallel
with shared variables acting as communication channels. Only one solution
to each call is computed.
A sequence of or-relation calls is evaluated sequentially but all the
solutions are found by a parallel exploration of the different evaluation
paths. A set constructor provides the main interface between and-relations
and or-relations. This wraps up all the solutions to a sequence of or-relation
calls in a list. The solution list can be concurrently consumed by an
and-relation call.
The and-parallel definitions of relations that will only be used in a
single functional mode can be given using conditional equations. This gives
PARLOG the syntactic convenience of functional expressions when non-determinism is
not required. Functions can be invoked eagerly or lazily; the eager evaluation
of nested function calls corresponds to and-parallel evaluation of conjoined
relation calls.
This paper is a tutorial introduction and semi-formal definition of
PARLOG. It assumes familiarity with the general concepts of logic programming.
∂23-Aug-83 1342 BMACKEN@SRI-AI.ARPA Important Events
Received: from SRI-AI by SU-AI with TCP/SMTP; 23 Aug 83 13:42:27 PDT
Date: Tue 23 Aug 83 13:38:08-PDT
From: BMACKEN@SRI-AI.ARPA
Subject: Important Events
To: csli.family: ;
SEPTEMBER 1 COLLOQUIUM
"CSLI and SDF: Where are we and where are we going?"
by
Jon Barwise and Betsy Macken
3 - 5 PM
Turing Auditorium, Polya Hall
- - - - - - - - - - - - - - -
"Gourmet Potluck Dinner"
by
Hans Uszkoreit, Susan Stucky, and CSLI Family
6 - 9:30 PM
Stanford Campus Recreation Association
on
Bowdoin between Campus Drive and Stanford Avenue
Families welcome
Pool, volleyball, etc.
-------
∂23-Aug-83 1651 ULLMAN@SU-SCORE.ARPA schedule
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 Aug 83 16:51:12 PDT
Date: Tue 23 Aug 83 16:55:45-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: schedule
To: super@SU-SCORE.ARPA
OK, I think I finally got the schedule straight:
Sept. 1: Keith Clark on parallel PROLOG. This seminar will take
place in MJH 252.
Sept. 15: A representative of TDC will talk on their proposal for
a parallel computation center.
Sept. 16: (FRIDAY) Jack Dennis will talk with the group at 3PM.
Oct. 6: Burton Smith of Denelcor. A public seminar will be
held the previous day.
Any volunteers for the gaps?
-------
∂23-Aug-83 1740 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA wallich@csnet-sh : 4th question
Received: from SU-SCORE by SU-AI with TCP/SMTP; 23 Aug 83 17:40:20 PDT
Received: from Shasta by Score with Pup; Tue 23 Aug 83 17:44:39-PDT
Date: Tuesday, 23 Aug 1983 17:46-PDT
To: ullman at Score, super at Score
Subject: wallich@csnet-sh : 4th question
From: Steve Lundstrom <slndstrm@Shasta>
Following are my comments about the message received. This has not yet
been transmitted to Wallich. Would appreciate comments first.
Wallich is, of course, asking the impossible question. The questions
"Who is going to sponsor next-generation research and how will the
resources get where they should go?" and
"How will work sponsored by different agencies be coordinated"
are almost purely political (ie. based on who is most persuasive and/or
holds the current "power" base).
I would like to make a few observations, based on four years trying to
make some progress in this area:
1. Some funds will come directly from industry, who will direct them where
they PERCEIVE the best payoff for their respective companies.
a. Major Companies: some of them will have (or already have)
v significant internal R&D projects underway. Since this means big
$'s, it is likely that the specific directions will be chosen to
augment or strengthen their current market positions (incremental
growth is percieved to be the least risky).
b. Startups: A few startups can be expected, but (in general) they
will have limited capital, and therefore will each concentrate on
only one, very constrained product area. The bright side of this is
that the startups will likely be where innovation leading to major
advances can be expected (reflecting the higher risk with some
failures.
c. Intercompany Cooperation: Is likely to be minimal, except for the
few research and development cooperatives to be started and for
situations such as CDC spinning out ETA Inc. The problem's do not
lie with the researchers - they would welcome working together.
The problems lie with the managers and with the law which, in
general, implies that something monopolistic or anti-trust related
is going on when competing companies talk to each other.
(NOTE: A similar problem often constrains government and industry
researchers from collaborating in a useful manner, especially when
contractural relationships are involved in addition.
2. Some funds will come from the government. I don't believe that we
can reasonably expect coordination within the government either. The
government itself is a "free-market" political system. By that,
I mean that all the various agencies, their departments, etc. are
themselves special interest groups, dedicated to their each
individual goals and objectives. They must each compete for
resources and support in order to accomplish their objectives.
Therefore, much of what occurs is VERY dependent on the strengths
and abilities of those who labor to justify the programs of their
individual groups.
[NOTE: This process is apparently not recognized by most researchers
who clamor at the doors of the various agencies for $, but do not
volunteer to help those agencies in the process of justifying the
resource and objective requests (which will be the source of the
requested $).]
3. For the same reason, agencies will coordinate their work if it is in
the best interests of each to do so. Otherwise, they are not likely
to (why should you help someone competing for the same resources you
need? unless you both have common objectives and projects.)
4. Another MAJOR problem is one of joint and continuing committments.
By this I mean that both government and industry commit to working
toward some common goal, but circumstances out of control of one
side or another can negate the well-intended committment.
An example should serve to highlight this problem. This example does
not delve into the depths of why various things happened, it only
shows the effects of a government/industry relationship on the industry
"partner".
a. In 1975, NASA initiated the planning for the Numerical Aerodynamic
Simulation Facility. When industry was approached with a Request for
Proposal for a Preliminary Study, there was an implication that the
effort was the first step of a phased procurement such that
only those who were awarded the first contract would be allowed to
bid on following contracts, eventually leading to the design,
implementation, and construction of a very large, state of the art,
supercomputer system (sustaining in excess of 1 GFLOP performance).
As the project evolved, the next step (Feasibility Study), was awarded
to the original vendors. However, the next step was the development of
a new specification by NASA, followed by another open competition,
again for a phased procurement, awarded to the original two vendors
(no one else bid). After the System Design Study, these two vendors
submitted proposals to NASA (one copy of a proposal was about 16"-20"
thick). Subsequently, NASA decided to terminate that procurement and
to proceed with a different approach to solving their problem.
b. During the 5 years covered by this scenario, many circumstances
changed, often out of control of anyone involved on either the
government or industry side. The result was that each vendor involved
spent over $1 million of their own development funds over 4-5 years,
and received an additional $1 million to cover additional costs (at no
profit).
Not many manufacturers can afford to commit major resources and
key innovators for 5 years to a venture where they do not have more
direct control over the outcome.
In this case, the funding of the projects were delayed significantly
by Congress, by changes of administration, by competition with many
other programs within the government, etc.
c. In the same project, one of the vendors began the relationship
with government fully intending that the work would lead to another
in a planned line of high-speed scientific processors (and would
therefore be willing to commit product development funds to augment
government funds). During the course of the project, new corporate
level leaders reviewed corporate objectives, and set new ones. Among
these was a decision NOT to be involved in such a product area.
The result was viewed from the government side as a move from having
access to the first of a new line of major computational systems to
another "orphan" (as Illiac IV was - only one built). An "orphan" is
clearly not a desirable research tool since the results cannot be
easily transferred to other users at other sites with their own
equipment.
NOTE: The above scenario is incomplete, and does not reflect all
circumstances, but does show some of the complexities which lurk
when trying to cope with Wallich's questions.
If anyone is interested in further details and circumstances, I will
try to clarify them, as best as I can recollect them.
Now, a note about using NETWORKS:
When trying to develop a very large project utilizing widely scattered
resources, existing networks are extremely difficult to use productively.
What is needed is some form of constrained distribution in order to
enforce the kinds of communication required in order to manage the
development efficiently. In the project described above, I had 50 people
working from 4-5 different sites including Pennsylvania, Alabama,
Santa Monica, Sunnyvale, and Camarillo. Unless a manager is provided
system development support tools which allow all on the project to
ignore the existance of a network and which enforce the relationships
and communications which much exist within the project, networks are
likely to create more work than they avoid. Such distributed system
development support tools (including the networks) are, of course,
necessary if a highly integrated system design is expected from a highly
distributed design group.
S.F.Lundstrom
∂24-Aug-83 0837 CL.BOYER@UTEXAS-20.ARPA Re: more on Stoyan
Received: from UTEXAS-20 by SU-AI with TCP/SMTP; 24 Aug 83 08:37:46 PDT
Date: Wed 24 Aug 83 10:38:32-CDT
From: Bob Boyer <CL.BOYER@UTEXAS-20.ARPA>
Subject: Re: more on Stoyan
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Thu 18 Aug 83 16:32:00-CDT
Hi.
I have forwarded your messages about Stoyan on to (a) Guy Steele,
the program chairman of the next LISP conference and (b)
people on the UT's CS recruiting committee.
You may be interested to know that David Russinoff, who got
Wilson's theorem through our theorem-prover, has recently
gotten the law of quadratic reciprocity through. That's
a much deeper theorem, I'm told.
I'm looking forward to your coming visit to UT in October,
and I hope you can be persuaded to come down here on a
more permanent basis.
-------
∂24-Aug-83 1325 DFH
To: "@FACCOM.[1,DFH]"@SU-AI
Next Facilities Committee Meeting
The next facilities committee meeting will be in the beginning of October.
I would like to get an idea of your schedules at this time so we can have as
many of you as possible come. Please let me know what dates and times you
could attend the meeting in the first week of October. Thanks.
∂24-Aug-83 1357 ARK New CF hire?
I noticed that someone was hiring Phillip Almquist, so I asked Ralph
about it. His response is below. Somehow, if CF is hiring people,
shouldn't the oversight committee at least be informed about it?
Arthur
∂24-Aug-83 1353 REG
Near as I can tell, Arthur, I'm too busy to answer your question.
He will be CF
∂24-Aug-83 1640 ATKINSON@SU-SCORE.ARPA 83-84 Student Support
Received: from SU-SCORE by SU-AI with TCP/SMTP; 24 Aug 83 16:40:50 PDT
Date: Wed 24 Aug 83 16:45:36-PDT
From: Amy Atkinson <Atkinson@SU-SCORE.ARPA>
Subject: 83-84 Student Support
To: Principal.Investigators: ;
Stanford-Phone: (415) 497-9354
Well, it's August 24th, so any of you who have not reported your student
support plans should do so, please. We really do need to get these done
soon so that the students will have tuition grant cards when they register,
and paychecks on payday.
Thank you,
Amy
-------
∂24-Aug-83 2025 CLT opera
Wed. 19 Oct 20:00 Midsummer
Thu. 1 Dec 20:00 Duchesse
Thu. 8 Dec 19:30 Boris
∂25-Aug-83 0155 ARK for your information
∂25-Aug-83 0029 ARK@SU-SIERRA.ARPA New rates for Sierra.
Received: from SU-SIERRA by SU-AI with TCP/SMTP; 25 Aug 83 00:28:54 PDT
Date: Fri 19 Aug 83 17:54:08-PDT
From: Stephen Hansen <Hansen@SU-SIERRA.ARPA>
Subject: New rates for Sierra.
To: System@SU-SIERRA.ARPA
ReSent-date: Thu 25 Aug 83 00:32:19-PDT
ReSent-from: Arthur Keller <ARK@SU-SIERRA.ARPA>
ReSent-to: ark@SU-AI.ARPA
Pending approval by the University Controller's Office, the
following rate schedule for the Sierra System will take effect on
September 1, 1983. Relative to the current rate schedule, the
definition of "A", "B", and "C" time has been modified, and the per
unit charges for disk allocation, CPU use, and connect time have been
increased.
The following time blocks are defined:
weekday weekend
---------------
0200-0900 | C | C |
|---------------|
0900-1300 | A | C |
|---------------|
1300-0200 | A | B |
---------------
"B" time is charged at 3/4 of the "A" rate.
"C" time is charged at 1/2 of the "A" rate.
No adjustment is made for holidays.
CPU charges ==> $1.69/"A"-CPU-minute
Disk charges ==> $0.10/page/month
Connect charges ==> $1.35 per "A" hour
-------
∂25-Aug-83 0204 ARK Query about rates
To: Hansen@SU-SIERRA
CC: JMC@SU-AI, ARK@SU-AI
Assuming that the information I am requesting is not confidential, I would
like some background data about the Sierra rates. I am on CSD's Computer
Facilities Oversight Committee, and we will hopefully have something to
say about SCORE's and SAIL's rates.
1. What is the total estimated annual expenses of operating Sierra?
(Please divide this into expendables, salaries+benefits, maintenance,
depreciation, and other.)
2. What is the amount of disk space you expect to sell on the average?
How much do you expect to collect for disk space on an annual basis?
3. For each of the three rate periods, what is the expected connect time
you expect to sell on the average? How much do you expect to collect
by rate period for connect time on an annual basis?
4. Please answer question 3 for CPU time instead of connect time.
Please be assured that I am not in any way being critical. It's just that
comparative figures for the various systems would be most useful. I hope
that information like this for SAIL and SCORE can be made public as well.
Thanks for your help.
Arthur
∂25-Aug-83 1637 ULLMAN@SU-SCORE.ARPA CIS panel session
Received: from SU-SCORE by SU-AI with TCP/SMTP; 25 Aug 83 16:37:43 PDT
Date: Thu 25 Aug 83 16:37:56-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: CIS panel session
To: feigenbaum@SUMEX-AIM.ARPA, dek@SU-AI.ARPA, jmc@SU-AI.ARPA,
reid@SU-SHASTA.ARPA
cc: meindl@SU-SIERRA.ARPA, ullman@SU-SCORE.ARPA
Jim Meindl has asked me to organize a panel session on the topic
"the next 10 years in software development," for the CIS annual
review, Nov. 9-10. Tentatively, the format would consist of
brief statements by the panelists and questions from the audience.
I hope that you four will agree to serve as the panel, for which
you will have the thanks of Jim and me, at the very least.
-------
I would be glad to take part in CIS panel.
∂25-Aug-83 1645 ULLMAN@SU-SCORE.ARPA seminar course
Received: from SU-SCORE by SU-AI with TCP/SMTP; 25 Aug 83 16:44:56 PDT
Date: Thu 25 Aug 83 16:44:12-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: seminar course
To: super@SU-SCORE.ARPA
At the brief (< 2 hrs.) meeting today, we discussed the formation of
a 1 credit seminar for the fall, with speakers from the "group"
but outside speakers whenever possible. I am willing to take the
responsibility for the quarter. The tentative time is 4:15 on Thursdays.
We would try to couple the public seminar with informal discussions
with outside speakers.
Does anybody have a serious conflict with the proposed time?
-------
∂25-Aug-83 1722 STUCKY@SRI-AI.ARPA Potluck
Received: from SRI-AI by SU-AI with TCP/SMTP; 25 Aug 83 17:21:44 PDT
Date: 25 Aug 1983 1720-PDT
From: Stucky at SRI-AI
Subject: Potluck
To: CSLI.Potluck:
f
CSLI POTLUCK
You are cordially invited to CSLI's first social
event: a potluck dinner.
Location: Thursday, September 1, 6-9:30 p.m.
Stanford Campus Recreation Association
on Bowdoin between Campus Drive and Stanford Avenue
Activities: Swimming, volleyball, eating
Bring: Children, food, spouses, swimwear, table service,
volleyballs
We suggest that you bring some item to contribute to the
potluck, a main dish, dessert etc., and we ask that you bring your own
table service. Beverages will be provided. In order to facilitate
planning, we request you RSVP to Susan Stucky (859-4593 or
Stucky@SRI-AI) by August 27, indicating how many people you are
bringing. Families welcome.
-------
∂25-Aug-83 1737 RESTIVO@SU-SCORE.ARPA Phillipe Roussel
Received: from SU-SCORE by SU-AI with TCP/SMTP; 25 Aug 83 17:37:39 PDT
Date: Thu 25 Aug 83 17:38:35-PDT
From: Chuck Restivo <RESTIVO@SU-SCORE.ARPA>
Subject: Phillipe Roussel
To: jmc@SU-AI.ARPA
cc: restivo@SU-SCORE.ARPA
Would like to come by Tuesday the 30th with Phillipe. Morning or
afternoon; what is your preference ?
I request an appointment for the 31st, 1st or 2nd. Mornings or
afternoons. Can you accomodate me ?
Thank you,
Chuck Restivo
-------
11:30 Tuesday would be fine for meeting with Roussel. Phone me about
other appointment any afternoon.
∂25-Aug-83 1803 GOLUB@SU-SCORE.ARPA Computer Science Library Technical Reports List
Received: from SU-SCORE by SU-AI with TCP/SMTP; 25 Aug 83 18:03:40 PDT
Mail-From: LIBRARY created at 24-Aug-83 16:44:15
Date: Wed 24 Aug 83 16:44:15-PDT
From: C.S./Math Library <LIBRARY@SU-SCORE.ARPA>
Subject: Computer Science Library Technical Reports List
To: Golub@SU-SCORE.ARPA
cc: Library@SU-SCORE.ARPA
ReSent-date: Thu 25 Aug 83 18:04:35-PDT
ReSent-from: Gene Golub <GOLUB@SU-SCORE.ARPA>
ReSent-to: dek@SU-AI.ARPA, jmc@SU-AI.ARPA, feigenbaum@SUMEX-AIM.ARPA,
WIEDERHOLD@SU-SCORE.ARPA
I have had Richard working on a study of our technical reports list as
compared to what is indexed by Government Reports Announcement and
available online through NTIS. So far his study has shown that only
17% of our reports are indexed in NTIS. We also show a slightly higher
circulation of the reports not indexed in NTIS. On the average, we
are 16.5 weeks ahead of NTIS in indexing.
I have always considered the technical reports an important part of the
primary literature in our collection and consider it part of our
collection development procedures as opposed to simply ordering on
request, although we do add to the collection when researchers request
a particular report. The computer science department with its support
of the automatic and selective exchange agreements has provided the
support of this part of our collection. I am aware that faculty and
students would like to have the entire file online. At this point, we
have been placing only the new technical reports list online. I am
writing you to ask you if you would support our move in this direction.
It is my understanding that Score has a data base management system
available which may be compatable to a program Richard has done to
convert our records to the Score system. If not, can you suggest
how we could do this. Would the department allocate us storage to
be able to place the file online: approximately 25,000 records at
300 characters per record. The science libraries as a whole would
like to have a technical reports file online either through CIT
or RLIN. However I would hope we could still have the Math/CS Library
section directly accessible to the computer science department.
If you feel the index is valuable enough to have online and the
department can provide the storage, this will be another step
toward providing researchers access to our files while in their
offices. We may be able to handle routine requests for reports
as we do with new technical reports: i.e. requests as to whether
they are available in the library and/or recalling material.
Having the index onlineould be of great help to the graduate
students in making them aware of the literature available and the
library.
I would appreciate hearing your opinions on this topic.
Thanks
Harry
-------
I strongly support Harry's proposals, and I think we can regard
the file space as a CSD-CF service to all its users.
∂26-Aug-83 1604 DFH
Messages
1. This is the text of a memo from Gene Golub:
Gentlemen:
Unfortunately, the Provost refuses to send the Rod Brooks papers on to
the Advisory Board without a more complete vote from the faculty.
There had been a faculty meeting in the spring where there was a
positive vote on Brooks but not all of you were able to attend.
IT IS URGENT THAT WE OBTAIN YOUR VOTE.
YES------- NO--------- ABSTAIN---------
Elyse has the completed file on Brooks in her office. Please get your
vote to me by Monday, August 29, noon.
Many thanks, Gene
2. Dr. McDonald of SF State Univ CS Dept. called. He is trying to find
someone to teach a graduate level AI course in Sept. phone (55) 469-1008.
3. J.A. Robinson, Syracuse Univ. He probably got ahold of you in Washington.
Alan Colmeraurer was to ask you at IJCAI about being on the Board of
Editors for the Journal of Logic. 315-423-3159
∂27-Aug-83 1138 @MIT-MC:TK@MIT-MC
Received: from MIT-MC by SU-AI with TCP/SMTP; 27 Aug 83 11:38:24 PDT
Date: Saturday, 27 August 1983, 14:27-EDT
From: Tom Knight <TK at MIT-MC>
To: JMC at SU-AI, TK at SU-AI
In-reply-to: The message of 18 Aug 83 15:47-EDT from John McCarthy <JMC at SU-AI>
Date: 18 Aug 83 1247 PDT
From: John McCarthy <JMC@SU-AI>
To: TK@SU-AI
Understand you're on West Coast. Can you give me a call?
Guess I missed this one. Presumably you wanted to ask me about machine
configurations. What I said, in summary, is that you want at least 2
memory cards so the damn thing doesn't page to death. A 167 mb drive is
ok if you are planning on using the machine in a networked configuration,
but not if you are using it standalone. A 474 mb drive is better, of
course, primarily from the standpoint of allowing you to save intermediate
system with large programs already loaded up. You'll need at least one
machine on the network with a cartridge tape drive and modem, so we can
get software to you and receive bug reports back. It was nice seeing you
at the conference.
∂28-Aug-83 0415 POURNE@MIT-MC Marconi
Received: from MIT-MC by SU-AI with TCP/SMTP; 28 Aug 83 04:15:22 PDT
Date: 28 August 1983 07:14 EDT
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: Marconi
To: JMC @ SU-AI
In-reply-to: Msg of 22 Jan 83 1502 PST from John McCarthy <JMC at SU-AI>
alas, this was the essay I wanted a hard copy of! I had thought
it lost. Ah, well. They tell me the committee decides in
September.
Best,
JEP
∂28-Aug-83 1152 EF@MIT-MC
Received: from MIT-MC by SU-AI with TCP/SMTP; 28 Aug 83 11:52:16 PDT
Date: 28 August 1983 14:53 EDT
From: Edward Fredkin <EF @ MIT-MC>
To: EF @ MIT-MC, jmc @ SU-AI
I hope to get you the text soon!
∂28-Aug-83 1159 EF@MIT-MC
Received: from MIT-MC by SU-AI with TCP/SMTP; 28 Aug 83 11:59:21 PDT
Date: 28 August 1983 15:00 EDT
From: Edward Fredkin <EF @ MIT-MC>
To: jmc @ SU-AI, ef @ MIT-ML
8-25-83
Brookline, Mass
Dear John,
I have written this, in a hurry, so that I could get it to
you right away. I urgently want your reaction, suggestions and
help. Most importantly, you and I should have everything settled
between us, and we should have communicated to the AAAI council
so that the situation is not left in limbo when we both leave on
September 3rd or 4th.
I would like to give this effort a best chance of success,
so that the field of AI might benefit as much as possible from
the opportunities we see. On the other hand, the window of
opportunity might pass us by if we choose to squabble about
details. There is no pressing need to see that we already have
perfect consensus on all issues, rather we should have the
confidence that we can reach such consensus.
The field of AI has grown and gotten better defined than in
the days when Marvin defined the field as the set of things that
interested him. There is the need to broaden the support to
include the kind of people you mentioned, bright post-docs, who
might have come from another field, who just might discover
something important that the rest of us would never think of.
There certainly is the need to get good hardware into the hands
of researchers who don't happen to be at MIT. Stanford or CMU.
Well, enough rambling, please take a look at the following
and call me or send me some net mail. I just got interrupted by
a very long phone call from Ken Olsen. DEC is having a little
sales meeting in Boston, 8,000 salesmen are here from all over.
Ken claims they organized the biggest company picnic ever held in
New England. In any case, he doesn't think AI or Robotics is
really that important. He feel that Gordon is just wrong about
both fields.
I guess I still rambled on!
PS Ken set things up for me to go to the show, I'll tell
you about it some time.
1
BACKGROUND AND HISTORY
We believe that the United State may be at a turning point,
facing either a great opportunity or a potential risk, depending
on the course of action taken now in the field of Artificial
Intelligence. The reason is not because of any startling new
research developments or breakthroughs, but rather because, for
the first time, there is a dramatically increased awareness of
the possibility of practical applications evolving from AI
research, especially in Japan. The Fifth Generation book is
actually the number one best seller (non-fiction) in Japan this
week, the last week in August. This comes at a time when the
rapid progress of computer technology finally allows AI programs
to run on relatively inexpensive systems. The confluence of
events may make it possible to do things now, that may not be
possible later. The reason that we are at a turning point is
because we face a window of opportunity that we must not miss.
For the past twenty years, the US government, through DARPA,
has supported research in the field of Artificial Intelligence.
While the work, which started at MIT and Stanford, was an obscure
area of basic research in the late fifties, it is now realized
that the practical applications of that research may have a very
large and increasingly important effect on the future strength of
this country, vis a vis its economic competitors. Japan, is
today the country with the most ambitious national program in the
field of AI. England has also embarked on a national effort.
While the US has been the leader in this field, we may not be
able to maintain that lead without embarking upon efforts of
greatly increased scope. The payoff, of maintaining our
preeminent position, appears to be so large as to justify very
great investments in this area. On the other hand, should we
fall behind in this technology, the risks we might face are not
merely economic ones, since, someday, advanced AI technology may
be an important factor in other aspects of national power. Most
importantly, the opportunity to precipitate large scale
governmental support appears to be present today, it was never
present in the past and, if this opportunity passes by, it may
not be present again for some time.
THE MEETING AND WHITE PAPER
The goals envisioned for the meeting would be to consider
certain aspects of AI research, development and applications
possibilities, consider what other nations are doing in that
context, and finally consider possible recommendations for
national action in the US. While it would be healthy for the
conferees to consider broadly the realm of future possibilities,
quite conservative extrapolations of what we already know should
be sufficient to allow one to conclude that the scale of prudent
activity for the US is greater than presently being considered.
If one looks at the range of practical applications, using
industry standard criteria such as projected return on
investment, a possible conclusion will be that the most
2
reasonable level of investment in all aspects of AI, from basic
research to practical applications of proven technology, may be
far greater than what is now being done. If the right group of
conferees did come to such a conclusion, and if it was able to
support its conclusions with reasoned arguments, then it is
likely that its recommendations might result in appropriate
governmental action.
