perm filename E80.IN[LET,JMC]2 blob
sn#544084 filedate 1980-10-18 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00431 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00042 00002 ∂01-Jul-80 0616 ENGELMORE at USC-ISI Additional support
C00045 00003 ∂01-Jul-80 1157 FFL Telephone call from Ned Irons of Colorado
C00046 00004 ∂01-Jul-80 1203 HAYES at SRI-KL pressures
C00051 00005 ∂01-Jul-80 1635 FFL
C00052 00006 ∂01-Jul-80 1640 FFL Luckham NSF
C00053 00007 ∂01-Jul-80 1825 JK addendum to sundays note
C00054 00008 ∂02-Jul-80 0931 HVA SRA Summer Support
C00055 00009 ∂02-Jul-80 0936 RAJ.REDDY at CMU-10B RECOMMENDATION LETTER
C00057 00010 ∂02-Jul-80 1040 doyle at MIT-AI (Jon Doyle) Mail coming
C00058 00011 ∂02-Jul-80 1048 LGC ARPA Report
C00059 00012 ∂02-Jul-80 1332 BALZER at USC-ISIB Re: AI and Philosophy session at AAAI conference
C00061 00013 ∂02-Jul-80 1502 ENGELMORE at USC-ISI FY81 funding
C00064 00014 ∂02-Jul-80 1545 LES Levinthal travel
C00066 00015 ∂03-Jul-80 0448 REM at MIT-MC (Robert Elton Maas) Analysis of Go endgames
C00071 00016 ∂03-Jul-80 0732 REG
C00072 00017 ∂03-Jul-80 0811 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Giant File Server
C00075 00018 ∂03-Jul-80 2039 RPG $$$ for Waltz
C00076 00019 ∂04-Jul-80 0124 CSD.GARDNER at SU-SCORE Hacker and Raz book
C00077 00020 ∂04-Jul-80 0348 REM via AMES-TIP
C00078 00021 ∂04-Jul-80 1254 CSL.FB at SU-SCORE Re: Giant File Server
C00080 00022 ∂06-Jul-80 0604 REM via AMES-TIP ENDGO1.WRU
C00082 00023 ∂06-Jul-80 1310 100 : clt susie
C00083 00024 ∂07-Jul-80 0000 JMC*
C00084 00025 ∂07-Jul-80 1623 FFL Wire service program to Mrs. Henson
C00085 00026 ∂07-Jul-80 1659 RMS at MIT-AI (Richard M. Stallman)
C00086 00027 ∂07-Jul-80 1708 SEK
C00087 00028 ∂07-Jul-80 1815 CSD.GOLUB at SU-SCORE
C00089 00029 ∂07-Jul-80 2019 JK nsf grant
C00091 00030 ∂08-Jul-80 0000 JMC*
C00092 00031 ∂08-Jul-80 1053 HVA Your signature needed
C00093 00032 ∂08-Jul-80 1436 CLT
C00094 00033 ∂09-Jul-80 0833 FFL
C00095 00034 ∂09-Jul-80 0900 JMC*
C00096 00035 ∂09-Jul-80 0900 FFL
C00097 00036 ∂09-Jul-80 0953 Buchanan at SUMEX-AIM AAAI panel on AI & Philosophy
C00099 00037 ∂09-Jul-80 1015 LES Keyboard krock
C00101 00038 ∂09-Jul-80 1026 Buchanan at SUMEX-AIM AAAI panel
C00102 00039 ∂09-Jul-80 1142 SJC Thanks and fiche
C00104 00040 ∂09-Jul-80 1406 BIS maclisp support after rpg
C00105 00041 ∂09-Jul-80 1451 RPG Malignment
C00109 00042 ∂09-Jul-80 1510 RPG
C00112 00043 ∂10-Jul-80 0951 FFL Second call from Keith Davidson of Boston, freelance journalist
C00113 00044 ∂10-Jul-80 1350 TOB
C00114 00045 ∂10-Jul-80 1706 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Gene/Joe/NA problem
C00115 00046 ∂10-Jul-80 1817 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Giant File Server (Galaxy-wide File System)
C00117 00047 ∂10-Jul-80 1733 RPG Plans
C00120 00048 ∂11-Jul-80 0002 DLW at MIT-AI (Daniel L. Weinreb)
C00122 00049 ∂12-Jul-80 1025 Sandelin at SUMEX-AIM giant file server
C00123 00050 ∂12-Jul-80 1645 AAM Optical character recognition
C00126 00051 ∂13-Jul-80 1908 REM at MIT-MC (Robert Elton Maas)
C00142 00052 ∂15-Jul-80 2035 REM at MIT-MC (Robert Elton Maas)
C00146 00053 ∂16-Jul-80 0130 POURNE at MIT-MC (Jerry E. Pournelle)
C00147 00054 ∂16-Jul-80 1019 FFL
C00148 00055 ∂16-Jul-80 1132 FFL
C00149 00056 ∂16-Jul-80 1138 SJC Fiche
C00151 00057 ∂16-Jul-80 1231 HAYES at SRI-KL variadic functions
C00153 00058 ∂16-Jul-80 2133 RWW
C00161 00059 ∂17-Jul-80 0148 POURNE at MIT-MC (Jerry E. Pournelle)
C00162 00060 ∂17-Jul-80 1042 HVA Tel. Call
C00163 00061 ∂17-Jul-80 1539 FFL
C00164 00062 ∂17-Jul-80 1817 Woods at PARC-MAXC Re: dired or the like
C00167 00063 ∂17-Jul-80 1823 BIS SAIL keyboards
C00170 00064 ∂17-Jul-80 2150 BIS
C00171 00065 ∂18-Jul-80 0813 DEK PUB???
C00172 00066 ∂18-Jul-80 0941 FFL Mr. Tang's salary
C00173 00067 ∂18-Jul-80 1359 FFL
C00174 00068 ∂18-Jul-80 1741 REM at MIT-MC (Robert Elton Maas) ENDGO1 >
C00179 00069 ∂18-Jul-80 1741 REM at MIT-MC (Robert Elton Maas) SciAm article
C00183 00070 ∂19-Jul-80 0208 POURNE at MIT-MC (Jerry E. Pournelle)
C00186 00071 ∂19-Jul-80 0321 POURNE@MIT-MC (Sent by COMSAT@MIT-MC)
C00187 00072 ∂19-Jul-80 0321 POURNE@MIT-MC (Sent by COMSAT@MIT-MC)
C00188 00073 ∂21-Jul-80 0146 POURNE at MIT-MC (Jerry E. Pournelle) yea,verily...
C00193 00074 ∂21-Jul-80 0151 ME
C00194 00075 ∂21-Jul-80 0157 POURNE at MIT-MC (Jerry E. Pournelle)
C00196 00076 ∂21-Jul-80 0335 REM at MIT-MC (Robert Elton Maas) Go endgame algorithm update
C00198 00077 ∂21-Jul-80 1545 Bmoore at SRI-KL Accomodations for FMAI
C00216 00078 ∂22-Jul-80 0910 JRA lisp paper
C00217 00079 Date: 21 Jul 1980 1323-PDT
C00234 00080 ∂23-Jul-80 0856 FFL
C00235 00081 ∂23-Jul-80 0900 JMC*
C00237 00082 ∂23-Jul-80 1258 LES TI terminal
C00238 00083 ∂24-Jul-80 0734 CSD.BSCOTT at SU-SCORE Les Earnest
C00251 00084 ∂24-Jul-80 0905 BALZER at USC-ISIB TITLE FOR YOUR ''CHALLENGES AND PERSPECTIVES'' TALK
C00252 00085 ∂24-Jul-80 1046 Bmoore at SRI-KL Formal Methods Workshop Update
C00258 00086 ∂24-Jul-80 1119 Wiederhold at SUMEX-AIM Re: giant file server
C00260 00087 ∂24-Jul-80 1326 TOB update
C00263 00088 ∂24-Jul-80 1349 Bmoore at SRI-KL Formal Methods Workshop Update
C00269 00089 ∂24-Jul-80 1601 Bmoore at SRI-KL Panel discussion on AI and Philosophy
C00271 00090 ∂24-Jul-80 1756 HEDRICK at RUTGERS Psychology Today article
C00284 00091 ∂24-Jul-80 1757 Roy Marantz <MARANTZ at RUTGERS> Re: Psychology Today article
C00287 00092 ∂24-Jul-80 1938 Mark Crispin <Admin.MRC at SU-SCORE> Re: Psychology Today article
C00295 00093 ∂24-Jul-80 1953 JK proof checker
C00297 00094 ∂24-Jul-80 2344 RMS at MIT-AI (Richard M. Stallman)
C00298 00095 ∂25-Jul-80 0108 REM via SU-TIP
C00299 00096 ∂25-Jul-80 0838 Sandelin at SUMEX-AIM giant file server
C00301 00097 ∂25-Jul-80 0913 REG
C00302 00098 ∂25-Jul-80 1513 Sandelin at SUMEX-AIM
C00303 00099 ∂25-Jul-80 1619 Sandelin at SUMEX-AIM queueing
C00304 00100 ∂25-Jul-80 1639 FFL
C00305 00101 ∂26-Jul-80 1918 HPM via parts unknown a belated thanks
C00306 00102 ∂27-Jul-80 0831 Richard Jay Solomon <Solomon at MIT-Multics> visiting SU-AI
C00309 00103 ∂28-Jul-80 1154 FFL
C00310 00104 ∂28-Jul-80 1321 FFL
C00311 00105 ∂28-Jul-80 1400 JMC*
C00312 00106 ∂28-Jul-80 1444 FFL
C00313 00107 ∂29-Jul-80 0924 Bmoore at SRI-KL Alan Thompson (?)
C00314 00108 ∂29-Jul-80 1300 JMC*
C00315 00109 ∂29-Jul-80 1627 CJS Classnotes
C00316 00110 ∂29-Jul-80 1811 REM
C00318 00111 ∂30-Jul-80 1048 FFL
C00319 00112 ∂30-Jul-80 1300 JMC*
C00320 00113 ∂30-Jul-80 1458 CJS Classnotes for next quarter
C00321 00114 ∂30-Jul-80 1523 CSL.CRC.LYD at SU-SCORE Mr. Hao Dong
C00323 00115 ∂30-Jul-80 1704 FFL
C00324 00116 ∂30-Jul-80 1835 LGC Talk Abstract
C00325 00117 ∂31-Jul-80 0033 MLB undecidable production system?
C00327 00118 ∂31-Jul-80 0709 Mark Crispin <Admin.MRC at SU-SCORE> Dialnet TELNET protocol
C00332 00119 ∂31-Jul-80 1140 LGC 1. Abstract Revision 2. Other Developments
C00337 00120 ∂31-Jul-80 1155 FFL
C00338 00121 ∂31-Jul-80 2200 JMC*
C00339 00122 ∂01-Aug-80 0106 ME auto-detach and auto-logout
C00347 00123 ∂01-Aug-80 0430 RPG Decidability
C00349 00124 ∂01-Aug-80 0715 JRA conference party
C00350 00125 ∂01-Aug-80 1000 CG thesis
C00351 00126 ∂01-Aug-80 1013 FFL
C00352 00127 ∂01-Aug-80 1606 FFL Alison Bishop called. 7-2203.
C00353 00128 ∂01-Aug-80 2245 ME NS dates
C00354 00129 ∂02-Aug-80 0900 JMC*
C00369 00130 ∂02-Aug-80 1337 Wiederhold at SUMEX-AIM Giant file server initial meeting.
C00376 00131 ∂03-Aug-80 1427 ME
C00377 00132 ∂03-Aug-80 1524 REM via AMES-TIP Go endgame
C00379 00133 ∂04-Aug-80 0446 RAHE at USC-ECL Your Human-nets piece
C00385 00134 ∂04-Aug-80 1226 kirkley at DNSRI chess.
C00396 00135 ∂05-Aug-80 2046 JK
C00398 00136 ∂05-Aug-80 2200 BUD ben
C00399 00137 ∂05-Aug-80 2304 JK
C00400 00138 ∂07-Aug-80 1340 FFL
C00401 00139 ∂07-Aug-80 1349 FFL
C00402 00140 ∂07-Aug-80 1355 FFL
C00403 00141 ∂07-Aug-80 1319 BIS MACLISP future
C00407 00142 ∂07-Aug-80 2008 JMC
C00408 00143 ∂07-Aug-80 1937 LLW CLEAN SWEEP CELEBRATION
C00410 00144 ∂08-Aug-80 0044 POURNE at MIT-MC (Jerry E. Pournelle) did you get this?
C00414 00145 ∂08-Aug-80 0051 POURNE at MIT-MC (Jerry E. Pournelle) survival of freedom
C00416 00146 ∂08-Aug-80 0058 POURNE at MIT-MC (Jerry E. Pournelle)
C00418 00147 ∂08-Aug-80 0059 POURNE at MIT-MC (Jerry E. Pournelle)
C00420 00148 ∂08-Aug-80 0113 POURNE at MIT-MC (Jerry E. Pournelle) getting together...
C00422 00149 ∂08-Aug-80 0215 DCL Maclisp
C00423 00150 ∂08-Aug-80 0918 RWW
C00442 00151 ∂08-Aug-80 1259 RMS at MIT-AI (Richard M. Stallman)
C00444 00152 ∂08-Aug-80 1335 FFL
C00445 00153 ∂08-Aug-80 1423 ME files on [1,3]
C00446 00154 ∂08-Aug-80 1429 Nilsson at SRI-KL AIJ Paper
C00447 00155 ∂08-Aug-80 1612 Wiederhold at SUMEX-AIM A proposal first draft for a giant file server
C00470 00156 ∂09-Aug-80 1132 Wiederhold at SUMEX-AIM (Response to message)
C00473 00157 ∂09-Aug-80 1255 REG
C00474 00158 ∂09-Aug-80 1759 LWE
C00475 00159 ∂10-Aug-80 0216 RPG
C00489 00160 ∂10-Aug-80 0235 RPG Value
C00491 00161 ∂10-Aug-80 2010 RPG Offer
C00493 00162 ∂11-Aug-80 1438 FFL
C00494 00163 ∂11-Aug-80 2201 LLW . . .The Tents Of The Wicked
C00495 00164 ∂12-Aug-80 0213 LWE
C00496 00165 ∂12-Aug-80 0828 FFL
C00502 00166 ∂12-Aug-80 1000 JMC*
C00503 00167 ∂12-Aug-80 1718 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
C00504 00168 ∂12-Aug-80 1926 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) meeting next Monday
C00505 00169 ∂12-Aug-80 2001 REG
C00506 00170 ∂12-Aug-80 2243 LWE saving digests
C00507 00171 ∂13-Aug-80 0843 JRA SAIL USAGE
C00516 00172 ∂13-Aug-80 1459 100 : REM via SU-TIP PCNET/DIALNET
C00517 00173 ∂13-Aug-80 1524 FFL
C00523 00174 ∂14-Aug-80 1221 CG Kreisel's address
C00524 00175 ∂14-Aug-80 1458 RWW working
C00525 00176 ∂14-Aug-80 1927 RWW
C00526 00177 ∂15-Aug-80 0245 LWE
C00527 00178 ∂15-Aug-80 0900 JMC*
C00528 00179 ∂15-Aug-80 1105 100 : REM via AMES-TIP Fountain of canon balls for levitating a skytower
C00531 00180 ∂15-Aug-80 1119 FFL
C00532 00181 ∂15-Aug-80 1358 FFL
C00533 00182 ∂15-Aug-80 1502 CJS classnotes for the Fall quarter
C00534 00183 ∂15-Aug-80 1654 OWE Visitors from Norway
C00536 00184 ∂16-Aug-80 1148 Kanerva at SUMEX-AIM Menus
C00541 00185 ∂17-Aug-80 0050 ME DM3025 questionaire
C00543 00186 ∂17-Aug-80 1606 MINSKY at MIT-AI (Marvin Minsky)
C00544 00187 ∂17-Aug-80 2059 RWG
C00545 00188 ∂17-Aug-80 2104 RWG
C00549 00189 ∂17-Aug-80 2218 RWG
C00551 00190 ∂17-Aug-80 2354 PMF Eschewing Kinky Tower Designs
C00553 00191 ∂18-Aug-80 0949 Feigenbaum at SUMEX-AIM visit of important person from ONR
C00556 00192 ∂18-Aug-80 1140 LGC PUB Problem
C00557 00193 ∂18-Aug-80 1848 Wiederhold at SUMEX-AIM For your info - data collected about our computing for NRC .
C00562 00194 ∂18-Aug-80 2106 REM at MIT-MC (Robert Elton Maas) Female person you asked about putting up
C00566 00195 ∂19-Aug-80 1134 FFL Phone call from Paul Raeburn of Associated Press
C00568 00196 ∂19-Aug-80 1219 TOB VAX
C00569 00197 ∂19-Aug-80 1425 FFL
C00570 00198 ∂19-Aug-80 1432 FFL
C00571 00199 ∂19-Aug-80 1441 FFL Two parties connected with alum association activities next week
C00572 00200 ∂19-Aug-80 1448 FFL
C00579 00201 ∂20-Aug-80 0037 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: VAX
C00580 00202 ∂20-Aug-80 0127 RMS at MIT-AI (Richard M. Stallman)
C00581 00203 ∂20-Aug-80 1307 DRUFFEL at USC-ISI VAX cycles
C00583 00204 ∂20-Aug-80 1726 FWH PV+A Seminar
C00585 00205 ∂20-Aug-80 2356 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: VAX cycles
C00603 00206 ∂21-Aug-80 0739 JRA
C00604 00207 ∂21-Aug-80 0814 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
C00605 00208 ∂21-Aug-80 1005 TOB
C00610 00209 ∂21-Aug-80 1144 FFL Professor Ketonen's apppointment
C00611 00210 ∂21-Aug-80 1438 Denny Brown <CSD.DBROWN at SU-SCORE> Scheduling of the AI courses.
C00614 00211 ∂21-Aug-80 1448 Denny Brown <CSD.DBROWN at SU-SCORE> The courses and degrees blurb re CSAI
C00622 00212 ∂21-Aug-80 1453 Denny Brown <CSD.DBROWN at SU-SCORE> Sample CSAI program
C00625 00213 ∂21-Aug-80 1528 Denny Brown <CSD.DBROWN at SU-SCORE> Course schedule issue resolved
C00627 00214 ∂ extract from Human-nets with ITT ad from Sept. Sci Am
C00642 00215 ∂22-Aug-80 1000 JMC*
C00643 00216 ∂22-Aug-80 2053 RPG Last Attempt
C00644 00217 ∂22-Aug-80 2113 Rindfleisch at SUMEX-AIM FYI RE FILE SERVER DISKS
C00649 00218 ∂22-Aug-80 2332 MINSKY at MIT-ML (Marvin Minsky) tower stability
C00655 00219 ∂23-Aug-80 1711 Wiederhold at SUMEX-AIM Re: FYI RE FILE SERVER DISKS
C00672 00220 ∂25-Aug-80 1426 FFL Call from Bennett Karmin, I.B.M. 408 256 5307
C00679 00221 ∂26-Aug-80 0851 Nilsson at SRI-KL (Response to message)
C00680 00222 ∂26-Aug-80 1640 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) [John McCarthy <JMC at SU-AI>:]
C00682 00223 ∂26-Aug-80 1730 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
C00698 00224 ∂27-Aug-80 1020 FFL
C00699 00225 ∂27-Aug-80 1645 REM via SU-TIP Deaf access to SU Comp.Sci. ?