WHO SHOULD BE INVITED
Attendance at this meeting should be limited to a small
number, perhaps 20 participants. The participants should consist
of two groups, those very familiar with AI, who are good at
conveying the concepts, possibilities and possible consequences
of various courses of action to a group of intelligent laymen
(laymen in the field of AI). And then the others, who should be
persons of great prestige who would be drawn from government,
science and industry.
ATTENDEES DRAWN FROM THE AI COMMUNITY
The members of the AI community, who attend the meeting,
will have the responsibility of educating the others, and
communicating the possible scenarios consequent to various
proposed actions. There is a great tendency for the experts in
an obscure field to talk over the heads of the most intelligent
of laymen, resulting in very little communication. In this case,
it will be very important to work hard at conveying the
information from expert to layman in ways that result in
understanding and appreciation for the most important issues.
There is, in this plan, the implication that it would be
sensible for the members of the AI community who attend the
meeting, to be of one mind on the important issues, and if that
is not possible, then there may be no point in holding the
meeting. The Japanese have made a great art of coming to a
consensus, where the natural tendency to display one's ego by
vociferous disagreement is suppressed. Some of the superstars in
the field AI are infamous for the opposite, never participating
in a meeting without showing strong disagreement with some point,
as a matter of principal. While the AI community can know and
love its heros for their weaknesses as well as for their
strengths, if we want to have an effect on the world, we will
have to be pragmatic, and hold back a little of our
individualism. The AI persons need not agree on all points, but
they must understand in advance what they disagree about, they
must be able to delimit and make clear the consequent
alternatives, and they must not allow such differences to get in
the way of communicating the broad areas of agreement. As a
consequence, the AI participants must get together before the
meeting to explore the issues and to come to a consensus.
THE NON-AI PARTICIPANTS
The main objective of the meeting will be to communicate the
rationale for greatly increasing government funding of various
3
aspects of AI research and technology. The audience should
consists of a group of prestigious movers and shaker who are
able to appreciate the possibilities. It will be important to
have representatives from three major groups: Government, because
they will best understand the process of turning such concepts
into actions; Industry, because much of the economic consequences
will be a result of what American Industry chooses to do, and
because of the practical experience of such persons; Science,
because the government is used to, for good reason, consulting
with certain trusted scientific advisors. Again, since the
objective is to give the best opportunity to the realization of
the goals of the meeting, the non-AI participants need to be very
carefully selected.
PERTINENT MATERIALS AND THE WHITE PAPER
A very important aspect of the conference would be the
distribution to the conferees, prior to the conference, of
materials that detail what is happening in the field of AI.
Examples might include the Fifth Generation by Feigenbaum and
McCorduck and papers such as "Artificial Intelligence: An
Assessment of the State-of-the-Art and Recommendation for Future
Directions" by David Waltz et al, in the Fall 83 issue of AI
Magazine. In addition, the organizers of the meeting should
write and distribute condensed background information that can be
easily and quickly digested. Subsequent to the conference, a
white paper should be prepared and distributed. That white paper
will be the focal point of consensus and a document that makes
specific recomendations for action.
THE SPECIFIC RECOMMENDATION TO THE AAAI COUNCIL
The council should agree that the AAAI should sponsor a
meeting, along the lines outlined above. While the AAAI would
sponsor the meeting, the opinions and conclusions expressed by
the participants would be their own, and not official positions
of the AAAI. A steering committee consisting of Raj Reddy,
Gordon Bell, Ed Fredkin, Bob Kahn (if he is willing) and perhaps
one or two others should be constituted. Ed Fredkin would be the
Chairman of the committee and meeting. The meeting will be at no
cost to the AAAI, it will pay its fair share of all expenses
incurred by the AAAI and return a modest profit. Claudia Mazzett
has indicated that she would be able and pleased to handle
various arrangements for the meeting, for which the AAAI would be
reimbursed. The most likely site would be in the Washington
area, and the meeting would take place sometime before the end of
the year.
4
Ok John, I hope you got it!!
∂28-Aug-83 1837 RESTIVO@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 28 Aug 83 18:37:41 PDT
Date: Sun 28 Aug 83 18:38:40-PDT
From: Chuck Restivo <RESTIVO@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Sun 28 Aug 83 18:20:00-PDT
Wednesday 10a, and Thursday afternoon then ?
-------
Either looks ok now. Can I make it definite tomorrow without inconvenience?
∂28-Aug-83 1934 RESTIVO@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 28 Aug 83 19:34:07 PDT
Date: Sun 28 Aug 83 19:35:09-PDT
From: Chuck Restivo <RESTIVO@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Sun 28 Aug 83 18:42:00-PDT
Absolutely, let me know what is most convenient.
-------
∂29-Aug-83 0006 ME XGP
∂28-Aug-83 2332 JMC
Is the xgp broken?
ME - Yes, and Don Coates has been on vacation. Perhaps he'll be back
this week.
∂29-Aug-83 0043 FEIGENBAUM@SUMEX-AIM.ARPA Re: CIS panel session
Received: from SUMEX-AIM by SU-AI with TCP/SMTP; 29 Aug 83 00:43:05 PDT
Date: Mon 29 Aug 83 00:45:12-PDT
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: Re: CIS panel session
To: ULLMAN@SU-SCORE.ARPA, dek@SU-AI.ARPA, jmc@SU-AI.ARPA, reid@SU-SHASTA.ARPA
cc: meindl@SU-SIERRA.ARPA
In-Reply-To: Your message of Thu 25 Aug 83 16:39:47-PDT
Jeff,
(and Jim too)
I would have been delighted to be on that panel, but unfortunately I will be
out of town on those two days (in fact all of that week). how about asking
Genesereth and/or Lenat to replace me?
Ed
-------
∂29-Aug-83 0933 KJB@SRI-AI.ARPA Sept 1 colloquium
Received: from SRI-AI by SU-AI with TCP/SMTP; 29 Aug 83 09:33:29 PDT
Date: Mon 29 Aug 83 08:20:30-PDT
From: KJB@SRI-AI.ARPA
Subject: Sept 1 colloquium
To: csli.people: ;
Dear all,
There are many developments over the past few days, developments
which effects us all. I think it is very important for everyone to be
at the meeting on Thursday, or to have someone there who will give them
a full account of the meeting. This meeting is not for the general
public, but rather for us to have a frank disucssion of csli.
Jon
-------
∂29-Aug-83 0944 DFH A draft of your book
∂26-Aug-83 2118 JMC A draft of your book
∂23-Aug-83 0521 TARNLUND@SRI-AI.ARPA A draft of your book
Received: from SRI-AI by SU-AI with TCP/SMTP; 23 Aug 83 05:21:44 PDT
Date: Tue 23 Aug 83 05:23:10-PDT
From: TARNLUND@SRI-AI.ARPA
Subject: A draft of your book
To: McCarthy@SU-AI.ARPA
May I send you a little reminder of our agreement in
Karlsruhe that you send me a draft of your book and
I send you a dollar per student that is using it.
-------
We sent him that didn't we?
--------------
Yes, I remember sending it, however, he may not have it yet.
I sent it to him at Uppsala University.
∂29-Aug-83 1529 CHURD@SU-SCORE.ARPA Score
Received: from SU-SCORE by SU-AI with TCP/SMTP; 29 Aug 83 15:29:28 PDT
Date: Mon 29 Aug 83 15:30:31-PDT
From: Cuthbert Hurd <CHURD@SU-SCORE.ARPA>
Subject: Score
To: McCarthy@SU-SCORE.ARPA
Now informed by Lynn Gotelli that my account is to be continued.
-------
Fine. They all say you misinterpreted a message aimed at the CSL and
OR accounts that are being terminated, because their own computer
(Sierra) is now working well enough so that there is no justification
for their continuing to have SCORE accounts.
∂29-Aug-83 1631 PACK@SU-SCORE.ARPA Re: discussing traffic lights
Received: from SU-SCORE by SU-AI with TCP/SMTP; 29 Aug 83 16:31:38 PDT
Date: Mon 29 Aug 83 16:32:39-PDT
From: Leslie E. Pack <PACK@SU-SCORE.ARPA>
Subject: Re: discussing traffic lights
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Sun 28 Aug 83 18:15:00-PDT
Yes. I'll come by on Wednesday or Thursday. Any preferences about
the time?
-------
∂30-Aug-83 1345 JMC*
Jacobsen
∂30-Aug-83 2251 minker%umcp-cs@UDel-Relay Mints
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 30 Aug 83 22:51:19 PDT
Date: 29 Aug 83 16:54:28 EDT (Mon)
From: JACK MINKER <minker%umcp-cs@UDel-Relay>
Return-Path: <minker%umcp-cs@UDel-Relay>
Subject: Mints
To: JMC@SU-AI
Via: UMCP-CS; 31 Aug 83 1:41-EDT
Dear John,
Thanks again for your participation in the Lerner panels. I
just wasnt to remid you to send me the information on Mints so that
I can include it in my next report.
Jack
The person who will have the best information about Mints is Vladimir Lifschitz
in the Computer Science or Mathematics Department of the University of
Texas at El Paso. He is a fairly recent emigre. He will also know
whether Mints wants publicity.
∂31-Aug-83 1012 TARNLUND@SRI-AI.ARPA
Received: from SRI-AI by SU-AI with TCP/SMTP; 31 Aug 83 10:11:59 PDT
Date: Wed 31 Aug 83 08:41:02-PDT
From: TARNLUND@SRI-AI.ARPA
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Mon 29 Aug 83 09:54:00-PDT
Thanks, I have got it !
-------
∂31-Aug-83 1018 RINDFLEISCH@SUMEX-AIM [Tom Binford <TOB@SU-AI>: LISP machines ]
Received: from SUMEX-AIM by SU-AI with PUP; 31-Aug-83 10:18 PDT
Date: Wed 31 Aug 83 10:07:19-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM>
Subject: [Tom Binford <TOB@SU-AI>: LISP machines ]
To: Feigenbaum@SUMEX-AIM, JMC@SAIL, Ullman@Diablo, CSL.JLH@Score, Bosack@Score,
BScott@Score
cc: Rindfleisch@SUMEX-AIM
FYI... Tom R.
---------------
1) 23-Aug Tom Binford LISP machines
2) 31-Aug To: TOB@SAIL Re: LISP machines
Message 1 -- ************************
Return-Path: TOB@SU-AI
Received: from SU-AI.ARPA by SUMEX-AIM.ARPA with TCP; Tue 23 Aug 83 11:45:41-PDT
Date: 23 Aug 83 1141 PDT
From: Tom Binford <TOB@SU-AI>
Subject: LISP machines
To: rindfleisch@SUMEX-AIM
Tom
It seems that robotics got left out of the DOD equipment
pot for LISP machines. We have a major need for LISP machines
and I would like that we get our share.
When can we talk?
tom
Message 2 -- ************************
Mail-From: RINDFLEISCH created at 31-Aug-83 10:03:43
Date: Wed 31 Aug 83 10:03:43-PDT
From: T. C. Rindfleisch <Rindfleisch@SUMEX-AIM.ARPA>
Subject: Re: LISP machines
To: TOB@SAIL
cc: Rindfleisch@SUMEX-AIM.ARPA
In-Reply-To: Your message of Tue 23 Aug 83 11:41:00-PDT
I'm back in operation now from AAAI, Tom, and am glad to talk whenever
you would like. I'm not sure what to say, though, in that I am only
directly responsible for the HPP equipment planning. The "DARPA
equipment proposal", for which a contract was recently awarded, was
formulated and submitted more than a year and a half ago (February
1982). It is not a "pot" but rather a combination of separate detailed
requests for HPP, Formal Reasoning, CSD-CF, and CSL. DARPA has
postponed funding of our equipment needs until this contract could be
awarded (1.5 years!) and the HPP part has long been fully committed.
Tom R.
-------
-------
RPG tells me that he talked to you while the proposal was being prepared,
and you expressed no interest at the time.
∂31-Aug-83 1100 DFH
I would like to be gone from about 3 - 4 pm this
afternoon to run an errand. Let me know if this
is a problem.
∂31-Aug-83 1124 RPG Minor points
Don't forget the senior research associate upgrade. Thanks.
Also, perhaps we ought to talk about Binford and Lisp machines before you
go. I talked to him at least 5 times about the equipment contract over a
period of 1 year vis-a-vis Robotics, and each time he either did not
answer or else he said he was getting Suns or Vaxes to fill his needs. I
think that Brooks has told him that he (Brooks) wants some 3600s and that
is why Binford is now complaining.
Since I think that I made a strong effort to accomodate Binford, I'm not
sympathetic to him now. If you think we ought to negotiate, I will do
it in your absence.
-rpg-
∂31-Aug-83 1428 Bonnie%UPenn%UPenn@UDel-Relay That resort near Poughkeepsie you mentioned to Ray Reiter & me
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 31 Aug 83 14:28:24 PDT
Date: Wed, 31 Aug 83 09:34 EDT
From: Bonnie Webber <Bonnie%UPenn@UDel-Relay>
Return-Path: <Bonnie%UPenn%UPenn@UDel-Relay>
Subject: That resort near Poughkeepsie you mentioned to Ray Reiter & me
To: jmc@su-ai
Via: UPenn; 31 Aug 83 17:28-EDT
Could you give me its name again, so that I may inquire about their facilities
for workshops.
1 hit on key "mohonk".
Exit
↑C
.
It is Mohonk Mountain House in New Paltz, NY (914)255-1000.
∂31-Aug-83 1432 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Burton Smith visit - Reschedule Request
Received: from SU-SCORE by SU-AI with TCP/SMTP; 31 Aug 83 14:32:16 PDT
Received: from Shasta by Score with Pup; Wed 31 Aug 83 14:21:54-PDT
Date: Wednesday, 31 Aug 1983 14:24-PDT
To: super at Score
Cc: slndstrm at Shasta, allison at Shasta
Subject: Burton Smith visit - Reschedule Request
From: Steve Lundstrom <slndstrm@Shasta>
Burton Smith was originally scheduled to visit on October 5 & 6 (talk
with our group on 6th). He requested to change dates to October 19 &
20. I need to reply to him tomorrow (Thurs.). Are there any problems
or conflicts with such a change? {Dennis: How does this work with
the EE380 Seminars?} [Jeff: How does this fit with the master
schedule?] Steve
∂31-Aug-83 2236 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
Received: from SU-SCORE by SU-AI with TCP/SMTP; 31 Aug 83 22:36:07 PDT
Received: from Shasta by Score with Pup; Wed 31 Aug 83 22:36:53-PDT
Date: Wednesday, 31 Aug 1983 22:40-PDT
To: Jeffrey D. Ullman <ULLMAN at SU-SCORE.ARPA>
Cc: super at SU-SCORE.ARPA, wallich at Uwisc, allison at Shasta
Subject: Re: [wallich@csnet-sh.ARPA (Paul Wallich): Fourth question]
In-reply-to: Your message of Wed 17 Aug 83 16:38:29-PDT.
From: Steve Lundstrom <slndstrm@Shasta>
8/31/83
Wallich is, of course, asking the impossible question. The questions
"Who is going to sponsor next-generation research and how will the
resources get where they should go?" and
"How will work sponsored by different agencies be coordinated"
are almost purely political (ie. based on who is most persuasive and/or
holds the current "power" base).
I would like to make a few observations, based on four years trying to
make some progress in this area:
1. Some funds will come directly from industry, who will direct them where
they PERCEIVE the best payoff for their respective companies.
a. Major Companies: some of them will have (or already have)
significant internal R&D projects underway. Since this means big
$'s, it is likely that the specific directions will be chosen to
augment or strengthen their current market positions (incremental
growth is percieved to be the least risky).
b. Startups: A few startups can be expected, but (in general) they
will have limited capital, and therefore will each concentrate on
only one, very constrained product area. The bright side of this is
that the startups will likely be where innovation leading to major
advances can be expected (reflecting the higher risk with some
failures.
c. Intercompany Cooperation: Is likely to be minimal, except for the
few research and development cooperatives to be started and for
situations such as CDC spinning out ETA Inc. The problem's do not
lie with the researchers - they would welcome working together.
The problems lie with the managers and with the law which, in
general, implies that something monopolistic or anti-trust related
is going on when competing companies talk to each other.
(NOTE: A similar problem often constrains government and industry
researchers from collaborating in a useful manner, especially when
contractural relationships are involved in addition.
2. Some funds will come from the government. I don't believe that we
can reasonably expect coordination within the government either. The
government itself is a "free-market" political system. By that,
I mean that all the various agencies, their departments, etc. are
themselves special interest groups, dedicated to their each
individual goals and objectives. They must each compete for
resources and support in order to accomplish their objectives.
Therefore, much of what occurs is VERY dependent on the strengths
and abilities of those who labor to justify the programs of their
individual groups.
[NOTE: This process is apparently not recognized by most researchers
who clamor at the doors of the various agencies for $, but do not
volunteer to help those agencies in the process of justifying the
resource and objective requests (which will be the source of the
requested $).]
3. For the same reason, agencies will coordinate their work if it is in
the best interests of each to do so. Otherwise, they are not likely
to (why should you help someone competing for the same resources you
need? unless you both have common objectives and projects.)
4. Another MAJOR problem is one of joint and continuing committments.
By this I mean that both government and industry commit to working
toward some common goal, but circumstances out of control of one
side or another can negate the well-intended committment.
An example should serve to highlight this problem. This example does
not delve into the depths of why various things happened, it only
shows the effects of a government/industry relationship on the industry
"partner".
a. In 1975, NASA initiated the planning for the Numerical Aerodynamic
Simulation Facility. When industry was approached with a Request for
Proposal for a Preliminary Study, there was an implication (at least
perceived by industry) that the effort was the first step of a phased
procurement such that only those who were awarded the first contract
would be allowed to bid on following contracts, eventually leading to
the design, implementation, and construction of a very large, state of
the art, supercomputer system (sustaining in excess of 1 GFLOP
performance). As the project evolved, the next step (Feasibility
Study), was awarded to the original vendors. However, the next step
was the development of a new specification by NASA, followed by another
open competition, again for a phased procurement, awarded to the
original two vendors (no one else bid). After the System Design Study,
these two vendors submitted proposals to NASA (one copy of a proposal
was about 16"-20" thick). Subsequently, NASA decided to terminate that
procurement and to proceed with a different approach to solving their
problem.
b. During the 5 years covered by this scenario, many circumstances
changed, often out of control of anyone involved on either the
government or industry side. The result was that each vendor involved
spent over $1 million of their own development funds over 4-5 years,
and received an additional $1 million to cover additional costs (at no
profit).
Not many manufacturers can afford to commit major resources and
key innovators for 5 years to a venture where they do not have more
direct control over the outcome.
In this case, the funding of the projects were delayed significantly
by Congress, by changes of administration, by competition with many
other programs within the government, and by concerns that appeals
or other legal maneuvering would hamper the progress of the project,etc.
c. In the same project, one of the vendors began the relationship
with government fully intending that the work would lead to another
in a planned line of high-speed scientific processors (and would
therefore be willing to commit product development funds to augment
government funds). During the course of the project, new corporate
level leaders reviewed corporate objectives, and set new ones. Among
these was a decision NOT to be involved in such a product area.
The result was viewed from the government side as a move from having
access to the first of a new line of major computational systems to
another "orphan" (as Illiac IV was - only one built). An "orphan" is
clearly not a desirable research tool since the results cannot be
easily transferred to other users at other sites with their own
equipment.
NOTE: The above scenario is incomplete, and does not reflect all
circumstances, but does show some of the complexities which lurk
when trying to cope with Wallich's questions.
If anyone is interested in further details and circumstances, I will
try to clarify them, as best as I can recollect them.
Now, a note about using NETWORKS:
When trying to develop a very large project utilizing widely scattered
resources, existing networks are extremely difficult to use productively.
What is needed is some form of constrained distribution in order to
enforce the kinds of communication required in order to manage the
development efficiently. In the project described above, I had 50 people
working from 4-5 different sites including Pennsylvania, Alabama,
Santa Monica, Sunnyvale, and Camarillo. Unless a manager is provided
system development support tools which allow all on the project to
ignore the existance of a network and which enforce the relationships
and communications which much exist within the project, networks are
likely to create more work than they avoid. Such distributed system
development support tools (including the networks) are, of course,
necessary if a highly integrated system design is expected from a highly
distributed design group.
While I have your ear (eyes?), just a note about MCC and other ambitious
multi-company projects. I hope that they will be successful. They will
have their hands full however. I have observed many restructurings/
reorganizations within corporations first hand. The establishment of
an organization on paper never seems to be enough for things to happen.
You must have everyone in the organization BELIEVE in it and WORKING
for it. About every 3-6 months, management weeds out the obstacles in
the next lower layer of management. If there are 5 levels of structure,
that means possibly 30 months (2.5 YEARS), before the organization is
running smoothly. The problems of communications, trust, personnal contact,
NIH, etc. can only be compounded in a multi-corporation enterprise.
On the other hand, these problems need to be faced up to and solved
if we are going to remain competitive. Why don't we turn some good
systems designers loose on organizational/social system designs rather
than always being satisfied with cold hardware/software systems?
S.F.Lundstrom
∂01-Sep-83 1000 JMC*
Payless
∂01-Sep-83 1024 DELAGI%SUMEX-AIM.ARPA@SU-SCORE.ARPA PPA Non-Disclosure
Received: from SU-SCORE by SU-AI with TCP/SMTP; 1 Sep 83 10:24:36 PDT
Received: from SUMEX-AIM by Score with Pup; Thu 1 Sep 83 10:25:28-PDT
Date: Thu 1 Sep 83 10:18:05-PDT
From: Bruce Delagi <DELAGI@SUMEX-AIM>
Subject: PPA Non-Disclosure
To: super@Score
cc: rindfleisch@SUMEX-AIM
If you haven't seen the agreement DEC is requesting, I'll have copies to hand
out at the seminar this (thursday) afternoon. It appears that the
folks who own the intellectual property involved in the PPA insist that
I not discuss the PPA details with those who have not signed.
/bruce
-------
∂01-Sep-83 1024 DFH Keith Clark
I have a phone message for him. Do you know where
I can contact him?
∂01-Sep-83 1031 DFH travel arrangements
To: JMC, CLT
Unless you have other changes in the flights, Dina Bolla
will do your tickets this afternoon (business class).
There are two problems, both with hotels. The Museum
Hotel in Amsterdam is sold out. The reservation at the
Bristol is confirmed, but they wanted a deposit. They
have refused to use the American express no. as a guarantee.
Dina Bolla has asked them to accept an agency guarantee,
but we have no word back on this. What they want is $175
by Sept. 3, which means the money would need to be wired.
∂01-Sep-83 1126 Bonnie%UPenn%UPenn@UDel-Relay
Received: from UDEL-RELAY by SU-AI with TCP/SMTP; 1 Sep 83 11:26:43 PDT
Date: Thu, 1 Sep 83 11:57 EDT
From: Bonnie Webber <Bonnie%UPenn@UDel-Relay>
Return-Path: <Bonnie%UPenn%UPenn@UDel-Relay>
Subject:
To: John McCarthy <JMC@SU-AI>
Via: UPenn; 1 Sep 83 14:27-EDT
Thanks for the name & address. I will get in touch with them.
∂01-Sep-83 1708 DFH Common[e83,jmc]
There are some blotches, so will try to print it
again or otherwise fix tomorrow
∂01-Sep-83 1834 ME Prancing Pony Bill
Prancing Pony bill of JMC John McCarthy 1 September 1983
Previous Balance 0.99
Monthly Interest at 1.5% 0.01
-------
TOTAL AMOUNT DUE 1.00
Please deliver payments to Diana Hall, room 358, Jacks Hall.
Make checks payable to: STANFORD UNIVERSITY.
To ensure proper crediting, please include your Pony account name on your check.
Bills are payable upon presentation. Interest of 1.5% per month will be
charged on balances remaining unpaid 25 days after bill date above.
You haven't paid your Pony bill since 12/82.
Accounts with balances remaining unpaid for more than 55 days are
considered delinquent and are subject to reduction of credit limit.
Please pay your bill and keep your account current.
∂02-Sep-83 1015 JMC*
Kanerva
∂02-Sep-83 1019 EF@MIT-MC
Received: from MIT-MC by SU-AI with TCP/SMTP; 2 Sep 83 10:19:28 PDT
Date: 2 September 1983 12:58 EDT
From: Edward Fredkin <EF @ MIT-MC>
To: jmc @ SU-AI
Hi, have you gotten any reaction from members of the AAAI council?
For the time being, you can send me stuff at either MC or OZ.
I will most likely be cancelling or postponing my trip to
Paris, something has come up. If you get a chance, you can call me at
617-277-4444.
Best Regards, Ed Fredkin
∂02-Sep-83 1027 @SU-SCORE.ARPA:NILSSON@SRI-AI.ARPA AAMSI
Received: from SU-SCORE by SU-AI with TCP/SMTP; 2 Sep 83 10:27:46 PDT
Received: from SRI-AI.ARPA by SU-SCORE.ARPA with TCP; Fri 2 Sep 83 09:24:46-PDT
Date: 2 Sep 1983 0922-PDT
From: Nilsson at SRI-AI
Subject: AAMSI
To: jmc at SU-AI, jmc at SU-SCORE
1 Wed 31 Ted Shortliffe <Short Medical AI in Austria
2 2 Sep To: Shortliffe at SU Re: Medical AI in Austria
1 -- ************************
Return-Path: <SHORTLIFFE@SUMEX-AIM.ARPA>
Received: from SUMEX-AIM.ARPA by SRI-AI.ARPA with TCP; Wed 31 Aug 83 16:54:28-PDT
Date: Wed 31 Aug 83 16:55:38-PDT
From: Ted Shortliffe <Shortliffe@SUMEX-AIM.ARPA>
Subject: Medical AI in Austria
To: nilsson@SRI-AI.ARPA
cc: walker@SRI-AI.ARPA
Office: Room TC-117, Stanford Med Center; Phone: (415) 497-6979
Nils,
thanks for sending along the description of the Austrian medical
computing group. I met Horn in Amsterdam at MEDINFO last week and heard
his paper. He is doing rather good work in medical AI, as is a group under
Roux and Fieschi in Marseilles.
Incidentally, I hear from Bruce Buchanan that the issue of AAAI
sponsorship of the annual San Francisco AAMSI Congress wasn't raised at
last week's AAAI Council meeting. Don Walker has the letter of inquiry from
Dr. Collen from AAMSI (the American Association for Medical Systems and
Informatics). Unlike the invitation last year, they are no longer requesting
a cash contribution -- just moral support. In return, AAAI members will
get reduced registration fees for the meeting. Is there any chance that
this matter could be considered by the new council in time for a decision
to be made before the next AAMSI Congress announcement is printed later
this fall? Bruce suggested that an electronic mail poll might be possible.
Don can fill you in on further details. If this is better handled by
someone else, please let me know or forward my message along accordingly.
Thanks,
Ted
-------
2 -- ************************
Date: 2 Sep 1983 0920-PDT
From: Nilsson
Subject: Re: Medical AI in Austria
To: Shortliffe at SUMEX-AIM
cc: NILSSON
Ted, Sorry we didn't discuss the AAMSI matter at the AAAI Council
meeting. We had a very long agenda and the meeting
lasted to about midnite as it was. I will forward your
msg to John McCarthy (who is the current president). Perhaps,
as you suggest, it would be appropriate to handle this matter
by electronic mail discussion with the Council. I note with concern though that
AAMSI wants us to guarantee some mailings and publicity--and we have
been chary in the past about too many entanglements of this
sort. If they merely want us to lend our name in THEIR publicity, I'm
all for it. (I'll send a copy of the AAMSI letter--from Collen-- to
John also. Best wishes, Nils
-------
John, Ted sent me this msg about AAMSI. I'll let you deal with it
as you think best. I'll send the Collen letter
by ordinary mail. -Nils
-------
∂02-Sep-83 1038 TRATTNIG@SU-SCORE.ARPA Keith Clark's visit
Received: from SU-SCORE by SU-AI with TCP/SMTP; 2 Sep 83 10:35:23 PDT
Date: Fri 2 Sep 83 10:36:31-PDT
From: Werner Trattnig <TRATTNIG@SU-SCORE.ARPA>
Subject: Keith Clark's visit
To: mccarthy@SU-SCORE.ARPA
John,
Keith Clark told me that he has provided you a copy of his paper on
PARLOG. Please could you provide me a copy of it as I am quite
interested in this subject.
Thanks, Werner-
-------
∂02-Sep-83 1107 GOTELLI@SU-SCORE.ARPA Re: directory
Received: from SU-SCORE by SU-AI with TCP/SMTP; 2 Sep 83 11:06:23 PDT
Date: Fri 2 Sep 83 11:07:26-PDT
From: Lynn Gotelli <GOTELLI@SU-SCORE.ARPA>
Subject: Re: directory
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Thu 1 Sep 83 21:33:00-PDT
John, Thanks for the information about your daughter's SAIL
account. I am charging her SAIL usage to your unrestricted
account 1-FCZ154.
Lynn
-------
∂02-Sep-83 1125 NILSSON@SRI-AI.ARPA AAMSI
Received: from SRI-AI by SU-AI with TCP/SMTP; 2 Sep 83 11:25:30 PDT
Date: 2 Sep 1983 0922-PDT
From: Nilsson at SRI-AI
Subject: AAMSI
To: jmc at SU-AI, jmc at SU-SCORE
1 Wed 31 Ted Shortliffe <Short Medical AI in Austria
2 2 Sep To: Shortliffe at SU Re: Medical AI in Austria
1 -- ************************
Return-Path: <SHORTLIFFE@SUMEX-AIM.ARPA>
Received: from SUMEX-AIM.ARPA by SRI-AI.ARPA with TCP; Wed 31 Aug 83 16:54:28-PDT
Date: Wed 31 Aug 83 16:55:38-PDT
From: Ted Shortliffe <Shortliffe@SUMEX-AIM.ARPA>
Subject: Medical AI in Austria
To: nilsson@SRI-AI.ARPA
cc: walker@SRI-AI.ARPA
Office: Room TC-117, Stanford Med Center; Phone: (415) 497-6979
Nils,
thanks for sending along the description of the Austrian medical
computing group. I met Horn in Amsterdam at MEDINFO last week and heard
his paper. He is doing rather good work in medical AI, as is a group under
Roux and Fieschi in Marseilles.
Incidentally, I hear from Bruce Buchanan that the issue of AAAI
sponsorship of the annual San Francisco AAMSI Congress wasn't raised at
last week's AAAI Council meeting. Don Walker has the letter of inquiry from
Dr. Collen from AAMSI (the American Association for Medical Systems and
Informatics). Unlike the invitation last year, they are no longer requesting
a cash contribution -- just moral support. In return, AAAI members will
get reduced registration fees for the meeting. Is there any chance that
this matter could be considered by the new council in time for a decision
to be made before the next AAMSI Congress announcement is printed later
this fall? Bruce suggested that an electronic mail poll might be possible.
Don can fill you in on further details. If this is better handled by
someone else, please let me know or forward my message along accordingly.
Thanks,
Ted
-------
2 -- ************************
Date: 2 Sep 1983 0920-PDT
From: Nilsson
Subject: Re: Medical AI in Austria
To: Shortliffe at SUMEX-AIM
cc: NILSSON
Ted, Sorry we didn't discuss the AAMSI matter at the AAAI Council
meeting. We had a very long agenda and the meeting
lasted to about midnite as it was. I will forward your
msg to John McCarthy (who is the current president). Perhaps,
as you suggest, it would be appropriate to handle this matter
by electronic mail discussion with the Council. I note with concern though that
AAMSI wants us to guarantee some mailings and publicity--and we have
been chary in the past about too many entanglements of this
sort. If they merely want us to lend our name in THEIR publicity, I'm
all for it. (I'll send a copy of the AAMSI letter--from Collen-- to
John also. Best wishes, Nils
-------
John, Ted sent me this msg about AAMSI. I'll let you deal with it
as you think best. I'll send the Collen letter
by ordinary mail. -Nils
-------
∂02-Sep-83 1253 TUCCI@SU-SCORE.ARPA project
Received: from SU-SCORE by SU-AI with TCP/SMTP; 2 Sep 83 12:52:54 PDT
Date: Fri 2 Sep 83 12:53:51-PDT
From: Christopher Tucci <TUCCI@SU-SCORE.ARPA>
Subject: project
To: ark@SU-AI.ARPA, pw@SU-AI.ARPA, jmc@SU-AI.ARPA
will be turned in Tuesday -- just fixing up some last details
in the code and the write-up!
Chris
-------
∂02-Sep-83 1305 SHARON@SU-SCORE.ARPA dina bola travel
Received: from SU-SCORE by SU-AI with TCP/SMTP; 2 Sep 83 13:05:25 PDT
Date: Fri 2 Sep 83 13:06:32-PDT
From: Sharon Bergman <SHARON@SU-SCORE.ARPA>
Subject: dina bola travel
To: jmc@SU-AI.ARPA
cc: dfh@SU-AI.ARPA
During the lunch hour, someone from Dina Bola Travel Agency dropped
off some tickets for you. I have them here at the reception desk.
Sharon
-------
∂02-Sep-83 1500 JMC*
kaufman
∂04-Sep-83 0150 ARK Happy birthday!
and many more, I'm sure!
Arthur
∂04-Sep-83 1331 JK
∂04-Sep-83 1327 JMC closing the consulate
Your idea is a good one. If the organizations protesting had signs suggesting
closing the consulate, it would make the Russians more nervous than any other
kind of sign and would have some probability of being picked up by the media
and actually happening.
--------
Yes.
Did you read the editorial in the SF Examiner Sunday issue?
It contains the same idea.
∂04-Sep-83 1940 RPG Congratulations
on fronts. I hope that you are both happy together.
-rpg-
∂04-Sep-83 2204 SHORTLIFFE%SUMEX-AIM.ARPA@SU-SCORE.ARPA forwarded messages regarding AAMSI and AAAI
Received: from SU-SCORE by SU-AI with TCP/SMTP; 4 Sep 83 22:04:44 PDT
Received: from SUMEX-AIM by Score with Pup; Sun 4 Sep 83 22:05:50-PDT
Date: Sun 4 Sep 83 22:07:19-PDT
From: Ted Shortliffe <Shortliffe@SUMEX-AIM>
Subject: forwarded messages regarding AAMSI and AAAI
To: jmc@Score
Office: Room TC-117, Stanford Med Center; Phone: (415) 497-6979
John,
Here is my exchange of messages with Nils regarding the AAMSI meeting.
Thought I should bring you up to date.
Thanks for your help,
Ted
---------------
1) 2-Sep Nilsson at SRI-AI Re: Medical AI in Austria
2) 4-Sep To: Nilsson@SRI-AI.A AAMSI Congress
Message 1 -- ************************
Return-Path: NILSSON@SRI-AI.ARPA
Received: from SRI-AI.ARPA by SUMEX-AIM.ARPA with TCP; Fri 2 Sep 83 15:27:30-PDT
Date: 2 Sep 1983 0920-PDT
From: Nilsson at SRI-AI
Subject: Re: Medical AI in Austria
To: Shortliffe at SUMEX-AIM
cc: NILSSON
Ted, Sorry we didn't discuss the AAMSI matter at the AAAI Council
meeting. We had a very long agenda and the meeting
lasted to about midnite as it was. I will forward your
msg to John McCarthy (who is the current president). Perhaps,
as you suggest, it would be appropriate to handle this matter
by electronic mail discussion with the Council. I note with concern though that
AAMSI wants us to guarantee some mailings and publicity--and we have
been chary in the past about too many entanglements of this
sort. If they merely want us to lend our name in THEIR publicity, I'm
all for it. (I'll send a copy of the AAMSI letter--from Collen-- to
John also. Best wishes, Nils
-------
Message 2 -- ************************
Date: Sun 4 Sep 83 22:04:42-PDT
From: Ted Shortliffe <Shortliffe@SUMEX-AIM.ARPA>
Subject: AAMSI Congress
To: Nilsson@SRI-AI.ARPA
Office: Room TC-117, Stanford Med Center; Phone: (415) 497-6979
In-Reply-To: Your message of Fri 2 Sep 83 09:20:00-PDT
Nils,
Sorry to have bothered you about this at all. I should have remembered
to contact John McCarthy as the new President. However, I do appreciate your
response and your help getting the necessary information to John. I had a
call from him on Friday, and I gathered he was going to ask Claudia to poll
the Council members about the matter.
Incidentally, the AAMSI meeting seems to be financially solid after
two good initial meetings, and that is why they weren't asking for up-front
money. Their emphasis now is on improving the scientific quality of the
meeting, and one method for doing this is to build ties to other societies
with a reputation for quality research and conferences. I would personally
like to see AAAI cosponsor the meeting, both to increase the AI presence
at the AAMSI Congress and to lower the registration rate for AAAI members who
choose to attend.
I understand your concern about the entanglements of promotional
guarantees, but this is exactly the kind of task that can and should be
passed on to the medical liaison in charge of the AAAI-M. For the last
two years I've sent information on the AAMSI meeting to all AAAI-M members,
so we have, in effect, already been helping with AAMSI publicity (at least
for that portion of the AAAI that is involved in medically-related work).
My last mailing will be this autumn (my term as liaison expires in October).
Pete Szolovits and B. Chandrasekaran are the two candidates to replace me
in the position, and either should do a reliable job keeping the ball going
for the next two years. The new liaison will be elected by the AAAI-M
membership this autumn. Incidentally, about 20% of the AAAI membership has
joined the AAAI-M.
Thanks again for your help and support.
Regards,
Ted
-------
-------
∂05-Sep-83 2008 ULLMAN@SU-SCORE.ARPA
Received: from SU-SCORE by SU-AI with TCP/SMTP; 5 Sep 83 20:08:07 PDT
Date: Mon 5 Sep 83 20:09:12-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
To: JMC@SU-AI.ARPA
In-Reply-To: Message from "John McCarthy <JMC@SU-AI>" of Fri 26 Aug 83 21:12:00-PDT
thanks.
-------
∂06-Sep-83 1042 BMACKEN@SRI-AI.ARPA Associations
Received: from SRI-AI by SU-AI with TCP/SMTP; 6 Sep 83 10:42:21 PDT
Date: Tue 6 Sep 83 10:41:47-PDT
From: BMACKEN@SRI-AI.ARPA
Subject: Associations
To: BMacken@SRI-AI.ARPA, BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA,
Briansmith@PARC-MAXC.ARPA, grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA,
JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA, kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA,
peters@SRI-AI.ARPA, rperrault@SRI-AI.ARPA, sag%SU-psych@SU-SCORE.ARPA,
stan@SRI-AI.ARPA, Wasow%SU-psych@SU-SCORE.ARPA, winograd@PARC-MAXC.ARPA,
appelt@SRI-AI.ARPA, desrivieres@PARC-MAXC.ARPA, eric@MIT-MC.ARPA,
ford@PARC-MAXC.ARPA, halvorsen@PARC-MAXC.ARPA, hans@SRI-AI.ARPA,
hobbs@SRI-AI.ARPA, kay@PARC-MAXC.ARPA, konolige@SRI-AI.ARPA,
pollard.hplabs@UDEL-RELAY.ARPA, pullum.hplabs@UDEL-RELAY.ARPA,
pcohen@SRI-KL.ARPA, pereira@SRI-AI.ARPA, shieber@SRI-AI.ARPA,
stickel@SRI-AI.ARPA, stucky@SRI-AI.ARPA, withgott@PARC-MAXC.ARPA
In order to allocate funds properly across projects and areas, we need
to associate each associate, student, and anyone else getting paid with
Program SL funds with particular projects.
In addition, Jon has decided to eliminate the need for the Publication
Committee by requiring anyone wishing to publish a paper through CSLI to
submit it for review first to the associated project manager and then to the
area leader. From there it will go to the Executive Committee and to Dianne
Kanerva for publishing. Thus, anyone wishing to publish papers through
CSLI must be assigned to a project whether or not they are getting paid by
Program SL funds.
Would everyone in these categories please decide which project or projects
you want to be associated with and let me know as soon as possible. If
you are connected with more than one project, let me know the approximate
percentages of time you expect to spend on each during this year. I need
to meet with people in the Sponsored Projects office later this week to
set up our accounts, so the sooner the better.
Thanks.
B.
-------
∂06-Sep-83 1151 ullman@Diablo Re: [Tom Binford <TOB@SU-AI>: LISP machines ]
Received: from SU-HNV by SU-AI with PUP; 06-Sep-83 11:50 PDT
Date: Tue, 6 Sep 83 11:51 PDT
From: Jeff Ullman <ullman@Diablo>
Subject: Re: [Tom Binford <TOB@SU-AI>: LISP machines ]
To: BScott@SU-Score, Bosack@SU-Score, CSL.JLH@SU-Score, Feigenbaum@Sumex-Aim,
JMC@Sail, Rindfleisch@Sumex-Aim
My understanding was that Robotics was included in the non--HPP
portion of the AI area, and that John McC's "pot" should cover
LISP machines for use by Robotics. Actually, thinking back to
the time the proposal was formulated, long before HPP planned to
move out of MJH, the issue was not one of dividing up the pot of
LISP machines, but of determining what dialects of LISP they
should run, and therefore, which suppliers should be used.
I think the intend was always that the LISP machines would be
available to all ARPA contractors needing them on an equal
basis. Now we have both HPP and Robotics leaving MJH, and
we probably should find some equitable way of dedicating
machines to the several groups. I hope that can be done by
negotiation between the parties, without getting the department
involved. IN NO SENSE WAS ROBOTICS "LEFT OUT" OF THE GRANT.
∂06-Sep-83 1527 ULLMAN@SU-SCORE.ARPA meeting place
Received: from SU-SCORE by SU-AI with TCP/SMTP; 6 Sep 83 15:26:50 PDT
Date: Tue 6 Sep 83 15:27:16-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: meeting place
To: super@SU-SCORE.ARPA
The meeting this Thursday, Sept. 8, 3PM, will be in AEL 102 instead
of the CIS conference room.
Bruce Delagi will cover the PPA architecture. All attendees will
be expected to sign the DEC nondisclosure agreement.
-------
∂07-Sep-83 1049 BSCOTT@SU-SCORE.ARPA N00039-82-C-0250
Received: from SU-SCORE by SU-AI with TCP/SMTP; 7 Sep 83 10:49:41 PDT
Date: Wed 7 Sep 83 10:42:15-PDT
From: Betty Scott <BSCOTT@SU-SCORE.ARPA>
Subject: N00039-82-C-0250
To: DCL@SU-AI.ARPA, TOB@SU-AI.ARPA, Wiederhold@SUMEX-AIM.ARPA, JMC@SU-AI.ARPA,
Manna@SU-AI.ARPA
cc: MAS@SU-AI.ARPA, DFH@SU-AI.ARPA, Atkinson@SU-SCORE.ARPA, RBL@SU-AI.ARPA,
BScott@SU-SCORE.ARPA
This is to let you know that the current ARPA contract extension through
1/31/84 has been officially approved.
Betty
-------
∂07-Sep-83 1600 AAAI-OFFICE@SUMEX-AIM [John McCarthy <JMC@SU-AI>:]
Received: from SUMEX-AIM by SU-AI with PUP; 07-Sep-83 16:00 PDT
Date: Wed 7 Sep 83 16:01:12-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM>
Subject: [John McCarthy <JMC@SU-AI>:]
To: AMAREL@RUTGERS.ARPA, BUCHANAN@SUMEX-AIM, BENGELMORE@SRI-KL.ARPA,
LERMAN@SRI-KL.ARPA, GENESERETH@SUMEX-AIM, GROSZ@SRI-AI.ARPA,
HART@SRI-KL.ARPA, JMC@SAIL, MINSKY@MIT-MC.ARPA, NILSSON@SRI-AI.ARPA,
REDDY@CMU-CS-A.ARPA, RICH@MIT-MC.ARPA, STAN@SRI-AI.ARPA, GJS@MIT-MC.ARPA,
TENENBAUM@SRI-KL.ARPA, WALKER@SRI-AI.ARPA, BONNIE.UPENN@UDEL-RELAY.ARPA,
DWALTZ@BBNG.ARPA, BOBROW@PARC-MAXC.ARPA, MCDERMOTT@YALE.ARPA,
STEFIK@PARC-MAXC.ARPA, KRD@MIT-MC.ARPA
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
John McCarthy is out of town until September 27th so, he's asked me to distribute Ed Fredkin's proposal to the AAAI's Executive Council. At this time, he makes no recommendation on this proposal.
--- Claudia Mazzetti
---------------
Return-Path: JMC@SU-AI
Received: from SU-AI.ARPA by SUMEX-AIM.ARPA with TCP; Sun 4 Sep 83 01:07:08-PDT
Date: 04 Sep 83 0104 PDT
From: John McCarthy <JMC@SU-AI>
To: aaai-office@SUMEX-AIM
Here is Ed Fredkin's (EF@MIT-MC) proposal for a meeting and white
paper on AI. I suppose the issue for the Council is whether we
consider a push for high level attention to AI at this time
appropriate and, if so, whether this is the right way to go about
it. Though Fredkin is not currently active in AI, he
is energetic and knows lots of people. I suppose
the Council should be represented more specifically on the Steering
Committee if we decide to proceed.
BACKGROUND AND HISTORY
We believe that the United State may be at a turning point,
facing either a great opportunity or a potential risk, depending
on the course of action taken now in the field of Artificial
Intelligence. The reason is not because of any startling new
research developments or breakthroughs, but rather because, for
the first time, there is a dramatically increased awareness of
the possibility of practical applications evolving from AI
research, especially in Japan. The Fifth Generation book is
actually the number one best seller (non-fiction) in Japan this
week, the last week in August. This comes at a time when the
rapid progress of computer technology finally allows AI programs
to run on relatively inexpensive systems. The confluence of
events may make it possible to do things now, that may not be
possible later. The reason that we are at a turning point is
because we face a window of opportunity that we must not miss.
For the past twenty years, the US government, through DARPA,
has supported research in the field of Artificial Intelligence.
While the work, which started at MIT and Stanford, was an obscure
area of basic research in the late fifties, it is now realized
that the practical applications of that research may have a very
large and increasingly important effect on the future strength of
this country, vis a vis its economic competitors. Japan, is
today the country with the most ambitious national program in the
field of AI. England has also embarked on a national effort.
While the US has been the leader in this field, we may not be
able to maintain that lead without embarking upon efforts of
greatly increased scope. The payoff, of maintaining our
preeminent position, appears to be so large as to justify very
great investments in this area. On the other hand, should we
fall behind in this technology, the risks we might face are not
merely economic ones, since, someday, advanced AI technology may
be an important factor in other aspects of national power. Most
importantly, the opportunity to precipitate large scale
governmental support appears to be present today, it was never
present in the past and, if this opportunity passes by, it may
not be present again for some time.
THE MEETING AND WHITE PAPER
The goals envisioned for the meeting would be to consider
certain aspects of AI research, development and applications
possibilities, consider what other nations are doing in that
context, and finally consider possible recommendations for
national action in the US. While it would be healthy for the
conferees to consider broadly the realm of future possibilities,
quite conservative extrapolations of what we already know should
be sufficient to allow one to conclude that the scale of prudent
activity for the US is greater than presently being considered.
If one looks at the range of practical applications, using
industry standard criteria such as projected return on
investment, a possible conclusion will be that the most
reasonable level of investment in all aspects of AI, from basic
research to practical applications of proven technology, may be
far greater than what is now being done. If the right group of
conferees did come to such a conclusion, and if it was able to
support its conclusions with reasoned arguments, then it is
likely that its recommendations might result in appropriate
governmental action.
WHO SHOULD BE INVITED
Attendance at this meeting should be limited to a small
number, perhaps 20 participants. The participants should consist
of two groups, those very familiar with AI, who are good at
conveying the concepts, possibilities and possible consequences
of various courses of action to a group of intelligent laymen
(laymen in the field of AI). And then the others, who should be
persons of great prestige who would be drawn from government,
science and industry.
ATTENDEES DRAWN FROM THE AI COMMUNITY
The members of the AI community, who attend the meeting,
will have the responsibility of educating the others, and
communicating the possible scenarios consequent to various
proposed actions. There is a great tendency for the experts in
an obscure field to talk over the heads of the most intelligent
of laymen, resulting in very little communication. In this case,
it will be very important to work hard at conveying the
information from expert to layman in ways that result in
understanding and appreciation for the most important issues.
There is, in this plan, the implication that it would be
sensible for the members of the AI community who attend the
meeting, to be of one mind on the important issues, and if that
is not possible, then there may be no point in holding the
meeting. The Japanese have made a great art of coming to a
consensus, where the natural tendency to display one's ego by
vociferous disagreement is suppressed. Some of the superstars in
the field AI are infamous for the opposite, never participating
in a meeting without showing strong disagreement with some point,
as a matter of principal. While the AI community can know and
love its heros for their weaknesses as well as for their
strengths, if we want to have an effect on the world, we will
have to be pragmatic, and hold back a little of our
individualism. The AI persons need not agree on all points, but
they must understand in advance what they disagree about, they
must be able to delimit and make clear the consequent
alternatives, and they must not allow such differences to get in
the way of communicating the broad areas of agreement. As a
consequence, the AI participants must get together before the
meeting to explore the issues and to come to a consensus.
THE NON-AI PARTICIPANTS
The main objective of the meeting will be to communicate the
rationale for greatly increasing government funding of various
aspects of AI research and technology. The audience should
consists of a group of prestigious movers and shaker who are
able to appreciate the possibilities. It will be important to
have representatives from three major groups: Government, because
they will best understand the process of turning such concepts
into actions; Industry, because much of the economic consequences
will be a result of what American Industry chooses to do, and
because of the practical experience of such persons; Science,
because the government is used to, for good reason, consulting
with certain trusted scientific advisors. Again, since the
objective is to give the best opportunity to the realization of
the goals of the meeting, the non-AI participants need to be very
carefully selected.
PERTINENT MATERIALS AND THE WHITE PAPER
A very important aspect of the conference would be the
distribution to the conferees, prior to the conference, of
materials that detail what is happening in the field of AI.
Examples might include the Fifth Generation by Feigenbaum and
McCorduck and papers such as "Artificial Intelligence: An
Assessment of the State-of-the-Art and Recommendation for Future
Directions" by David Waltz et al, in the Fall 83 issue of AI
Magazine. In addition, the organizers of the meeting should
write and distribute condensed background information that can be
easily and quickly digested. Subsequent to the conference, a
white paper should be prepared and distributed. That white paper
will be the focal point of consensus and a document that makes
specific recomendations for action.
THE SPECIFIC RECOMMENDATION TO THE AAAI COUNCIL
The council should agree that the AAAI should sponsor a
meeting, along the lines outlined above. While the AAAI would
sponsor the meeting, the opinions and conclusions expressed by
the participants would be their own, and not official positions
of the AAAI. A steering committee consisting of Raj Reddy,
Gordon Bell, Ed Fredkin, Bob Kahn (if he is willing) and perhaps
one or two others should be constituted. Ed Fredkin would be the
Chairman of the committee and meeting. The meeting will be at no
cost to the AAAI, it will pay its fair share of all expenses
incurred by the AAAI and return a modest profit. Claudia Mazzetti
has indicated that she would be able and pleased to handle
various arrangements for the meeting, for which the AAAI would be
reimbursed. The most likely site would be in the Washington
area, and the meeting would take place sometime before the end of
the year.
-------
∂07-Sep-83 1705 KJB@SRI-AI.ARPA Description of research projects
Received: from SRI-AI by SU-AI with TCP/SMTP; 7 Sep 83 17:04:49 PDT
Date: Wed 7 Sep 83 16:59:41-PDT
From: KJB@SRI-AI.ARPA
Subject: Description of research projects
To: BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA, Briansmith@PARC-MAXC.ARPA,
grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA, JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA,
kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA, peters@SRI-AI.ARPA,
rperrault@SRI-AI.ARPA, sag%SU-psych@SUMEX-AIM.ARPA, stan@SRI-AI.ARPA,
wasow@PARC-MAXC.ARPA, Wasow%SU-psych@SUMEX-AIM.ARPA,
winograd@PARC-MAXC.ARPA
Ray, We really need to get the description of the SL program
out. I end up needing to give out copies of it several times
a day. Could you give it to me next Monday, with whatever
changes the project managers have suggested? I will read it over
and give it to D. Kanerva. Thanks, Jon
-------
∂07-Sep-83 2144 POURNE@MIT-MC HELP
Received: from MIT-MC by SU-AI with TCP/SMTP; 7 Sep 83 21:44:30 PDT
Date: 8 September 1983 00:45 EDT
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: HELP
To: jmc @ SU-AI
cc: POURNELLE @ MIT-MC
Deadlines coming. Pathway to judge established. Need words,
comments, critiques, other stuff. regarding packet sent to you
by Snail last month.