C00701 00226 ∂27-Aug-80 2353 JK disk allocation
C00702 00227 ∂28-Aug-80 0111 Boyer at SRI-KL (Bob) Meeting
C00710 00228 ∂28-Aug-80 1652 DES
C00712 00229 ∂28-Aug-80 2315 LLW Raining Momentum On The Just And The Unjust
C00715 00230 ∂28-Aug-80 2354 GG leaving
C00724 00231 ∂29-Aug-80 0954 FFL Visit of Prof. Iwata on Sept. 25 and 26
C00725 00232 ∂29-Aug-80 1002 CSD.BARB at SU-SCORE Visitor from ONR
C00727 00233 ∂29-Aug-80 1055 FFL Ketonen, Doyle, Goad, Talcott office arrangements
C00728 00234 ∂29-Aug-80 1304 rwg at MIT-AI (Bill Gosper)
C00730 00235 ∂29-Aug-80 1324 FFL REMINDER
C00731 00236 ∂29-Aug-80 1325 FFL REMINDER
C00732 00237 ∂29-Aug-80 1450 CSD.BARB at SU-SCORE ONR Person
C00733 00238 ∂29-Aug-80 1952 LLW Vacuum Balloonery
C00739 00239 ∂31-Aug-80 0242 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #60
C00761 00240 ∂31-Aug-80 1655 JK ekl
C00762 00241 ∂31-Aug-80 2308 JK
C00765 00242 ∂01-Sep-80 0000 JMC*
C00766 00243 ∂01-Sep-80 0640 SOL via MIT-TIP ap digests
C00767 00244 ∂01-Sep-80 0742 SOL via MIT-TIP news saving
C00781 00245 ∂01-Sep-80 0802 RPG
C00784 00246 ∂01-Sep-80 1733 RPG Pattern matcher inverse
C00786 00247 ∂02-Sep-80 0229 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #62
C00807 00248 ∂02-Sep-80 0858 FFL Sigma Xi
C00808 00249 ∂02-Sep-80 1056 FFL re papers to give to Barbara Roberts
C00810 00250 ∂02-Sep-80 1102 FFL
C00811 00251 ∂02-Sep-80 1300 JMC*
C00812 00252 ∂02-Sep-80 1308 teitelman at PARC-MAXC Re: FLIP
C00814 00253 ∂02-Sep-80 1353 Masinter at PARC-MAXC Re: pattern matching
C00816 00254 ∂02-Sep-80 1608 FFL
C00818 00255 ∂02-Sep-80 1712 Masinter at PARC-MAXC
C00822 00256 ∂02-Sep-80 2023 JK ekl
C00824 00257 ∂02-Sep-80 2239 JK ekl
C00826 00258 ∂03-Sep-80 1050 Denny Brown <CSD.DBROWN at SU-SCORE> Re: Department funds for Waltz' expenses
C00830 00259 ∂03-Sep-80 1344 FFL
C00832 00260 ∂03-Sep-80 1308 FFL
C00833 00261 ∂03-Sep-80 1215 TOB
C00838 00262 ∂03-Sep-80 1615 RPG My Reading Committee:
C00840 00263 ∂03-Sep-80 1644 CSD.ULLMAN at SU-SCORE
C00842 00264 ∂03-Sep-80 2242 RPG I don't want to appear impertinent, but...
C00844 00265 ∂03-Sep-80 2249 TOB request for display purchase
C00848 00266 ∂04-Sep-80 0033 RPG Load
C00849 00267 ∂04-Sep-80 0114 ME DMs from CASBS getting new modems
C00850 00268 ∂04-Sep-80 0141 ME modems for DMs from CASBS
C00852 00269 ∂04-Sep-80 0856 TOB Happy birthday
C00853 00270 ∂04-Sep-80 0912 CSD.ULLMAN at SU-SCORE Bibliographies
C00855 00271 ∂04-Sep-80 0918 FFL
C00856 00272 ∂04-Sep-80 1100 JMC*
C00857 00273 ∂04-Sep-80 1631 CSD.FRIEDL at SU-SCORE FILE SERVER PROPOSED PROJECT
C00858 00274 ∂04-Sep-80 2039 MRC
C00859 00275 ∂04-Sep-80 2258 LES
C00860 00276 ∂05-Sep-80 0046 POURNE at MIT-MC (Jerry E. Pournelle)
C00861 00277 ∂05-Sep-80 0056 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #65
C00884 00278 ∂05-Sep-80 0438 Robert Elton Maas <REM at MIT-MC>
C00888 00279 ∂05-Sep-80 0439 REM at MIT-MC (Robert Elton Maas)
C00894 00280 ∂05-Sep-80 0923 FFL
C00895 00281 ∂05-Sep-80 1237 TOB trip
C00896 00282 ∂05-Sep-80 1311 RPG `Research activity'
C00897 00283 ∂05-Sep-80 1319 Nilsson at SRI-KL Book
C00898 00284 ∂05-Sep-80 1815 REM via SU-TIP
C00900 00285 ∂06-Sep-80 0316 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #66
C00922 00286 ∂07-Sep-80 0047 ME ns
C00924 00287 ∂07-Sep-80 0114 ME NS remembers...
C00926 00288 ∂07-Sep-80 0248 RWG at MIT-MC (Bill Gosper) evening juggling of differential eqns
C00928 00289 ∂08-Sep-80 0404 REM via SU-TIP Your ideas on concentrating on immediate instead of sheer-guesswork
C00929 00290 ∂08-Sep-80 1100 FFL
C00930 00291 ∂08-Sep-80 1415 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Re: Faculty interests file
C00931 00292 ∂08-Sep-80 2353 ROB AP vs NYT
C00947 00293 ∂09-Sep-80 1612 RWG
C00948 00294 ∂09-Sep-80 1634 RWG at MIT-MC (Bill Gosper)
C00950 00295 ∂09-Sep-80 1748 DPB Offices and desks in MJ
C00953 00296 ∂09-Sep-80 2232 kirkley at sri-unix Hello... but...
C00956 00297 ∂10-Sep-80 0353 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #70
C00965 00298 ∂11-Sep-80 0034 POURNE at MIT-MC (Jerry E. Pournelle) deadlines
C00978 00299 ∂11-Sep-80 0835 FFL
C00979 00300 ∂11-Sep-80 1650 JRA
C00983 00301 ∂12-Sep-80 1121 FFL
C00984 00302 ∂13-Sep-80 0132 POURNE@MIT-MC
C00987 00303 ∂13-Sep-80 0722 HUMAN-NETS@MIT-AI HUMAN-NETS AM Digest V2 #73
C01008 00304 ∂13-Sep-80 0718 REM via SU-TIP Progress on XGP-->PRESS program
C01011 00305 ∂13-Sep-80 0900 JMC*
C01012 00306 ∂13-Sep-80 1533 JK
C01014 00307 ∂14-Sep-80 0331 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #74
C01025 00308 ∂14-Sep-80 1532 VRP dialnet
C01026 00309 ∂14-Sep-80 1616 Buchanan at SUMEX-AIM Natl Rsch Council position paper
C01034 00310 ∂14-Sep-80 2100 JMC*
C01040 00311 ∂15-Sep-80 0846 REM via SU-TIP Status of XGP/PRESS program
C01042 00312 ∂15-Sep-80 1102 CSL.LAB.SALLY COmputer Forum Memo
C01047 00313 ∂15-Sep-80 1112 Denny Brown <CSD.DBROWN at SU-SCORE> Richard's office
C01049 00314 ∂15-Sep-80 1115 FFL
C01050 00315 ∂15-Sep-80 1421 PHY McCarthy book
C01052 00316 ∂15-Sep-80 1616 kirkley at sri-unix tnx for the reply.
C01055 00317 ∂16-Sep-80 0328 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #76
C01072 00318 ∂16-Sep-80 0656 REM via SU-TIP Progress
C01073 00319 ∂16-Sep-80 1311 FFL
C01074 00320 ∂16-Sep-80 1343 FFL
C01075 00321 ∂16-Sep-80 1418 Denny Brown <CSD.DBROWN at SU-SCORE> New PhD student summary info.
C01079 00322 ∂16-Sep-80 1421 CSD.ULLMAN at SU-SCORE Travel for thesis committee
C01081 00323 ∂16-Sep-80 1506 CSD.ULLMAN at SU-SCORE
C01082 00324 ∂16-Sep-80 1520 ME macros upon login
C01083 00325 ∂16-Sep-80 1741 Buchanan at SUMEX-AIM Natl Rsch Council Panel
C01084 00326 ∂16-Sep-80 1742 Buchanan at SUMEX-AIM NRC draft (4 pages)
C01093 00327 ∂16-Sep-80 1842 PLS POP2
C01096 00328 ∂16-Sep-80 2000 JMC*
C01097 00329 ∂16-Sep-80 2107 CSL.FB at SU-SCORE
C01099 00330 ∂17-Sep-80 0023 kirkley at sri-unix
C01101 00331 ∂17-Sep-80 0722 CSL.FB at SU-SCORE
C01102 00332 ∂17-Sep-80 0852 FFL
C01103 00333 ∂17-Sep-80 1604 CSD.JEANIE at SU-SCORE Faculty Meeting
C01105 00334 ∂17-Sep-80 1643 CSD.GENESERETH at SU-SCORE Siglunch, Friday the 19th
C01109 00335 ∂17-Sep-80 2343 100 : kspace via AMES-TIP
C01110 00336 ∂17-Sep-80 2347 100 : kspace via AMES-TIP
C01111 00337 ∂18-Sep-80 0336 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #78
C01140 00338 ∂18-Sep-80 0900 JMC*
C01141 00339 ∂18-Sep-80 0920 CSD.GOLUB at SU-SCORE Re: Faculty Meeting
C01142 00340 ∂18-Sep-80 1141 ROB Pies and Wires
C01143 00341 ∂18-Sep-80 1630 FFL
C01145 00342 ∂19-Sep-80 0014 100 : John Kelley (kspace) via AMES-TIP
C01146 00343 ∂19-Sep-80 0103 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #79
C01170 00344 ∂19-Sep-80 0819 Kanerva at SUMEX-AIM DEC-20 business plan
C01172 00345 ∂19-Sep-80 1123 FFL
C01173 00346 ∂19-Sep-80 1531 FFL
C01174 00347 ∂19-Sep-80 2108 RMS at MIT-AI (Richard M. Stallman) buyer for emacs
C01176 00348 ∂20-Sep-80 0146 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #80
C01198 00349 ∂20-Sep-80 1133 Kanerva at SUMEX-AIM Business plan for a text system
C01200 00350 ∂21-Sep-80 0813 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #81
C01230 00351 ∂22-Sep-80 0201 LWE debate
C01232 00352 ∂22-Sep-80 0540 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #82
C01255 00353 ∂22-Sep-80 1102 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Honeywell wants a speaker in Phoenix
C01257 00354 ∂22-Sep-80 1050 FFL
C01258 00355 ∂22-Sep-80 2042 Kanerva at SUMEX-AIM DEC-20 configuration
C01260 00356 ∂23-Sep-80 0337 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #83
C01283 00357 ∂23-Sep-80 1020 FFL
C01284 00358 ∂23-Sep-80 1258 RPG Damages
C01285 00359 ∂23-Sep-80 1439 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
C01286 00360 ∂23-Sep-80 1449 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Evans & Sutherland
C01289 00361 ∂23-Sep-80 1524 PAT lib.lst
C01290 00362 ∂23-Sep-80 1536 Buchanan at SUMEX-AIM Natl Rsch Council mtg
C01292 00363 ∂23-Sep-80 1900 JMC*
C01293 00364 ∂24-Sep-80 0321 REM at MIT-MC (Robert Elton Maas)
C01296 00365 ∂24-Sep-80 0359 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #84
C01321 00366 ∂24-Sep-80 0806 FFL
C01322 00367 ∂24-Sep-80 0810 FFL
C01323 00368 ∂24-Sep-80 0908 Baskett at PARC-MAXC Re: Evans & Sutherland
C01325 00369 ∂24-Sep-80 1032 FFL
C01326 00370 ∂24-Sep-80 2252 JK
C01327 00371 ∂25-Sep-80 0131 RWG at MIT-MC (Bill Gosper)
C01330 00372 ∂25-Sep-80 0539 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #85
C01356 00373 ∂25-Sep-80 0900 JMC*
C01357 00374 ∂25-Sep-80 0956 PJH via ROCHESTER
C01359 00375 ∂25-Sep-80 1107 CSD.JEANIE at SU-SCORE Faculty Meeting Agenda
C01361 00376 ∂25-Sep-80 1155 FFL
C01362 00377 ∂25-Sep-80 1317 CSD.BSCOTT at SU-SCORE NSF Supplement for Doyle
C01364 00378 ∂25-Sep-80 1443 CT
C01365 00379 ∂25-Sep-80 1441 JK Nozaki's paper
C01366 00380 ∂25-Sep-80 1859 Denny Brown <CSD.DBROWN at SU-SCORE> CS200 - request for speakers
C01370 00381 ∂25-Sep-80 2003 TOB
C01377 00382 ∂26-Sep-80 0354 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #86
C01402 00383 ∂26-Sep-80 0803 CSD.BSCOTT at SU-SCORE
C01403 00384 ∂26-Sep-80 1030 JRA
C01405 00385 ∂26-Sep-80 1330 FFL
C01406 00386 ∂26-Sep-80 1412 CSL.CRC.LYD at SU-SCORE Visitors
C01411 00387 ∂26-Sep-80 1646 FFL
C01412 00388 ∂27-Sep-80 0330 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #87
C01439 00389 ∂28-Sep-80 0123 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #88
C01465 00390 ∂28-Sep-80 1000 JMC*
C01467 00391 ∂28-Sep-80 2336 RMS at MIT-AI (Richard M. Stallman)
C01471 00392 ∂29-Sep-80 0841 FFL
C01472 00393 ∂29-Sep-80 0928 CSD.ULLMAN at SU-SCORE arpa contract
C01474 00394 ∂29-Sep-80 0930 FFL SEMINAR - John MccCarthy
C01475 00395 ∂29-Sep-80 0940 FFL Seminar
C01476 00396 ∂29-Sep-80 0941 FFL
C01477 00397 ∂29-Sep-80 1018 FFL SEMINAR -JOHN MCCARTHY
C01478 00398 ∂29-Sep-80 1105 Keith A. Lantz <CSL.LANTZ at SU-SCORE> Re: arpa contract
C01480 00399 ∂29-Sep-80 1139 CSL.FB at SU-SCORE Re: arpa contract
C01481 00400 ∂29-Sep-80 1343 CSD.ULLMAN at SU-SCORE
C01482 00401 ∂29-Sep-80 1628 CG
C01483 00402 ∂29-Sep-80 1952 AAM Fenaughty & Griffith
C01484 00403 ∂30-Sep-80 0107 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #90
C01508 00404 ∂30-Sep-80 0851 CSD.FROG at SU-SCORE Welcoming Lunch for New CSD Graduate Students
C01509 00405 ∂30-Sep-80 1411 Nilsson at SRI-KL TEX work
C01510 00406 ∂30-Sep-80 1421 Nilsson at SRI-KL Book
C01511 00407 ∂30-Sep-80 1426 DRH seminar schedule
C01512 00408 ∂30-Sep-80 1704 SEK phone number
C01513 00409 ∂30-Sep-80 2356 VRP via Dialnet dialnet
C01514 00410 ∂01-Oct-80 0023 VRP via Dialnet dialnet
C01515 00411 ∂01-Oct-80 0327 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #91
C01543 00412 ∂01-Oct-80 0909 VRP via Dialnet Dialnet performance measurement
C01545 00413 ∂01-Oct-80 1014 JMC*
C01546 00414 ∂01-Oct-80 1012 VRP via Dialnet Dialnet performance
C01547 00415 ∂01-Oct-80 1108 FFL
C01548 00416 ∂01-Oct-80 1227 Boyer at SRI-KL (Bob) Recommendation Request
C01549 00417 ∂01-Oct-80 1238 VRP via Dialnet Dialnet
C01551 00418 ∂01-Oct-80 1339 FFL
C01552 00419 ∂01-Oct-80 1755 FWH PV+A Seminar
C01554 00420 ∂01-Oct-80 2258 MINSKY at MIT-AI (Marvin Minsky) hole through world
C01557 00421 ∂01-Oct-80 2309 RWG at MIT-MC (Bill Gosper)
C01559 00422 ∂02-Oct-80 0024 POURNE at MIT-MC (Jerry E. Pournelle) article
C01560 00423 ∂02-Oct-80 0234 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #92
C01583 00424 ∂02-Oct-80 0809 FFL Expense report for David Waltz
C01584 00425 ∂02-Oct-80 0811 FFL Meeting day and time for Seminar
C01585 00426 ∂02-Oct-80 0813 FFL
C01586 00427 ∂02-Oct-80 0948 FWH PV+A Seminar
C01588 00428 ∂02-Oct-80 1001 SOL via MIT-TIP we be inBay Area oct 16-17
C01589 00429 ∂02-Oct-80 1006 JMC*
C01590 00430 ∂02-Oct-80 1011 TOB ARPA
C01591 00431 ∂02-Oct-80 1100 JMC*
C01592 ENDMK
C⊗;
∂01-Jul-80 0616 ENGELMORE at USC-ISI Additional support
Date: 1 Jul 1980 0609-PDT
Sender: ENGELMORE at USC-ISI
Subject: Additional support
From: ENGELMORE at USC-ISI
To: LES at SU-AI, JMC at SU-AI
Message-ID: <[USC-ISI] 1-Jul-80 06:09:50.ENGELMORE>
John and Les,
Following the phone call from Les yesterday, I talked to Bob Kahn
about the two alternatives for addtional support, i.e. new task with new
proposal vs. accelrated funding in FY81. He felt it really wasn't worth the
paperwork to process a new proposl that would essentially support one
person, and was content to go with the second alternative, provided I
can find the money in the budget. Since I have that problem in either case,
we might as well go with the accelerated funding idea, as that involves so
much less paperwork for you and for me. Just so we understand what that
option means, I will, assuming I can find the money, put a total of $510K
out on the contract for the work in AI and Image Understanding, rather than
$410K as originally planned ($200K is for the IU work), and you will be out
of funds as of Sept. 30, 1981.
As to the availability of funds, I will try to determine that as quickly as
I can. It will require some juggling, as I don't have any uncommitted funds
for FY81 right now. That maneuver will require Kahn's approval.
Bob
∂01-Jul-80 1157 FFL Telephone call from Ned Irons of Colorado
To: JMC, FFL
Will call again late this afternoon, but if you are not in at tat time,
would you please call him at 303 586 3834.
∂01-Jul-80 1203 HAYES at SRI-KL pressures
Date: 1 Jul 1980 1131-PDT
From: HAYES at SRI-KL
Subject: pressures
To: mccarthy at SU-AI
thanks for your interesting note. my immediate reaction was that
your formulation of the dependence looked more or less right, but
on reflection i am not sure what stateof is supposed to mean, or what
the entity called s is supposed to be. my reformulation would be
something like this, exhibiting the temporal nature of the
dependence explicitly:
all g) gas g imp
exists f) all x) sample(x,g) imp all t) density(x,t)=f(press(x,t),temp(x,t))
where i have put the existential quantifier at the beginning, since this
functional relationship is presumably lawlike, ie constant over many samples. ;
however, i agree this minor twist does not adress the serious problem
wrt default reasoning.
to apply circumscription one would need a formulation which said, there
exist these two dependencies, ie density depends on temperature and also
on pressure. One can just grasp the nettle at this point and introduce
an explicit theory of dependency, with a depends-on relation between
two fluents. One needs the axiom scheme
[all d) depends-on(h,d) iff d=f1ord=f2or...d=fn]
iff [exists f) all t) h(t) = f(f1(t),...,fn(t)) ]
in order to get from that to the functional formulation, but this is
replaceable by a simple enumerable set of firstorder axionms with a
regular structure.
the problem is, what does depends-onn MEAN, or equivalently, what axioms
can one write about it? here are a few vague suggestions.