Please?
JEP
∂07-Sep-83 2157 GENESERETH@SUMEX-AIM ai course
Received: from SUMEX-AIM by SU-AI with PUP; 07-Sep-83 21:56 PDT
Date: Wed 7 Sep 83 17:24:03-PDT
From: Michael Genesereth <GENESERETH@SUMEX-AIM>
Subject: ai course
To: jmc@SAIL
cc: genesereth@SUMEX-AIM
John,
You may recall that I mentioned a call I received from Dr. Robert
Rector of UCLA. He is interested in having us do a repeat of the
AI course for him in LA sometime next Spring. I told him we'd talk
iyt over and get back to him this week. He offered 25% of the gross
as compensation.
While I thought the course in SC went all right and I enjoyed the
chance to chat with you, I am not enthusiastic about spending too
much time teaching outside of Stanford. I am enthusiastic about
spending more time chatting about "prevention" and other things,
but of course we could do that here.
The compensation ius not bad. If they get 60 people as before at
$800 per person that would be $6000 for each of us.
Would you like to take him up on the offer. If so, I'd be pleased
to do it. In either case I need to tell him soon.
mrg
Incidentally, Mandel wants us to do a repeat in SC next year also.
-------
∂08-Sep-83 0010 HST
i hope you came home in a good way.do you know the net address of
ken kennedy at rice university? h.erbert
∂08-Sep-83 0400 @LBL-CSAM.ARPA:uw-beaver!ubc-vision!reiter@LBL-CSAM n.m.workshop
Received: from LBL-CSAM by SU-AI with TCP/SMTP; 8 Sep 83 04:00:25 PDT
Date: Thu, 8 Sep 83 04:05:09 PDT
From: uw-beaver!ubc-vision!reiter@LBL-CSAM
Return-Path: <uw-beaver!ubc-vision!reiter@LBL-CSAM>
Message-Id: <8309081105.AA08998@LBL-CSAM.ARPA>
Received: by LBL-CSAM.ARPA (3.347/3.35)
id AA08998; Thu, 8 Sep 83 04:05:09 PDT
To: uw-beaver!lbl-csam!JMC@SU-AI, uw-beaver!lbl-csam!bonnie.upenn@UDEL-RELAY
Subject: n.m.workshop
Dear John,
Bonnie and I are beginning to plan for the workshop, to be held in Oct.
1984 in some eastern woodsy setting. When you suggested the workshop you
thought that you would be able to convince AAAI to sponsor it. Do you
know anything about this possibility now? What would they provide? Travel
money? Food and accommodation for participants? Proceedings?
As you know, IEEE wants to sponsor it also? Does dual sponsorship create
problems for AAAI?
My net address is reiter@rutgers.
Hope all is well with you. If you have written up any of your more recent
stuff on circumscription I'd like to see it.
Regards, Ray.
∂08-Sep-83 1145 NILSSON@SRI-AI.ARPA Fredkin Meeting
Received: from SRI-AI by SU-AI with TCP/SMTP; 8 Sep 83 11:43:44 PDT
Date: 8 Sep 1983 1142-PDT
From: Nilsson at SRI-AI
Subject: Fredkin Meeting
To: aaai, jmc at SU-AI
cc: Nilsson
After reading the Fredkin proposal for a AAAI sponsored meeting, I
am in favor of AAAI supporting the first meeting he talks about, namely
the one in which various AI "heavies" get together to see if 1) there
ought to be a larger meeting (including govt. and other attendees), and 2)
if these heavies can reach any sort of consensus about what the AI position
ought to be at such a larger meeting. This preliminary meeting could
really be anywhere that's convenient to most of the people who will be
attending. -Nils
-------
∂08-Sep-83 1544 JRP@SRI-AI.ARPA VISITING SCHOLAR CARD
Received: from SRI-AI by SU-AI with TCP/SMTP; 8 Sep 83 15:44:26 PDT
Date: Thu 8 Sep 83 15:43:37-PDT
From: JRP@SRI-AI.ARPA
Subject: VISITING SCHOLAR CARD
To: CSLI-folks: ;
cc: JRP@SRI-AI.ARPA
Those of you who are not now on the STanford faculty, and do
not have any other official Stanford connection, need Visiting
Scholar cards. This will permit you to buy a parking stic ker and have
access to the libraries. If you need such, please send me a
message right away. For convenience, include in your message your
official title and your address; we will need the former to apply
for the card and the latter to get it to you. The visiting scholar
card will be useful as an immediate way of getting parking and
library privileges, even to those who will have some other status
later on.
-------
∂08-Sep-83 1831 Stefik.PA@PARC-MAXC.ARPA Re: [John McCarthy <JMC@SU-AI>:]
Received: from PARC-MAXC by SU-AI with TCP/SMTP; 8 Sep 83 18:31:03 PDT
Date: Thu, 8 Sep 83 18:29 PDT
From: Stefik.PA@PARC-MAXC.ARPA
Subject: Re: [John McCarthy <JMC@SU-AI>:]
In-reply-to: "AAAI-OFFICE@SUMEX-AIM.ARPA's message of Wed, 7 Sep 83
16:01:12 PDT"
To: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
cc: AMAREL@RUTGERS.ARPA, BUCHANAN@SUMEX-AIM.ARPA,
BENGELMORE@SRI-KL.ARPA, LERMAN@SRI-KL.ARPA, GENESERETH@SUMEX-AIM.ARPA,
GROSZ@SRI-AI.ARPA, HART@SRI-KL.ARPA, JMC@SAIL.ARPA, MINSKY@MIT-MC.ARPA,
NILSSON@SRI-AI.ARPA, REDDY@CMU-CS-A.ARPA, RICH@MIT-MC.ARPA,
STAN@SRI-AI.ARPA, GJS@MIT-MC.ARPA, TENENBAUM@SRI-KL.ARPA,
WALKER@SRI-AI.ARPA, BONNIE.UPENN@UDEL-RELAY.ARPA, DWALTZ@BBNG.ARPA,
BOBROW.PA@PARC-MAXC.ARPA, MCDERMOTT@YALE.ARPA,
STEFIK.PA@PARC-MAXC.ARPA, KRD@MIT-MC.ARPA
I'm not close enough to the institutional funding situation to
understand this. I don't know that the amount of money available now
for the people and computing environments is a major limiting factor to
progress in AI?
Not knowing the answer, here are some arguments needing evaluation.
-----------
Arguments against the proposal.
(1) The money is already there.
It is in the existing DARPA & NSF budgets for AI as well as the
strategic computing program.
(2) The budget has just been increased.
The US is already responding to the "fifth generation" work in Japan ,
(or perhaps to the same underlying opportunities) with the "strategic
computing" program at DARPA. This program is already in the stages of
funding, and is coupled with the primary sources of funding in AI. It
would be more productive to work to shape this program and make concrete
proposals than to start another round of meetings and discussions to
generate funding support.
(3) Crash programs in fundamental research don't work.
Example -- the binge on research funds for the "War on Cancer" which
quietly faded away.
Arguments for the proposal.
(1) "Advances in AI" have been used to generate interest and support for
work on computer architecture. But this has not resulted in incresases
in money allocated for basic research on AI.
∂09-Sep-83 0348 Tenenbaum@SRI-KL.ARPA Re: [John McCarthy <JMC@SU-AI>:]
Received: from SRI-KL by SU-AI with TCP/SMTP; 9 Sep 83 03:48:31 PDT
Date: Thu 8 Sep 83 22:34:01-PDT
From: Marty Tenenbaum <Tenenbaum at SRI-KL at SU-DSN>
Subject: Re: [John McCarthy <JMC@SU-AI>:]
To: AAAI-OFFICE at SUMEX-AIM
cc: AMAREL at RUTGERS at SU-DSN, BUCHANAN at SUMEX-AIM, BENGELMORE at SRI-KL at SU-DSN,
LERMAN at SRI-KL at SU-DSN, GENESERETH at SUMEX-AIM, GROSZ at SRI-AI at SU-DSN,
HART at SRI-KL at SU-DSN, JMC at SU-AI, MINSKY at MIT-MC at SU-DSN,
NILSSON at SRI-AI at SU-DSN, REDDY at CMU-CS-A at SU-DSN, RICH at MIT-MC at SU-DSN,
STAN at SRI-AI at SU-DSN, GJS at MIT-MC at SU-DSN, WALKER at SRI-AI at SU-DSN,
BONNIE.UPENN at UDEL-RELAY at SU-DSN, DWALTZ at BBNG at SU-DSN,
BOBROW at PARC-MAXC at SU-DSN, MCDERMOTT at YALE at SU-DSN,
STEFIK at PARC-MAXC at SU-DSN, KRD at MIT-MC at SU-DSN
In-Reply-To: Your message of Thu 8 Sep 83 16:50:47-PDT
Fredkin's proposal is a timely one. A national policy on AI research,
one that exploited our country's unique strengths and resources, would
make far more sense than our current paranoiac, copy-cat responses to
the Japanese and other national programs. One possibility: a
concerted national effort to develop a truly intelligent system by the
turn of the century. Such a program, thoughtfully conceived, could
be far more productive (and justifiably better funded) than current
fragmented efforts.
The pros and cons of a national AI policy, at the least, deserve
serious examination by our own community. This purpose could be served
by the preliminary meeting proposed by Fredkin. The results of that
meeting would then determine whether to proceed with the second
meeting.
-------
∂09-Sep-83 1035 @SU-SCORE.ARPA:TRATTNIG@SU-SIERRA.ARPA guest speakers
Received: from SU-SCORE by SU-AI with TCP/SMTP; 9 Sep 83 10:35:13 PDT
Received: from SU-SIERRA.ARPA by SU-SCORE.ARPA with TCP; Fri 9 Sep 83 10:35:59-PDT
Date: Fri 9 Sep 83 10:40:33-PDT
From: Werner Trattnig <TRATTNIG@SU-SIERRA.ARPA>
Subject: guest speakers
To: ullman@SU-SCORE.ARPA
cc: super@SU-SCORE.ARPA
Jeff,
I have just talked to Ian Watson from Manchaster University. He could
give a presentation of the Manchaster Data Flow Machine on Thursday,
September 22, at 3 pm in our seminar. Is this o.k. with your schedule?
There are also some other researchers in this area who would like to
join our group. However, I told them that the "membership fee" is a
presentation. In particular, Al Davis from Fairchild, Bob Keller from
LLL, David Warren from SRI, and Ed Ashcroft from SRI would like to
join our group. As an initial step, I'd suggest to add them to our
mailing list. Their mail addresses are:
adavis@sri-krl, ashcroft@sri-csl, warren@sri-ai,
lll-crg!keller@LBL-CSAM,
W-
-------
∂09-Sep-83 1106 bobrow.PA@PARC-MAXC.ARPA Re: [John McCarthy <JMC@SU-AI>:]
Received: from PARC-MAXC by SU-AI with TCP/SMTP; 9 Sep 83 11:06:45 PDT
Date: Fri, 9 Sep 83 11:01 PDT
From: bobrow.PA@PARC-MAXC.ARPA
Subject: Re: [John McCarthy <JMC@SU-AI>:]
In-reply-to: "AAAI-OFFICE@SUMEX-AIM.ARPA's message of Wed, 7 Sep 83
16:01:12 PDT"
To: AAAI <AAAI-OFFICE@SUMEX-AIM.ARPA>
cc: AMAREL@RUTGERS.ARPA, BUCHANAN@SUMEX-AIM.ARPA,
BENGELMORE@SRI-KL.ARPA, LERMAN@SRI-KL.ARPA, GENESERETH@SUMEX-AIM.ARPA,
GROSZ@SRI-AI.ARPA, HART@SRI-KL.ARPA, JMC@SAIL.ARPA, MINSKY@MIT-MC.ARPA,
NILSSON@SRI-AI.ARPA, REDDY@CMU-CS-A.ARPA, RICH@MIT-MC.ARPA,
STAN@SRI-AI.ARPA, GJS@MIT-MC.ARPA, TENENBAUM@SRI-KL.ARPA,
WALKER@SRI-AI.ARPA, BONNIE.UPENN@UDEL-RELAY.ARPA, DWALTZ@BBNG.ARPA,
BOBROW.PA@PARC-MAXC.ARPA, MCDERMOTT@YALE.ARPA,
STEFIK.PA@PARC-MAXC.ARPA, KRD@MIT-MC.ARPA
I think that rather than launching directly into a meeting, perhaps we
could have a brainstorming session over the network trying to suggest
what things might be done if the US were to change the current scale of
investment in AI. If a number of good ideas came out of such
brainstorming, then a conference could be gathered. It's job would be
to put together more detailed plans and the sets of arguments that would
be needed to convince the appropriate powers that there was a good place
to expend some additional money. If, however, this group cannot come up
with significant suggestions which require increased funding, then it
will be apparent that a conference will serve no useful purpose at this
time.
Please feel free to solicit suggestions from those you know not on the
mailing list above. I will be glad to gather the suggestions and
compile a summary, including arguments for and against proposals and
people's rough cost estimates.
Some frequently voiced suggestions:
1) Lisp machines for every one in the field
(potential cost $30,000 times ~1000 active researchers
amortized over three years
~$3 Million / year)
2) Graduate fellowship program (includes tuition and stipend)
(100 positions @~$30,000
~3 Million/ year)
3) Postdoctoral fellowship at Universities
(15 position@ ~$60,000 -- includes overhead and stipend)
~1 Million / year
Questions:
Are these figures (costs and people) roughly correct?
How much of this funding is currently available?
∂09-Sep-83 1107 ULLMAN@SU-SCORE.ARPA seminar progress report
Received: from SU-SCORE by SU-AI with TCP/SMTP; 9 Sep 83 11:02:20 PDT
Date: Fri 9 Sep 83 11:00:21-PDT
From: Jeffrey D. Ullman <ULLMAN@SU-SCORE.ARPA>
Subject: seminar progress report
To: super@SU-SCORE.ARPA
I have a number, CS440, for the seminar.
The time is 4:15 Thursdays, and we're working on a room.
I have set up a file <ullman>schedule that is readable AND WRITABLE
by anyone. If you have a speaker for any of the sessions, just
edit the file. I'll talk at the first meeting where no one else
is willing to talk.
As for the "private" meeting, I suggest we plan to continue these
at 3PM. We have the TDC representative this Thursday (15th) and the next
day (Friday, 16th) we have a special meeting with Jack Dennis.
Werner has just told me he arranged for Ian Wilson of Manchester
to speak on the 22nd of September to the private groups.
That bumps Steve's talk on the Los Alamos meeting for the fourth time;
I hope Steve will agree to hold that meeting at 3PM Thurs. 29th,
just before the first CS440 talk, which I assume will be an in-house
speaker. We also have Burton Smith to talk to on Oct. 22,
although he will be talking in EE385, not the CS440 seminar.
Other dates for the "private" meeting are open.
-------
∂09-Sep-83 1322 LIBRARY@SU-SCORE.ARPA AI and Lisp Books
Received: from SU-SCORE by SU-AI with TCP/SMTP; 9 Sep 83 13:22:07 PDT
Date: Fri 9 Sep 83 13:23:10-PDT
From: C.S./Math Library <LIBRARY@SU-SCORE.ARPA>
Subject: AI and Lisp Books
To: mccarthy@SU-AI.ARPA
Prof. McCarthy:
The following titles have come to my attention: LISP; a gentle introduction to
symbolic computation by David S. Touretzky and Artificial Intelligence; tools,
techniques and applications by Marc Eisenstadt and Tim O'Shea. Neither title
has been published yet. From their titles they sound like books we would
want. However the authors involved do not sound real familiar (only have
one tech. rept. by Touretzky) and the publisher is Harper and Row which
really surprises me. Do you know anything abouth these titles or authors?
Would they be an appropriate level and quality for our library?
Harry
-------
Harry:
I have heard praise of the Touretzky book that suggests the library should
have it. I would gamble on the Eisenstadt and O'Shea book also.
I've been away so the issue may be moot.
∂09-Sep-83 1419 KJB@SRI-AI.ARPA EMACS on IBM-PC?
Received: from SRI-AI by SU-AI with TCP/SMTP; 9 Sep 83 14:19:26 PDT
Date: Fri 9 Sep 83 14:18:47-PDT
From: KJB@SRI-AI.ARPA
Subject: EMACS on IBM-PC?
To: csli.people: ;
John Perry heard recently that there is an editor for the IBM-PC
that is very similar to emacs. Does anyone know anything about
this? What it is called, e.g.? Jon
-------
∂09-Sep-83 1441 WUNDERMAN@SRI-AI.ARPA "Outer" Colloquium Series
Received: from SRI-AI by SU-AI with TCP/SMTP; 9 Sep 83 14:41:19 PDT
Date: Fri 9 Sep 83 14:40:29-PDT
From: WUNDERMAN@SRI-AI.ARPA
Subject: "Outer" Colloquium Series
To: CSLI-folks@SRI-AI.ARPA
cc: wunderman@SRI-AI.ARPA
From John Etchemendy via Wunderman at CSLI:
I will be the chairman of this year's "outer" colloquium committee.
As Jon explained last Thursday, the idea of the outer colloquium
series is to bring in speakers whose work bears on language and
information, but whose approach to these topics is somewhat different
from ours. This is rather vague, and meant to be. Basically, we
want speakers whose talks would be interesting, for whatever reason,
to a goodly portion of the CSLI people, though not ones whose work
is directly relevant to any of the projects on Program SL. (The latter
sort will fall under the aegis of Ivan's inner colloquium series.)
What I need from you are LOTS of suggestions; the more, the better.
We'll be bringing in from twelve to eihteen speakers for this series.
For budgetary reasons, at least half will have to be locals (Bay Area
residents or passers-through), and most will probably be academicians
(since they tend to come cheaper than popular authors, media personal-
ities, and so forth). Still, with everyone's input, we should be able
to come up with a varied and interesting collection of speakers.
I'll need your suggestions fairly quickly, since this Fall's speakers
should really have been invited last Spring. So in the next day or
two, send me names, addresses, and descriptions of anyone you think
might be a good bet. Since I don't receive mail at Ventura, send the
suggestions to the Philosophy Department.
Thanks.
-------
∂09-Sep-83 1624 WUNDERMAN@SRI-AI.ARPA ANNOUNCEMENT
Received: from SRI-AI by SU-AI with TCP/SMTP; 9 Sep 83 16:24:14 PDT
Date: Fri 9 Sep 83 16:23:38-PDT
From: WUNDERMAN@SRI-AI.ARPA
Subject: ANNOUNCEMENT
To: CSLI-folks@SRI-AI.ARPA
cc: wunderman@SRI-AI.ARPA
JOIN US FOR THE FIRST INNER COLLOQUIUM EVENT:
PRESENTER: Stanley Peters
TITLE: SID: The History of an Integrated Research Project
DATE: Thursday, September 15, 1983
TIME: 4:00 PM
PLACE: Turing Auditorium, Polya Hall, S.U.
(near Pine Hall, Jordan Quad)
For further information, contact wunderman@sri or call 497-0939,
or sag%su-psych@score
-------
∂09-Sep-83 1739 BMACKEN@SRI-AI.ARPA associations
Received: from SRI-AI by SU-AI with TCP/SMTP; 9 Sep 83 17:39:16 PDT
Date: Fri 9 Sep 83 17:38:32-PDT
From: BMACKEN@SRI-AI.ARPA
Subject: associations
To: BMacken@SRI-AI.ARPA, BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA,
Briansmith@PARC-MAXC.ARPA, grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA,
JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA, kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA,
peters@SRI-AI.ARPA, rperrault@SRI-AI.ARPA, sag%SU-psych@SU-SCORE.ARPA,
stan@SRI-AI.ARPA, Wasow%SU-psych@SU-SCORE.ARPA, winograd@PARC-MAXC.ARPA,
appelt@SRI-AI.ARPA, desrivieres@PARC-MAXC.ARPA, eric@MIT-MC.ARPA,
ford@PARC-MAXC.ARPA, halvorsen@PARC-MAXC.ARPA, hans@SRI-AI.ARPA,
hobbs@SRI-AI.ARPA, kay@PARC-MAXC.ARPA, konolige@SRI-AI.ARPA,
pollard.hplabs@UDEL-RELAY.ARPA, pullum.hplabs@UDEL-RELAY.ARPA,
pcohen@SRI-KL.ARPA, pereira@SRI-AI.ARPA, shieber@SRI-AI.ARPA,
stickel@SRI-AI.ARPA, stucky@SRI-AI.ARPA, withgott@PARC-MAXC.ARPA
Just a reminder that we need your project associations and percentages
if you are on more than one project by Tuesday afternoon. We need the
information for principals as well as associates. Project managers should
agree to the associations before I get them. (If you can't reach a manager
let me know what you think it will be and that it's not for sure.)
Thanks to all of you who have already responded.
When I get the list together, I'll send you copies to check over.
B.
-------
∂09-Sep-83 1809 KJB@SRI-AI.ARPA p.s. to Announcment
Received: from SRI-AI by SU-AI with TCP/SMTP; 9 Sep 83 18:09:02 PDT
Date: Fri 9 Sep 83 16:45:49-PDT
From: KJB@SRI-AI.ARPA
Subject: p.s. to Announcment
To: csli-folks@SRI-AI.ARPA
Ron Kaplan has voluntered to give a demonstration of the dandelion
at Ventura in the early afternoon next Thursday.
-------
∂09-Sep-83 2223 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: seminar progress report
Received: from SU-SCORE by SU-AI with TCP/SMTP; 9 Sep 83 22:22:52 PDT
Received: from Shasta by Score with Pup; Fri 9 Sep 83 22:22:13-PDT
Date: Friday, 9 Sep 1983 22:29-PDT
To: Jeffrey D. Ullman <ULLMAN at SU-SCORE.ARPA>
Cc: super at SU-SCORE.ARPA
Subject: Re: seminar progress report
In-reply-to: Your message of Fri 9 Sep 83 11:00:21-PDT.
From: Steve Lundstrom <slndstrm@Shasta>
Unfortunately, I will be out of town on the 29th. Perhaps
Joe Oliger, Werner Trattnig, and Ken Stevens could sub since
between them they covered all the meetings.
∂09-Sep-83 2229 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Re: seminar progress report
Received: from SU-SCORE by SU-AI with TCP/SMTP; 9 Sep 83 22:29:24 PDT
Received: from Shasta by Score with Pup; Fri 9 Sep 83 22:28:46-PDT
Date: Friday, 9 Sep 1983 22:35-PDT
To: Jeffrey D. Ullman <ULLMAN at SU-SCORE.ARPA>
Cc: super at SU-SCORE.ARPA
Subject: Re: seminar progress report
In-reply-to: Your message of Fri 9 Sep 83 11:00:21-PDT.
From: Steve Lundstrom <slndstrm@Shasta>
CORRECTION: Burton Smith will be talking to the EE380 Seminar on
Wed. Oct. 19. I have invited him to participate in an informal
discussion with us at our "private" meeting on Thurs. Oct. 20.
At this point, I have not invited him to give any other formal
presentations, and expect he will have departed by Sat. Oct. 22.
I am trying to arrange Burton's schedule for Thurs. Oct. 20.
Does anyone in the group want to meet privately, or to have
a session between your students and Burton?
Steve
∂12-Sep-83 1110 DOYLE@CMU-CS-C.ARPA Long-term AI research
Received: from CMU-CS-C by SU-AI with TCP/SMTP; 12 Sep 83 11:09:41 PDT
Received: ID <DOYLE@CMU-CS-C.ARPA>; Mon 12 Sep 83 14:11:09-EDT
Date: Mon 12 Sep 83 14:11:04-EDT
From: DOYLE@CMU-CS-C.ARPA
Subject: Long-term AI research
To: McCarthy@SU-AI.ARPA, Nilsson@SRI-AI.ARPA, Minsky%MIT-OZ@MIT-MC.ARPA,
Newell@CMU-CS-A.ARPA
cc: carbonell@CMU-CS-C.ARPA
Dear John, Nils, Marvin, and Allen,
Jaime Carbonell and I were recently discussing the newly admitted
graduate students here at CMU, and arrived at a tentative, but
disturbing conclusion: that the students interested in AI are not very
interested in fundamental questions, open problems, and long term
research, but instead are eager to get in on big, build-it-now
projects in expert systems and natural language interfaces. If
anything similar is the case at Stanford or MIT, it is cause for
alarm, since we cannot hope that many students at smaller schools will
press on areas they see receive little attention from the biggest
schools.
While it may be that CMU's admission policies favor engineering rather
than scientific interests, I think the principal cause of this
difficulty is that scientific work in AI is not very prominent. The
long-term scientific work (theoretical and empirical) reported in
conferences and journals is quite sparse compared to the massive
reports on expert systems, applied language understanding, and AI
hardware. In other fields, there is a similar preponderance of
engineering and commercial work over basic research, but the basic
research is made prominent with journals explicitly devoted to
theoretical or experimental work. This concentration is all that AI
lacks to make such work visible and respectable to students. Since
the term ``artificial intelligence'' connotes no orientation toward
science or engineering, the AI Journal, AI Magazine, IJCAI, and AAAI
conferences only display these two literatures in their relative
proportions --- which at the moment hides basic work among the mass of
more immediate projects.
I would like to suggest a few possible solutions to this difficulty.
The simplest is to have AI Journal and AI Magazine classify their
papers into sections as does JACM, but I do not think that would
really be effective. A better approach would be to create a new
journal under AAAI auspices, perhaps called something like Theoretical
Artificial Intelligence. (This name leaves out the long term
psychological experimental work, but I think that Cognitive Science
probably could cover that literature easily.) Another possibility is
to create one or more series of ``Memoirs,'' occasional issues or
booklets which would present either semi-monographs or collections of
papers. (I have in mind as models the AMS Memoirs and the Archive for
the History of the Exact Sciences.) This solution would allow having
theoretical, experimental, expert system, etc. series to make everyone
happy, and their separation would give each equal prominence
regardless of the respective page-counts.
I would be grateful for comments or advice you might have on these
suggestions. I am not quite sure what to do about this, other than to
inform others who I think might be interested, and especially the AAAI
executive committee. Could you please forward this to anyone else you
think might be interested?
Thank you very much,
Jon Doyle
-------
I agree there's a problem, and my first "President's message" in the
AI Magazine will be a harangue on the subject.
∂12-Sep-83 1250 LAWS@SRI-AI.ARPA Re: Long-term AI research
Received: from SRI-AI by SU-AI with TCP/SMTP; 12 Sep 83 12:50:30 PDT
Date: Mon 12 Sep 83 12:45:13-PDT
From: Ken Laws <Laws@SRI-AI.ARPA>
Subject: Re: Long-term AI research
To: DOYLE@CMU-CS-C.ARPA
cc: McCarthy@SU-AI.ARPA, Nilsson@SRI-AI.ARPA, Minsky%MIT-OZ@MIT-MC.ARPA,
Newell@CMU-CS-A.ARPA, carbonell@CMU-CS-C.ARPA
In-Reply-To: Message from "DOYLE@CMU-CS-C.ARPA" of Mon 12 Sep 83 12:06:43-PDT
An alternative to a new theoretical journal is to have one or
more new engineering journals. This would leave the theoretical
papers for JACM and the AI Journal. IJCAI would also have a
higher percentage of theoretical papers if other conferences
siphoned off the engineering papers.
I think we will see the engineering conferences very soon. Already
we have seen the NBS Trends and Applications conference, and SPIE
is having its AI bash soon. The IEEE Computer Society, through the
PAMI Technical Committee, is planning to start a series of major
AI conferences in addition to the workshops it has already scheduled;
these conferences would be oriented toward the papers that are too
applied for AAAI.
The IEEE Computer society may also be planning an AI magazine,
although I have not heard of any definite plans. The emphasis
would undoubtably be on the engineering side of AI. Other
applied journals will spring up to cover AI in medicine, robotics,
vision, speech, art, etc. I do not expect this to take long.
-- Ken Laws
-------
∂12-Sep-83 1545 slndstrm%SU-SHASTA.ARPA@SU-SCORE.ARPA Thursday Meeting (9/15, 3pm)
Received: from SU-SCORE by SU-AI with TCP/SMTP; 12 Sep 83 15:45:21 PDT
Received: from Shasta by Score with Pup; Mon 12 Sep 83 15:44:12-PDT
Date: Monday, 12 Sep 1983 15:51-PDT
To: super at Score
Subject: Thursday Meeting (9/15, 3pm)
From: Steve Lundstrom <slndstrm@Shasta>
Our visitors this week will be from Technology Development.
Harry Head will discuss the Parallel Processing Center proposal.
Dick Wilson will profile the activities (and plans) in their Advanced
Computer Architecture R&D Group.
Thanks. Steve Lundstrom
∂12-Sep-83 1808 MRC@SU-SCORE.ARPA removal of pre-1983 Score mailboxes
Received: from SU-SCORE by SU-AI with TCP/SMTP; 12 Sep 83 18:08:50 PDT
Date: Mon 12 Sep 83 17:59:19-PDT
From: Mark Crispin <MRC@SU-SCORE.ARPA>
Subject: removal of pre-1983 Score mailboxes
To: "Pre-83 Score users": ;
Postal-Address: 725 Mariposa Ave. #103; Mountain View, CA 94041
Phone: (415) 497-1407 (Stanford); (415) 968-1052 (residence)
If you received a copy of this message, it is because you
had a Score account prior to 1983 with the old prefixes (ADMIN,
CSD, CSL, ICL, OR). When the Score system was broken up early in
'83, mail forwardings were established for all the users who
migrated to Sierra (CSL, ICL, and OR). We did a similar thing
when we removed the CSD prefixes later that year.
We intend to remove all of these mail forwardings, effective
January 1, 1983. After that time, Score will no longer recognize
mailboxes beginning with ADMIN, CSD, CSL, ICL, or OR. We've
allocated a year to let you notify your electronic mail
correspondents of your new mailboxes.
This will *not* affect any other forwardings. In
particular, any aliases or forwardings in existance for a current
Score user will not be removed, irregardless of your having
received this message. Only the old prefixes will go away.
If you are a Score user, you should instruct your
correspondents to use your ID name without the prefix. In other
words, if your user name used to be CSD.JONES and it is JONES
today, if anybody still mails to CSD.JONES it will stop working
after the end of this year. People mailing to you should use
JONES instead.
If you are a Sierra user, Score will no longer explicitly
know about your mailbox name unless you have a Score account.
You can still get mail forwarding through Score, however. If
your former Score account was ICL.SMITH and your present Sierra
account is SMITH, you can have your correspondents send mail
either to SMITH@SU-SIERRA or (if you don't trust the Internet
gateway) to SMITH%Sierra@SU-SCORE.
Please note that this does not change anything on the Sierra
computer. If your account name is ICL.SMITH at Sierra your
account and mailbox will not be affected at Sierra.
If you have any questions about how this affects you
individually, please send mail to me (MRC@SU-SCORE) and I'll try
to explain. If you feel you should be exempted from this, please
send mail to me and your request will be considered on a
case-by-case basis.
-------
∂13-Sep-83 0638 Bossu.GIA@MIT-MULTICS.ARPA
Received: from MIT-MULTICS by SU-AI with TCP/SMTP; 13 Sep 83 06:38:05 PDT
Date: 13 September 1983 09:31 edt
From: Bossu.GIA at MIT-MULTICS
To: JMC at SU-AI
Do you know the name of the project account of Pat Hayes
at Multics Stanford University?
thank you
Genevieve Bossu (gia of Luminy).
Pat Hayes is PJH@SU-AI, but he hasn't logged in since May.
∂13-Sep-83 1312 @SRI-AI.ARPA:BrianSmith.PA@PARC-MAXC.ARPA Area C Meeting
Received: from SRI-AI by SU-AI with TCP/SMTP; 13 Sep 83 13:11:41 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Tue 13 Sep 83 09:09:20-PDT
Date: 13 Sep 83 09:07 PDT
From: BrianSmith.PA@PARC-MAXC.ARPA
Subject: Area C Meeting
To: CSLI-Folks@SRI-AI.ARPA
cc: BrianSmith.PA@PARC-MAXC.ARPA
(Not yet having Betsy's allocation charts, I am sending this to everyone;
a "CSLI-Area-C" mailing list will soon be set up.) Summary:
Time: 2:00 p.m. Thursday, Sept. 15.
Place: Seminar room at Ventura
What: Meeting to plan area C activities for the fall.
One of the most important things we have to do during the first year
(before the review) is to build up our expertise in area C. I would like
those of us who are intending to work in this area to get together, to
plot and scheme about how we want to spend the next few months. I
know some people are out of town this week, but I would nonetheless
like to have a first meeting as soon as possible.
The "Approaches to Computation" seminar is going to be at 2:00 every
Thursday, starting on the 29th. So why don't we meet at that time for
the next two weeks (the 15th and 22nd) to plan things. As well as
talking about people's own research, we need to plan and coordinate
the research seminar, talk about visitors, possible workshops, etc.
There are so many potential area C topics we could deal with that we
could easily get lost; one thought I had was to select a couple of themes
or issues to focus on during this first year. I would like them to be
general enough to provide a rallying point for both semanticists and
language designers. Two suggestions:
a. The shift, in recent years, from imperative/instructional expressions
towards declarative/assertional ones. This has to do with the
relationship between the behavioural (what happens) vs. the
referential (what it is about) import of computational expressions.
b. The relationship between internal (within the machine) and external
(from machine to the embedding world) semantic relationships.
In both cases, studies of natural languages have focused rather squarely
on one side of the dichotomy, whereas traditional computer languages
have focused on the other (at least I think they have: it isn't trivial to
decide). So they seem good topics to compare and contrast the two
disciplines. Plus, sorting them out seems to me important in order to
understand both LISP and PROLOG.
These are just by way of suggestion; you may have other candidates.
And we need to figure out what it would mean to focus on them, in
concrete terms.
Brian
∂13-Sep-83 1614 HANS@SRI-AI.ARPA CSLI mailing lists, directory, etc.
Received: from SRI-AI by SU-AI with TCP/SMTP; 13 Sep 83 16:13:20 PDT
Date: Tue 13 Sep 83 16:09:07-PDT
From: Hans Uszkoreit <Hans@SRI-AI.ARPA>
Subject: CSLI mailing lists, directory, etc.
To: csli-friends@SRI-AI.ARPA
There are now several mailing lists and distribution addresses for
CSLI people who have access to machines on the ARPANET.
There is, for instance, an address CSLI-FOLKS@SRI-AI. Every message
you send to this address will be forwarded to all people who are
affiliated with CSLI (principals, staff, associates, postdocs, predocs)
and have an account on an ARPANET machine.
So far the following addresses exist:
address: distribution:
CSLI-FOLKS@SRI-AI CSLI affiliates
CSLI-PEOPLE@SRI-AI CSLI affiliates (same as CSLI-FOLKS)
CSLI-FRIENDS@SRI-AI CSLI affiliates + people who will be
invited for colloquia etc.
CSLI-PRINCIPALS@SRI-AI SL principals
CSLI-ADMINISTRATION@SRI-AI CSLI directors and staff
CSLI-BUILDING@SRI-AI CSLI building committee
CSLI-COMPUTING@SRI-AI CSLI computing committee
CSLI-EXECUTIVES@SRI-AI CSLI executive committee and directors
CSLI-REQUESTS@SRI-AI address changes, additions, gripes
Clearly, more addresses of this kind will be needed. If you have a list
of net-addresses that will be used as a distribution list by several
CSLI people, please send it to CSLI-REQUESTS@SRI-AI.
The corresponding address files to the distribution addresses can be
found in the directory <CSLI> at SRI-AI. The distribution file for
the address CSLI-FOLKS is <CSLI>CSLI.FOLKS, the one for CSLI-FRIENDS
is <CSLI>CSLI.FRIENDS, etc.
Mabry has set up a user group for CSLI. All CSLI affiliated users of
the SRI-AI machine can read (and therefore also copy) the files in
<CSLI>. They can also create files in the directory. The directory
should be used for files that will be accessed by several members of
CSLI. Examples are software and initiation files to be shared,
announcements, drafts to be commented on, etc.
It will depend on the individual file protection code whether the other
CSLI people are able to change (or delete) the file or only to read it.
The default protection is 777700, which means that the file can be
changed, deleted, renamed by any member of the user group. There are
reasons for setting the default protection to such a permissive value.
If it doesn't work out, we can still change it. For the time being,
please protect files individually and/or keep copies in your
directories. (You can find out more about protection codes by the
command HELP PROTECTION.)
Thanks, again to Mabry Tyson who was essential in setting up the
distribution addresses and the user group.
Gripes, suggestions, requests to CSLI-REQUESTS@SRI-AI.
-------
∂14-Sep-83 0843 BMACKEN@SRI-AI.ARPA Advisory Panal
Received: from SRI-AI by SU-AI with TCP/SMTP; 14 Sep 83 08:43:11 PDT
Date: Wed 14 Sep 83 08:45:14-PDT
From: BMACKEN@SRI-AI.ARPA
Subject: Advisory Panal
To: BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA, Briansmith@PARC-MAXC.ARPA,
grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA, JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA,
kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA, peters@SRI-AI.ARPA,
rperrault@SRI-AI.ARPA, sag%SU-psych@SU-SCORE.ARPA, stan@SRI-AI.ARPA,
Wasow%SU-psych@SU-SCORE.ARPA, winograd@PARC-MAXC.ARPA, Bmacken@SRI-AI.ARPA
cc: bmacken@SRI-AI.ARPA
Jon would like to agree on the Advisory Panal at the Executive meeting
tomorrow. The following names have been suggested. Do you have any
comments or additions?
Nils Nillson
one of George Pake, Bob Ritchie, J.S. Brown
one person from Bell Labs: Phil Anderson has been suggested as first choice
and Max Mathews as second choice
one person in the semantics of programming languages but who?
Jerry Fodor of MIT
Dan Dennet
J. Keyser
George Miller
Thanks.
B.
-------
∂14-Sep-83 1044 AAAI-OFFICE@SUMEX-AIM 1984 Conference Facilities
Received: from SUMEX-AIM by SU-AI with PUP; 14-Sep-83 10:43 PDT
Date: Wed 14 Sep 83 10:45:23-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM>
Subject: 1984 Conference Facilities
To: Tenenbaum@SRI-KL.ARPA
cc: Brachman@SRI-KL.ARPA, Walker@SRI-AI.ARPA, Reddy@CMU-CS-A.ARPA, JMC@SAIL
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
Marty,
I'd like to review the situation of the meeting rooms at UT. We have the
following rooms reserved:
ROOM LOCATON SIZE USE
---------------------------------------------------------------------------
Concert Hall Performing Art Ctr 3,000 Technical sessions
(this room is divided into three levels: orchestra = 1747 seats;1st balcony =
659 seats; 2nd balcony = 623 seats)
Opera Lab Theatre PAC 400 technical sessions/
tutorials
Bates Recital Hall Music Bldg. 700 "
Recital Studio Music Bldg. 194 tutorials
Rehersal Rm. PAC 200 tutorials
CH Green Room PAC 70 tutorials
LBJ Auditorium 1000 technical sessions
(1 block away)
OLT Green Rm PAC 35 Press Room
Lobby PAC Conference registration
It is not clear yet what space will be used for the exhibit program-UT has
suggested using the back stage of the Concert Hall for the exhibits. They
plan to install sound baffling behind the Curtain. It is not clear how
people are expected to get to this area (ie is there a back door entrance?).
How it will possibly interfere with the presentations/panels of the
technical program? The other option is Irwin Center which is several
blocks away. As you can see, neither option is very attractive.
Let me also express another reservation about the rooms. Presently, there
is an expectation that the entire Concert Hall (+3000 seats) will be used.
Do we expect people to be sitting in the first and second balconies to see
and listen to the presentations? It's not clear yet if people can even see any
projection from those two levels (should we tell people to take their opera
glasses with them to the conference?)Also, we have no idea if the people in the
back of the orchestra level can see any projection from the stage.
Other considerations: the seats in the PAC rooms do not have any writing
area-this year people complained that they had no area to write down
their notes. Also, we may be able to use classrooms in the Performing
Arts Center; however, the Local Arrangements Chair, has no idea what the
sizes of the classrooms are in that building.
Because of these limitations, I like to discuss with you the option of limiting
the size of the technical part of the conference to 2,000 people. Now, this
proposal may be a non-issue, because it is expected that fewer people will
attend next year's conference. If we do proceed with this recommendation,
we should have statement, such as "Space is limited-register early," mentioned
in the conference brochure frequently.
I look forward to hearing your thoughts on this matter soon.
Regards,
Claudia
-------
∂14-Sep-83 1101 KJB@SRI-AI.ARPA Good news
Received: from SRI-AI by SU-AI with TCP/SMTP; 14 Sep 83 11:01:02 PDT
Date: Wed 14 Sep 83 11:02:12-PDT
From: KJB@SRI-AI.ARPA
Subject: Good news
To: csli-folks@SRI-AI.ARPA
The executive committee met with Charlie Smith today about the
proposal which went before the SDF board yesterday for compu}iting
support and short range building}i help. IT was approved! This
is great news. Now we can all get down to real work.
Jon
-------
∂14-Sep-83 1226 DOYLE@CMU-CS-C.ARPA Re: Long-term AI research
Received: from CMU-CS-C by SU-AI with TCP/SMTP; 14 Sep 83 12:25:46 PDT
Received: ID <DOYLE@CMU-CS-C.ARPA>; Wed 14 Sep 83 15:25:05-EDT
Date: Wed 14 Sep 83 15:25:00-EDT
From: DOYLE@CMU-CS-C.ARPA
Subject: Re: Long-term AI research
To: Laws@SRI-AI.ARPA
cc: McCarthy@SU-AI.ARPA, Nilsson@SRI-AI.ARPA, Minsky%MIT-OZ@MIT-MC.ARPA,
Newell@CMU-CS-A.ARPA
In-Reply-To: Message from "Ken Laws <Laws@SRI-AI.ARPA>" of Mon 12 Sep 83 15:52:09-EDT
Dear Mr. Laws,
While I agree that engineering journals and conferences are needed
and will appear, I do not believe that that will address the problem
I raised in my message, namely the visibility of theoretical and long
term work. There needs to be an outlet that is neutral in orientation,
at least to serve those people who disagree with their editors about
whether their papers are suitably theoretical, applied, etc. That is
the role of the neutrally-named AI Journal, AI Magazine, AAAI, and IJCAI.
But the very neutrality of those titles offers no recognition of theoretical
work, no matter what the less visible editorial policy may be. I do not
see your suggestion doing more than slightly reducing the proportions of
papers involved---without addressing at all the recognition problem.
Please do not post my message to the bulletin boards. I do not wish
to receive volumes of mail on this topic, unless it is from the people I
addressed the letter to. My purpose in asking them to forward it to others
(which, I assume, is how you happened to see it) was to offer it to the
appropriate group of people for discussion. I have absolutely no interest
in being a moderator, and am not myself involved in the direction of the
AAAI, IJCAI, AI Journal, etc.
But thank you for your thoughts.
Jon Doyle
-------
∂15-Sep-83 1121 Winograd.PA@PARC-MAXC.ARPA Re: Advisory Panal
Received: from PARC-MAXC by SU-AI with TCP/SMTP; 15 Sep 83 11:17:59 PDT
Date: 15 Sep 83 11:07 PDT (Thursday)
From: Winograd.PA@PARC-MAXC.ARPA
Subject: Re: Advisory Panal
In-reply-to: BMACKEN@SRI-AI.ARPA's message of Wed, 14 Sep 83 08:45:14
PDT
To: BMACKEN@SRI-AI.ARPA
cc: BMoore@SRI-AI.ARPA, Bresnan.PA@PARC-MAXC.ARPA,
Briansmith.PA@PARC-MAXC.ARPA, grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA,
JRP@SRI-AI.ARPA, Kaplan.PA@PARC-MAXC.ARPA, kjb@SRI-AI.ARPA,
lauri@SRI-AI.ARPA, peters@SRI-AI.ARPA, rperrault@SRI-AI.ARPA,
sag%SU-psych@SU-SCORE.ARPA, stan@SRI-AI.ARPA,
Wasow%SU-psych@SU-SCORE.ARPA, winograd.PA@PARC-MAXC.ARPA
The obvious person on programming languages is Dana Scott
I am a little worried about Fodor, Dennett, and Miller (unless that is an OR, not
an AND) because having all of them would be a very heavy weight towards the
cognitive psychology end of things, a direction that we very consciously have
not slanted the center. I think we'd be better off with more philosophers of
language (I'll let Jon and John recommend) instead. If they are really to serve
as advisors, we don't want them pushing in a crossways direction.
--t
∂15-Sep-83 1334 ELYSE@SU-SCORE.ARPA fac. int.
Received: from SU-SCORE by SU-AI with TCP/SMTP; 15 Sep 83 13:33:55 PDT
Date: Thu 15 Sep 83 13:35:11-PDT
From: Elyse Krupnick <ELYSE@SU-SCORE.ARPA>
Subject: fac. int.
To: McCarthy@SU-SCORE.ARPA
Stanford-Phone: (415) 497-9746
AI, math theory of comp. computing with symbolic expressions, time-sharing.
-------
∂15-Sep-83 1455 @SU-SCORE.ARPA:TRATTNIG@SU-SIERRA.ARPA Special Supercomputer Seminar
Received: from SU-SCORE by SU-AI with TCP/SMTP; 15 Sep 83 14:55:04 PDT
Received: from SU-SIERRA.ARPA by SU-SCORE.ARPA with TCP; Thu 15 Sep 83 14:52:47-PDT
Date: Thu 15 Sep 83 14:52:40-PDT
From: Werner Trattnig <TRATTNIG@SU-SIERRA.ARPA>
Subject: Special Supercomputer Seminar
To: super@SU-SCORE.ARPA, luckham@SU-SIERRA.ARPA, owicki@SU-SIERRA.ARPA
SPECIAL SUPERCOMPUTER SEMINAR
Friday, September 16, 3 p.m.
Stanford University
Room ERL 232
CIS Conference Room
DATA FLOW CONCEPTS IN PROGRAMMING LANGUAGE AND COMPUTER ARCHITECTURE
Jack B. Dennis
Laboratory of Computer Science
M.I.T.
-------
∂15-Sep-83 1651 BUCHANAN@SUMEX-AIM Comtex Scientific
Received: from SUMEX-AIM by SU-AI with PUP; 15-Sep-83 16:51 PDT
Date: Thu 15 Sep 83 16:54:29-PDT
From: Bruce Buchanan <BUCHANAN@SUMEX-AIM>
Subject: Comtex Scientific
To: jmc@SAIL
cc: genesereth@Score
John,
Dr. Huang from Comtex will be here next MOnday to talk with Mike Genesereth,
Claudia Mazetti and me about collaborating with the AAAI in offering an
online search service of titles & abstracts immediately, and of full text
eventually. We are planning to meet at 1:30 Monday. Do you want to join
us?
Dr. Huang also would like to talk about a contract between Stanford
and Comtex to cover the SAIL memos. When you and I discussed this briefly
at AAAI, you asked me to try to work out details. Mostly they seem fine to
me, but I want to talk with her more. Do you want royalties to go to
(a) a fund for your research
(b) the CSD
(c) the CS/Math library?
(I include the latter because they did the work in collecting the materials,
and we could begin to "endow" an AI book fund.)
bgb
-------
∂15-Sep-83 1714 @SRI-AI.ARPA:PULLUM%HP-HULK.HP-Labs@Rand-Relay West Coast Conference on Formal Linguistics, 1984
Received: from SRI-AI by SU-AI with TCP/SMTP; 15 Sep 83 17:14:31 PDT
Received: from rand-relay.ARPA by SRI-AI.ARPA with TCP; Thu 15 Sep 83 17:14:57-PDT
Date: 15 Sep 1983 1457-PDT
From: PULLUM.HP-HULK@Rand-Relay
Return-Path: <PULLUM%HP-HULK.HP-Labs@Rand-Relay>
Subject: West Coast Conference on Formal Linguistics, 1984
Received: by HP-VENUS via CHAOSNET; 15 Sep 1983 14:56:28-PDT
To: CSLI-FRIENDS@SRI-AI
Cc: gpsg-dist..HP-LABS@Rand-Relay, PULLUM%;.HP-LABS@Rand-Relay
Message-Id: <432510989.1703.hplabs@HP-VENUS>
Via: HP-Labs; 15 Sep 83 16:31-PDT
The third annual West Coast Conference on Formal Linguistics (WCCFL 3) has now
been scheduled for Friday March 16 - Sunday March 18, 1984, at the University
of California, Santa Cruz. This conference is the premier west coast forum
for work relating to generative grammar (including phonology) and formal
semantics. A call for abstracts will go out by mail soon. Anyone interested
in being on the mailing list should write to WCCFL 3, Syntax Research Center,
Cowell College, UCSC, Santa Cruz, CA 95064. Also watch this space: I will put
further announcements on the net as and when further arrangements are made.
--Geoff Pullum
-------
∂16-Sep-83 0806 KJB@SRI-AI.ARPA CSLI annoncements
Received: from SRI-AI by SU-AI with TCP/SMTP; 16 Sep 83 08:06:34 PDT
Date: Fri 16 Sep 83 08:03:55-PDT
From: KJB@SRI-AI.ARPA
Subject: CSLI annoncements
To: csli-folks@SRI-AI.ARPA
Dear all,
I have gotten a lot of complaints about announcments not getting to
people who do not have electronic mailboxes. (Hmm, this is a self-
referential message, I fear.) We have to be careful not to exclude
people who do not have such accounts. I propose that we do two things:
a) Make all announcments of events AT LEAST a week in advance.
b) Make sure that a CSLI secretary collects hard copies of
these and mails them out to those on the approprate mailing list that
do not have electronic mail.
Now campus mail is extremely slow here, which is why we need to make
sure that we give at least a weeks time for announcements. Even then,
people are likely to receive the notice only two days before the
event, which is not really enough time for planning.
Of course all this will be better when we have our opwn computer network
and can make sure that more people have access to accounts.
A footnote: Right now, Pat Wunderman is the only secretary here, and is
terrrible overworked. We are in the process of hiring more secretaries,
and when we do, Pat will not be the one to do this sort of thing.
Jon
-------
∂16-Sep-83 0854 KJB@SRI-AI.ARPA More on csli announcements
Received: from SRI-AI by SU-AI with TCP/SMTP; 16 Sep 83 08:54:05 PDT
Date: Fri 16 Sep 83 08:50:24-PDT
From: KJB@SRI-AI.ARPA
Subject: More on csli announcements
To: csli-folks@SRI-AI.ARPA
Pat Wunderman tells me that the printer here (an ancient diablo)
is not working much of the time these days. Thus, we should
all make sure a hard copy of all announcment gets to her until
further notice.
In the long run, we would like to be able to publish a monthly
calendar of CSLI events. But this seems impractical for the
first few months, both because people do not have project and
area events set up that far in advance yet, and because we do
not have the staff to do it here at Ventura. Maybe we can get
this going by January 1.
Susan and Hans did a great job with the Sept 1 event. Now we
need a couple of people to plan something for early in October,
a sort of official function to which we invite officals of SU,
SRI Xerox, etc. at the official opening of the Center. I hope
two or three of you will volunteer to Betsy to set this up. Otherwise
she will have to draft people, and might not draft people who know
how to do something like this with style.
-------
∂16-Sep-83 0937 KJB@SRI-AI.ARPA Computer language people
Received: from SRI-AI by SU-AI with TCP/SMTP; 16 Sep 83 09:37:27 PDT
Date: Fri 16 Sep 83 09:35:08-PDT
From: KJB@SRI-AI.ARPA
Subject: Computer language people
To: jmc@SU-AI.ARPA
cc: briansmith@PARC-MAXC.ARPA, winograd@PARC-MAXC.ARPA
John, It is important for CSLI to get more computer language people
involved. It has occurred to me that Carolyn has not been around, yet.