1. f depends-on g means that ,ceterus paribus, a change in g leads
to a change in f
2. .....causes a change in f
3. .....means that there is some causal chain ("mechanism") linking
changes in g to changes in f
The trouble with the first two is that c.p. clause, and the last is an attempt
to get around it. notice that we can just assert the existence of the
causal link, knowing nothing od its nature, and also that other factors
might interfere with the operation of such a causal chain, which swallows ;
th cetpar condition into a general feature of causality statements.
the ontological committment here is beginning to get even me a little
dizzy, but i cant see any other way to make statements of functional
dependency accessible to circumscription or default reasoning. we have
bring them out into the axiomatisation. or else, maybe, modify the
circumscription machinery especially for this case? it might be worth
doing that, since after all the number of arguments a function has is
arguably a logical question.
another thought: suppose our logic had variadic function symbols
(ie variables over functions of many arities) and the machinery
for talking about arities, eg arity(f) and argument(n,f). Then
couldnt one use circumscription to conclude, from the the fact
that th arity of f was at least 2, that it was exactly 2? and would
that solve the problem?
id be interested in your comments.
pat
≤
-------
∂01-Jul-80 1635 FFL
To: JMC, FFL
I will be out of the office Wednesday and Thursday. I have tried repeatedly
today to reach Prof. Connolly's office and the ME office, but there has been
no answer, so I don't know where to deliver the 60 Minute material. If he
needs it before Monday, it is on your desk. If not, can try again to get it
to him next week.
∂01-Jul-80 1640 FFL Luckham NSF
To: JMC, FFL
eport
The report has been forwarded to
Washington by Sponsored Projects today.
∂01-Jul-80 1825 JK addendum to sundays note
I expect to be able to do a simple proof in ekl by the end of this week -
I have re-written the documentation for it several times; the most recent one is
on doc.ekl[ekl,jk].I have also given a copy of this to Fred Abramson in the
math department (he is a visiting logician at there). Let me know if you have any
further comments.
∂02-Jul-80 0931 HVA SRA Summer Support
To: JMC, HVA
John: Do you want to support Ben Moskowski at 100% for Summer on your NSF Grant?
Carolyn has informed me that she will not need Summer Support here.
∂02-Jul-80 0936 RAJ.REDDY at CMU-10B RECOMMENDATION LETTER
Date: 2 Jul 1980 1235-EDT
From: RAJ.REDDY at CMU-10B
Subject: RECOMMENDATION LETTER
To: JMC at SU-AI, DEK at SU-AI, TOB at SU-AI
cc: RAJ.REDDY
John, Don, & Tom,
Beverly Howell sent you mail June 25 asking for recommendation letters to
help obtain an H-1 visa for Hans Moravec. I would personally apprediate
if you could expedite the writing and sending them. His visa is in a desperate
state at present and he is in the hospital and cant't do much about it
himself.
Just write a To Whom It May Concern letter and lean heavily on his research
abilities and send them to Bev Howell, Carnegie-Mellon U., Robotics Inst.,
Pittsburgh, PA 15213.
Thanks,
Raj
I sent "To whom ..." letter on Monday to Beverly Howell.
∂02-Jul-80 1040 doyle at MIT-AI (Jon Doyle) Mail coming
Date: 2 JUL 1980 1338-EDT
From: doyle at MIT-AI (Jon Doyle)
Sent-by: TK at MIT-AI
Subject: Mail coming
To: jmc at SU-AI, eaf at SUMEX-AIM
CC: DOYLE at MIT-AI
Dear John and Ed,
I'm not sure who to tell this to, but I have begun to
tell people of my new address in the computer science
department. I would greatly appreciate it if your mail-
handling personnel could be warned to hold any mail that
arrives for me before I get there.
Thank you very much.
Jon Doyle
I have asked Frances to be sure Doyle's mail is held.
∂02-Jul-80 1048 LGC ARPA Report
The file ARPLGC.TXT[1,LGC] contains the material you requested for the ARPA
report. In its present form it takes up 30 lines; if this is too long for
your purpose, I can cut it down in accordance with any suggestions you may have.
Difficult as it may seem, the new ideas that were developed should be
stated in single sentences. One or two will be enough.
∂02-Jul-80 1332 BALZER at USC-ISIB Re: AI and Philosophy session at AAAI conference
Date: 2 Jul 1980 1313-PDT
From: BALZER at USC-ISIB
Subject: Re: AI and Philosophy session at AAAI conference
To: BMOORE at SRI-KL
cc: jmc at SAIL, BUCHANAN at SUMEX-AIM
In response to your message sent 25 Jun 1980 1011-PDT
BOB AND JOHN,
THE STUDY GROUP YOU TWO PARTICIPATED IN AT THE CENTER FOR ADVANCED STUDY IN
THE BEHAVIORAL SCIENCES IS THE BASIS FOR THIS (TENTATIVE) SESSION. BRUCE
BUCHANAN MENTIONED THIS STUDY AND SUGGESTED THAT IT MIGHT MAKE AN INTERESTING
PANEL SESSION. FURTHERMORE, HE OFFERED TO ORGANIZE IT BY CONTACTING
THE PEOPLE INVOLVED TO SEE IF IT MADE SENSE AND TO DETERMINE WHO SHOULD
PARTICIPATE.
PRESUSMABLY BY YOUR MESSAGE HE HAS CONTACTED NEITHER OF YOU. BY APPLYING
(MONOTONIC) LOGIC RULES, I CONCLUDE THAT EITHER HAS HASN'T CONTACTED ANYONE
YET, OR HE HAS STARTED WITH SOME OTHER MEMBER OF THE STUDY. IN EITHER CASE,
I THINK YOU SHOULD EXPECT TO HEAR FROM HIM SOON.
REGARDS,
BOB
-------
∂02-Jul-80 1502 ENGELMORE at USC-ISI FY81 funding
Date: 2 Jul 1980 1453-PDT
Sender: ENGELMORE at USC-ISI
Subject: FY81 funding
From: ENGELMORE at USC-ISI
To: JMC at SU-AI, LES at SU-AI
Message-ID: <[USC-ISI] 2-Jul-80 14:53:31.ENGELMORE>
John and Les,
I had a talk with Bob Kahn yesterday concerning my funding shortfalls
for the next fiscal year. At the moment things look pretty grim: our budget
for next year is just the basic budget, with no room for growth in the 6.1
area (unless, of course, I just cut somebody off in the middle of a multi-
year contract, which I'm not going to do). The ARPA budget that's finally
approved by congress may be higher than we now expect and base plans on,
which is a small ray of hope, but we won't know about that for a few months
at least. Bob also says that some money in the office budget tends to
get liberated around January of each year, as we get a better picture of
what projects are unlikely to be funded for the remainder of the year.
What all this means for you is that I will not be able to augment
your contract, at least initially, but will keep it as a high priority
item for later in the fiscal year (it's easy to add funds to a contract
in mid-stream, providing the funds are there). I realize this isn't
much help for your planning, but I don't see what else I can do right now.
Bob
∂02-Jul-80 1545 LES Levinthal travel
cet pract
CC: JMC
In response to your inquiry about paying Elliot Levinthal's travel
to the ARPA automation meeting, any expenditures from the ARPA
account basically come out of your hide. As you know, we are already
short of funds in that area.
We recently put in a request via Bob Engelmore for accelerated funding to
ease the problem. Specifically, we asked for funding to be completed so
that work on the current contract could end on 1 October 1981. If
accepted, this would effectively increase current funding by $100k.
A few minutes ago I received a message from Engelmore saying that Kahn
declined to fund this plan, though there is a possibility that something
could be done around January 1981. Unfortunately, it would be too risky
to plan on that.
In summary, it appears that Bob Kahn is inviting us to stretch an inadequate
budget but is unwilling or unable to commit funds to fix the shortage.
∂03-Jul-80 0448 REM at MIT-MC (Robert Elton Maas) Analysis of Go endgames
Date: 3 JUL 1980 0747-EDT
From: REM at MIT-MC (Robert Elton Maas)
Subject: Analysis of Go endgames
To: REM at SU-AI, JMC at SU-AI, LES at SU-AI, RWW at SU-AI
CC: RWG at MIT-MC, FFM at MIT-MC
I've figured out how to handle certain difficult cases like
5 - - - O X - - -
4 - - - O X - - -
3 - - - O X - - -
2 - - O X - - - -
1 - - - - - - - -
F G H J K L M N
where X can play in two different ways (J1 in gote but with very large sente
threat on the first ply, or K1 in gote but with delayed sente). My current
algorithm tells for any local skirmish both what the best move is for each
side in this skirmish (considering that gote loses a move compared with
sente, and moves generally decrease in value as the game winds to a close)
and what the value of each best-move is so that the player/computer can
decide whether to move in this skirmish or play elsewhere. Thus as of
tonight I claim to have completely solved the problem of making a rigorous
algorithm for analyzing endgames of Go and playing optimally according
to the commonly-accepted heuristic of always making the largest available
move. The expected value of the game as a whole can be computed from
(a) the sum of all the expected values of the individual skirmishes as
computed by my algorithm, (b) the current value of the largest gote move
available, and (c) whose turn it is now. The expected value, i.e. mean,
is S+G/2 or S-G/2 depending on whose turn it is; the distribution is
approximately binomial with stepsize 1 and spread G. Skirmishes can
be classified as gote or sente according to whether the values of the
first-ply moves by the two players are equal or nonequal respectively.
Given that each skirmish is analyzed in vacuuo by my algorithm, the
algorithm for "optimal" (according to the commonly-accepted heuristic)
will almost never have a reverse-sente move made because the person
having sente always gets there first, except when a situation has been
temporarily activated so that replies are larger than the opening move
in that skirmish and best play is to keep playing there until the
value of a move drops below the opening-play value.
More complete description of algorithm can be written if anybody
is interested. One of the intermediate calculations of the completely
general method invented tonight involves formulas involving MIN and
MAX, such as MIN(0,MAX(T-7,MIN(T-3,-5/3))) which represent the expected
value of the skirmish for different values of T which is the threshold
which is the value of the largest gote moves elsewhere on the board.
Only the largest value of T where playing here is worth more than
playing elsewhere, is important, so these complicated formulas simplify
to simple sente or gote declarations like I generated by a less
complicated but slightly less general method earlier this year.
I, for one, would be interested in seeing a more detailed exposition.
∂03-Jul-80 0732 REG
What I meant about uninstalled data disks was that there are some allocated
to various places that haven't been installed yet for various reasons.
I believe the aggregate demand for them already exceeds the number we can
install. I'll look into what we have allocated to see where I can find one
more.
∂03-Jul-80 0811 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Giant File Server
Date: 3 Jul 1980 0811-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: Giant File Server
To: jmc at SU-AI
cc: reg at SU-AI, csl.fb at SU-SCORE
PLEASE KEEP THE FOLLOWING MESSAGE CONFIDENTIAL. THANKS, ED
The wheels are rolling, ponderously. I spoke with Paul Friedl. I would
say he was cautiously positive about the whole idea. We may be talking about
a campus-wide Giant File Server, hooked to CIT's 3033, since he was skeptical
that the 4331 would hack it; and anyway the deal might go through politically
better if handled by CIT (for complex reasons). Since CIT will be on the Ethernet,
do you see any problems with CIT running the File Server?
For this to happen, we will have to be doing "research and experimentation"
on Giant File SErvers, not just USING them (IBM can not give us things to do
the latter because of antitrust pressures). I assured him that there was just
tons of reseach and experimentation to be done relative to our conception of
Giant File Servers (e.g. virtual files).
Well, now the ball is mostly in our court. Paul is checking on
what ties into to what machine. But we have to design what we
want (perhaps with help of CIT). And in the normal way we have to
propose some Joint Study research to cover what we want. Hence there
is an opportunity, but there is also work ahead. Who will do the
work?
The work will need to confront technical reality and political sensitivity.
We will need a technical design and a well-thought-through proposal.
I repeat, who will do the work? John, do you volunteer? With the
resignations of Seamonds and Bizarri we are down to bone in systems
staff.
Ed
-------
∂03-Jul-80 2039 RPG $$$ for Waltz
To: JMC at SU-AI, DPB at SU-AI
CC: LES at SU-AI
Perhaps only abot $500 will do since he can stay with me and/or he
can split it with Illinois somehow.
-rpg-
∂04-Jul-80 0124 CSD.GARDNER at SU-SCORE Hacker and Raz book
Date: 3 Jul 1980 1530-PDT
From: CSD.GARDNER at SU-SCORE
Subject: Hacker and Raz book
To: jmc at SU-AI
I'm about ready to look at that defeasibility paper again, if you're
finished with the book. --Anne
-------
∂04-Jul-80 0348 REM via AMES-TIP
Ok, will be writing up a more precise description including examples
of how to handle some of the more difficult cases.
<Go endgame-skirmish analysis>
∂04-Jul-80 1254 CSL.FB at SU-SCORE Re: Giant File Server
Date: 4 Jul 1980 1251-PDT
From: CSL.FB at SU-SCORE
Subject: Re: Giant File Server
To: CSD.FEIGENBAUM at SU-SCORE
cc: jmc at SU-AI, admin.gorin at SU-SCORE, csl.fb at SU-SCORE
In-Reply-To: Your message of 3-Jul-80 0811-PDT
John could contribute a lot to this effort. I know the 3033 has the power
to serve as a giant file server but there is a lot of klutzy software in the
way. One thing to note is that CIT now has a SPARE 3033 (for system
development!). As to systems staff, we (I) are in the process of hiring 3
new people (one hardware wizard and two software wizards) so things are not
so destitute as they might seem. And we have TWO new systems faculty coming
in the fall, Brian Reid and Keith Lantz, both of whom might have a lot to
contribute to this sort of effort.
Forest
-------
∂06-Jul-80 0604 REM via AMES-TIP ENDGO1.WRU
I've tonite written about half the document (paper to publish?) describing
the easy part of my endgame Go algorithm. It will take some study just
to understand this part, so I'm telling you about it even before I write
the second half which has the general algorithm spelled out and examples
of how it works. Expect the rest of the document in a few days or so,
although I may postpone it in order to catch upon the work I'm getting
paid for. But knowing me, I'll probably find some excuse for finishing
it soon (like if IMSSS crashes, or I don't feel like working, or whatever).
Anyway feel free to ask questions about anything covered in this first
half of the writeup/paper.
∂06-Jul-80 1310 100 : clt susie
call her
∂07-Jul-80 0000 JMC*
waltuch
∂07-Jul-80 1623 FFL Wire service program to Mrs. Henson
To: JMC, FFL
Lester says there is a manual on the wire service program, but he
wonders if it is more technical than you wish for Mrs. Henson. You
refer to a "write up". Is there an article or material other than
the manual which you want me to send, or is it in fact the manual?
Yes, the manual.
∂07-Jul-80 1659 RMS at MIT-AI (Richard M. Stallman)
Date: 7 JUL 1980 1957-EDT
From: RMS at MIT-AI (Richard M. Stallman)
To: jmc at SU-AI
Hello. Have you made any arrangements for funding for my visit?
No, and now it isn't clear that I can. ARPA lured me into offering
someone a job and then didn't come up with the money. This
disappointment is only a few days old, so maybe I'll be able
to do something about it, but if you can find alternate financing,
it would be good.
∂07-Jul-80 1708 SEK
OK, I'll make a point of being around afternoons this week. (Have to run
right now).
∂07-Jul-80 1815 CSD.GOLUB at SU-SCORE
Date: 7 Jul 1980 1410-PDT
From: CSD.GOLUB at SU-SCORE
To: Feigenbaum at SUMEX-AIM
cc: csd.bscott at SU-SCORE, csd.tajnai at SU-SCORE, reg at SU-AI,
CSD-Faculty: ;
Dear Ed:
This is in response to your memo of June 6 regarding charges for
the departmental computing facilities. We have carefully considered
the proposal and explored various alternatives. We have come to the
conclusion that it is not possible to support a large number of
accounts under the new accounting system.
Effective immediately please cancel all accounts associated with the
NA group with the exception of those of EHG at SAIL, DLB at SAIL, and
CSD.TREFETHEN at SCORE. These students are currently preparing theses
using TEX.
The department facilities have been of great use to us for scientific
and administrative activities, and we regret that we are unable to
continue to use them. Our main use of the systems has been for mail
especially associated with departmental activities and TEX. We
cannot justify the present charges to our contract monitors. Our
scientific computing is performed at SLAC and really cannot be done
on the department system.
Gene Golub
Joe Oliger
-------
∂07-Jul-80 2019 JK nsf grant
I thought about the matter further; I think Suppes would be the best
person to work with. The reason why I hesitated at first was that I
don't think I could develop a truly powerful system on a halftime
basis: In other words I would want work fulltime on it thru SAIL
while Suppes would have a group working on adapting this kind of
system for educational purposes. Thus they would be involved with
issues like informal mathematical english etc. The other reason
why I think this approach would be better is that it would not
be incocnsistent (and, in fact would continue) the nsf grant that
I applied with Suppes and MacDonald thru IMSSS. The additional
plus would be in us being able to sell it partly thru its
educational virtues (and thus sidetracking some of the another-FOL
type criticism). Whats more I find the prospect of working with a
logician (be it Barwise or Feferman) rather dreary.
I think that selling this as a joint effort between IMSSS and SAIL
would increase its chances of acceptance: It will not be seen as
just another theorem-proving project. If you want to I can talk to
Suppes about all this: I will be seeing him this friday.
Ok. See how he reacts.
∂08-Jul-80 0000 JMC*
chair
∂08-Jul-80 1053 HVA Your signature needed
To: JMC, HVA
Will you be in today? Although TOB sent in a letter and budget (a few months
ago), requesting support for NSF coference here this Sept., NSF needs the
official, signed forms (going thru' Spons. Projects); so I have several things
requiring your signature. Thank you.
∂08-Jul-80 1436 CLT
I will be working here at the house the rest of the afternoon.
If you wish to go to the lumber store, pick me up on the way.
[UC seems to work still]
OK, around 5
∂09-Jul-80 0833 FFL
To: DBL at SU-AI, FFL at SU-AI, JMC at SU-AI
We do not have Paul Gilmore's phone number in this office. His address is
U. of British Columbia, 2075 Wesbrook Mall, Vancouver, B.C., Canada
V6T1W5.
∂09-Jul-80 0900 JMC*
lisp
∂09-Jul-80 0900 FFL
To: JMC, FFL
Keith Davidson, a freeland Australian journalist, called at the suggestion
of Dr. Minsky. Doing an article on AI and would like to speak with you for
5 or l0 minutes. Asks that you call him collect: 617 322 8204.
∂09-Jul-80 0953 Buchanan at SUMEX-AIM AAAI panel on AI & Philosophy
Date: 9 Jul 1980 0952-PDT
From: Buchanan at SUMEX-AIM
Subject: AAAI panel on AI & Philosophy
To: jmc at SU-AI
cc: bmoore at SRI-KL, hayes at SRI-KL, balzer at USC-ISIB
John,
One of the suggestions for a panel discussion at the AAAI mtg in
August was a summary of the year's activities at the Center. If we could
get Bob Moore to agree to chair such a session, would you participate?
I believe it would be valuable for the AI community to know what
topics were considered, what questions were raised, and
what difficulties the philosophers saw in AI work. Do you agree?
regards,
Bruce
-------
I would be happy to participate.
∂09-Jul-80 1015 LES Keyboard krock
To: REG at SU-AI, ME at SU-AI
CC: JMC at SU-AI, MO at SU-AI, ROY at SU-AI, ROB at SU-AI
I do not understand why you have permuted the numeric keys or why you did
so without public discussion.