I would think she would be a natural for a lot of our activities. And
ther must be others that I don't know. Would you think who they might
be and let me know their addresses so that I can get them on the
mailing lists? Jon
-------
∂16-Sep-83 1024 AAAI-OFFICE@SUMEX-AIM Presidential address
Received: from SUMEX-AIM by SU-AI with PUP; 16-Sep-83 10:24 PDT
Date: Fri 16 Sep 83 10:27:27-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM>
Subject: Presidential address
To: JMC@SAIL
cc: aaai-office@SUMEX-AIM
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
John,
We will need your Presidential address by Oct 7 for the
upcoming issue of the AI Magazine.
Thanks,
Claudia
-------
∂16-Sep-83 1052 JRP@SRI-AI.ARPA fall educational program
Received: from SRI-AI by SU-AI with TCP/SMTP; 16 Sep 83 10:51:49 PDT
Date: 16 Sep 1983 1052-PDT
From: Jrp at SRI-AI
Subject: fall educational program
To: csli-folks at SRI-AI
In addition to our Thursday events, a number of departmental offerings
are pretty directly relevant to the csli program. There are roughly
two categories here, courses that provide the necessary background in
relevant subjects, and advanced courses and seminars. An example of
the latter is the Suppe-Almog seminar on context. I'd like to get
out a notice of all things in the latter category, so please send me
messages concerning such offerings in your departments. In the future,
of course, the program will be less haphazard, but we dont want to wait
until then for everyone to know what is going on.
-------
∂16-Sep-83 1132 @SRI-AI.ARPA:BRESNAN.PA@PARC-MAXC.ARPA syntactic constraints on discourse
Received: from SRI-AI by SU-AI with TCP/SMTP; 16 Sep 83 11:29:49 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Fri 16 Sep 83 10:26:16-PDT
Date: Fri, 16 Sep 83 10:19 PDT
From: BRESNAN.PA@PARC-MAXC.ARPA
Subject: syntactic constraints on discourse
To: csli-people@sri-ai.ARPA, lauri@sri-ai.ARPA
cc: BRESNAN.PA@PARC-MAXC.ARPA
I would like to have a preliminary planning session of all those who
want to work on CSLI project A.2: "Syntactic and Phonological
Constraints on Discourse". It appears that there are two subgroups of
this project--one focussing initially on the phonology of intonation,
the other focussing initially on syntax and discourse. I've asked Meg
Withgott to organize the phonology subgroup, and I will be organizing
the syntax subgroup. I've already spoken with Marilyn Ford, Jane
Simpson (now in Australia), Hans Utzkoreit, and Susan Stucky, who will
all be part of this project. We expect that Lauri Karttunen will be
part of this group when he returns later this academic year.
Among the topics of interest to this group are the following:
- How are discourse functions such as TOPIC, FOCUS syntactically
encoded?
- How does discourse structure, and the planning of spontaneous speech,
vary as syntactic structure varies from highly configurational (English)
to highly nonconfigurational (Warlpiri)?
- What can syntactic phenomena (e.g. anaphora, logophoricity) reveal
about the structure of discourse representations?
This message is addressed to csli-people so that everyone who would like
to work on this project will contact me. After I've found out who is
interested, I will try to arrange a mutually agreeable time for us to
start up.
Joan Bresnan
∂16-Sep-83 2044 Brachman@SRI-KL.ARPA AAAI-84 Technical Program
Received: from SRI-KL by SU-AI with TCP/SMTP; 16 Sep 83 20:43:36 PDT
Date: Fri 16 Sep 83 15:44:34-PDT
From: Ron Brachman <Brachman at SRI-KL at SU-DSN>
Subject: AAAI-84 Technical Program
To: Amarel at RUTGERS at SU-DSN, Bledsoe at UTEXAS-20 at SU-DSN,
Brachman at SRI-KL at SU-DSN, Buchanan at SUMEX-AIM, KRD at MIT-MC at SU-DSN,
BEngelmore at SRI-KL at SU-DSN, LErman at SRI-KL at SU-DSN,
CSD.Genesereth at SU-SCORE, Grosz at SRI-AI at SU-DSN, Hart at SRI-KL at SU-DSN,
JMC at SU-AI, Minsky at MIT-MC at SU-DSN, Nilsson at SRI-AI at SU-DSN,
Reddy at CMU-CS-A at SU-DSN, Stan at SRI-AI at SU-DSN, Stefik at PARC-MAXC at SU-DSN,
GJS at MIT-MC at SU-DSN, Tenenbaum at SRI-AI at SU-DSN, Walker at SRI-AI at SU-DSN,
Bonnie.UPenn at UDEL-RELAY at SU-DSN
cc: AAAI-Office at SUMEX-AIM
September 16, 1983
Dear AAAI Councilor,
Now that this year's conference is safely and successfully behind us,
it's time to start thinking seriously about AAAI-84. As you know, I
have (perhaps naively) agreed to be Program Chairman for next year's
conference. From the sound of it, there is quite a bit of work to do,
but surely getting started early will be of some help.
To that end, I would like to solicit your opinion on a number of issues
pertaining to the conference. Below I will list a set of items that
need decisions, many in the very near future. Please feel free to offer
your feelings on any or all of these matters. I cannot promise that I
will abide by those feelings on each issue, but I think it is very
important at least to get a sense of the organization before any
important decisions are taken, and I think that this is best achieved
through the Executive Council. Also, at any time during the course of
the year, please feel free to get in touch with me if you have any
thoughts pertaining to the technical program; I would especially like to
hear about interesting panel, debate, and other plenary session ideas
you might have.
I can be reached easily by ARPANET message or letter (look inside the
front cover of the SIGART Newsletter for my addresses), or by phone
(415-858-4273).
Thanks very much for your help,
Ron Brachman
!
Philosophy of the conference:
Much ado has been made about the emphasis of the conference -
should it be a scientific gathering, or should the conference be an
attempt to promote AI in the commercial world? I have a strong personal
opinion on the matter, but am willing to listen to reasonable
suggestions. My feeling is that Mike Genesereth took exactly the right
approach last year - the conference highlighted currently
breaking research, and attempted to promote research in the
field. Unless convinced that there is something wrong with that
emphasis, I will do my best to see it continue with AAAI-84.
Length of the conference:
With no competing IJCAI this year, the number of submissions to
AAAI may be overwhelming. There may consequently be strong incentive to
lengthen the conference, in order to avoid too many parallel sessions.
I could imagine at most one more day (i.e., include at least part of
Tuesday for technical sessions). However, unless there is very strong
support in favor of lengthening the technical program, I would try to
keep it to the three days (overlapping the Tutorials would probably be a
mistake).
Number of parallel sessions:
I have heard opinions both ways on this issue. Personally, I
would like to keep the number of papers to a minimum, without excluding
any top-quality work. Mike felt that the three parallel sessions this
year worked out well.
Presentations:
How do you feel about the length of time allotted to
presentations? Further, we have had a perennial problem of bad
presentations, and any advice you can give on getting presenters to
improve their public speaking would be greatly appreciated. Another
issue is the visibility of transparencies - with the increasing
conference attendance, the rooms for talks are larger, and viewgraphs
have traditionally been poor. Should we limit visual aids to 35mm
slides? Finally, do you like the idea of having a session chairperson?
Should that person be a mainstream AI person, or a volunteer who simply
keeps time and keeps a very low profile?
Call for papers:
The stringent length requirement for AAAI papers has been a bit
of a problem for some writers, who try to cram complete system
descriptions into ten pages. However, having served on the Program
Committee, I have found that the only thing making quality control
possible is this length limitation. Is there strong support for
changing the length of submitted papers?
Occasionally the set of categories offered for paper submission
presents a problem for authors - they cannot decide which of two
categories to submit to, or find that none of the stated ones are
appropriate. I would happily entertain suggestions for improving this
part of the submission system or for a reasonable set of categories. My
current plan is to make some modifications to last year's set, and to
add a bit of explanation for categories that are potentially
confusing. One further suggestion that has been made is that, in
addition to the topic of the paper, "applied" or "theoretical" be added.
Mike did not do this last year because that split was not appropriate
for all categories. Is it worth thinking about?
Finally, we have the multiple authorship problem. We certainly
want to keep the number of submissions manageable, but we do not want to
exclude excellent papers for arbitrary reasons. One extension
to the requirement has been suggested: allow a given person to be FIRST
author on only one paper. What is your opinion of this or other
suggestions?
Ancillary events:
Were the pre-conference workshops a good idea? If so, can you
suggest a set (along with a set of organizers) for this year? What is
your opinion of the trade show and the technology transfer symposium?
Unless we get a strong opposing sentiment, we plan to leave the trade
show more or less as is, and plan NOT to continue the technology
transfer symposium.
Concrete suggestions for panels and other plenary sessions:
We need concrete ideas soon - it is becoming clear that these
sessions must be organized well in advance of the Program Committee
meeting. Please help us out with your ideas, and suggestions of people
to be responsible.
Improvements to poster, conference brochure, program booklet:
Most people seemed to like the conference announcement brochure
this year; the poster got mixed reviews. Claudia Mazzetti has a sample
for a national education conference that she would like to use as a
model for next year's program. It looked extremely nice to me, but any
additional suggestions would be greatly appreciated.
How important is it to have rooms with sessions close together?
...how close? This seems to many to be critically important.
If you find it as important, it will in part dictate the size and
structure of the program, because of the somewhat limited facilities at
Texas.
Also, how important is it to have writing surfaces available
with the seats in the lecture rooms?
And certainly any thoughts on other issues pertaining to the program are
welcome... Thank you in advance for your help.
-------
∂18-Sep-83 1226 DIETTERICH@SUMEX-AIM AAAI Post-doctoral fellowships
Received: from SUMEX-AIM by SU-AI with PUP; 18-Sep-83 12:26 PDT
Date: Sun 18 Sep 83 12:29:43-PDT
From: Tom Dietterich <DIETTERICH@SUMEX-AIM>
Subject: AAAI Post-doctoral fellowships
To: jmc@SAIL
cc: aaai-ofFICE@SUMEX-AIM, DIETTERICH@SUMEX-AIM, buchanan@SUMEX-AIM
I am expecting to graduate next June, and I would like to find some
source of funding for postdoctoral research starting in September of
1984. Bruce tells me that you are coordinating an effort by the AAAI
to establish some post-doctoral fellowships. He suggested that I send
a message to you.
I need a fellowship that will enable me to follow up on my
dissertation research concerning the role of experimentation in
learning and theory formation. It looks like I'll be located at
Oregon State University in Corvallis (because my future spouse has
accepted a position there as an assistant professor), so this would
need to be a fellowship that is tenable at any institution. Hence, my
questions to you are: Do you expect that the AAAI fellowships will
be available by Sept. of 1984? Would they support me at Oregon State?
Has the application and review procedure been worked out?
Thanks very much for your help,
--Tom
-------
∂19-Sep-83 0004 POURNE@MIT-MC Great White Cabin
Received: from MIT-MC by SU-AI with TCP/SMTP; 19 Sep 83 00:04:19 PDT
Date: 19 September 1983 03:06 EDT
From: Jerry E. Pournelle <POURNE @ MIT-MC>
Subject: Great White Cabin
To: POURNELLE @ MIT-MC, LLW @ SU-AI, jmc @ SU-AI
One of our troops has appointment for week of 2x Sept, and will
be carryng report. Looks good to get a few pages into good
places.
Any comments or revisions of t he papers I sent around last m
onth?
Anything new you want in report?
Hans liked the stuff well enough.
llw is there any chance of getting a revised edition of the
Threat Assessment Report of '81?
∂19-Sep-83 0907 AAAI-OFFICE@SUMEX-AIM post-doctoral fellowships
Received: from SUMEX-AIM by SU-AI with PUP; 19-Sep-83 09:07 PDT
Date: Mon 19 Sep 83 09:10:26-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM>
Subject: post-doctoral fellowships
To: JMC@SAIL
cc: aaai-office@SUMEX-AIM
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
John,
It's not clear that the Exec Council made a final decision
about the post-doctoral fellowships - number per year, review
procedures, etc. Should we send a msg to EXECOM re this recommendation?
--Claudia
-------
∂19-Sep-83 1025 RPH BS
I am interested in completing my BS. I will write to the registrar to find
out what the requirements would be at this time, but I would like to have
an advisor (if that's the right thing). Would you be willing to help me?
If so, let me know. You can call me at 303-594-2009 or 303-495-4954. Or
you can get Bosack to figure out the network route (good luck).
∂19-Sep-83 1128 WALKER@SRI-AI.ARPA [Derek Sleeman <SLEEMAN@SUMEX-AIM.ARPA>: AI Research Notes series]
Received: from SRI-AI by SU-AI with TCP/SMTP; 19 Sep 83 11:27:56 PDT
Date: Mon 19 Sep 83 11:21:56-PDT
From: Don Walker <WALKER@SRI-AI.ARPA>
Subject: [Derek Sleeman <SLEEMAN@SUMEX-AIM.ARPA>: AI Research Notes series]
To: rperrault@SRI-AI.ARPA, evens@SRI-AI.ARPA, mann@USC-ISIB.ARPA,
lauri@UTEXAS-20.ARPA, weischedel@UDEL-RELAY.ARPA, acl@ROCHESTER.ARPA,
joshi.upenn@UDEL-RELAY.ARPA, bonnie.upenn@UDEL-RELAY.ARPA,
sondheimer@USC-ISIF.ARPA, jrobinson@SRI-AI.ARPA, amarel@RUTGERS.ARPA,
bledsoe@UTEXAS-20.ARPA, brachman@SRI-KL.ARPA, buchanan@SUMEX-AIM.ARPA,
krd@MIT-MC.ARPA, bengelmore@SRI-KL.ARPA, lerman@SRI-KL.ARPA,
csd.genesereth@SU-SCORE.ARPA, grosz@SRI-AI.ARPA, hart@SRI-KL.ARPA,
jmc@SU-AI.ARPA, nilsson@SRI-AI.ARPA, reddy@CMU-CS-A.ARPA, stan@SRI-AI.ARPA,
stefik@PARC-MAXC.ARPA, gjs@MIT-MC.ARPA, tenenbaum@SRI-KL.ARPA,
walker@SRI-AI.ARPA, bonnie.upenn@UDEL-RELAY.ARPA,
aaai-office@SUMEX-AIM.ARPA, uw-beaver!ubc-vision!mack@LBL-CSAM.ARPA,
ubc-vision!mack@UW-BEAVER.ARPA, joshi.upenn@UDEL-RELAY.ARPA,
amarel@RUTGERS.ARPA, hayes@ROCHESTER.ARPA, raj.reddy@CMU-CS-A.ARPA
This message is being sent to the relevant people in ACL, AAAI, and IJCAI.
As noted, ACL has recently set up its own monograph series with Cambridge
University Press. It may be reasonable for AAAI and IJCAI to consider
working out similar arrangements, perhaps with Kaufmann. Reactions???
Don
---------------
1) 18-Sep Derek Sleeman AI Research Notes series
2) 19-Sep To: SLEEMAN@SUMEX-AI Re: AI Research Notes series
Message 1 -- ************************
Return-Path: <@SRI-KL.ARPA:SLEEMAN@SUMEX-AIM.ARPA>
Received: from SRI-KL.ARPA by SRI-AI.ARPA with TCP; Sun 18 Sep 83 15:10:23-PDT
Received: from SUMEX-AIM.ARPA by SRI-KL.ARPA with TCP; Sun 18 Sep 83 15:10:15-PDT
Date: Sun 18 Sep 83 15:10:36-PDT
From: Derek Sleeman <SLEEMAN@SUMEX-AIM.ARPA>
Subject: AI Research Notes series
To: walker@SRI-KL.ARPA
cc: sleeman@SUMEX-AIM.ARPA, sridharan@RUTGERS.ARPA
As you may have heard Pitman's will shortly start publishing an AI Research Notes
series. Sridharan and I are currently the main editors. The plan is to publish:
a) monographs b) collections of companion papers c) polished versions of theses.
Under b), we include selected, and usually enhanced, papers from conferences.
I would like to discuss with you the possibility of doing just that for IJCAI,
AAAI and ACL conferences.
I have put several copies of the prospectus in the mail to you at SRI, and hope
we can get in touch shortly,
Derek Sleeman
[Now physically at Stanford]
-------
Message 2 -- ************************
Date: Mon 19 Sep 83 10:27:23-PDT
From: Don Walker <WALKER@SRI-AI>
Subject: Re: AI Research Notes series
To: SLEEMAN@SUMEX-AIM
cc: sridharan@RUTGERS
In-Reply-To: Your message of Sun 18 Sep 83 15:10:26-PDT
Derek,
Thanks for the information about your new series. In direct
competition with you is the new series ACL has established with
Cambridge University Press, with the title "Studies in Natural Language
Processing." However, AAAI and IJCAI are less well organized, although
some discussions have taken place with William Kaufmann, Inc., the
publisher that handles distribution for proceedings from both organizations.
I will forward your message to the appropriate people in both AAAI and
IJCAI; I will also let people in ACL know of your plans. It is likely
that there is more than enough material to satisfy a range of publication
strategies. One policy position likely to be taken by everyone is that
if a substantial portion of the papers for a volume come from proceedings,
there should be some financial reimbursement to the organization. I'll
be glad to discuss these matters further following some feedback.
In any case, good luck,
Don
-------
-------
∂19-Sep-83 1146 DFH Re: space
∂13-Sep-83 1212 YEARWOOD@SU-SCORE.ARPA Re: space
Received: from SU-SCORE by SU-AI with TCP/SMTP; 13 Sep 83 12:08:49 PDT
Date: Tue 13 Sep 83 11:09:33-PDT
From: Marlene Yearwood <YEARWOOD@SU-SCORE.ARPA>
Subject: Re: space
To: DFH@SU-AI.ARPA
In-Reply-To: Message from "Diana Hall <DFH@SU-AI>" of Tue 6 Sep 83 11:15:00-PDT
Stanford-Phone: (415) 497-2266
We have assigned a visiting professor, Allen Tucker,
to that office for autumn quarter. I will let you know if
it is available after that.
marlie
-------
The above is in response to my query to Marlie about getting the other half
of Creary/Wieneke office. -- Diana.
∂19-Sep-83 1501 @MIT-MC:DAVIS%MIT-OZ@MIT-MC Re: [John McCarthy <JMC@SU-AI>:]
Received: from MIT-MC by SU-AI with TCP/SMTP; 19 Sep 83 15:01:33 PDT
Date: 19 Sep 1983 1758-EDT
From: Randy Davis <DAVIS%MIT-OZ@MIT-MC>
Subject: Re: [John McCarthy <JMC@SU-AI>:]
To: AAAI-OFFICE@SUMEX-AIM
cc: AMAREL@RUTGERS, BUCHANAN@SUMEX-AIM, BENGELMORE@SRI-KL, LERMAN@SRI-KL,
GENESERETH@SUMEX-AIM, GROSZ@SRI-AI, HART@SRI-KL, JMC@SU-AI, MINSKY@MIT-MC,
NILSSON@SRI-AI, REDDY@CMU-CS-A, RICH@MIT-MC, STAN@SRI-AI, GJS@MIT-MC,
TENENBAUM@SRI-KL, WALKER@SRI-AI, BONNIE.UPENN@UDEL-RELAY, DWALTZ@BBNG,
BOBROW@PARC-MAXC, MCDERMOTT@YALE, STEFIK@PARC-MAXC, KRD@MIT-MC,
feigenbaum@SUMEX-AIM
In-Reply-To: Your message of 7-Sep-83 2005-EDT
Danny: Msg received this time.
All: As some of you have no doubt also discovered, the Office of
Technology Assessment (in the person of Earl Dowdy) has been trying
to organize a 1-day conference on AI so that they can advise various
Senate and House committees on how important the technology is, what
to do to promote it, etc. etc.
I mentioned the possibility of EF's meeting as a way to minimize the
"oh god yet another hi level meeting" syndrome. He seemed interested
but (I'm reading in here) still seemed to want to run his 1-day show.
In any case, I think we might want to take advantage of this
opportunity to get some impact, if (as Danny suggests) we can come up
with sharp issues and suggestions.
Dowdy wants to run his thing Oct 17, so there ain't a whole lot of
time. Perhaps better to let this one go by than to make a hash of it,
but if we can get our act together in time this might be a good
high-level route to make the suggestions effective.
Is there someone with political experience who knows whether OTA is
indeed an effective voice? Has anyone else suggested to Dowdy that
the meetings be combined? Is there agreement that this might be
a good idea?
Randy
-------
If we decide the Fredkin meeting is appropriate, we shouldn't combine
it with Dowdy's show. We can suppose that Dowdy already knows what
he wants to come out of it, and it may not co-incide with what we would
want.
∂20-Sep-83 0707 PETTY@RUTGERS.ARPA 1983 abstract mailing
Received: from RUTGERS by SU-AI with TCP/SMTP; 20 Sep 83 07:07:09 PDT
Date: 20 Sep 83 10:05:57 EDT
From: PETTY@RUTGERS.ARPA
Subject: 1983 abstract mailing
To: Lerman@SRI-KL.ARPA, SHORTLIFFE@SUMEX-AIM.ARPA, DREIFUS@WHARTON-10.ARPA,
bennett@SU-SCORE.ARPA, MITTAL@PARC-MAXC.ARPA, CHANDRASEKARAN@RUTGERS.ARPA,
JSMITH@RUTGERS.ARPA, WILKINS@SRI-KL.ARPA, BRUCE@BBNA.ARPA,
FRIEDLAND@SUMEX-AIM.ARPA, PLONDON@USC-ISIB.ARPA, RDG@SU-AI.ARPA,
PRESSBURGER@KESTREL.ARPA, CSD.GARDNER@SU-SCORE.ARPA, FAGAN@SUMEX-AIM.ARPA,
FIKES@PARC-MAXC.ARPA, JMC@SU-AI.ARPA, CLANCEY@SUMEX-AIM.ARPA,
KRD%MIT-OZ@MIT-ML.ARPA, HAMILTON.ES@PARC-MAXC.ARPA, AMSLER@SRI-AI.ARPA,
ChiNguyen.ES@PARC-MAXC.ARPA, lisa@UTEXAS-11.ARPA, kwh%MIT-OZ@MIT-ML.ARPA,
UTGOFF@RUTGERS.ARPA, TUROCK@RUTGERS.ARPA, ECG.RICH@DEC-MARLBORO.ARPA
cc: petty@RUTGERS.ARPA, loungo@RUTGERS.ARPA
Below is a list of our newest technical reports.
The abstracts for these are available for access via FTP with user account
<anonymous> with any password. The file name is:
<library>tecrpts-online.doc
If you wish to order copies of any of these reports please send mail via the
ARPANET to LOUNGO@RUTGERS or PETTY@RUTGERS. Thank you!!
[ ] CBM-TR-128 EVOLUTION OF A PLAN GENERATION SYSTEM, N.S. Sridharan, J.L.
Bresina and C.F. Schmidt.
[ ] CBM-TR-133 KNOWLEDGE STRUCTURES FOR A MODULAR PLANNING SYSTEM, N.S.
Sridharan and J.L. Bresina.
[ ] CBM-TR-134 A MECHANISM FOR THE MANAGEMENT OF PARTIAL AND INDEFINITE
DESCRIPTIONS, N.S. Sridharan and J.L. Bresina.
[ ] DCS-TR-126 HEURISTICS FOR FINDING A MAXIMUM NUMBER OF DISJOINT BOUNDED
BATHS, D. Ronen and Y. Perl.
[ ] DCS-TR-127 THE BALANCED SORTING NETWORK,M. Dowd, Y. Perl, L. Rudolph and
M. Saks.
[ ] DCS-TR-128 SOLVING THE GENERAL CONSISTENT LABELING (OR CONSTRAINT
SATISFACTION) PROBLEM: TWO ALGORITHMS AND THEIR EXPECTED
COMPLEXITIES, B. Nudel.
[ ] DCS-TR-129 FOURIER METHODS IN COMPUTATIONAL FLUID AND FIELD DYNAMICS,
R. Vichnevetsky.
[ ] DCS-TR-130 DESIGN AND ANALYSIS OF PROTECTION SCHEMES BASED ON THE
SEND-RECEIVE TRANSPORT MECHANISM, (Thesis) R.S. Sandhu. (If
you wish to order this thesis, a pre-payment of $15.00 is
required.)
[ ] DCS-TR-131 INCREMENTAL DATA FLOW ANALYSIS ALGORITHMS, M.C. Paull and B.G.
Ryder.
[ ] DCS-TR-132 HIGH ORDER NUMERICAL SOMMERFELD BOUNDARY CONDITIONS: THEORY AND
EXPERIMENTS, R. Vichnevetsky and E.C. Pariser.
[ ] LCSR-TR-43 NUMERICAL METHODS FOR BASIC SOLUTIONS OF GENERALIZED FLOW
NETWORKS, M. Grigoriadis and T. Hsu.
[ ] LCSR-TR-44 LEARNING BY RE-EXPRESSING CONCEPTS FOR EFFICIENT RECOGNITION,
R. Keller.
[ ] LCSR-TR-45 LEARNING AND PROBLEM SOLVING, T.M. Mitchell.
[ ] LRP-TR-15 CONCEPT LEARNING BY BUILDING AND APPLYING TRANSFORMATIONS
BETWEEN OBJECT DESCRIPTIONS, D. Nagel.
-------
∂20-Sep-83 1153 DKANERVA@SRI-AI.ARPA Prospective papers, tech reports, monographs
Received: from SRI-AI by SU-AI with TCP/SMTP; 20 Sep 83 11:53:26 PDT
Date: Tue 20 Sep 83 08:38:47-PDT
From: DKANERVA@SRI-AI.ARPA
Subject: Prospective papers, tech reports, monographs
To: csli-folks@SRI-AI.ARPA
In trying to get a picture of the kind and amounts of
things CSLI will be publishing this first year, it would help
if any of you who have papers, tech reports, monographs, or
some other possibly appropriate writing in the works would
send me a brief message about it. Please feel free to consult
me also in the early stages of your writing about any questions
you may have.