The SAIL configaration is clearly superior to that of conventional
typewriters. The argument that it should be stadardized the older way to
simplify training overlooks the point that almost no one touch-types
numbers. This issue was carefully examined and resolved locally long ago
(correctly, I believe).
From your viewpoint, I would guess that all calculators have been improperly
designed in that they put "0" next to "1", contrary to standards established
by The Telephone Company.
If you believed that a change was needed for some reason, you should have
had the guts to discuss it openly, rather than undertaking a sneaky coup.
The only remarks about this that I heard came from ME, just after he
returned from that hotbed of conservatism, MIT. I am aware of no concrete
proposals or discussions of this issue.
I am disappointed both in what has been done and the manner of execution.
I recommend that you restore the keyboards to the prior configuration,
then make a proposal and defend it.
∂09-Jul-80 1026 Buchanan at SUMEX-AIM AAAI panel
Date: 9 Jul 1980 1024-PDT
From: Buchanan at SUMEX-AIM
Subject: AAAI panel
To: jmc at SU-AI
John,
If you would care to chair the panel, then we would be delighted.
I put my question in the light of Bob Moore's chairing it so that you could
respond first on whether it was something you would
like to have happen, regardless of who organized it.
bgb
-------
∂09-Jul-80 1142 SJC Thanks and fiche
To: "@S1[DIS,S1]" at SU-AI
∂09-Jul-80 1134 SJC Thanks and fiche
To: "@S1[DIS,S1]" at SU-AI
A belated thanks to all who contributed to the infamous greenbook; your
efforts have finally been packed off to the printer's.
I'm creating a file area [FIC,S1] to collect all program sources, data
files, or other material you'd like to put in the microfiche that
accompanies the greenbook. (We'd like to include everything we expect to
give out to the world, so that it's equally available to all readers and
relieves us of the hassle of answering individual requests--but remember
that this is a record of fiscal 1979 work, not of all work to date.) On 16
July 80 I'll take whatever I find there and commit it to plastic, and then
remove the area. Let me know if that date poses any problems, or if
somebody with a likely interest in fiche is unlikely to see this message
by then.
--Steve
∂09-Jul-80 1406 BIS maclisp support after rpg
To: DCL, JMC, LES, REG, TOB
Dick Gabriel will be having his orals this summer, probably next month.
What will happen to MACLISP after his finishing?
∂09-Jul-80 1451 RPG Malignment
To: REG
CC: LES, DCL, JMC, TOB, BIS, RPG
Hm, let's set the record straight. For several years I have maintained
MacLisp at no charge to Stanford (except for machine time) since the
department has chosen to not support me anymore. Before this, the Verification
Group paid my RA for such services. Currently, Verification, Hand-eye, and
the McCarthy crew are MacLisp users, as well as many random people. I would
guess my user community numbers around 30-40 at the moment.
The current (as of last week) arrangement is that for 1000k disk space and
300 bams, all questions and bugs, and some new features will be considered
and acted upon quickly, and that consultation on user bugs etc will be
entertained. This arrangment, involving what is traditionally is referred
as ``privileges'' is a barter.
I think that an appropriate way of looking at things is that this
is a small price to pay for a service that is otherwise free. Much of
my disk utilitzation comes from keeping sources for various debugging aids,
which was a habit gotten into when this wasn't a problem. To convert to
having them in the LSP area would take a bit of time and tracking down of references
which I cannot afford while I do my thesis. I plan to get MacLisp into
some kind of better shape when I am finished, but since the MIT maintainance
crowd is unreliable, there will be a period of several weeks tracking down
bugs introduced by these zealots.
I originally justified my higher SL because I had to portion them among my normal
work (thesis related things) and MacLisp work. I think that payment in
this form is better than payment of my time for having done this (and
continuing to do this) at fair market value. My position is that I am
providing a valuable service for a very, very small consideration,
and that this barter is a good deal or all concerned.
DCL, earlier this summer, offered to continue to ``pay'' for my machine
usage, but since his group represents only 1/3 of my user community, I
decided that this was not a fair arrangement.
Therefore, I would like to extend at this time my offer to continue to provide
MacLisp to Stanford (including, by the way, the SCORE version) for the
usual, uninflated rate of 1000k and 300 whams (bams?).
-rpg-
∂09-Jul-80 1510 RPG
Here is the original malignment (I should have changed the title
when I decided not to include this exchange in my message, but I
forgot):
∂09-Jul-80 1317 LES Exorbitant service level
To: REG
CC: RPG
As you know, I have been trimming accounts and priveleges for some time,
so as to give you a slightly smaller mess to deal with. Many moons ago,
DCL negotiated an inflated service level for RPG on the grounds that he
was maintaining MACLISP for the Verification group. As usually happens,
RPG seems to have incorporated this privilege into his standard of living.
During a recent slaughter, I trimmed RPG back again to a mere full-time
allocation (200 bams). I now leave it to you to either hold the line
or be generous (and then defend this position against all the others who
would like more bams).
Bon apatite.
---------
∂07-Jul-80 0135 RPG
∂29-May-79 1149 LES
∂28-May-79 1603 RPG Service level
Could you please put me back to 300 as we agreed several months ago.
I will continue to maintain MacLisp for the Lab gratis in return for
a few minor concessions such as this.
-rpg-
Accepted.
Les, could we re-establish the above agreement, seeing that I hope
to graduate this summer and need all the crunch I can get at this time,
in addition to my Lisp duties.
-rpg-
∂10-Jul-80 0951 FFL Second call from Keith Davidson of Boston, freelance journalist
To: JMC, FFL
I tired to turn him over to Les or to Robotics, but he has been referred to
you by Prof. Minsky and wants tto speak with you. He says it won't take over
5 minutes. Call collect, if you please, 617 322 8204.
∂10-Jul-80 1350 TOB
John
I get Assembly Engineering.
Thanks.
Tom
∂10-Jul-80 1346 JMC
I forget. Do you get Assembly Engineering and do you want mine?
∂10-Jul-80 1706 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Gene/Joe/NA problem
Date: 10 Jul 1980 1702-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: Gene/Joe/NA problem
To: CSD-Faculty:
Dear everybody,
I met with Gene and Joe, and the problem they perceived is well on the
way to solution. We'll use some "community" funds to ease their
problem. It takes some patience and time to "work into" a new system
of doing things.
Ed
-------
∂10-Jul-80 1817 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Giant File Server (Galaxy-wide File System)
Date: 10 Jul 1980 1758-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: Giant File Server (Galaxy-wide File System)
To: csd.wiederhold at SU-SCORE, reg at SU-AI, jmc at SU-AI, sandelin at SUMEX-AIM
cc: csd.feigenbaum at SU-SCORE
I would like to get a Steering Committee going that can plan
Stanford's proposal for a Joint Study with IBM on the Giant File
Server.
I propose a committee consisting of:
Wiederhold (chairman), representing data base and file system constructors
Gorin representing CSD-CF and LOTS
McCarthy representing the faculty
A representative from CIT (Jon, would you provoke Shaw/Roberts into
appointing one?)
I would suggest that the committee begin its plan activity just as soon
as possible.
Anybody accept?
Ed
-------
∂10-Jul-80 1733 RPG Plans
To: JMC
CC: RPG
As you might know, I am planning to graduate in the not too distant future
(such as early October). Essentially I have no plan for what I will be doing
at that time. Between now and then I will not be doing much in the way
of general improvements to MacLisp, though we are about 2 or 3 releases
behind MIT at the moment. Given the shoddy state of the newer versions vis a vis
SAIL, I would guess about a month or so to get us back up to par with them
and to get MacLisp to not depend on my directory and constant companionship.
There may be some mutual advantage to coming to a medium length agreement
about how this continued maintainance might come to happen. If I work
for LLL, naturally, there shouldn't be much in the way of a problem.
If I go away (for instance, to CMU), then there would be a problem,
but one which could be solved to mutual happiness with appropriate
access arrangements.
I also ask this as a sidelong way of inquiring what, if any, recent
plans you have about the proposed Vol. II of your LISP book.
-rpg-
If I understand what you are proposing - some maintenance in exchange
for good access, there should be no problem. Scott Kim, who will TA
CS206 in the Fall, may have some questions and ask some advice. We have
no current plan for volume II. We will not modify the book much before
Fall, but we will work on it then and hopefully finish it off.
∂11-Jul-80 0002 DLW at MIT-AI (Daniel L. Weinreb)
Date: 11 JUL 1980 0258-EDT
From: DLW at MIT-AI (Daniel L. Weinreb)
To: JMC at SU-AI
Date: 10 Jul 1980 1442-PDT
From: John McCarthy <JMC at SU-AI>
Is there a later LISP Machine Manual than Novermber 1978?
Yes; there is a "Second Preliminary Edition" dated January 1979.
However, it contains only minor corrections to the November 1978
version. Chapters 21 to 23 of both versions, describing processes,
jobs, windows, and friends, are now completely obsolete; there is a
person at MIT writing documentation on their replacements (known as the
New Window System). I may integrate this into a new manual, updating
everything else, when I return to MIT in September, if I get time.
∂12-Jul-80 1025 Sandelin at SUMEX-AIM giant file server
Date: 12 Jul 1980 1026-PDT
From: Sandelin at SUMEX-AIM
Subject: giant file server
To: csd.fiegenbaum at SU-SCORE
cc: reg at SAIL, jmc at SAIL, csd.wiederhold at SU-SCORE
I will pass the word to shaw/roberts ...will report results...I would like
to help in any way that would be useful...jon
-------
∂12-Jul-80 1645 AAM Optical character recognition
To: JMC
CC: HPM
A little while ago we had a conversation about reading text in a fixed font by
computer. At that time I was doing quite a number of things [RA, AI qual, ...]
so I didn't have time to follow up on the discussion.
However, I found out that I'm working quite near III this summer, so I called
them up and talked to someone about the "Graphics-1" system which is used for
reading characters in a fixed font. I will be getting some advertising literature
for this system and will try to go and see it or at least talk to someone who
is involved with it in person.
I'm having a little trouble getting access to the ARPAnet in LA, but I will try
to keep you informed about what I find out [I'll be returning in late September].
Allan Miller
(the "red-haired grad student")
Even though I am a member of the Board of Directors of III, I will be quite
interested in the information you obtain. III put big hopes into selling
Graphics-1 to the Social Security Administration but eventually lost out to
much more primitive methods that do less. Graphics-1 is very expensive
and based on obsolete technology, and III isn't motivated to modernize it
without customers. However, I believe that some of the ideas in Graphics-1
together with new ideas might lead to a cost-effective system, but I am
not personally knowledgable about such technologies.
I will be interested in discussing it when you come back.
∂13-Jul-80 1908 REM at MIT-MC (Robert Elton Maas)
Date: 13 JUL 1980 2208-EDT
From: REM at MIT-MC (Robert Elton Maas)
To: MRC at SU-AI, JMC at SU-AI
PC Net has gotten out of hand. I am considering implementing DIALNET
instead of PC Net protocol when I get a fullblown microprocessor system
later this month. One problem is that I won't be getting a 1200 baud
modem, rather I'll be getting a 300 baud modem. If I were to implement
DIALNET, would there be anybody to talk to? I.e. if I dial up a 300 baud
port into SU-AI is there a way to run DIALNET as a user program and
talk to it just as if I were calling in on the 1200 baud port? If not
I'll have to reject DIALNET too and hack up something on my own that
is PC Net like but not the latest broken protocol.
∂15-Jul-80 2035 REM at MIT-MC (Robert Elton Maas)
Date: 15 JUL 1980 2334-EDT
From: REM at MIT-MC (Robert Elton Maas)
To: RWG at MIT-MC, REM at MIT-MC, JMC at SU-AI
In addition to the reduction of the endgame of Go to mathematics, I have
now found a simple algorithm for solving the simultaneous MIN,MAX equations
that arise from recursive loops. Previously the math here was a handwave
(it's just math) whereas now I specify a way of doing the math also.
I have just finished using Macsyma to verify a rather difficult case
involving a ko connecting to an avalance if one side wins the ko and
the other doesn't defend. There are three major situations plus two
more in a small ko that develops if the threatened player defends the
major threat. There are a total of 8 states plus the 8 of the small ko
itself. The simultaneous MIN,MAX equations that must be solved are:
B16 = W16 MAX 1+W17F2-T
B16E2 = W16
W16 = B16 MIN -1+T
B17 = W17 MAX 1+W18-T
W17 = B17 MIN -1+B16E2+T MIN B19+T
W17F2 = B17 MIN B19+T
B18 = W18 MAX 22
W18 = B18 MIN 5+T
B19 = -1/3 MAX -T ;Small ko solved a priori
W19 = -1/3 MIN -1+2*T ;Small ko solved a priori
The solution obtained from Macsyma is:
T=0 T=1/3 T=5/3 T=19/3 T=17
B16 1-T 2/3 2/3 2/3 2/3
B16E2=W16 T-1 T-1 2/3 2/3 2/3
B17 6 6 6 6 6
W17 2T-2 2T-2 T-1/3 6 6
W17F2 0 T-1/3 T-1/3 6 6
B18 22 22 22 22 22
W18 5+T 5+T 5+T 5+T 22
B19 -T -1/3 -1/3 -1/3 -1/3
W19 2T-1 -1/3 -1/3 -1/3 -1/3
Note that for T>17 all states for a given situation (16 17 18 or 19)
are the same. When T>17 none of these situations are playable. When
T is between 19/3 and 17 only #18 is playable, but since it occurs only
as a result of #16 being played followed by #17, #18 will occur only
in a temporarily-activated state. Anyway, without looking at that
table of results, see if you can solve those simultaneous MIN,MAX
equations. Note that whenever a variable is unconstrained, it takes
the value of the last T where it was constrained, for example for all
values of T greater than 1/3 #19 is unconstrained, but players choose
to play elsewhere until T drops to 1/3, so that is its expected value.
Each of these results can of course be expressed compactly in MIN,MAX
notation. For example W17F2 = (0 MAX T-1/3) MIN 6.
∂16-Jul-80 0130 POURNE at MIT-MC (Jerry E. Pournelle)
Date: 16 JUL 1980 0429-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
To: POURNELLE at MIT-MC, JMC at SU-AI
What is your US Snails address? I want to send you
something.
Also: I am putting together a collection of fiction and
non-fiction entitled THE SURVIVAL OF FREEDOM. I could use a
short essay from you. Have you anything?
Jerry
John McCarthy, Computer Science Dept.,Stanford CA 94305
I remember about Survival of Freedom and hope to have some
alternatives for you in a few days. Have you seen Knowledge
and Decisions by Thomas Sowell? It's great.
∂16-Jul-80 1019 FFL
To: JMC, FFL
Mike Eaton of Dizcomp Corp. of Menlo Park called re his letter to you of
June 23, relating to intelligent modum and concerning Dialnet protocol.
He says it has something to do with loaning equipment to us in exchange
of advice. Telephone; 854-5434.
∂16-Jul-80 1132 FFL
To: JMC, FFL
Please call Dr. Balzar, ISI, at 213 822 1511.
∂16-Jul-80 1138 SJC Fiche
To: "@S1[DIS,S1]" at SU-AI
Here are answers to some questions I've received about greenbook microfiche
since my earlier message asking that contributions be placed in [FIC,S1] by
today:
1. Can I submit current files instead of old, buggy ones even if the new
ones include some work done since FY1979?
Yes, particularly if the new ones are more easily accessible.
2. Can I give you pointers to files instead of putting copies on [FIC,S1]?
Yes.
3. How do I tell whether my files belong on fiche?
Perhaps it would help to list the contents of last year's fiche. If you
are responsible for significant modifications to any of last year's
material, or if you are responsible for similar material, you probably
should consider immortalization on Mylar:
FASM.FAI FSIM.FAI SOPAS09.PAS PASRUN.S1
NUMER1.S1 PCPW5.PAS PCFOR0.PAS LINK.PAS
OBS1.S1 OPS.IN OPS.OUT UYK.S1
TEN.S1 DDT.S1 M.PAS SORT.PAS
LOCMAP.PAS R.PAS HINT.PAS ECO.PAS
TRL.PAS WR.PAS BREF.PAS DRW
LAYOUT RTOM DEFS CBLS
IOS MICRO.PAS F.MIC D.S1
P.MIC I.MIC E.MIC EDIAG.S1
MDDT.PAL
∂16-Jul-80 1231 HAYES at SRI-KL variadic functions
Date: 16 Jul 1980 1210-PDT
From: HAYES at SRI-KL
Subject: variadic functions
To: jmc at SU-AI
your message about dependency left me thinking. The use in 1pc of
function symbols to encode dependency seems too glib and too restrictive,
making it dfficult to, fo example, discuss the ways in which one
thing might depend on another. Maybe some fairly easy extension to 1pc
would help. here is one: treat all functions as having a single argument
which is in general a vector or sequencwe of entities. This immediately
allows 'variadic' functions, since there is now no particular reason
why all the arguments of a functon must have the same length. Also,
one can ask what a thing depends on by a question of the form
?exist v. a=f[v]
without being forced to beg th question of how many things it depends
on.
I am pretty sure that it the usual tarskian semantics can be adapted
to this variant logic without any real difficulty.
would this help with your problem?
pat
-------
∂16-Jul-80 2133 RWW
∂16-Jul-80 0040 JMC
Is Chris back? Please ask him to phone.
I believe he will return tomorrow. As soon as I see him I'll tell him.
∂17-Jul-80 0148 POURNE at MIT-MC (Jerry E. Pournelle)
Date: 17 JUL 1980 0444-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
To: JMC at SU-AI
I remember about Survival of Freedom and hope to have some
alternatives for you in a few days.
-------------
Good. I'm getting the book together now. It should be
a good one.
----------------------------
Have you seen Knowledge
and Decisions by Thomas Sowell? It's great.
No. Thanks, i"ll look out for it.
Best, Jerry
∂17-Jul-80 1042 HVA Tel. Call
Pls. call Jack Blackburn at (202) 389-6972.
∂17-Jul-80 1539 FFL
To: JMC, FFL
Pat Suppes asks that you call him at home, 94l-l803, if he has not reached
you very recently.
∂17-Jul-80 1817 Woods at PARC-MAXC Re: dired or the like
Date: 17 Jul 1980 5:52 pm PDT (Thursday)
From: Woods at PARC-MAXC
Subject: Re: dired or the like
In-reply-to: JMC's message of 17 Jul 1980 1302-PDT
To: John McCarthy <JMC at SU-AI>
The present DIRED has one facility that may be of some use, but for the most
part it is not set up for your purpose. The facility is the ⊗Xfoonly command (foo
for short, or output/foo for long), which outputs the current DIRED page into a
file in "foonly" format, i.e., <name>.{<ext>}[prj,prg].
As for merging the list into your FILES file, I should think that, once you've
added descriptions for the new lines, you could merge them rather easily using
the SSORT or MSORT programs, or even E's QSORT macro (page 11 of
EINIT[1,3]). DIRED could presumably do the merging if all the old files still
exist -- just include them all in the directory list and put out a new foonly file;
you're then faced with the problem of transferring the descriptions, however!
For the future, one of the long-range plans for DIRED is to add a description
field, essentially taking over the remaining functions provided by PRUNE. This
will let you maintain descriptions of the files as they are created, and
presumably there will some way to get the descriptions into a file in your
favorite format. (In fact, since the descriptions will be saved in a special text
file, it'll probably be simple to write an E macro that transforms it into your
FILES format.) None of this, however, is going to happen in the foreseeable
future. (There are little things like my work at Xerox and my thesis to be
considered . . . )
-- Don.
∂17-Jul-80 1823 BIS SAIL keyboards
To: JMC
CC: REG
Were you there when they crucified our Keyboard?