Thanks. Dianne Kanerva
CSLI Editor
-------
∂20-Sep-83 1212 @MIT-MC:Pool@MIT-MULTICS Jackson paper
Received: from MIT-MC by SU-AI with TCP/SMTP; 20 Sep 83 12:12:13 PDT
Date: 20 September 1983 1441-edt
From: Ithiel de Sola Pool <Pool @ MIT-MULTICS>
Subject: Jackson paper
To: JMC @ MIT-MC
.nf
Jackson's is a superb analysis of the available technologies
of communications, which at the end falls into an ideological
trap. What this paper does well, is something that needs to
be done more often. Too much of our public policy is made by
lawyers and economists who do not understand the technological
alternatives with which they are playing. As Jackson notes,
the history of broadcasting regulation is a star example. The
"scarcity" that exists and that for 60 years has resulted in a
tightly regulated, politicized, oligopolistic, and indeed
trashy system of broadcasting, is the result of regulatory
policies adopted by political decision makers, not a result of
technological constraints to which the decision makers thought
they were responding. There were alternatives of which they
never dreamed.
Indeed, the connection that no central decision maker is
likely to be as wise as is the collective process of many
small decisions made by millions of people over time, is a
shared conviction of most of us at this conference. The
errors that have been made by many well-intentioned regulators
which could have been avoided in a diffused system of decision
making are repeatedly illustrated in the papers that deal with
past regulation. I join with Charles Jackson in warning that
most of the proposed schemes for regulating cable are
ill-advised, short-sighted, and based on lack of understanding
of where the technology is going. In particular the bills
currently being proposed in both houses of Congress would
saddle us with a disasterous quarter or half century of
stupidity much as we have had with broadcasting, until
technology worked us out of a politicized system that we need
not get into. And the bills proposed by the cable industry
and by its critics are equally disasterous in these respects.
What then is wrong with Jackson's simple formula, "don't
regulate"? It assumes that there is always a choice between
doing nothing (allowing the incremental processes of pluralistic
decisions to work) and making a central decision. Political
processes favor that kind of simple two-valued formulation of
alternatives -- and that is one of the things that is wrong with
politics as a decision process. We get into silly debates as to
whether we are for or against regulation -- as though such an
on-off choice existed. We insist on characterizing people as
"conservatives" or "liberals" as though serious substantive
decisions can be made by any such ideological formulas.
The fact of the matter is that pluralistic "unregulated"
decisions are always made within the context of an accepted
system of rules of the game, which for shorthand we label by
such terms as "the law" or "property". For reasons of stable
planning, these are not designs that should be changed
frequently or lightly, but they ←λa←λr←λe matters of decision. It
is not the case that there is just one natural system of law
or property; there are structural choices that have to be made
about what the legal or property system should be. Those are
political decisions.
So all of us here would agree that in designing a system of
cable, or of telecommunications in general, one goal should be
to minimize the amount of governmental day-by-day decision
making, administrative supervision, licensing, or rate
regulation. We should, as Jackson argues, create as fluid as
possible a situation for entrepreneurs to experiment with
novel marketing ideas and novel technologies. It is indeed a
marvelously fertile field for innovation and progress, and
furthermore it is one that touches on the special issues of
the First Amendment.
However, in this field, as in every field of policy, there
arise certain decisions which cannot be made pluralistically.
The reasons for that are several; Jackson lists some of them
such as natural monopoly, essential service affected by the
public interest, etc. There are others too, such as the one I
have stressed, the need to define the set of enforcable rights
in which people trade. In these comments I am not going to
address the issue of why such decision situations arise. The
point is that ←λd←λe ←λf←λa←λc←λt←λo there are situations in which some
central decision maker does have to reach a decision which is
going to affect the interests and welfare of millions of
people. In our kind of democratic society two principles
ought to govern such decisions over and above the cost/benefit
considerations for the particular subject matter. One is that
such decisions are properly made by democratically chosen
government authorities, not private monopolists (hence the
anti-trust laws.) The other is that the decision be made in
such a way as to minimize regulatory processes and encourage
pluralistic decision making.
All of this bears vitally on the present situation of cable.
We are moving in an ill-considered and thoughtless way into a
system that depends in detail on government granted monopoly
franchises, in which the present political debate is about
whether the FCC should take over regulatory powers from the
cities, not over what the structure of the system should be.
There is nothing in the law of nature that says a cable
franchise should be for a 100 channel system. It could be
separate channel by channel, separate for the carrier plant
from the channels, sold at auction or granted by formula.
There are arguments for and against every solution. The one
thing that in this instance is not possible is to make no
decision and leave it to pluralistic processes. For various
reasons it is not acceptable to allow anyone who wants to
string cables over or under the city streets to do so where
they want, when they want, and to whom they want. That being
the case, there are decisions that have to be made
politically.
By failing to face up to what those essential political
decisions are, and how they can be made in ways that minimize
government administration and maximize the prospects for
pluralistic entrepreneurial use of the available new
technologies, Jackson misses a great opportunity, and ends up
with a weak set of conclusions from what starts out as a
brilliant paper.
He makes a passing remark about the possible desirability of a
new common carrier to operate conduits under the street, and
leaves it there unconsidered except for expressing a perhaps
well founded distaste for making that the job of present cable
companies.
He casually concludes that the major problem of cable
regulation over the next 10 years will be determining how far
cable companies will be allowed to offer services that have
traditionally been offered by the phone systems. I beg to
differ; that will be a problem, but far less important and
controversial than another one. The number one problem is
that if we continue going as we are, cable systems will
acquire ←λd←λe ←λf←λa←λc←λt←λo a politically unacceptable degree of
monopolistic control over the contents of what people can
communicate in this country.
I am not arguing for a moment with Jackson's cafeteria list of
all the technological possibilities for a pluralistic
communications system. Furthermore I believe that in the end
the technology will overwhelm the monopolistic powers that
cable systems are acquiring. Twenty-five years hence we may
find ourselves in a new movement of deregulation and
divestiture (though we will use new-fashioned words) of which
the cable systems will be the victims as we allow other new
means of communication to flourish. The fact remains that in
between now and that future we are moving inadverantly into a
very undesirable and politically non-viable situation of
concentration of political power over communications.
Whatever the prospects for the long run, the picture that is
emerging for the next couple of decades is that the dominant
successful second carrier (besides the phone company) in
American cities will be a cable system. As Jackson points
out, it will not be the most efficient system for each
function, but as a package it seems likely to be the dominant
alternative. This major communications facility will be run
as a monopoly in each city, and the monopolist who runs it
will be chosen by a highly political franchising process.
That is bad enough, but now we come to the really ominous
prospect. The cable monopolist will make most of his money by
choosing and vending programming. He will therefore become a
self-interested controller of what any publisher or producer
can put on cable in that market. Jackson makes fun of the
unemployed producers driving their Mercedes' to the breadline,
and so he obviously misses the point. Americans will not long
tolerate a situation in which a politically chosen company in
each city can decide which videotex service, or movie series,
or news programs can get effectively published in that city.
To argue, as Jackson does, that there are some more expensive
ways or new ways out in the future to bypass the enfranchised
cable monopolist is no answer. Would we in this country allow
city governments to choose and enfranchise the city's monopoly
newspaper and dismiss objections by saying there are other
ways of delivering the news? That would clearly be
intolerable.
Jackson and I agree that the wrong solution to this
prospective problem would be the establishment of a system of
government administered fairness and balance requirements or
even a rigid prohibition on cable companies experimenting with
content offerings. But it is highly irresponsible not to
recognize that the pattern that seems to be evolving is one of
a highly politicized system with substantial control over
effective publication of major types of material. It is also
bad policy analysis not to ask what the alternative structures
of cable property rights can be to minimize those dangers and
to maximize the amount of pluralistic unregulated decision
making over cable use.
This is not the place to put forward panaceas, but it must be
noted that most people who are thinking about these matters
tend to look for systems in which by one means or another
entrepreneurs other than the enfranchised monopolist can lease
cable time to provide competitive and alternative services to
the public. To label such property schemes "regulation", and
property schemes in which the city or some other political
authority bestows an exclusive control of content on their
chosen agent "non-regulation" is clearly obfuscation.
∂21-Sep-83 0116 ARK Chris Tucci's Comprehensive Project
To: CT@SU-AI, PW@SU-AI, DFH@SU-AI, JMC@SU-AI
CC: ARK@SU-AI
After a cursory inspection, I have decided that Chris's comp project
should be sent on to the committee after Paul Wieneke has checked the
output, since it has changed from last time, and I have now way to verify
it. I've left it on Diana's desk. Thanks.
Arthur
∂21-Sep-83 0950 DFH
To: "@FACCOM.[1,DFH]"@SU-AI
Facilities Committee Meeting
I am scheduling the facilities committee meeting at 11 am on Monday, October 3,
in room 252 MJH. My usual apologies to everyone for whom this is inconvenient--
you are a difficult group to get together.
∂21-Sep-83 1006 REG
To: DFH, JMC
∂21-Sep-83 0952 @SU-SCORE.ARPA:DFH@SU-AI
Received: from SU-SCORE by SU-AI with TCP/SMTP; 21 Sep 83 09:51:52 PDT
Received: from SU-AI.ARPA by SU-SCORE.ARPA with TCP; Wed 21 Sep 83 09:53:03-PDT
Date: 21 Sep 83 0950 PDT
From: Diana Hall <DFH@SU-AI>
To: "@FACCOM.[1,DFH]"@SU-AI
Facilities Committee Meeting
I am scheduling the facilities committee meeting at 11 am on Monday, October 3,
in room 252 MJH. My usual apologies to everyone for whom this is inconvenient--
you are a difficult group to get together.
----
I don't think I have any formal relationship with the computer facilities
committee after September 30. But if John wants me to come, I'll be there
(a little late, as I teach from 10 to 11).
Ralph
∂21-Sep-83 1421 KJB@SRI-AI.ARPA postdocs
Received: from SRI-AI by SU-AI with TCP/SMTP; 21 Sep 83 14:20:51 PDT
Date: Wed 21 Sep 83 14:22:44-PDT
From: KJB@SRI-AI.ARPA
Subject: postdocs
To: csli-principals@SRI-AI.ARPA
Dear all, Below is an announcement I sent out to the secretary of the
association for symbolic logic for inclusion in their November
newletter. I would be happy to send out similar announcements
to other organizations if you will draft such a thing, with all the
relevant information, and send it to me. We also will need to have
postdoc posters printed up and distributed, I suppose.
Jon
September 19, 1983
Professor Ward Henson
Secretary, ASL
DEpartment of Mathematics
University of Illinois
Urbana, ILL. 61801
Dear Ward,
The announcement below is for the next issue of the ASL Newsletter.
I am pleased that the newsletter, which I started, is really turning
into what I hoped it would become, by the way.
Regards,
Jon Barwise
Postdoctoral Fellowships:
The Center for the Study of Language and Information anticipates
awarding several postdoctoral fellowships for the next academic year.
One or more of these could be in areas of interest to members of the ASL,
specifically in the areas of the semantics of human and computer
langueages, and in the area of the logic of information and
computation. Further information can be obtained from Dr. E. Macken,
CSLI, Ventura Hall, Stanford University, Stanford, CA. 94305, USA
-------
∂21-Sep-83 2133 TRATTNIG@SU-SCORE.ARPA Tomorrow's Seminar
Received: from SU-SCORE by SU-AI with TCP/SMTP; 21 Sep 83 21:33:51 PDT
Date: Wed 21 Sep 83 21:33:37-PDT
From: Werner Trattnig <TRATTNIG@SU-SCORE.ARPA>
Subject: Tomorrow's Seminar
To: super@SU-SCORE.ARPA
Tomorrow, Thursday, Ian Watson from Manchaster University will give a
seminar.
We shall meet as ususal at 3 p.m. in the CIS conference room. Please
try to arrive on time. It is insulting to the speaker when people
drop in around 3:30.
Werner-
-------
THE MANCHESTER DATAFLOW MACHINE - PRINCIPLES AND PERFORMANCE
The manchester dataflow machine is based on the tagged dataflow model
of computation. It has been operational for almost two years, has
twelve processors and runs at a speed of approximately 1.7 MIPS.
This seminar will explain why the tagged model is a significant
improvement over the static model for the construction of real
machines and discuss some of the implementation details.
Results of a preliminary performance evaluation on a set of realistic
programs will also be presented showing the speedup which results as a
funtion of the number of processors.
Further developments will also be discussed.
-------
∂21-Sep-83 2140 @SRI-AI.ARPA:BrianSmith.PA@PARC-MAXC.ARPA No Computation Seminar Meeting
Received: from SRI-AI by SU-AI with TCP/SMTP; 21 Sep 83 21:40:25 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Wed 21 Sep 83 21:41:39-PDT
Date: 21 Sep 83 21:40 PDT
From: BrianSmith.PA@PARC-MAXC.ARPA
Subject: No Computation Seminar Meeting
To: CSLI-Folks@SRI-AI.ARPA
cc: BrianSmith.PA@PARC-MAXC.ARPA
We will not have a meeting tomorrow afternoon in preparation for
the fall quarter research seminar on computation, last week's message
notwithstanding.
Brian
∂22-Sep-83 1107 @SRI-AI.ARPA:TW@SU-AI New seminar in CS Dept.
Received: from SRI-AI by SU-AI with TCP/SMTP; 22 Sep 83 11:07:03 PDT
Received: from SU-AI.ARPA by SRI-AI.ARPA with TCP; Thu 22 Sep 83 11:07:20-PDT
Date: 22 Sep 83 1103 PDT
From: Terry Winograd <TW@SU-AI>
Subject: New seminar in CS Dept.
To: csli-people@SRI-AI
I will be organizing a weekly seminar in the computer science Dept. this
fall on a new area I am currently developing as a research topic: the
theory of "talkware". This area deals with the design and analysis of
languages that are used in computing, but are not programming languages.
These include specification languages, representation languages, command
languages, protocols, hardware description languages, data base query
languages, etc. There is currently a lot of ad hoc but sophisticated
practice for which a more coherent and general framework needs to be
developed. The situation is analogous to the development of principles of
programming languages from the diversity of "coding" languages and methods
that existed in the early fifties.
The seminar will include outside speakers and student presentations of
relevant literature, emphasizing how the technical issues dealt with in
current projects fit into the development talkware theory. It will meet
at 2:15 every Wednesday in Jacks 301. The first meeting will be Wed.
Sept. 28. All CSLI people are welcome to attend, and urged to volunteer
as a speaker if they are doing relevant work. For a more extensive
description, see {SCORE or MAXC}<WINOGRAD>TALKWARE or {SAIL}TALKWA[1,TW].
∂23-Sep-83 1030 @SRI-AI.ARPA:Barwise.PA@PARC-MAXC.ARPA Working group on Programming Language Semantics
Received: from SRI-AI by SU-AI with TCP/SMTP; 23 Sep 83 10:30:24 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Fri 23 Sep 83 10:31:31-PDT
Date: Fri, 23 Sep 83 10:30 PDT
From: Barwise.PA@PARC-MAXC.ARPA
Subject: Working group on Programming Language Semantics
To: CSLI-Folks@SRI-AI.ARPA
cc: Barwise.PA@PARC-MAXC.ARPA
Working Group in C1: Semantics of Computer Languages
Brian Smith and I plan to have a working group that meets
Monday afternoons from 4 to 6 at PARC. We have decided that this
will not be an introductory reading group, but rather a working group
for people with some familiarity with the basic material.
The first meeting will be on Monday, October 3, 1983. We will
assume that people have read much of Gordon's book "The Denotational
Description of Programming Languages" (Springer Verlag, 1979). At
that meeting will will decide what we will work through next, but the
plan is to read Dana Scott's recent paper "Domains for Denotational
Semantics". Other things we might pursue are a draft of a new book
by Plotkin, Stoy's book, or a paper by Martin-Lof.
Let me know if you plan to attend. More detailed instructions
on room, etc., will be sent out later to those who repond.
Jon Barwise
∂23-Sep-83 1738 KJB@SRI-AI.ARPA Secretaries at CSLI
Received: from SRI-AI by SU-AI with TCP/SMTP; 23 Sep 83 17:38:01 PDT
Date: Fri 23 Sep 83 17:24:57-PDT
From: KJB@SRI-AI.ARPA
Subject: Secretaries at CSLI
To: csli-folks@SRI-AI.ARPA
Policy on secretaries:
Within the next few weeks we will have three secretaries, an office
assistant for running errands and making tea in the afternoon, a
receptionist who will act as center contact for those with offices
elsewhere that need something done here, and an administrative assistant.
This is in addition to the Assistant Director (Betsy), Administrator,
Editor and editorial assistant, and the Director of Computing Systems
and his assistants.
However, for now we are severely understaffed at the
secretarial and office assistant level. In particular, people have
gotten used to using Pat Wunderman as the main contact. As of a
couple of weeks ago, she became an Executive Secretary, working for
Betsy and me. However, because people have come to expect her to do
various things, she has not had the time either to do the things we
needed done, or to work on learning to use the computer efficiently. I
would like to ask everyone's cooperation in letting her get on with
her new job by using Sandra for the things you have used Pat for. As
the other secretaries are hired, these tasks will get divided up.
Since Sandra will not have a terminal, this will involve you using the
phone more than you have recently. Also, don't forget that there are
many things that Bach-Hong and JOyce can help us with.
Phone nos: Sandra: 497-0939
Bach-Hong: 497-1249
Joyce: 497-1249
Also, please remember that we ARE understaffed now, so help out in any
way you can by having things done elsewhere when possible.
Thanks,
Jon
-------
∂23-Sep-83 1755 GROSZ@SRI-AI.ARPA Seminar Announcement--Context
Received: from SRI-AI by SU-AI with TCP/SMTP; 23 Sep 83 17:55:27 PDT
Date: Fri 23 Sep 83 17:51:50-PDT
From: GROSZ@SRI-AI.ARPA
Subject: Seminar Announcement--Context
To: csli-friends@SRI-AI.ARPA
cc: almog@SRI-AI.ARPA
PHILOSOPHY OF LANGUAGE SEMINAR
P.Suppes and J.Almog(with the cooperation of CSLI)
Time: Tuesday 3:15, Ventura Hall
SUBJECT: WHY CONTEXT WON'T GO AWAY
For many years, in fact since Frege's pioneering work in
formal semantics in the 19th century, it was assumed among formal
semanticists that we can do without IT ("Without what?", well,YOU tell
me, YOU said we can do without IT). The dispensability thesis was due
not only to the fact that logicians dealt primarily with mathematical
logic where context-dependence is negligible (or so it seems). Rather,
even when logicians and philosophers took notice of indexical
locutions in natural language, they came to embrace the following
"ELIMINATION THEOREM":
EVERY INDEXICAL EXPRESSION CAN BE ELIMINATED IN FAVOR OF A PURELY
QUALITATIVE AND ETERNAL DESCRIPTIONAL EXPRESSION.
The list of believers in this "dogma" is very impressive:
Frege, Russell, Wittgenstein, Carnap, Quine, and Davidson. But, in the
last ten years two very important "impossibility theorems" have been
put forward in this context: (i) A semantic result, due to D. Kaplan,
stating that an indexical object language (e.g. English) CANNOT be
translated (by a linguistic-meaning preserving mapping) into an
eternal metalanguage, and (ii) A closely related cognitive result, due
to J. Perry, stating that indexical characterizations of mental states
CANNOT be systematically reduced to cognitively-equivalent purely
descriptive characterizations.
On the other side of this mountain, there is the outlook of AI
and computational research into natural language understanding. Not
only did no one here embrace the "elimination-theorem", but rather the
"impossibility" results may have been assumed as a common-sensical
STARTING point.
Our seminar aims to explore the interaction between the
philosophical and computational outlooks on the question "why context
won't go away?". We will start with some overview talks on the state
of the art in both frameworks. Having set the scene, we plan to invite
speakers from various ends of the spectrum to tell us why context is
here to stay-- with a vengeance--and what might be done to account for
it. We plan to focus on some paradigmatic areas of research:
1. The semantics and pragmatics of definite noun phrases.
2. The notion of salience and strategies of interpretation.
3. New frontiers: the discourse (vs.the sentence), an honest notion
of context (vs.the highly austere notion in current philosophical
theories and the less well-grounded notions in AI).
4. The semantics of names and its interaction with the
semantics of mental states (so as to cover some well known
puzzles of attitudinal semantics and see whether the context holds
the key here).
Tentative speakers: B.Grosz, J.Perry, J.Hobbs, B.Smith,
C.Peacocke, S.Stitch, B.Moore, P.Suppes, H.Wettstein,
J.Etchemendy, S.Soames, J.Moravcsik, J.Almog. We plan on continuing
through other quarters if demand persists.
N.B. We start on Tuesday 10.4.83.
-------
∂24-Sep-83 1538 KJB@SRI-AI.ARPA Newsletter
Received: from SRI-AI by SU-AI with TCP/SMTP; 24 Sep 83 15:38:07 PDT
Date: Sat 24 Sep 83 15:33:50-PDT
From: KJB@SRI-AI.ARPA
Subject: Newsletter
To: csli-friends@SRI-AI.ARPA
CSLI Newsletter
9/24/83
In the following I will summarize various developments that some of
you may not be aware of. I will also add messages that I have
received from others that seem relevant. If the newsletter seems
useful, we will continue it, but I do not plan to spend every Saturday
writing it. It will eventually be taken over by having the
receptionist put together items that everyone sends her. -Jon Barwise
* * * * * * *
Research Seminars
The research seminars begin this coming Thursday, September 29. I have
not had word on what will happen at these first meetings. However, I
do want to remind you that all we all agreed in July, before the SDF
Board Meeting, that this first quarter all principals and associates
should attend both seminars, and that these initial seminars would be
oriented toward getting us all on the same wavelength, both with
respect to human and computer languages.
* * * * * * *
Outer Colloquium Announcement
"Prescriptive Grammar"
by
Dr. Geoffrey Nunberg
(Nunberg has an article on this subject that is going to appear in the
Atlantic in the near future.)
* * * * * * *
Working Group in C1 (Semantics of Computer Languages)
There will be a working group in C1 meeting Monday afternoons, at
Xerox PARC, from 4-6 pm, starting October 3. The first meeting will
discuss the initial chapters of Gordon's book "The Denotation
Description of Programming Languages" and will make plans for the
future. Anyone who is interested should contact Jon Barwise or Brian
Smith.
* * * * * * *
The size of CSLI
As you will recall from the September 1 meeting, there is a
considerable gap between the funds we need for four years, and the
funds we have received -- computers and building aside. The Executive
Committee has been wrestling with this, and has decided to compromise
between the two extremes, the one being to plan from right now to live
within our means, the other being to plan to be the size of program
indicated in the proposal and to count on SDF or someone else to come
up with he funds to make this possible.
After analysis of the budget, it seems that there really is no way to
trim enough this year to make it possible to stay at the same size in
years to come. There are just too many commitments for this year
already. However, by being very careful this year, we expect to be
able to save 10% to 20% of our budgeted funds. If we combine this
savings with extra funds brought in by continuing NSF (etc.) grants
and some increase by SDF next June, we should be able to proceed with
a very strong program, if not as large as originally foreseen. Here
are the relevant numbers:
This year Proposed to SDF New Target
Principals 17 17 17
Associates 17 17 15
Affiliates 8 8 8
Visitors 4 8 4
Postdocs 1 20 10
Grad Students 10 40 20
Staff 15 19 16
SSRA's 0 12 0
-----
Total 72 141 90
I must stress that this new target will only be possible, even for the
first four years, if we are frugal and also bring in extra funds.
This year people should be sure to renew grants, but concentrate on
research. Next summer and fall, however, we will all have to work on
new proposals. By then we should also have the staff necessary to do
the paperwork involved in proposal writing.
* * * * * * *
The organization of CSLI
The research projects are going to serve as the central organizational
structure for CSLI. At least, we are going to try running things that
way. Thus, for example, if there is some special task that need to be
done we will ask one of the projects to be in charge of it. We will
take into account the total FTE's allocated to various projects in
determining how much they should be asked to do. As a first example,
we have asked Project A1 to be responsible for organizing the Opening
Ceremony this fall.
* * * * * * *
Computing at CSLI
One of the most time consuming tasks this fall has been getting the
computational environment started. Things seem to be falling into
place, thanks to the efforts of the Computing Committee (Brian Smith,
Stan Peters, Terry Winograd, Ray Perrault and Mabry Tyson), the new
grant from SDF for equipment (roughly $4,000,000, depending on what we
can get as gifts), and the efforts of our new Director of Computing
Systems, Eric Ostrom.
Ostrom comes to CSLI from MIT, where he is currently Director
of Computing Systems for the MIT EE/CS Departments, manager of the MIT
AI Lab, and manager of the MIT Local Area Network. He is currently
serving as our consultant, but will be here full time as of
mid-October or early November. He has already been invaluable in our
dealing with DEC and in getting things started around here.
* * * * * * *
CSLI Technical Reports
Diane Kanerva has been hard at work editing our research
program announcement (which is now available for those of you who want
to send it to others) and getting the Technical Reports Series
organized (among other things we have asked her to do that she
probably did not anticipate having to do).
Initially, at least, we will distribute the CSLI Technical
Reports free of charge, since it is a way for us to advertise our
existence. All technical reports will be refereed internally before
being issued. The mechanism is that each report will be read by a
project manager (or his or her representative), then by an area
manager, and finally approved by the executive committee. The idea is
not to require adherence to any particular view, but to insure that
only first-rate papers come out with the CSLI imprint.
There have been discussions of a monograph series, but at
present it is not clear how many people will have monograph length
works in the near future.
* * * * * * *
Requests for Funds
Requests for funds for travel and short term visitors will be
handled through the sixteen projects. Anyone requesting funds should
pick up the appropriate form from Betsy and have it approved by her
with recommendations from the appropriate project manager or managers
and area coordinator. After Betsy has approved it, it will go to
Bach-Hong Tran or the corresponding person at SRI or Xerox. Area
coordinators have been given target figures for their areas, and will
translate those into target figures for each project.
* * * * * * *
New Principal
We have invited Martin Kay to become a Principal of CSLI, with
whatever rights and duties that entails. He has agreed, without even
finding out what those rights and duties are. We are all very
pleased. We have already appointed him to the Education Committee.
* * * * * * *
Graduate Students
CSLI is has committed itself to support the following
excellent graduate students for the following year. They are each
supposed to become involved with one of the research projects, so you
should welcome them to the projects and help them get started.