As I understand it, the SAIL keyboard was finalised after much thought and
discussion. Those who put the 0-key on the left did not do so by whim;
they had their reasons, probably better than those advanced recently.
No doubt these reasons appear in the minutes of a Systems Meeting unknown
to recent generations. John, I presume you were there. Ralph is trying
to decide whether or not to move the 0-key to the right of the 9.
Could you shed some historical light on this?
I was not consulted at any time.
Perhaps I misunderstood your message. The original design of our
keyboards was the work of Steve Russell and Dave Poole and others whom I
don't remember. Poole is still in the neighborhood. I don't recall
caring very much.
I was not consulted in the present change. It was discussed in
the BBOARD file, but I wasn't reading it at the time or I would have
vetoed the change. Many people including me who touch time have been
incommoded by moving not only the numbers but also the special characters
above them. I have the further problem of also using Imlac and Datamedia
keyboards that haven't been modified.
The change doesn't make our keyboards conventional, so I think it
is somewhat harmful on the whole, but I don't want to involve myself in
such matters.
∂17-Jul-80 2150 BIS
There is a myth that the SAIL keyboards were designed by you personally.
Support comes from the presence of `∀' and `∃' as opposed to, say, THETA
and PHI, under the presumption that the keyboard was designed to
facilitate work in logic rather than, say, geometry. You have dispelled
the myth.
I don't blame you for not wanting to get involved in discussions of it
all. I personally think that `0' on the left is better; I mostly resent
the autocratic way in which the keyboard change was effected.
Now that you mention it, that part of the myth is correct.
∂18-Jul-80 0813 DEK PUB???
The other editor of J. of Algorithms, Herb Wilf, will be spending several
months here, and during that time I imagine new policies will be worked out.
However, for the moment we are sticking to tried and true methods in order to
establish a high standard of quality for the journal
∂18-Jul-80 0941 FFL Mr. Tang's salary
To: JMC, FFL
Yes, $l2,000 is too much. The paper work seems to show the amount agreed
upon for him was $500 per month from 4-l-80 through 6-30-8l or $7,500.
I am holding the papers as you may wish to verify this with Hersche. She
is out of the office today but will be in on Monday.
∂18-Jul-80 1359 FFL
To: JMC, FFL
The Hutchinson book on Population Ecology is in your office.
∂18-Jul-80 1741 REM at MIT-MC (Robert Elton Maas) ENDGO1 >
Date: 18 JUL 1980 1823-EDT
From: REM at MIT-MC (Robert Elton Maas)
Subject: ENDGO1 >
To: RWG at MIT-MC
CC: REM at MIT-MC, MANN at BBNA, JMC at SU-AI
Regarding ko threat state, it turns out that except for large kos (which
occur in middle game, not very endgame), i.e. those with value more than
about 10 points (equivalent to "20 points in gote" as usually described),
it is better NOT to assume any particular ko state. The reason is that
there are plenty of gote situations of similar value, so whoever loses
the ko can make a gote move of nearly the same value and thus not lose
more than about 1 point due to not having enough ko threats. Thus treating
each of the three moves difference between winning and losing ko:
X O X O X O X O
X O O O <=== X O - O <===> X - X O ===> X X X O
X O Ofill X O capture X O Xfill X O
as equally-valuable moves seems to be the correct way even if it is known
a priori who has the most ko threats. For example suppose the life of
a 5-stone group requires winning the ko to make two eyes or to connect with
neighboring live group, so that the total point difference between the
two sides filing the Ko is say 12 points (twice number of stones plus
a couple points in the connecting chains). Then each of the three
moves would be worth 4 points. Suppose ko threats are lopsided so
that white will win the ko. So at the end of the ko fight white captures
in ko, getting that 4 points, black has only a 3.5 point ("7 points in
gote") move, white fills the ko for 4 points, and black makes another
3.5 point move. White has gained a total of only 1 point compared
with the expected value. There should be no more than one (maybe two)
of these during an entire endgame so most of the time it won't affect
the outcome. This 1-point bias is masked by the statistical fluctuation
due to all those little 1-point or half-point glitches each time one
size of move is used up and one or the other player has only the next
smaller size move to play next. (At the point when 4-point moves are
available, the person to move next can expect to gain about 2 points
compared with the expected value, but with spread of plus/minus 2 points,
i.e. anywhere from 0 to 4 points but with peak around 2; binomial
distribution roughly 1,4,6,4,1/16 or at half-point intervals
1,8,28,56,70,56,28,8,1/256) At the begin of the endgame the value of
moves and hence the spread are larger, perhaps 8 or 10, and the
1-point bias becomes insignificant. However if one sees a medium-sized ko
coming up, and if it is obvious who will win the ko, then perhaps the estimate
of final score should be fudged by 1 point just to have the most accurate
estimate available (like 3 points in blacks favor instead of 4, plus or
minus 8 points worst case or plus or minus 3 points standard deviation).
∂18-Jul-80 1741 REM at MIT-MC (Robert Elton Maas) SciAm article
Date: 18 JUL 1980 1853-EDT
From: REM at MIT-MC (Robert Elton Maas)
Subject: SciAm article
To: RP at MIT-MC
CC: RWG at MIT-MC, JMC at SU-AI
RP@MIT-MC 07/12/80 09:57:59 Re: SciAm article
REM@MIT-MC 07/11/80 02:33:45 Re: SciAm article
I agree that limitations are necessary in the paper but...
[Known definite integrals from CRC tables]
I miss the point of this. Clearly, there are many things one can find in
lookup tables which MACSYMA cannot do. NUSUM cannot do all of Jolley's
sums.
Correct. This is a basic design weakness of Macsyma. Whereas Macsyma
has lots of powerful methods for things that have been sufficiently
generalized, it has almost NO ability to match up user-requested problems
with a library of known results and to use those results to solve the
user-requested problem. Most of the world's knowledge is inaccessible
to people because there is too much of it too poorly indexed. If a
program could intelligently match problems against known results, people
wouldn't have to spend years looking for an answer that is already known.
I propose (1) an A.I. method that matches problems against answers,
performing automatically all the generalization and specialization needed
to make the problem exactly match the problem-with-known-solution, (2)
a mathematical method that takes that structure and unmaps the known
solution back to the posed question, (3) a training mode in a Macsyma-like
program that lets the user type in a question, the program tries to solve
it by methods and specific results it already has available, if successful
the trainer compares the method described by the program with the method
described in the research paper the trainer is taking the problem from
to see which is the better method or if they are equivalent (perhaps the
program found a brand new way to prove an old result, perhaps not), but
if not successful the trainer gives the answer to the program and
enough logical argument to convince the program it is correct, and the
program stores the answer and method for future reference. As more
and more known results are added to the program, it becomes the
standard reference for all mathematics, mostly replacing the usual
mathematical library when specific results are needed.
∂19-Jul-80 0208 POURNE at MIT-MC (Jerry E. Pournelle)
Date: 19 JUL 1980 0506-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
To: MINSY at MIT-MC, POURNELLE at MIT-MC, JMC at SU-AI
Date: 23 Jun 1980 1242-PDT
From: John McCarthy <JMC at SU-AI>
A phenomenon I call survival of the unfittest has certainly
occurred in NASA. I often wonder whether in organizations like NASA
and also in the nuclear power industry, it wouldn't be better if there
were a few years of almost zero budget to drive out the losers leaving
only a few fanatics followed by a big expansion rather than a steady
budget.
←
---
DO YOU MEAN THAT? We have ways to get that idea to
reagan. He may not accept it, but do we want him to think about
it?> I have had enough brandy to not be sure just now.
JEP
I said, "I often wonder ...". I think there is something in the idea,
but I don't think four years is long enough to implement both phases
of the process. However, it may be feasible to create new organizations
for certain purposes rather than expand old ones that happen to have the
right names. This might be most applicable to government agencies, but
it appears that Reagan hasn't (at least so far) the political strength
to actually propose wiping out non-functional agencies. Anyway he must
first be elected, and there's many a slip ...
∂19-Jul-80 0321 POURNE@MIT-MC (Sent by COMSAT@MIT-MC)
From: POURNE@MIT-MC (Sent by COMSAT@MIT-MC)
Date: 07/19/80 05:27:21
POURNE@MIT-MC (Sent by COMSAT@MIT-MC) 07/19/80 05:27:21
To: JMC at MIT-MC
[COMSAT: This was a failing QSEND.]
POURNE@MIT-MC 07/19/80 05:27:05
what are you doing up at this ghastly hour?
Listing my computer files, but I had just decided to give up and
go to bed.
∂19-Jul-80 0321 POURNE@MIT-MC (Sent by COMSAT@MIT-MC)
From: POURNE@MIT-MC (Sent by COMSAT@MIT-MC)
Date: 07/19/80 05:25:27
POURNE@MIT-MC (Sent by COMSAT@MIT-MC) 07/19/80 05:25:27
To: JMC at MIT-MC
[COMSAT: This was a failing QSEND.]
POURNE@MIT-MC 07/19/80 05:25:24
I suspect Reagan will win. At the moment anyway, JimmY carter
wouldn't be able to beat NIXON..⊗
∂21-Jul-80 0146 POURNE at MIT-MC (Jerry E. Pournelle) yea,verily...
Date: 21 JUL 1980 0445-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
Subject: yea,verily...
To: POURNELLE at MIT-MC, JMC at SU-AI
I fully agree. Most of the pro-technology strength in
the US comes from young people, and is not strictly speaking,
rational. It stems from (1) a naive acceptance of
high-technology, including space ships and such like, coupled
with (2) romantic notions derived largely from SF movies and
stories. And, immodest as it may be, I have to say that a good
bit of the intellectual defense of high technology among that
group has come from me; that to the extent they are willing to
defend their position with rational arguments, they got a
surprising number of those arguments from my columns.
But--their hearts are in the right place, and think very
little of human history has been the result of wide-=spread
rational decisions; when reason has triumphed it has done so
because of "authority"; authority has a bad press nowadays, but
perhaps it is coming back a bit; I have noticed at least that
many of my fans are more and more willing to "take my word for
it" on some tricky technical point, partly because I do my best
not to feed them any bullshit, and to say I don't know when I
don't.
L-5 is also interesting. You'll know more next year
when we invite you to be on the board of the blooming thing. It
is undergoing a transformation now, and there is some murkiness
in my vision of where it is going. I was for a while persuaded
last year that I would necessarily become President of L-5
because there was no possible other. Gerald Driggers was then
found, and moreover, Barbara Marx Hubbard donated enough money
to pay him about 30K / year to be full time President. (She
will not keep that up forever). He has used that time to try to
raise more money for continuing the "professional president"
idea; my view is that it might have been better had that year of
grace been used to build committee structures and get L-5
members doing a lot of the society's work, decentralizing and
delegating and building structure.
No matter; we wil shortly know if Gerry was right and
that large sums of money can be raised for l-5; if not, we have
no choice but to build organizational structure andx live on our
(not inconsiderable) own internal resources.
We'll see.
Meanwhile, any place that rational defenses of
technology can be published is one more place--that is, the more
the better. It isn't that rational defense is going to win the
game, but without it, we won't have the priesthood...
My own pet project now is a Lunar Colony before 2000.
Note I said Colony, not a Base.
More on this later. But yes, I agree ...
∂21-Jul-80 0151 ME
To: ROB, REG
CC: RPG, LES, JMC, BIS, TOB
∂21-Jul-80 0048 JMC
To: RPG, ROB, ME, LES
αl<space> works fine, but αk<space> kills one char at a time.
ME - It's not surprising that the system does strange things when a few
extra bits get turned on in various places in system code, for instance,
turning some JRSTs into JFCLs. The CTY explains all. KBUS 0 seems to
have been at fault.
∂21-Jul-80 0157 POURNE at MIT-MC (Jerry E. Pournelle)
Date: 21 JUL 1980 0456-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
To: POURNELLE at MIT-MC, JMC at SU-AI
After Reagan's acceptance speech, with its clear implication
that this is NOT the era of limits, and so forth, I think he
MUST be committed to high technology; and I would think he could
use some dramatic ideas that would have the effect of focussing
effort in high tech and also capturing the US imagination.
The Lunar Colony is obviusly the right idea for that; it gets
some research going that will spill out into the general economy
(teleoperated equipment, good hands for manipulation--I saw
Vic's Stanford Arm the other day and the arm and brain are
really good, but the hand is primitive; and I talked to an Alvin
operator last week, and found to my horror that we spend all
that money to keep Alvin on the sea bottom, but give him really
terrible hands and even worse control actuators...)
Anyway, a lunar colony would, it seems to me, focus
attetnion on intrinsically valuable techn9ology for investments,
as well as have a dramatic (and properly romantically appealing)
payoff.
JEP
∂21-Jul-80 0335 REM at MIT-MC (Robert Elton Maas) Go endgame algorithm update
Date: 21 JUL 1980 0634-EDT
From: REM at MIT-MC (Robert Elton Maas)
Subject: Go endgame algorithm update
To: RWG at MIT-MC, REM at MIT-MC, JMC at SU-AI
Well, thanks to Macsyma I've completed the final test of my method for
solving fixed-point of infinite-recursion-loop needed for ko situations.
This last and most complicated test involved a triple-ko and all the
double-kos and single-kos that it generated. There are 4 major states
to the triple ko, 3 to each of the double kos, etc. Here's the starting
situation:
... - X O -
... X X X O
... - X O -
... X O - O
... X O O O
... ... ...
After X captures twice and it is O's move but he can't recapture immediately,
the expected value is 7/3 MAX 5-2*T MIN 19/9+T MAX 3. (At least it is
if all the formulas I fed to Macsyma were correct; If and when I have
a program that computes these directly from the board, without my having
to manually type in all the equations and manipulate them, these expected
scores will be more reliable.)
Now I guess I'd better get back to writing this all up...
∂21-Jul-80 1545 Bmoore at SRI-KL Accomodations for FMAI
Date: 21 Jul 1980 1125-PDT
From: Bmoore at SRI-KL
Subject: Accomodations for FMAI
To: jmc at SAIL, clt at SAIL
cc: bmoore
I have just seen the layout of the rooms we will have for the Formal Methods
worshop, and it turns out that most of them are doubles. May I assume that
the two of you want to share a room with each other?
Bob
-------
Yes, we'll share?
∂22-Jul-80 0910 JRA lisp paper
john,
paper is off to printers. thanks. gee, did i notice a request for
an ansii-standard lisp in there? --tsk, tsk.
i stuck a blurb about my tlc-lisp in your mail box --its lisp machine
lisp for the z-80, with the small version about a 7090, and the large version
about 1/3 of a ka-10 in space. both versions run about 1/3 ka-10 speed.
john
Date: 21 Jul 1980 1323-PDT
From: Billy <BRACKENRIDGE at USC-ISIB>
Subject: The Hacker Papers
I would like to point everybody to "The Hacker Papers" in the
August issue of Psychology Today. The article is in the form of
a computer mail or BBOARD discussion at Stanford. It advances the
thesis that computer hacking is sick, addictive, and antisocial
behavior. It raises some valid points. I am sure everybody in
Human-Nets will have opinions on this subject.
The most important thing about this article is that it is in the
public press. Just what we need, more Computer Hackers are somehow
less than human dribble. This theme has come up in the telephone
rate discussion. I have heard "We can't tell them we are using
the phones for computer terminals" as though this is some sort of
pornographic behavior.
One of the people here was interviewed by someone from USC's
School of Communications. The interviewer was doing his Phd on
the dehumanization of people who communicate through computer
terminals. It blew his small mind to find a woman old enough
to have raised a family who insisted that quite to the contrary
people not terminals were dehumanizing. Had the interviewer
visited the Department of Motor Vehicles recently? Would he
rather deal with a computer terminal or one of the bored,
tired, and nasty DMV employees?
As a last thought I leave you with the words of a UCLA coed on
learning that I was "into" Computer Science. "Oh, an engineer. I
used to go over to the engineering building. When you ball an
engineer they really appreciate it. But man, They can be locked."
∂23-Jul-80 0856 FFL
To: JMC, FFL
John Doyle has arrived. He will be using my phone today, but it wondering
about his office space. Had any arrangements been made. I assume that
I speak with Betty Scott about such space unless you have already done so
or unless you think your personal request will have more force. He will be
around tthis morning for a while. He is presently staying with relatives
in San Jose. 408 24l 694l.
Jon Doyle should get Martin Brooks's desk for now, and we'll do better
on 1st Sept.
∂23-Jul-80 0900 JMC*
Cate
∂23-Jul-80 1258 LES TI terminal
Would like to borrow the little TI terminal till August 4.
Thanks.
∂24-Jul-80 0734 CSD.BSCOTT at SU-SCORE Les Earnest
Date: 24 Jul 1980 0732-PDT
From: CSD.BSCOTT at SU-SCORE
Subject: Les Earnest
To: JMC at SU-AI
cc: CSD.BScott at SU-SCORE
John, this is a reminder that you should write a letter to Les about his
August 31 termination. You will recall that when I mentioned this before
you were going to consider whether his leaving is a resignation or a lay-off.
He really should have formal notice of a lay-off by the end of this month.
Betty
-------
∂24-Jul-80 0905 BALZER at USC-ISIB TITLE FOR YOUR ''CHALLENGES AND PERSPECTIVES'' TALK
Date: 24 Jul 1980 0905-PDT
From: BALZER at USC-ISIB
Subject: TITLE FOR YOUR "CHALLENGES AND PERSPECTIVES" TALK
To: JMC at SU-AI
JOHN,
IF YOU COULD GIVE ME A TITLE TODAY FOR YOUR 1/2 HOUR TALK AT AAAI
I WILL GET IT INCLUDED IN THE CONFERENCE PROGRAM.
HERB DECLINED THE INVITATION, BUT RAJ ACCEPTED AND HIS TALK IS
ENTITLED "TOWARDS A THEORY OF KNOWLEDGE".
REGARDS,
BOB
-------
∂24-Jul-80 1046 Bmoore at SRI-KL Formal Methods Workshop Update
Date: 24 Jul 1980 1047-PDT
From: Bmoore at SRI-KL
Subject: Formal Methods Workshop Update
To: rbrachman at BBND, pcohen at BBND, doyle at AI,
To: hayes at SRI-KL, hendrix at SRI-KL, joshi at WHARTON,
To: jmc at SAIL, nilsson at SRI-KL, rperrault at SRI-KL,
To: bsmith at BBND, brian at MIT-ML, waldinger at SRI-KL,
To: webber at BBND, rww at SAIL, grm at RAND-UNIX, clt at SAIL,
To: hlevesque at SRI-KL, disrael at BBND, rick at RAND-UNIX,
To: appelt at SRI-KL, konolige at SRI-KL
cc: bmoore, stan
This is a status report on the upcoming workshop on Formal Methods in
Artificial Intelligence, August 15-17 at the Monterey Dunes Colony in
Moss Landing, California. (The previous announcement mistakenly said
Moss Beach.) The conference site is about half-way between Santa Cruz
and Monterey. For those of you who may not have settled your travel
arrangements, Moss Landing is about a one hour drive south of the San
Jose Airport or a twenty minute drive north of the Monterey airport.
For those of you coming down from the Bay Area Friday morning, we will
attempt to organize carpools. You will be directed to the meeting
place (probably the clubhouse) when you arrive at the gate, and we
hope to get the first session underway at 1:00 PM.
We would like to remind you that you may give an individual
presentation if you wish, but we would like to have a one page
abstract before the workshop begins so we can schedule talks and
discussions of related topics together. We would also very much like
to have suggestions for general discussion topics. Stan Rosenschein
has just joined SRI, so please direct all messages about the program
to either of us at the AI Center here. As a final point about the
arrangements, most of the rooms will be doubles, with a very few
singles, so if you want to arrange to share with a particular person,
please let us know.
Below is what we believe is the final list of attendees. We are
looking forward to a most interesting workshop and are glad that you
will be there.
Bob Moore
Stan Rosenschein
Attendees:
Doug Appelt Stanford University
James Allen University of Rochester
Ron Brachman Bolt Beranek and Newman Inc.
Phil Cohen Bolt Beranek and Newman Inc.
Jon Doyle Massachusetts Institute of Technology
Pat Hayes University of Essex
Rick Hayes-Roth The Rand Corporation
Gary Hendrix SRI International
Kurt Konolige SRI International
David Israel Bolt Beranek and Newman Inc.
Aravind Joshi University of Pennsylvania
Hector Levesque University of Toronto
Gary Martins The Rand Corporation
John McCarthy Stanford University
Drew McDermott Yale University
Bob Moore SRI International
Nils Nilsson SRI International
Ray Perrault University of Toronto
Ray Reiter University of British Columbia
Stan Rosenschein SRI International
Brian Smith Massachusetts Institute of Technology
Carolyn Talcott Stanford University
Richard Waldinger SRI International
Bonnie Lynn Webber University of Pennsylvania
Richard Weyhrauch Stanford University
Scott Weinstein University of Pennsylvania
-------
I know it's late, and I should have mentioned it sooner, but Creary
(lgc here) should be invited to the Formal Methods meeting.
He has some worthwhile if not great results.
∂24-Jul-80 1119 Wiederhold at SUMEX-AIM Re: giant file server
Date: 24 Jul 1980 1116-PDT
From: Wiederhold at SUMEX-AIM
Subject: Re: giant file server
To: Sandelin, csd.fiegenbaum at SU-SCORE
cc: reg at SAIL, jmc at SAIL, csd.wiederhold at SU-SCORE
In response to the message sent 12 Jul 1980 1026-PDT from Sandelin at SUMEX-AIM
We are looking forward to Ed Shaw designating someone soon.
Then we could get together for an initial plan.
Tuesday afternoons are traditional CSD meeting times.
So I7d like to have comments if next Tuesday, say 3pm would be suitable
(thats the 29th of July).
Ed mentioned that since he is leaving soon, any help we want to obtain from
him should be scheduled before Aug. 11th.
Gio
-------
∂24-Jul-80 1326 TOB update
John
This is an update on recent events.
I received a call from Paul Friedl, IBM Palo Alto,
concerning a possible joint study in robotics. We
had a meeting Monday, and he went away to see whether
there was IBM interest in setting up a joint study
within the umbrella of the current joint study in
distributed computing. We talked about a scale of
an IBM arm with a series 1 computer, two graduate
students, and IBM staff members.
I arranged to talk with the manager at Hewlett-Packard
for their automation program. He was interested and
said that under other circumstances they would not
be interested, they would do it all with their own
people. In this case, they don't have the background.
He is formulating a program now; he does not want to
cut things he has had in his deprtment, so he is looking
for funding as an add-on. Summary: there may be something
there, but it is too early to tell.
Charlie Rosen and Earl Sacerdoti at Machine Intelligence
asked whether the robotics group at Stanford is interested
in a university/industry proposal on some of our vision/robotics,
for example object description and recognition using laser
ranging (now called "structured light"). They say that there
are large amounts of money. I knew about the special program.
There are a couple issues I raise. I have discussed development
of VLSI vision elements with Hughes and planned to propose with
them. And while Machine Intelligence is innovative for industry
and in the right technology area, is it too small to win on a
proposal?
What do you think about IBM? Any reservations?
∂24-Jul-80 1349 Bmoore at SRI-KL Formal Methods Workshop Update
Date: 24 Jul 1980 1324-PDT
From: Bmoore at SRI-KL
Subject: Formal Methods Workshop Update
To: rbrachman at BBND, pcohen at BBND, doyle at AI, hayes,
To: hendrix, joshi at WHARTON, jmc at SAIL, nilsson, rperrault,
To: bsmith at BBND, brian at MIT-ML, waldinger, webber at BBND,
To: rww at SAIL, grm at RAND-UNIX, clt at SAIL, hlevesque at BBNA,
To: disrael at BBND, rick at RAND-UNIX, appelt, konolige
cc: bmoore, stan
(Some of you are getting this message for the second time because of
trouble in sending the original message.)
This is a status report on the upcoming workshop on Formal Methods in
Artificial Intelligence, August 15-17 at the Monterey Dunes Colony in
Moss Landing, California. (The previous announcement mistakenly said
Moss Beach.) The conference site is about half-way between Santa Cruz
and Monterey. For those of you who may not have settled your travel
arrangements, Moss Landing is about a one hour drive south of the San
Jose Airport or a twenty minute drive north of the Monterey airport.
For those of you coming down from the Bay Area Friday morning, we will
attempt to organize carpools. You will be directed to the meeting
place (probably the clubhouse) when you arrive at the gate, and we
hope to get the first session underway at 1:00 PM.
We would like to remind you that you may give an individual
presentation if you wish, but we would like to have a one page
abstract before the workshop begins so we can schedule talks and
discussions of related topics together. We would also very much like
to have suggestions for general discussion topics. Stan Rosenschein
has just joined SRI, so please direct all messages about the program
to either of us at the AI Center here. As a final point about the
arrangements, most of the rooms will be doubles, with a very few
singles, so if you want to arrange to share with a particular person,
please let us know.
Below is what we believe is the final list of attendees. We are
looking forward to a most interesting workshop and are glad that you
will be there.
Bob Moore
Stan Rosenschein
Attendees:
Doug Appelt Stanford University
James Allen University of Rochester
Ron Brachman Bolt Beranek and Newman Inc.
Phil Cohen Bolt Beranek and Newman Inc.
Jon Doyle Massachusetts Institute of Technology
Pat Hayes University of Essex
Rick Hayes-Roth The Rand Corporation
Gary Hendrix SRI International
Kurt Konolige SRI International
David Israel Bolt Beranek and Newman Inc.
Aravind Joshi University of Pennsylvania
Hector Levesque University of Toronto
Gary Martins The Rand Corporation
John McCarthy Stanford University
Drew McDermott Yale University
Bob Moore SRI International
Nils Nilsson SRI International
Ray Perrault University of Toronto
Ray Reiter University of British Columbia
Stan Rosenschein SRI International
Brian Smith Massachusetts Institute of Technology
Carolyn Talcott Stanford University
Richard Waldinger SRI International
Bonnie Lynn Webber University of Pennsylvania
Richard Weyhrauch Stanford University
Scott Weinstein University of Pennsylvania
-------
∂24-Jul-80 1601 Bmoore at SRI-KL Panel discussion on AI and Philosophy
Date: 24 Jul 1980 1600-PDT
From: Bmoore at SRI-KL
Subject: Panel discussion on AI and Philosophy
To: hayes, jmc at SAIL, les at SAIL
cc: bmoore
I first contacted Dan Dennett, and he said that he would like to come,
but just couldn't manage it. John Haugeland, however, agreed to come and
be on the panel. He said that a travel advance would be very helpful, so
we agreed that when he made his plane reservation he would tell us the price
of the ticket and we would send him a check. John said he cold not come for
the whole conference, and would probably arrive midday on Tuesday, so I said
I would arrange for the airport limo to bring him to some convenient spot
for me to meet him at the conference.
Bob
-------
∂24-Jul-80 1756 HEDRICK at RUTGERS Psychology Today article
Mail-from: ARPANET site RUTGERS rcvd at 23-Jul-80 2110-PDT
Date: 24 Jul 1980 0002-EDT
From: HEDRICK at RUTGERS
Subject: Psychology Today article
To: admin⊗.mrc at SU-SCORE, admin⊗.jqj at SU-SCORE
cc: gopstein at RUTGERS, josh at RUTGERS, marantz at RUTGERS
Remailed-date: 24 Jul 1980 1751-PDT
Remailed-from: Mark Crispin <Admin.MRC at SU-SCORE>
Remailed-to: Les at SU-AI, JMC at SU-AI
2The last thing I want to do is add another flame to your already
masterful collection. But where did you get that Psychology professor?
Aside from an astounding collection of bloopers (or do only Wheels
and maintenance personnel use HRROI at LOTS?), it seems to me that
that set of comments deserved a more considered response. I had
hoped to see somebody who had some relevant analysis, either
related empirical results or theory. Instead we got another
hand-wringing essay about the wretched state of American Youth.
The one interesting point in the article was his results on attitudes
and behavior of shy and non-shy people. This seemed to me to suggest
several interesting hypotheses, but he didn't pick up on any of them.
While I guess we can't fault him for technical errors in his computer
science, it seems to me that the least we can expect is that he would
not introduce any other areas that he doesn't understand. Thus I
found it amusing that he rejected the concept that the Devil tempts
men to sin, and proposed as an alternative diabolical goal,
a set of actions that is almost a classical definition of what sin
is. His comments about the character of Mr. Spock were not well
received among the local Trekies, either.
We might suggest to him that the quality of his article is not
going to encourage computer scientists to take courses in non-
technical areas, since it will reinforce their prejudices that
social sciences are mainly hot air. I think this is a shame, because
I have seen people treat similar subjects in very interesting ways.
For example, a few years ago there was concern when a number of cases
occured of people getting mugged in New York and bystanders didn't
do anything. There were the usual sets of articles condemning
people for being uncaring. Then somebody did some research designed
to find out what circumstances lead people to intervene when something
bad is happening, and what ones don't. It turns out that there is
an amazing ambiguity in interpretation of events, and that people
actually were not sure that what was going on was something where
intervention was appropriate. It also turned out that even a very
brief aquaintance with the victim beforehand caused them to be far
more likely to detect that there was something bad going on.
One would have hoped for some equally illuminating piece of analysis
here. I think a bit more care would show that what is going on in
the hacker community is almost the direct opposite of what he
accepts. Of course I don't know LOTS personally, but from your
messages, it sounds similar to Rutgers (though unless you have
exaggerated, it is possibly a bit more narrow, since we don't have
any hackers who completely lack outside interests).
Let us consider first the evidence for how human relationships come
about. (In an attempt to avoid psychological jargon, I will hereafter
refer to meaningful human relationships as friendship.) There seem to
be two basic things known about friendships:
- that they are difficult to form
- that they tend to form when people are brought together for
other purposes, in a way that is not personally threatening
His discussion of shyness supports the first point. Basically he
claims that the students that he finds fall into one of two
categories:
- shy ones, who want close friendships but don't find enough to
satisfy them
- unshy ones, who are satisfied with superficial relationships
The second point is supported by general observation - namely that
most friendships tend to be among people who are thrown together
because they work together, bowl together, etc. There is also a
rather interesting study of relationships in college dorms that
showed that people tended to know people who lived near them.
Now let us consider the current social situation, which he finds
completely dark. In the dim past, we had people living in small
tribes, with male and female lodges. These people spent their
whole lives together, working and playing, and (in some tribes)
living from day to day. Thus we have a small group of people, all
of whom are around each other enough to become friends. While no
doubt they didn't all become friends, at least one could have a
reasonable proportion. In contrast, we have a modern city, where
you can't know a substantial proportion of the people around you.
Those friendships that do develop tend to do so on the basis of
chance encounters at work, and in various spare-time activities.
(How else could they happen? The people who live in the apartment
above me won't even answer the door.) I claim that psychologically
things aren't all that different - friendships are formed the same
way - it is just that we have less time to do so, and often have to
take more initiative, since we don't have a single set of people
with whom all of our friendships will naturally develop.
Now, my alternate model of a hacker community is that it is as close
as we will ever come (outside of certain specialized communes) to
the primitive environment, where all the men spend all their time
together, and there is a natural environment for building up
friendships. I guess I agree with the analyses in the article of
how it gets this unique status. Computers are uniquely attractive,
and it is probably hard to think of any other artifact around
which work, play, and study will all congregate. However I am not
sure that I agree at all about the social or psychological
consequences.
- the uninterest in the rest of the people on campus is not a
sign of atrophy of the soul. It is rather a sign that
the person is a member of a very tight community, and
that the best opportunities for human contact are within
this community
- the fact that close friendships are too rare is probably not
an indication that computers are dehumanizing. The
evidence in the "hacker papers" is that indeed there are
a reasonable number of friendships within LOTS. In
contrast, the evidence presented by your psychologist
is that friendships are rather rare in the university
as a whole. My hypothesis is that a systematic study
would show that friendships among the hackers are
in fact more common than among randomly-chosen
Stanford students. What you are seeing is that basic
fact that friendships are hard to form, even under the
best of circumstances.
- it may well be true that friendships are somehow different
in kind within LOTS. This is because the friends share not
just a small set of common interests, but their whole
lives are both involved in the same thing.
Note that I agree that the hacker environment is something
unusual and worthy of psychological study. But I do think that it
is something more interesting than just a case of computers
having a general dehumanizing effect. Actually I would prefer
for people to have their lives centered around something a bit
more appropriate than a machine (as a Christian I believe that
only God is appropriate for this role). But it is not clear to
me that the actual effects are much different than a group of
fishermen who live their lives together, entirely tied up in
fishing.
-------
∂24-Jul-80 1757 Roy Marantz <MARANTZ at RUTGERS> Re: Psychology Today article
Mail-from: ARPANET site RUTGERS rcvd at 24-Jul-80 0846-PDT
Date: 24 Jul 1980 1142-EDT
From: Roy Marantz <MARANTZ at RUTGERS>
Subject: Re: Psychology Today article
To: JOSH at RUTGERS, HEDRICK at RUTGERS, admin⊗.mrc at SU-SCORE, admin⊗.jqj at SU-SCORE
cc: gopstein at RUTGERS
In-Reply-To: Your message of 24-Jul-80 0011-EDT
Remailed-date: 24 Jul 1980 1751-PDT
Remailed-from: Mark Crispin <Admin.MRC at SU-SCORE>
Remailed-to: Les at SU-AI, JMC at SU-AI
c As a former (and not so former) member of clique type groups (which a
hacker group tends to be), I think that these groups are not bad persay,
but that they can get to "intense" for some of there members. By this I
mean that by the fact that there members have so much in common they
tend to have more trouble with people who don't have these common
experiences-knowledges.
I find that the easiest way to get together with other people (I hope
I don't sound too much like a psychologist) is to develope some other
interest you have. In my case, I've gotten into folk dancing in a big
way. What concerns (maybe not the best word) me is that unless the
activity is intense I find I get boarded with it rather quickly. At
this point I don't know if that is "normal" or not. (I tend to think it
is, am not too concerned if it isn't)
Roy
-------
∂24-Jul-80 1938 Mark Crispin <Admin.MRC at SU-SCORE> Re: Psychology Today article
Date: 24 Jul 1980 1926-PDT
From: Mark Crispin <Admin.MRC at SU-SCORE>
Subject: Re: Psychology Today article
To: Marantz at RUTGERS, Josh at RUTGERS, Hedrick at RUTGERS,
Admin.JQJ at SU-SCORE, Gopstein at RUTGERS
cc: JMC at SU-AI, Les at SU-AI
In-Reply-To: Your message of 24-Jul-80 0842-PDT
Chuck, are you aware that the fact that you consider the
blooper about HRROI to be such implies that you are a hopeless
emotional cripple? Or so the PT article might have one believe.
I am told that the glossary was dictated over the telephone long
distance from a phone in the (noisy) LOTS computer room, and that
the connection was bad to begin with. The fact that PT was
satisfied with such indicates the depth of their research.
I must say, I was expecting something a lot deeper in that
Psychology Today article than what came out. As the presentation
and the article stand now, it looks like something from out of
the National Star or some similar sensationalistic journal. I
should note that among psychologists, PT is considered to be just
that; sensationalism for the lay public. The interesting
research and papers are printed in J.APA, etc.
Besides being the Stanford TOPS-20 system programmers, both
JQ Johnson and myself were students in the social sciences; I as
an undergraduate majoring in psychology, JQ as a graduate student
in political science. I would certainly take offense if somebody
claimed that the social sciences are a lot of hot air; however,
it is true that a lot of hot air exists when either social
scientists or "hard" scientists take a superficial look at the
other. Clearly this is what Zimbardo has done.
It is an exaggeration to say that the LOTS hackers
completely lack outside interests. I'd say that in general, the
"LOTS kiddies" (as I jokingly nickname them) tend to be if
anything a bit more well-rounded than most denizens of university
computer centers. Part of this may be the Stanford environment;
there are certainly a lot more opportunities for "nerds" to come
in contact with "normal people" at Stanford than, say, at
Stevens. Similar influences probably exist at Rutgers, although
the Rutgers campus is more spread out and isolation is more
likely to occur.
I theorize that the ability to form superficial
relationships is a critical pre-requisite to being able to have a
lasting "deep" relationship. While people may reject superficial
relationships on an intellectual basis, they seek them on an
instinctive basis. Consider the "Hey, we're going to get some
pizza, wanna come along?" type of dinner invitation. Cliques are
a good means of establishing such "superficial" relationships;
not to mention being of great benefit to a shy person needing to
develop his or her social skills. Common-interest cliques form
on the basis of the shared interest, and no emotionally painful
(for the shy) friendship-building process is needed to become an
insider. It's a chance to BELONG in a world where there are
intense pressures to belong.
Like Roy, with varying intensities I'm part of a number of
cliques/fannish groups -
. fan of several rock groups (especially HEART and Blondie!)
. Rocky Horror Picture Show (twice a week!)
. ice skating (twice a week ditto)
. eating well
. fantasy art collecting
. dragons (!!!)
. computer hacking (which pays for all of the above)
just to list my most intense activities. In each activity I am
often sharing my experiences with others who share the relevant
interest. I have no idea where I'd find the time to fit in
anything else if I had to "expand" my interests.
Which reminds me. I'm surprised that PT hasn't gone after
RHPS fandom, since I would expect it to be far more vulnerable to
this sort of sensationalism than hacking. After all, RHPS has
transvestism, cannibalism, bisexuality, sadism, murder... A lot
of people take a superficial look at RHPS fans and consider them
wierd; but I can tell you that on Friday and Saturday nights RHPS
fans are the happiest people in the world. They are with other
people like them -- who they may never see at any other time --
and have a great time. Actually, I'm glad PT hasn't. They may
pass a law against RHPS.
Maybe Zimbardo should investigate religion and its
dehumanizing effects. After all, religion has been a killer of
humanity for millenia; holy wars, suicide cults (People's Temple,
etc.), the Moonies(!!)... People are made subservient to the
good of the religion, cult, god, whatever. As an atheist,
clearly I should be alarmed at this and would want to look into
ways to redirect these poor religious believers into normal
activities! If this sounds absurd (does it?), it seems to be
that this is what PT wants people to think of hackers.
-- Mark --
-------
∂24-Jul-80 1953 JK proof checker
The proof checker is fairly functional now. It can be played with by
aliasing under ekl,jk and typing lisp at the monitor level. The last
four or so pages in doc.ekl tell how to use it. The files test.lsp and
test.ppr contain sample proofs. Note, however, that the pretty printed
versions don't show the use of the fexpr form of the commands; this,
makes life easier. The top level is rather kludgy - also the sorting
mechanisms don't exist yet (this is patched over in the present code)
not to say anything about the more complicated procedures.
I am going away for a few days - I expect to be back at the latest
within a week.
∂24-Jul-80 2344 RMS at MIT-AI (Richard M. Stallman)
Date: 25 JUL 1980 0243-EDT
From: RMS at MIT-AI (Richard M. Stallman)
To: jmc at SU-AI
Any chance of funds for me?
∂25-Jul-80 0108 REM via SU-TIP
To: JMC, RWG
First complete draft of endgame-go algorithm writeup:
MIT-MC REM;ENDGO1 >
SU-AI ENDGO1.WRU[1,REM]
Lots of examples of use of my algorithm are given.