Akira Ishikawa (Linguistics)
Michael Barlow (Linguistics)
Kathleen Kells (Computer Science)
Terry Au (Psychology)
Deanna Wilkes-Gibbs (Psychology)
John Lamping (Computer Science)
There are still some application to be acted on, if suitable
arrangement can be worked out. Thanks to the Education Committee for
sorting this all out at short notice.
* * * * * * *
Remote Affiliates
CSLI is going to have a position of "remote affiliate". This
term has been thrown around over the past few months in a number of
ways. We are going to use it to designate a person or group with whom
we have a close working arrangement, for example, frequent visits to
or from. Being a remote affiliate, for example, should serve as a
permanent endorsement in requests for travel funds.
Several people and groups have already been proposed as remote
affiliates. Rather than do this on some ad hoc basis, we would like
each project to suggest one or two names, to be reviewed by the area
manager. There will, presumably, be overlap between projects. We want
to come out with a total of at most sixteen, at present, four from
each area. Later, we will consider adding more.
* * * * * * *
Announcements
Once we have our receptionist hired, he or she will coordinate all
announcements. Until then, send them to me (kjb@sri-ai, Barwise at
Ventura Hall) marked clearly as announcements. Whatever comes in by
Wednesday noon will make into a newsletter or announcement that is
mailed out by US and Campus Mail early Friday a.m.
It is very important, as we get things started, that all events be
announced to all our friends. Each project manager, seminar leader,
colloquium person, etc., must take responsibility for getting
announcements out. It causes a lot of extra confusion to have up to a
hundred people calling here asking what is going to happen next week.
The following is given as a good example of an announcement, as well
as for its content.
* * * * * * *
Philosophy of Language Seminar
P.Suppes and J.Almog (with the cooperation of CSLI)
Time: Tuesday 3:15, Ventura Hall
SUBJECT: WHY CONTEXT WON'T GO AWAY
For many years, in fact since Frege's pioneering work in
formal semantics in the 19th century, it was assumed among formal
semanticists that we can do without IT ("Without what?", well,YOU tell
me, YOU said we can do without IT). The dispensability thesis was due
not only to the fact that logicians dealt primarily with mathematical
logic where context-dependence is negligible (or so it seems). Rather,
even when logicians and philosophers took notice of indexical
locutions in natural language, they came to embrace the following
"ELIMINATION THEOREM": EVERY INDEXICAL EXPRESSION CAN BE ELIMINATED IN
FAVOR OF A PURELY QUALITATIVE AND ETERNAL DESCRIPTIONAL EXPRESSION.
The list of believers in this "dogma" is very impressive:
Frege, Russell, Wittgenstein, Carnap, Quine, and Davidson. But, in the
last ten years two very important "impossibility theorems" have been
put forward in this context: (i) A semantic result, due to D. Kaplan,
stating that an indexical object language (e.g. English) CANNOT be
translated (by a linguistic-meaning preserving mapping) into an
eternal metalanguage, and (ii) A closely related cognitive result, due
to J. Perry, stating that indexical characterizations of mental states
CANNOT be systematically reduced to cognitively-equivalent purely
descriptive characterizations.
On the other side of this mountain, there is the outlook of AI
and computational research into natural language understanding. Not
only did no one here embrace the "elimination-theorem", but rather the
"impossibility" results may have been assumed as a common-sensical
STARTING point.
Our seminar aims to explore the interaction between the
philosophical and computational outlooks on the question "why context
won't go away?". We will start with some overview talks on the state
of the art in both frameworks. Having set the scene, we plan to invite
speakers from various ends of the spectrum to tell us why context is
here to stay-- with a vengeance--and what might be done to account for
it. We plan to focus on some paradigmatic areas of research:
1. The semantics and pragmatics of definite noun phrases.
2. The notion of salience and strategies of interpretation.
3. New frontiers: the discourse (vs.the sentence), an honest
notion of context (vs.the highly austere notion in current
philosophical theories and the less well-grounded notions in AI).
4. The semantics of names and its interaction with the
semantics of mental states (so as to cover some well known puzzles of
attitudinal semantics and see whether the context holds the key here).
Tentative speakers: B.Grosz, J.Perry, J.Hobbs, B.Smith,
C.Peacocke, S.Stitch, B.Moore, P.Suppes, H.Wettstein, J.Etchemendy,
S.Soames, J.Moravcsik, J.Almog. We plan on continuing through other
quarters if demand persists.
N.B. We start on Tuesday 10.4.83.
* * * * * * *
Policy on secretaries
Within the next few weeks we will have three secretaries, an office
assistant for running errands and making tea in the afternoon, a
receptionist who will act as center contact for those with offices
elsewhere that need something done here, and an administrative
assistant. This is in addition to the Assistant Director (Betsy),
Administrator, Editor and editorial assistant, and the Director of
Computing Systems and his assistants.
However, for now we are severely understaffed at the secretarial and
office assistant level. In particular, people have gotten used to
using Pat Wunderman as the main contact. As of a couple of weeks ago,
she became an Executive Secretary, working for Betsy and me. However,
because people have come to expect her to do various things, she has
not had the time either to do the things we needed done, or to work on
learning to use the computer efficiently. I would like to ask
everyone's cooperation in letting her get on with her new job by using
Sandra for the things you have used Pat for. As the other secretaries
are hired, these tasks will get divided up. Since Sandra will not
have a terminal, this will involve you using the phone more than you
have recently. Also, don't forget that there are many things that
Bach-Hong and JOyce can help us with.
Phone nos: Sandra: 497-0939
Bach-Hong: 497-1249
Joyce: 497-1249
Also, please remember that we ARE understaffed now, so help out in any
way you can by having things done elsewhere when possible, at least
for the time being.
* * * * * * *
Advisory Panel
Seven leaders from Europe and the US have been invited to serve on
CSLI's Advisory Panel. The invitations were sent out this past week.
We have invited these men and women to serve for two years, and to
meet here first on November 10-12. (I will refrain from naming these
individuals until after they have accepted.) We will put one of the
projects in charge of planning this event.
* * * * * * *
CSLI Grand Opening
This will be held in October and is being planned and executed by
Project A1.
* * * * * * *
Daily Tea at CSLI Begins!
CSLI now owns one electric tea kettle, two tea pots, and two pounds of
Mr. Peet's tea. Tea will begin Monday at 3:30 and will be held daily
in the Ventura Lounge. Everyone is urged to come whenever they can.
They are encouraged to bring some cookies, brownies, or other treats
from time to time. The tea is free. Many thanks to the Tea Committee
for getting this going.
* * * * * * *
Thursday schedule (reminder)
8:30-9:30 Executive Committee
10:00-11:30 Seminar: "Approaches to Human Languages"
12:00- 1:30 TINLunch
2:00- 3:30 Seminar: "Approaches to Computer Languages"
3:30- 4:00 Tea (daily!)
4:00- 5:00 Colloquium
The seminars and colloquium are scheduled to be held in Room 175
Redwood Hall, which is right across the street from Ventura, next to
the "C" parking lot.
-------
∂26-Sep-83 1007 BMACKEN@SRI-AI.ARPA Grand Opening
Received: from SRI-AI by SU-AI with TCP/SMTP; 26 Sep 83 10:07:20 PDT
Date: Mon 26 Sep 83 10:07:32-PDT
From: BMACKEN@SRI-AI.ARPA
Subject: Grand Opening
To: BMacken@SRI-AI.ARPA, BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA,
Briansmith@PARC-MAXC.ARPA, grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA,
JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA, kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA,
peters@SRI-AI.ARPA, rperrault@SRI-AI.ARPA, sag%SU-psych@SU-SCORE.ARPA,
stan@SRI-AI.ARPA, Wasow%SU-psych@SU-SCORE.ARPA, winograd@PARC-MAXC.ARPA,
appelt@SRI-AI.ARPA, desrivieres@PARC-MAXC.ARPA, eric@MIT-MC.ARPA,
ford@PARC-MAXC.ARPA, halvorsen@PARC-MAXC.ARPA, hans@SRI-AI.ARPA,
hobbs@SRI-AI.ARPA, kay@PARC-MAXC.ARPA, konolige@SRI-AI.ARPA,
pollard.hplabs@UDEL-RELAY.ARPA, pullum.hplabs@UDEL-RELAY.ARPA,
pcohen@SRI-KL.ARPA, pereira@SRI-AI.ARPA, shieber@SRI-AI.ARPA,
stickel@SRI-AI.ARPA, stucky@SRI-AI.ARPA, withgott@PARC-MAXC.ARPA
Jon suggested that we rotate the responsibilty for organizing our
various functions across projects. It seems a good way of sharing the
work without a hardship on anyone. Project A1, then, will be
responsible for the CSLI Grand Opening.
Later today, when I have updated my list with information I received this
morning, I will send everyone my list of project associations. I'll need
you to send me corrections and additions so the complete version
won't be available until later in the week. However, there are already 7
people on A1 so a working group for the Opening is available.
The Opening should be a very nice function that will serve as a thank you
for all the people who helped on the proposal as well as an opening event
for all thoes currently associated with the project. I will send out the
names I have so you can let me know of others that should be invited.
B.
-------
∂26-Sep-83 1400 rod@Whitney CS200 departmental lectures
Received: from SU-WHITNEY by SU-AI with PUP; 26-Sep-83 14:00 PDT
Date: 26 Sep 1983 13:56-PDT
From: Rod Brooks <rod@Whitney>
Subject: CS200 departmental lectures
To: dek@sail, rwf@sail, acy@sail, ullman@score, jmc@sail, zm@sail
Message-Id: <83/09/26 1356.850@Whitney>
There are still plenty of slots open in the CS200 departmental
lecture series. Its 2:45-4:00 pm Thursdays in Skilling Auditorium.
All of October and the start of November are still free, so
you can name your day if you register early. (You can even
take this Thursday if you want.) Let me know.
Rod
∂26-Sep-83 1522 @SRI-AI.ARPA:TW@SU-AI Semantics of interrogatives and imperatives
Received: from SRI-AI by SU-AI with TCP/SMTP; 26 Sep 83 15:21:58 PDT
Received: from SU-AI.ARPA by SRI-AI.ARPA with TCP; Mon 26 Sep 83 15:22:30-PDT
Date: 26 Sep 83 1519 PDT
From: Terry Winograd <TW@SU-AI>
Subject: Semantics of interrogatives and imperatives
To: csli-people@SRI-AI
I was just asked by a first-year student for the best reference(s) on
the semantics of interrogatives and imperatives in a truth-theoretical
framework. Any suggestions? Thanks --t
∂26-Sep-83 1610 @SRI-AI.ARPA:Kaplan.pa@PARC-MAXC.ARPA Project A1: Phonology, Morphology, Synax
Received: from SRI-AI by SU-AI with TCP/SMTP; 26 Sep 83 16:09:52 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Mon 26 Sep 83 16:10:40-PDT
Date: 26 Sep 83 16:06 PDT
From: Kaplan.pa@PARC-MAXC.ARPA
Subject: Project A1: Phonology, Morphology, Synax
To: CSLI-PEOPLE@SRI-AI.ARPA
cc: Kaplan.pa@PARC-MAXC.ARPA
I haven't received a list of everybody who is planning on participating
in this project, but I know that many of us will be working on these
topics.
The year is about to begin in earnest, so I would like to set up an
initial meeting, sometime next week, of the people who will be
contributing to (and getting resources from) the various subparts of
this project.
Please reply to me if you expect to be involved, and indicate what your
time constraints for a meeting next week are.
--Ron
∂26-Sep-83 1613 @SRI-AI.ARPA:Kaplan.pa@PARC-MAXC.ARPA Project A3: Strategies and tactics in the processing of
Received: from SRI-AI by SU-AI with TCP/SMTP; 26 Sep 83 16:13:34 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Mon 26 Sep 83 16:14:16-PDT
Date: 26 Sep 83 16:11 PDT
From: Kaplan.pa@PARC-MAXC.ARPA
Subject: Project A3: Strategies and tactics in the processing of
utterances
To: CSLI-PEOPLE@SRI-AI.ARPA
cc: Kaplan.pa@PARC-MAXC.ARPA
It is time to get this project going also. This project is concerned
with the heuristic, semantic, and pragmatic factors that influence how
utterances are processed. I expect far fewer people will be involved in
this than in Project A1, but I would like to know who you are and also
to set up a meeting time next week.
Please reply, and again indicate constraints on possible meeting times.
--Ron
∂27-Sep-83 0001 JMC Expired plan
Your plan has just expired. You might want to make a new one.
Here is the text of the old plan:
I will be in Europe from Sept. 4 to Sept. 27. Diana Hall (DFH) knows
my schedule if something is urgent.
∂27-Sep-83 0934 BMACKEN@SRI-AI.ARPA associations
Received: from SRI-AI by SU-AI with TCP/SMTP; 27 Sep 83 09:33:35 PDT
Date: Tue 27 Sep 83 09:34:55-PDT
From: BMACKEN@SRI-AI.ARPA
Subject: associations
To: BMacken@SRI-AI.ARPA, BMoore@SRI-AI.ARPA, Bresnan@PARC-MAXC.ARPA,
Briansmith@PARC-MAXC.ARPA, grosz@SRI-AI.ARPA, jmc@SU-AI.ARPA,
JRP@SRI-AI.ARPA, Kaplan@PARC-MAXC.ARPA, kjb@SRI-AI.ARPA, lauri@SRI-AI.ARPA,
peters@SRI-AI.ARPA, rperrault@SRI-AI.ARPA, sag%SU-psych@SU-SCORE.ARPA,
stan@SRI-AI.ARPA, Wasow%SU-psych@SU-SCORE.ARPA, winograd@PARC-MAXC.ARPA,
appelt@SRI-AI.ARPA, desrivieres@PARC-MAXC.ARPA, eric@MIT-MC.ARPA,
ford@PARC-MAXC.ARPA, halvorsen@PARC-MAXC.ARPA, hans@SRI-AI.ARPA,
hobbs@SRI-AI.ARPA, kay@PARC-MAXC.ARPA, konolige@SRI-AI.ARPA,
pollard.hplabs@UDEL-RELAY.ARPA, pullum.hplabs@UDEL-RELAY.ARPA,
pcohen@SRI-KL.ARPA, pereira@SRI-AI.ARPA, shieber@SRI-AI.ARPA,
stickel@SRI-AI.ARPA, stucky@SRI-AI.ARPA, withgott@PARC-MAXC.ARPA
The list of people and projects is now on <csli>associations; Part I is
by name and Part II is by projects. Please check your name in each part
for the following:
percent of time at CSLI
correct project
correct percentages (note that the percentages sum to
the percent of time at CSLI)
Let me know the corrections as well as any missing information. It will
get confusing if you update the file directly, so send the info to me, and
I'll update the file.
Thanks.
B.
-------
∂27-Sep-83 1017 @SRI-AI.ARPA:withgott.pa@PARC-MAXC.ARPA Re: Project A1: Phonology, Morphology, Synax
Received: from SRI-AI by SU-AI with TCP/SMTP; 27 Sep 83 10:16:37 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Tue 27 Sep 83 10:15:18-PDT
Date: 27 Sep 83 10:13 PDT
From: withgott.pa@PARC-MAXC.ARPA
Subject: Re: Project A1: Phonology, Morphology, Synax
In-reply-to: Kaplan.pa's message of 26 Sep 83 16:06 PDT
To: Kaplan.pa@PARC-MAXC.ARPA
cc: CSLI-PEOPLE@SRI-AI.ARPA
Time constraints for next week:
Monday - ok
Tuesday - not noon (Johan & Kurt's group);
not 3 (Suppes)
Wednesday -
Ron- I've invited Bush over here at 1 p.m. on Wednesday. Would you
like to talk with her or show her anything any time that afternoon?
(Thursday)
Friday - ok
Meg
∂27-Sep-83 1031 @SRI-AI.ARPA:withgott.pa@PARC-MAXC.ARPA Re: Project A1: Phonology, Morphology, Synax
Received: from SRI-AI by SU-AI with TCP/SMTP; 27 Sep 83 10:31:24 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Tue 27 Sep 83 10:31:58-PDT
Date: 27 Sep 83 10:30 PDT
From: withgott.pa@PARC-MAXC.ARPA
Subject: Re: Project A1: Phonology, Morphology, Synax
In-reply-to: withgott.pa's message of 27 Sep 83 10:13 PDT
To: withgott.pa@PARC-MAXC.ARPA
cc: Kaplan.pa@PARC-MAXC.ARPA, CSLI-PEOPLE@SRI-AI.ARPA
Sorry about that last message folks. (maybe you can tell me what I left
off my schedule.... ah, the dangers of not reading your headers)
mmw
∂27-Sep-83 1049 WUNDERMAN@SRI-AI.ARPA Announcement of Colloquium, Thurs. Sept. 29
Received: from SRI-AI by SU-AI with TCP/SMTP; 27 Sep 83 10:49:07 PDT
Date: Tue 27 Sep 83 10:48:00-PDT
From: WUNDERMAN@SRI-AI.ARPA
Subject: Announcement of Colloquium, Thurs. Sept. 29
To: CSLI-Friends@SRI-AI.ARPA
cc: wunderman@SRI-KL.ARPA
Our second colloquium will be this Thursday, Sept. 29,
from 4:00-5:00 in the classroom of Redwood Hall (please
note this is a change from the previous building, Polya
Hall). Redwood is close to Polya and Pine Halls, in the
center of Jordan Quad; parking is available in the "C"
lot adjacent to the Quad.
Our second colloquium speaker will be Robert C. Moore of
SRI. An abstract of his presentation follows:
DEDUCTIVE METHODS FOR COMMONSENSE REASONING
By Robert C. Moore
A-I Center, SRI International
Automatically drawing conclusions from a base of
commonsense, factual information has always been considered one
of the central problems of artificial intelligence. This talk is
an overview of a broad spectrum of issues and ideas that have
arisen in attempting to solve this problem. It will attempt to
sketch out some of the connections between various notions,
including: rule-based systems, forward and backward chaining,
first-order logic, resolution theorem-proving, natural deduction,
pattern matching, unification, horn-clause logic, the
closed-world assumption, reasoning from incomplete knowledge,
connection graphs, logic programming, multi-sorted logic,
property inheritance, higher-order logic, modal and intensional
logic.
-------
∂27-Sep-83 1059 @SRI-AI.ARPA:withgott.pa@PARC-MAXC.ARPA Re: Project A2 : Phonology, Syntax, and Discourse S.
Received: from SRI-AI by SU-AI with TCP/SMTP; 27 Sep 83 10:58:48 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Tue 27 Sep 83 10:58:28-PDT
Date: 27 Sep 83 10:56 PDT
From: withgott.pa@PARC-MAXC.ARPA
Subject: Re: Project A2 : Phonology, Syntax, and Discourse S.
To: CSLI-PEOPLE@SRI-AI.ARPA
cc: withgott.pa@PARC-MAXC.ARPA
If you are interested in the phonological side of discourse structure,
please let me know who you are.
Also, a reminder that Bresnan is holding an organizational meaning on
syntax and discourse structure. Effective integration suggests that
phonologists be there, too. Let her know you are interested
(<BRESNAN@parc>).
Meg
∂27-Sep-83 1106 @SRI-AI.ARPA:halvorsen.pa@PARC-MAXC.ARPA Research Seminar, Natural Language, Thursday Sept. 29
Received: from SRI-AI by SU-AI with TCP/SMTP; 27 Sep 83 11:06:30 PDT
Received: from PARC-MAXC.ARPA by SRI-AI.ARPA with TCP; Tue 27 Sep 83 11:05:46-PDT
Date: 27 Sep 83 11:01 PDT
From: halvorsen.pa@PARC-MAXC.ARPA
Subject: Research Seminar, Natural Language, Thursday Sept. 29
To: csli-friends@sri-ai.ARPA
The first CSLI research seminar in natural language will
be given by Joan Bresnan, who will speak on issues in syntactic theory.
Place: Redwood Hall, classroom
Time: Thursday, September 29, 10:00 am
Redwood Hall is close to Ventura Hall on the Stanford Campus. It can
be reached from Campus Drive or Panama Street. From Campus Drive follow
the sign for Jordan Quad. Parking is in the C-lot between
Ventura and Jordan Quad.
The next seminar, to be held on Thursday, October 6, will be
given by Bob Moore, and will have as its topic "Problems in Semantic
Analysis of Natural Language".
∂27-Sep-83 1411 KJB@SRI-AI.ARPA newsletter.two
Received: from SRI-AI by SU-AI with TCP/SMTP; 27 Sep 83 14:11:09 PDT
Date: Tue 27 Sep 83 14:11:39-PDT
From: KJB@SRI-AI.ARPA
Subject: newsletter.two
To: csli-folks@SRI-AI.ARPA
Don't forget to get any items for next weeks newsletter to me
by tomorrow noon. Thanks, Jon
-------
∂27-Sep-83 1451 TUCCI@SU-SCORE.ARPA Form signed
Received: from SU-SCORE by SU-AI with TCP/SMTP; 27 Sep 83 14:51:23 PDT
Date: Tue 27 Sep 83 14:52:32-PDT
From: Christopher Tucci <TUCCI@SU-SCORE.ARPA>
Subject: Form signed
To: pw@SU-AI.ARPA
cc: dfh@SU-AI.ARPA, ark@SU-AI.ARPA, jmc@SU-AI.ARPA
Hi--
I signed the form this morning. Is there anything else I
should do?
Chris
-------
∂27-Sep-83 1504 ARK ∂27-Sep-83 1451 TUCCI@SU-SCORE.ARPA Form signed
To: Tucci@SU-SCORE
CC: DFH@SU-AI, JMC@SU-AI, PW@SU-AI
Received: from SU-SCORE by SU-AI with TCP/SMTP; 27 Sep 83 14:51:23 PDT
Date: Tue 27 Sep 83 14:52:32-PDT
From: Christopher Tucci <TUCCI@SU-SCORE.ARPA>
Subject: Form signed
To: pw@SU-AI.ARPA
cc: dfh@SU-AI.ARPA, ark@SU-AI.ARPA, jmc@SU-AI.ARPA
Hi--
I signed the form this morning. Is there anything else I
should do?
Chris
-------
ARK - Give the program-documentation package with the form to Marilynn Walker.
∂27-Sep-83 1554 DFH Facilities Committee Meeting
I scheduled this for Oct 3, Monday, 11 am, room 252.
This was the only time open on both Brian Reid's and Ed Feigenbaum's
schedule. Unfortunately, neither Ernst Mayr or Joe Oliger can attend
at this time. Ralph Gorin will come if you want him to -- he sent
you a separate message about this.
∂27-Sep-83 1612 DFH Travel arrangements
1. Brazil. I told them to make the return reservation
so that you arrived back here on Tues, Oct. 18.
If they use Varig, that actually means departing
Sao Paulo on the evening of Mon., Oct. 17. There
is a telex from Renteria with other details.
2. Korea. I have an open ticket on Pan Am which was brought by
Professor Chwa. I have made tentative reservations which
I entered on your calendar, but which I'm sure will need
adjustment. Please let me know. Also, Prof. Chwa will
be in the area again on Sept. 29 or 30, and will call to
arrange an appointment with you then to discuss your
arrangements while in Korea. He also would like a copy of
your presentation at that time, or at least a title and
abstract.
.27-Sep-83 1653 DFH
Messages
9/7 Margy, Harriet Kiebin's office, Harper & Rowe, 212-207-7234. Do you have any
comments on the Hacker's Dictionary galleys.
9/12 Cheng San Wu called to say he put your name down as a local reference for
an apartment he is renting.
9/13 Carol Kreck, The Denver Post, 303-820-1333, is doing a story on Robotics
and wants to interview you.
9/19 David Chudnovsky called
9/19 Earl Dowdy, U.S. Congress Office of Technology Assesment, 202-226-3953.
Planning to have a 1-day conference on AI Oct. 17. I explained you
had another commitment. He would like to talk to you anyway, as the
date is somewhat tentative.
9/20 Prof. Chwa stopped by with your airline ticket for the Korea conference.
He will try to arrange a time to see you on Sept. 29 or 30, and would
like a copy of your presentation at that time, or at least title and
abstract. Also see my travel arrangements message.
9/21 Tom Mahon, 415-686-3908. He is writing a book on Silicon Valley, wants to
interview you about AI at Stanford.
9/21 Jack Sklansky, home:714-644-2229, work: 714-856-6726. Chairing panel for
American Society for Cybernetics, wants to talk with you about it. Will
also be mailing you information.
---- Psychology Today called, will send a track of the piece involving you
that will appear in the Dec. issue.
9/26 Tyler Johnson, Jack Morton Productions, 546-0800. Working on tape -- needs
help.
9/26 Richard Vistnes stopped by. He had arranged with you that he would
re-take AI Qual first week of Oct, and now wants to set up a time with
you. He probably sent you a separate message about this also.
9/27 Dennis Bark, Assoc. Director of Hoover Institute, 7-2216. Please phone.
9/27 The ultrasuede coat you ordered in April is now in at Weidemans. Phone
Thurs., 328-4170.
Hope you had a good trip. Congratulations to you and Carolyn.
∂27-Sep-83 1704 AAAI-OFFICE@SUMEX-AIM Agenda
Received: from SUMEX-AIM by SU-AI with PUP; 27-Sep-83 17:04 PDT
Date: Tue 27 Sep 83 17:07:10-PDT
From: AAAI <AAAI-OFFICE@SUMEX-AIM>
Subject: Agenda
To: JMC@SAIL
cc: aaai-office@SUMEX-AIM
Telephone: (415) 328-3123
Postal-Address: 445 Burgess Drive, Menlo Park, CA 94025
John,
Hope your trip was enjoyable! As soon as you recover from the
trip, we will need to discuss a few things which have developed
while you were gone. The points on the agenda are:
* the possibility of the AAAI co-sponsoring the IEEE-PAMI
Applied AI Conference;
* post-doctoral fellowships;
* responses about the Fredkin Proposal; and
* changes in the AAAI Secretariat.
--- Claudia
P.S. I've also prepared a first draft of the questionaire to be distributed
to our members.
-------
I agree that nothing had been decided about post-doctoral fellowships.
We need to decide whether to do it and if yes decide how a program should
be administered. My inclination would be to try to get an organization
with an existing fellowship program to administer it for us. I also
want to pursue the electronic library of AI reports that Mike Genesereth
studied a bit. You can phone me at Stanford or at home as soon as it is
convenient for you.