∂25-Jul-80 0838 Sandelin at SUMEX-AIM giant file server
Date: 25 Jul 1980 0828-PDT
From: Sandelin at SUMEX-AIM
Subject: giant file server
To: csd.wiederhold at SU-SCORE, csd.fiegenbaum at SU-SCORE
cc: reg at SAIL, jmc at SAIL, sandelin
this project has been discussed..some confusion exists as to IBM
flexability under joint agreement to participate..IBM was to discuss
options with Fiegenbaum..Shaw is awaiting outcome before proceeding...I
suggest a direct Fiegenbaum/Shaw link to resolve...a problem is that new
IBM disks do not attach to 4331, although it may be possible to develop
software around 3370 disks ...I am officially away for a month starting
monday, but will have terminal with me and could come to meeting at
Stanford either 8/5 or 8/12 if you think it useful
-------
∂25-Jul-80 0913 REG
To: csd.Wiederhold at SU-SCORE, csd.Fiegenbaum at SU-SCORE,
Sandelin at SUMEX-AIM, REG at SU-AI, JMC at SU-AI
Re: file server meeting
I'm planning to be absent as much as possible during the week of 7/28. I
could come to a meeting 7/29 but I would prefer to avoid it. As Jon is
away then too, I would rather we delay meeting until 8/5 or 8/12 as Jon
suggests.
∂25-Jul-80 1513 Sandelin at SUMEX-AIM
Date: 25 Jul 1980 1454-PDT
From: Sandelin at SUMEX-AIM
To: jmc at SAIL
messages to fiegenbaum get queued, then timed out, but i was told they do
reach Feigenbaum (i hope)..i do know that messages to siegman do reach
Siegman
-------
∂25-Jul-80 1619 Sandelin at SUMEX-AIM queueing
Date: 25 Jul 1980 1544-PDT
From: Sandelin at SUMEX-AIM
Subject: queueing
To: jmc at SAIL
did you get my message on feigenbaum, in which i mentioned they also seem
to get queued, then timed out..the reason i ask is that that message to
you got queued, the timed out
-------
Yes, I got your message on Feigenbaum. I prefer the 5th to the 12th,
because there are several reasons why promptness is important in this
matter.
∂25-Jul-80 1639 FFL
To: JMC, FFL
Lutz Erbring of the U. of Chicago called. He wanted to locate a project
at Stanford which includes information on the UPI news wire accumulated
in computer readable form. 312 753 ll28
Other number 312 753-1357
account 1,lwe, password = media
∂26-Jul-80 1918 HPM via parts unknown a belated thanks
for the letter of recommendation. The completed visa application
was sent off to the government last week. -- Hans.
∂27-Jul-80 0831 Richard Jay Solomon <Solomon at MIT-Multics> visiting SU-AI
Date: 25 July 1980 1950-edt
From: Richard Jay Solomon <Solomon at MIT-Multics>
Subject: visiting SU-AI
To: JMC at SU-AI
This is by way of an introduction: I am a colleague of Ithiel Pool's at
the MIT Research Program on Communications Policy. I will be in the Bay
Area from August 14 through August 20, (attending the AI meeting for
part of the time). Our program is beginning a long-term project to
study the impacts and requirements for telecommunications systems during
various likely future crisis scenarios, such as a prolonged energy
crunch and other emergency situations. We noticed your interest in this
subject in the press accounts on the recent energy meeting at Stanford.
We may find it mutually useful for us to discuss the work at MIT while I
am at Stanford during that week.
I can be reached at 413 267-5171 (home) or 617 253-3124, 3144 (MIT).
On the assumption that we don't check ous mailboxes too often, I may
call you next week about an appointment.
Richard Solomon
It sounds interesting. Call when convenient to specify meeting time.
My phones are 415 497-4430 work and 857-0672 home. I'm usually home
till late morning.
∂28-Jul-80 1154 FFL
To: JMC, FFL
John Kelly, 328-6077, referred to you by John Sandeling of CIT, wishes to
have appointment to talk with you about ideas for research on Text Network
program. May I suggest a time to him or someone else?
Late this afternoon would be ok.
∂28-Jul-80 1321 FFL
To: JMC, FFL
John Kelly cannot be here this afternoon. He asks if it is possible to
see you either Tuesday between l:30 and 3, or Wednesday after 3 p.m. I
am to call him back.
Tuesday at 1:30 will be ok.
∂28-Jul-80 1400 JMC*
Ask FFL about filing reports.
∂28-Jul-80 1444 FFL
To: JMC, FFL
John Kelly will be here at l:30 tomorrow, Tuesday, to speak with you.
∂29-Jul-80 0924 Bmoore at SRI-KL Alan Thompson (?)
Date: 29 Jul 1980 0923-PDT
From: Bmoore at SRI-KL
Subject: Alan Thompson (?)
To: jmc at SAIL
John,
I was asking Bob Balzer about the composition of the non-monotonic logic
panel at AAAI, and he told me that you had recomended Alan Thompson. I'm a
bit puzzled: who is Alan Thompson and what does he know about non-monotonic
logic?
Bob
-------
I have no idea who Alan Thompson is or what he knows about non-monotonic
logic. Must be some mistake.
∂29-Jul-80 1300 JMC*
bbs,cate,bookcase,ear
∂29-Jul-80 1627 CJS Classnotes
To: CLT at SU-AI, JMC at SU-AI
CC: CSD.TAJNAI at SU-SCORE, CJS at SU-AI
I have received an order from the Bookstore for 60 copies of
"LISP: Programming & Proving" by McCarthy. If you can supply
me with the master copy, I will see that they are printed and
delivered to the Bookstore by the due date (8/25/80).
Connie
The notes will mainly be the same as last year's, but I am tinkering
with them. I would like as much time as possible.
∂29-Jul-80 1811 REM
∂12-Jul-80 1231 JMC
June 30 Computer World (math library) has articles on commercial file
compression. [REM - I looked at these articles. They claim
sample-then-crunch programs (like CRU2 and CRU3=SPINDL) obtain from 2:1 to
5:1 compression, but the only method actually cited is the
single-character null-context Huffman code, which crunches SU-AI files
only about 3:2 at best. I conclude the files they start out with are
incredibly redundant, with 80-column records that are mostly spaces, and
the rest upper-case and numbers, so that single character compression is
all that is needed for the ratios they cite. My methods should do
considerably better, although using more memory than they reccommend (10k
bytes which is 2.5k words).
∂30-Jul-80 1048 FFL
To: JMC, FFL
John Kelly called to ask if you had material on dialectical reasoning
which he may have to read.
Nothing for Kelly now.
∂30-Jul-80 1300 JMC*
Cate, BBS, cabinet, Grant, waltuch
∂30-Jul-80 1458 CJS Classnotes for next quarter
To: JMC at SU-AI, CLT at SU-AI, csd.ullman at SU-SCORE
CC: csd.tajnai at SU-SCORE, CJS at SU-AI
The Bookstore doesn't need these until August 25th. Unfortunately,
that is the day I leave for 2 weeks vacation. In order for me to
get the classnotes processed, printed, and over to the Bookstore
in time, I'll need to have the master copies by August 14. Can
you forsee any problems in meeting the August 14 deadline? If so,
let me know as soon as possible so I can make alternate arrangements.
Connie
∂30-Jul-80 1523 CSL.CRC.LYD at SU-SCORE Mr. Hao Dong
Date: 30 Jul 1980 1513-PDT
From: CSL.CRC.LYD at SU-SCORE
Subject: Mr. Hao Dong
To: CSD-Faculty: ;
July 28, 1980
CSL, EE and CSD Faculty, Staff and Students
E. J. McCluskey
Mr. Hao Dong
Mr. Hao Dong is a visiting scholar from Harbin Shipbuilding
Engineering Institute, Harbin, People's Republic of China and he ~
will be here for approximately one year. He is interested in
computer-aided electronic circuits design.
If you are interested in meeting with him and discussing his work,
his office is ERL 226D, ext. 71448.
!
-------
∂30-Jul-80 1704 FFL
To: JMC, FFL
About 4:l5 p.m. (system was down at the time) Kenneth Parker called
you. 7-2121.
∂30-Jul-80 1835 LGC Talk Abstract
The file NEOFF.ABS[1,LGC] contains the abstract (possibly too long, though
RCM said one page was right) of a talk I plan to give at the formal methods
workshop Aug. 15-17. This describes part of what I've been up to lately;
I'll try to send you a message before I leave giving the gist of some of
the other things. Comments on the abstract would be welcome.
∂31-Jul-80 0033 MLB undecidable production system?
To: "@GOSPER[1,MLB]" at SU-AI, JMC at SU-AI, DEK at SU-AI,
DON at SU-AI
A long time ago a friend showed me a very simple undecidable system,
which consisted of a small alphabet of symbols, and around seven "productions"
most of which were like ab=ba, or c=dd with a few like ae=fdb. The system was
said to have been proven uundecidable in the sense that in general one coldnt
tell if two strings composed of symbols in the alphabet cold be transformed into
one another via the prodctions.
Does anyone know where I cold find a reference for this or a similarly simple system?
Thanks (with apologies for the dropped Us from this losing terminal)...
Hopcroft and Ullman might have such a system, and Ullman probably knows
that literature.
∂31-Jul-80 0709 Mark Crispin <Admin.MRC at SU-SCORE> Dialnet TELNET protocol
Date: 31 Jul 1980 0704-PDT
From: Mark Crispin <Admin.MRC at SU-SCORE>
Subject: Dialnet TELNET protocol
To: G.Eldre at SU-SCORE, G.Milligan at SU-SCORE
cc: JMC at SU-AI, Les at SU-AI
I have implemented Dialnet support in my TOPS-20 TELNET. If
either of you guys have a working Dialnet modem and monitor yet,
you can snarf the files:
PS:<SUBSYS>TELNET.EXE
and PS:<SYSTEM>HOSTS2.BIN
from SCORE, install them on the same place at your system, and
try it out.
TELNETing to SAIL should work, and get you to a Dialnet NVT
at SAIL.
-------
∂31-Jul-80 1140 LGC 1. Abstract Revision 2. Other Developments
The abstract in NEOFF.ABS[1,LGC] has been revised (I had forgotten to mention
the connection with your notion of an abstract language).
This summer has been very fruitful for me (perhaps even as fruitful as last
summer, though with a different set of ideas). My ideas on representation
and reasoning (especially in connection with the planning of actions) have
come together in a satisfying way; this was the last unfulfilled prerequisite
for writing down formalized versions of the planning arguments for my
proposed initial version of the Advice Taker. During late August and early
autumn I intend to get a lot of writing done; currently projected papers
include:
1. A description of my approach to commonsense (non-monotonic) reasoning,
with some discussion of other appproaches.
2. An expansion of my workshop talk on neo-Fregean foundations for
representation theory.
3. A major update of SNAPOV.TXT (snapshot and overview of the Advice Taker
proposal), including detailed treatment of at least one example problem
solution.
4. A philosophical solution to John Perry's "problem of the essential
indexical" (this is a further development of the ideas in my
IJCAI-79 paper).
With these items in hand, I then plan to see what can be done about
getting some help with implementing a small initial version of the Advice
Taker, using some representation software recently developed here at
Stanford (a glorified property-list mechanism, with swapping of property
lists to provide virtually unlimited space for them). Also, I plan to see
what can be done about interesting certain members of the Philosophy
department in some sort of collaborative effort (such as the development
of the world's first piece of philosophical software along the lines of my
current proposals for representation theory). I believe that item 4 above
(which I hope to be able to present at a philosophy department colloquium)
will be of help to me in convincing philosophers that AI and philosophy
can fruitfully interact in ways that go far beyond merely listening to one
another's papers. Ideally, I would like to find a way of making it
possible for an appropriate graduate student in philosophy to write a
Ph.D. thesis on action theory or causal reasoning or semantics or
epistemology or ... , based on his active collaboration in the Advice
Taker project. I think it would be very beneficial for both philosophy
and AI if such collaborations were to become fairly routine, and I think
that I am uniquely qualified to be instrumental in making such a thing
happen here at Stanford. Also, in view of your own long-standing and
publicly expressed views on the need for interaction between philosphy and
AI, I naturally hope to stimulate your interest and possibly, your
participation, in attempts to get something started along these lines.
I'd like to discuss these possibilities with you further sometime after I
return from my vacation. -- Lew
∂31-Jul-80 1155 FFL
To: JMC, FFL
Julie Smith of Stanford Alumni Assn (7-2021) called to ask for a subject
for your speech to the Stanford Alumni on Friday, Aug. 29, 10:30 a.m.
∂31-Jul-80 2200 JMC*
calif on nuclear energy
∂01-Aug-80 0106 ME auto-detach and auto-logout
∂30-Jul-80 1522 JMC easing autologout
Now that we have plenty of job slots, we could replace autologout on
Datadiscs by autodetach. There could be an autologout after (say)
eight hours or when some limit of jobs was reached. I see we have
been up to 79 jobs. Is there a good reason for the current limit
of 80? What does an inactive job cost in core?
ME - The reason we hit the 79 job limit was from a bug and, I believe, a
runaway TTY (the SAIL/VAX connection). I don't expect us to hit that
limit normally, and if we ever do seem to be hitting it, we can raise the
limit (not dynamically, but simply by recompiling and reloading the
system). Each job slot, with no job logged in, permanently uses up one
page of physical core (out of the 3800 or so pages of user core), and
something like 85 words of job tables in the system. In addition, with an
(idle, say) job logged in, various pieces of free storage are used up; for
a detached job, probably about half a page per job and for a job attached
to a display about 2 pages of free storage (mostly for display storage).
All this free storage used by jobs comes out of a somewhat limited amount
of virtual address space in the system, where we typically run with about
30 or 40, but sometimes less, free pages. I estimate that we would normally
be completely out of free storage if there were about 220 job slots and
about 60 actual jobs logged in. When the system runs out of free storage,
it generally crashes fatally, since it depends heavily on having free
storage available. We would also probably run out of free storage with
about 85 jobs logged in with most on displays. I have another scheme
to revise system job data storage so that not all of it is always in
the exec virtual address space, but that is probably a bigger project
than the recently completed removal of the old job limit. It also would
only provide a savings of about 200 words of virtual memory per job,
although we might be able to save as much as 500 words per job. Possibly
a removal of some display storage from virtual space to physical space
could be achieved to provide some more headroom.
Anyway, getting back to autologout, we already provide the user option of
auto-detach instead of autologout, which users can enable with the DET (or
something like that) login option and the DET "privilege", which anyone
can have. However, the 2-hour limit for all jobs still applies to such
jobs, as does autologout at any time when we are actually running short on
job slots. I would not want to do away with the 2-hour limit, for various
reasons. For instance, when you are trying to find someone, it is
somewhat misleading to discover they are logged in, even though they
haven't run in 8 hours (which you might or might not discover right away).
Secondly, it does take up some free storage just to keep the job logged
in. And I don't see ANY reason for letting people stay logged in
semi-permanently like this. There is no advantage to the user, short of
not having to type "L PRG<cr>", and maybe a password, a couple of times a
day. Context for editing and program running can be saved from logout to
login by the RPGSAV features of logout and login. The only exception is
the case where the user leaves a program in the middle of execution,
whether it is the editor or some other program. But to leave such a
program unattended for more than 2 hours indicates to me a lack of
interest in the actual state of the program (or else a complete forgetting
of it), and in any case we can't guarantee to keep such inactive programs
around since there may be other active users who need the resources (job
slot, free storage, display, core). So I don't think users should be
encouraged to leave idle programs around for long periods of time. (Other
users are likely to kill such a program anyway, although they could be
taught not to kill such wasteful programs.)
Your practical objections to greater kindness to users may still be decisive,
but
1. Anyone can learn to check how long someone has been idle.
2. People are often interrupted and leave edits hanging. If they
forget to do αβ., they lose. This happens especially when the system
halts and they have to leave before it comes back to life.
3. If one can stay indefinitely logged in even when doing something
else, one sees one's messages as they appear.
4. The urge to discipline others is one that should be denied.
How about a login option that will do αβ. when autologout occurs
including when it is occasioned by SYSDOWN? It would be somewhat more
cautious if it put the hanging page in a file with a different name.
∂01-Aug-80 0430 RPG Decidability
To: MLB at SU-AI, gosper at PARC-MAXC, DON at SU-AI, DEK at SU-AI,
JMC at SU-AI
It was I who showed you this simple system. It is over the alphabet
{a,b,c,d,e,f} with productions:
ac=ca, bc=cb, ad=da, bd=db, ea=ace, eb=bde, f=fe, f=ef.
It is undecidable whether any two words over this alphabet are equal
under these productions. The theorem is called the Scott-Seyton theorem,
with Scott = Dana Scott. Seyton is Russian, and I don't think there is
an english language proof of it around, except that I proved it at Illinois
a long time ago and can roughly speaking reproduce it for you. My proof is
different than Seyton's, the latter which I feel is not a believable
proof. William Boone (who was at Princeton when JMC was there) was my
advisor at Illinois when I proved it and he seemed to think Seyton's
proof was in error too. In any case, he was talking about a book then
which may have been written and may have this proof in it.
The literature to look in is decidability theory in groups, semi-groups,
and Thue systems.
-rpg-
∂01-Aug-80 0715 JRA conference party
To: "@PA.DIS[1,JRA]" at SU-AI
We are planning a conference committee party for august 28 (evening) at our
house in redwood estates. Please let us know whether or not you can attend
and if so, how many will be coming (we have to tell the caterer soon)
john and ruth
Accept party invitation with pleasure for one maybe two.
∂01-Aug-80 1000 CG thesis
Any time you'd like to discuss the thesis is fine with me. If I'm not
at the department I can be reached at HP (857-8478). I can come down from
there to the department in about 10 minutes.
∂01-Aug-80 1013 FFL
To: JMC, FFL
Robert Grant called. His phone is 328-656l
∂01-Aug-80 1606 FFL Alison Bishop called. 7-2203.
To: JMC, FFL
∂01-Aug-80 2245 ME NS dates
∂01-Aug-80 1727 JMC
Do you remember in which year NS was first put up?
ME - The original APE came to life in summer 1972, and NS replaced it
around May of 1974.
∂02-Aug-80 0900 JMC*
cathy
∂02-Aug-80 1337 Wiederhold at SUMEX-AIM Giant file server initial meeting.
Date: 2 Aug 1980 1221-PDT
From: Wiederhold at SUMEX-AIM
Subject: Giant file server initial meeting.
To: Giant File Server (Galaxy-wide File System):
Although I do not yet have a reeesponsee from ED Shaw, I would like to
go ahead and meet Tuesday Aug. 5 at 3 pm, in the second floor conf.
room
(Jayne can you reserve it for us?)
in Margaret Jacks.
Agenda items :
Desirabilty and participants
Expected problems, means and costs to oveercome them
Subtask assignments for problem reesolution ( so we dont get hung in long loops)
Gio
-------
∂03-Aug-80 1427 ME
How about a login option that will do αβ. when autologout occurs
including when it is occasioned by SYSDOWN? It would be somewhat more
cautious if it put the hanging page in a file with a different name.
ME - This automatic αβ. before autologout has been discussed many times.
It is very difficult to implement, whether it writes out the current
file or just saves the text in a temporary file.
∂03-Aug-80 1524 REM via AMES-TIP Go endgame
I have obtained hardcopy of Jil Westcott's thesis on computing the safety
of groups at the "end" of a game without requiring more than one ply
of lookahead. I expect to be obtaining MIT-MC online files of her program.
When I do, I might embark upon interfacing her code to my code, and
upgrading everything in various trivial ways, and also implement my
algorithm for playing endgames optimally. Meanwhile, have you looked
through my ENDGO1.WRU yet? Any questions or comments? (Forgive the
style. Think of it as a cross between a HAKMEM and a table of results,
not really ready for publishing but with enough info to satisfy the
curious reader and to answer most questions.
Not yet.
∂04-Aug-80 0446 RAHE at USC-ECL Your Human-nets piece
Date: 4 AUG 1980 0445-PDT
From: RAHE at USC-ECL
Subject: Your Human-nets piece
To: JMC at SU-AI
cc: rahe
Applause, applause, applause. But I'll bet this note is the
minority! That comment on nuclear power should be good for at least
100 pages of Jane Fonda rhetoric.
-------
∂04-Aug-80 1226 kirkley at DNSRI chess.
From: kirkley at DNSRI
Date: Mon, 4 Aug 80 12:26-PDT
To: jmc@su-ai
cc: kirkley
Subject: chess.
file:chess, su-ai
rsvp: kirkley@dnsri
John -
Do you still have some good chess analysis or playing programs available ?
I'm Turing someone on the network with a program on the SRI Unix.
Regards,
Kirk
**
Somewhere we have the old M.I.T. program. Dave Wilkins at SRI would
know how to find it.
∂05-Aug-80 2046 JK
The NSF grant that I applied for through IMSSS will come through; in fact,
Pat tells me that grant total can be increased by about 20% and that this
might be enough to cover my entire academic salary for the next year.
Before we go this far I think that some issues should be clarified; maybe
you and Pat have already discussed this. Namely, being employed fulltime
through IMSSS would seem to imply a definite change in my relationship
with the AI lab: I would have to stop the work that I am currently doing
and devote all my time for improving the EXCHECK system at IMSSS within
the confines outlined in the grant proposal that I wrote last
February. As I said, maybe this has already been discussed, but it would
be important to get this clarified from the point of view of the work
that I am presently doing.
I will talk to Pat, but I need to know your preferences, because
maybe I can temporize about money or maybe I can persuade Pat of something.
∂05-Aug-80 2200 BUD ben
Ben had a relaps of his disease and depression, and won't be coming up after
all. Bud
Sorry to hear that. Anything useful I can do?
∂05-Aug-80 2304 JK
∂05-Aug-80 2118 JMC
I will talk to Pat, but I need to know your preferences, because
maybe I can temporize about money or maybe I can persuade Pat of something.
When would be a good time to talk about this?
∂07-Aug-80 1340 FFL
To: JMC, FFL
Please call Al Fenaughty at 21l 390 86ll.
∂07-Aug-80 1349 FFL
To: JMC, FFL
You may have noted that the luncheon honoring Les and Hersche is set for
Thursday, August 28, at ll:45 at the Holiday Inn. It is on your calendar.
∂07-Aug-80 1355 FFL
To: JMC, FFL
Ann Marinaro, referred to you by Kay Wax, asks if you would please call
her at 348-5793.
∂07-Aug-80 1319 BIS MACLISP future
To: BIS, DCL, JMC, TOB
CC: RPG
I have spoken with Dick Gabriel about the current state of MACLISP and its
possible futures. Dick will have finished his thesis in mid-October, and
there is an 80% probability that he will move on to CMU sometime after
that. What will become of MACLISP?
STATE OF MACLISP AT SAIL.
Because of the crunch on RPG to finish his thesis, we are now 100 edits
behind MIT's version of MACLISP. Our compiler is 3 releases behind MIT's.
PLISP has some very serious bugs and only ROD and RPG have been able to
use it without tickling them. Files need to be moved from RPG's area to a
system area and reorganised. Documentation is needed in how to make up
LISP. (Note: LISP cannot be assembled here; it must be done at MIT using
ITS paging.)
MIT is going to drop MACLISP on the 10 in favor of NIL or the LISP
Machine; thus, MACLISP will stabilise. Extensive systems work will not be
needed to keep up with MIT changes.
Dick estimates that 3 months full-time work will be required to bring MACLISP
into shape so that it can be maintained by a non-wizard or remotely.
A MODEST PROPOSAL.
Support Dick full-time for 3 months (mid-October thru mid-January) to
bring MACLISP into order. Salary should be that of a Systems Programmer.
After that, exchange some quantity of SAIL usage for Dick's maintaining
our MACLISP, perhaps remotely.
WHO SHOULD PAY?
Nobody keeps statistics on who uses MACLISP how much at SAIL. A good
guess is that the groups of JMC, DCL, and TOB are the heaviest users.
CS206, a Department course, uses MACLISP. Several unattached graduate
students have used or are using MACLISP for their theses. My feeling is
that MACLISP is a community resource and should be supported by funds of
CSD-CF. Alternatively, the burden of paying for MACLISP support may fall
on our projects.
Note: various HPPeople use MACLISP on the DEC-20 (e.g. Genesereth), and
this has also been maintained by RPG, who is willing to continue that
support as well. This argues further for CSD-CF support.
REACTIONS PLEASE.
Since we are the heaviest users of MACLISP, I think we should form a
unified front on this before confronting the powers of the CSD-CF. May I
have your reactions, please?
Many thanks for taking the initiative on a proposal for the support of
MACLISP, and I agree with your proposal entirely.
∂07-Aug-80 2008 JMC
Stanfor Barn tick. 327-7722 or CAPA 329-2623
1:30 to 6:30 today tomorrow or Wednesday
∂07-Aug-80 1937 LLW CLEAN SWEEP CELEBRATION
To: "@S1P[1,LLW]" at SU-AI
IN ORDER TO (SEMI-)PROPERLY CELEBRATE AND COMMEMORATE THIS WEEK'S
EXCEPTIONALLY GLORIOUS NEWS FROM CAPITOL HILL, THERE WILL BE A BURGER
FEAST EXTRAORDINAIRE AT FIRMWEDGE MANOR THIS SATURDAY AFTERNOON (9
AUGUST), COMMENCING AT 1700. PUBLISHABLE ASPECTS OF THE PROJECT'S RECENT
UNPREDECENTED POLITICAL/ECONOMIC TRIUMPHS WILL BE DISSEMINATED AT THIS
TIME, AS THANKSGIVING SACRIFICES ARE OFFERED, A TE DEUM SUNG, ETC.
PLEASE COME BY YOURSELF, OR WITH A SPOUSE/FRIEND (AND ANY CHILDREN EXCESS
TO YOUR PRESENT REQUIREMENTS FOR SAME), AND PARTICIPATE IN THIS GALA EVENT
(WHICH IS SLATED TO BE THE LAST DRESS REHEARSAL FOR THE PARTY TO CELEBRATE
THE ARRIVAL AT LLL OF OUR FY81 FUNDING).
RSVP TO CINDY (422-7286) OR CHRIS (422-0758), OR BY NET MAIL (OR RISK
GETTING LAST PARTY'S EXCESS HAMBURGER FOR YOUR PORTION!).
LOWELL
(FOR THE S-1 POLITICAL ACTION COMMITTEE)
Congratulations.
∂08-Aug-80 0044 POURNE at MIT-MC (Jerry E. Pournelle) did you get this?
Date: 8 AUG 1980 0346-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
Subject: did you get this?
To: POURNE at MIT-MC, JMC at SU-AI
I find this in my archives but I think no answer; not
that one is really needed. For a longer explication on the
Pajaro Dunes study conference, see mc:pourne;opport > which
will be published eventually, but maybe you oughta see it before
it comes out.
JEP
Date: 21 JUL 1980 0456-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
After Reagan's acceptance speech, with its clear implication
that this is NOT the era of limits, and so forth, I think he
MUST be committed to high technology; and I would think he could
use some dramatic ideas that would have the effect of focussing
effort in high tech and also capturing the US imagination.
The Lunar Colony is obviusly the right idea for that; it gets
some research going that will spill out into the general economy
(teleoperated equipment, good hands for manipulation--I saw
Vic's Stanford Arm the other day and the arm and brain are
really good, but the hand is primitive; and I talked to an Alvin
operator last week, and found to my horror that we spend all
that money to keep Alvin on the sea bottom, but give him really
terrible hands and even worse control actuators...)
Anyway, a lunar colony would, it seems to me, focus
attetnion on intrinsically valuable techn9ology for investments,
as well as have a dramatic (and properly romantically appealing)
payoff.
JEP
I suppose a Reagan administration is likely to be somewhat better
disposed towards NASA than the anti-technology crew that has substantial
influence among the Democrats and Carter himself. Maybe he
or his NASA nominee will be interested in a major long term goal like
a lunar colony. However, his technology advisers (and I hope they
are good) will be heavily involved in getting going again various
earthbound enterprises like nuclear energy that have been stymied by
the present administration. It can't hurt to propose, but if you
have a way of reaching him or his close advisers, beware of going
off half-cocked. Between now and the election, of course, he will
want to keep it simple, since that strategy has brought him a
tremendous lead. Most likely a lot of effort will go into trying
to win at least one house of Congress.
∂08-Aug-80 0051 POURNE at MIT-MC (Jerry E. Pournelle) survival of freedom
Date: 8 AUG 1980 0353-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
Subject: survival of freedom
To: JMC at SU-AI
Date: 16 Jul 1980 1110-PDT
From: John McCarthy <JMC at SU-AI>
John McCarthy, Computer Science Dept.,Stanford CA 94305
I remember about Survival of Freedom and hope to have some
alternatives for you in a few days. Have you seen Knowledge
and Decisions by Thomas Sowell? It's great.
This is a reminder. My deadline for material is about
mid←september so I can do my essays and the like. the absolute
deadline for everything is Mid october when the Publisher gets
It...
JEP
I plan to meet your September deadline.
∂08-Aug-80 0058 POURNE at MIT-MC (Jerry E. Pournelle)
Date: 8 AUG 1980 0400-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
To: JMC at SU-AI
We are much heartened here. A local TV pundit, former
Nixon advisor (Bruce herschensohn) is high up in Reagan
councils; and he has recently delivered himself of a channel 7
editorial commentary that wquoted five lines from my Step
farther Out, and made Charles Sheffield (he of presidency of AAS
who was staying here at the time) shout for joy.
We ddo, ity seems, have some influence. And true,
Reagan wants to Keep it simple for the election; but in my only
two contacts with him, he was not so simple as it might seem.
We also have ways to him through Possony and through Buckley,
and thus, although I cannot say we have any control, we do have
a route for rational argument to a man swayed by rational
argument. That's gotta beat what we got now!!!
∂08-Aug-80 0059 POURNE at MIT-MC (Jerry E. Pournelle)
Date: 8 AUG 1980 0402-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
To: JMC at SU-AI
Date: 08 Aug 1980 0055-PDT
From: John McCarthy <JMC at SU-AI>
I plan to meet your September deadline.
Groovy. It really should be an influential book. I hope to get
up your way in reasonable future; meanwhile, if you get down
here, do let us know.
On 20th, there is another III board meeting, and it would be nice to
see you that evening, though there is a possibility that I'll have
to scoot back to some AI conference thing. It would be interesting
to discuss technological initiatives.
∂08-Aug-80 0113 POURNE at MIT-MC (Jerry E. Pournelle) getting together...
Date: 8 AUG 1980 0416-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
Subject: getting together...
To: JMC at SU-AI
Alas, by 20 Aug my wife will have me off in the Carribbean,
assuming that the hurricanes have left anything of them; from
thence to Boston for the World SF Con, after which we return in
about second week September.
But do let me know of your comings to LA as we do have
much to discuss, including optimum strategy if indeed we do get
a repub Pres and one House of Congress; at least we have a
chance to push high technology to someone who thinks well of the
concept and has an advisor who reads my books!
OK, have a good trip to the Carribbean.
∂08-Aug-80 0215 DCL Maclisp
To: BIS
CC: JMC, TOB, RPG
The proposal to bring Maclisp up to a stable state and get rid of the
PLISP bugs is fine with me. We would like that.
- D
∂08-Aug-80 0918 RWW
To: ME, JMC, CLT
∂26-Jul-80 1939 JMC
To: RWW, CLT
I went through the list of files DON is proposing to delete
from [1,3]. I will hold out to save REDUCE and RLISP, and the
others have something to do with FOL or various versions of LISP.
If you think some of them should be saved communicate with DON
or ME. The complete list is on BBOARD.TXT[2,2].
BFOL DMP 128964 23-Oct-78 1142 1CLT BFOL 4-Sep-79 1 4
*BIGNUM LSP 256 15-Mar-70 0534 4-May-80 1 4
*BIGNUM REL 1536 25-Jul-72 1044 2ACH 4-May-80 0 1
*COMPLR DMP 30660 1-Mar-78 0145 W78RWW COMPLR 24-Jul-80 6 27
*COMPLR FAP 7676 25-Nov-77 1914 NETSYS F-DEC 3-Apr-80 1 3
*FOL SEG 46080 15-Jun-79 0030 S79RWW FOL 15-Jun-79 0 0
*FOL2 SEG 46080 13-Jun-79 1708 1 ME FOL 13-Jun-79 0 0
FOLML DMP 47044 8-Jun-78 1025 W78RWW FOLML 19-May-79 0 0
*FOLMRG DMP 10947 18-Oct-78 2134 1REF FOLMRG 29-Jul-79 0 1
LSPFNT DMP 4102 27-Feb-79 1637 1 JP COPY 13-Apr-79 0 0
*MLISP UCI 41984 19-Dec-74 0004 L70DAV MLISP 10-Mar-80 1 3
*MLISP DMP 24516 13-Nov-74 1743 ENGHJE MLISP 18-Jun-80 6 28
*MLISP2 DMP 34756 31-May-74 2114 1HJE MLISP2 2-Jun-80 1 4
*MLISPC DMP 34756 13-Nov-74 1745 ENGHJE MLISPC 10-Jun-80 3 15
*REDUCE DMP 51140 17-Nov-77 1321 1ACH REDUCE 16-Jul-80 7 29
*RLISP DMP 25540 22-Nov-77 0909 1ACH RLISP 17-Jul-80 1 6
Marty please do not delete from the system those files from John's list
that I put stars in front of. They are used by or useful to the FOL group.
∂08-Aug-80 1259 RMS at MIT-AI (Richard M. Stallman)
Date: 8 AUG 1980 1558-EDT
From: RMS at MIT-AI (Richard M. Stallman)
Sent-by: DGSHAP at MIT-AI
To: JMC at MIT-AI
It appears that Minsky can fund my trip, so I will be coming out after
all.
I don't seem to be having much luck finding a place to stay, out
there. Is there someone there who you can ask to help me?
I hear that SAIL is ceasing to exist???
What's happening?
I will see what I can do about helping find a place to stay. For how
long will you be coming and when?
Scientifically, there isn't much change. Administratively, SAIL is
being absorbed by the Computer Science to save money, although the
ARPA contract remains.
∂08-Aug-80 1335 FFL
To: JMC, FFL
The bookstore carries MODELS OF THOUGHT but does not have a copy at present.
She would not predict when another copy would come in.
∂08-Aug-80 1423 ME files on [1,3]
To: JMC, CLT
CC: RWW
RWW says he doesn't care about these two files from [1,3]. Do you?
BFOL DMP 128964 23-Oct-78 1142 1CLT BFOL 4-Sep-79 1 4
FOLML DMP 47044 8-Jun-78 1025 W78RWW FOLML 19-May-79 0 0
∂08-Aug-80 1429 Nilsson at SRI-KL AIJ Paper
Date: 8 Aug 1980 1428-PDT
From: Nilsson at SRI-KL
Subject: AIJ Paper
To: JMC at SU-AI
cc: Nilsson
John, I just finished reading your AI Journal paper on
circumscription. It's excellent! It was well written and
not hard to understand. (I had a little trouble with the
earlier version.) I really think it's important for more
AI-ers (of the "knowledge engineer" variety) to understand some
of the issues that you've been worrying about for many years.
Papers written as well (with examples) as your AIJ paper can have
an impact on them. -Nils
-------
∂08-Aug-80 1612 Wiederhold at SUMEX-AIM A proposal first draft for a giant file server
Date: 8 Aug 1980 1553-PDT
From: Wiederhold at SUMEX-AIM
Subject: A proposal first draft for a giant file server
To: Giant File Server (Galaxy-wide File System):
cc: csd.tan at SCORE, csd.craven at SCORE, csd.olumi at SCORE,
cc: csd.chen at SCORE, ark at SAIL
A GIANT FILE SUPPORT SYSTEM FOR STANFORD
Draft: Gio Wiederhold
This note contains two parts:
1. A sketch of a proposal for a large File Support System
for Stanford University.
2. Notes from an initial meeting on this topic held
August 5, 1980, 3:00 PM in MJH 252.
A proposal for a large file support system for Stanford University,
GIANT
Gigabyte Information Access for a Network [Tomorrow]).
A. The Problem
The demands for the storage, retrieval, and sharing of large amounts
of data are increasing at a rapid rate, and all computer systems at
Stanford are continuously under pressure to increase their storage
capacity, provide means to archive data that appears too expensive
to be kept on-line, and to help with the moving of files from one
machine to the other.
As services and research at Stnford grows these demands are expected
to increase at a rate faster than the demands for computing power
itself. There appears to exist a commonality of needs among otherwise
diverse computing environments. One of the needs is the backup
processing for long term storge reliability, now carried out
separately and differently at each of the Stanford Computer systems.
B. Objectives
[ John McCarthy : Scientific value of large storage capacity ]
We propose to develop and install a system to serve Stanford University's
computer storage requirements with advanced hardware and software
technology. This system should be directed towards the following
goals:
1. Make storage, using the most effective large storage devices,
available to all connected users.
2. Assure that future improvements of storage technology will be
transparent to the user.
3. Operate the storage services as a pool for Stanford.
4. Provide a high level, consisten interface to the users programs.
5. Provide reliability through backup, monitoring, and restoration
services.
While the motivation for this work has been triggered by the
announcement of very economical high-speed disk units, with gigabyte
capacities, a significant and important development effort will be
in the software interface for such a system. Some details will be
given in Section D.
!
The objective of the software will be to fetch data from the mass
storage device, select the units that actually carry information
for the user, group them appropriately, and transmit them. These
units may range from records or textlines, to entire files or sets
of many pages.
The intent of this approach is to reduce the bandwidth requirement
in the communication channels, and as an important by-product, to
also reduce the work done within the served computers file systems,
which conventionally carry out the data selection process themselves.
C. Current Development and Status
This proposal depends on ongoing efforts at Stanford and industry.
1. High Bandwidth Local Networks
The sharing of large quantities of data could not be proposed unless
rapid and flexible communication links are available. The network
development at Stanford, using Ethernet technology to link a large
variety of machines, is a precondition for this project. The current
ethernet has a file server node, serving the ALTO and VAX computers
and we expect to learn from this experience. A recent Stanford
CS thesis (John Shoch) has demonstrated the non-degrading behavior
of the Ethernet under conditions of extreme competition for
communication service (several factors over available bandwidths).
2. Common Database Facilities
A project, currently in progress, is mounting an identical relational
database management system, RIMS (originally developed by SAI for
EPRI) on three distinct and different computers (the SCORE DEC-20/60,
the VLSI project DEC VAX 780, and the SU/IBM joint Distributed
Computing study 4331).
If high level interfaces are to become a reality, identical facilities
have to be available locally, for checkout, independent development,
and confidence that programs and programmers will have a choice
independent of services from a remote system.
Related to this work is the development of a modern transportable
file system, FLASH. FLASH is written in PASCAL to enhance transportability,
and has been designed to serve PASCAL, Interlisp, FORTRAN, COBOL, and
PL/1 users with comprehensive and compatible file services. The
motivation for this project were the difficulties encountered in the
application of multiple experimental database techniques to
conventional databases, since each language made distinct assumptions
about file structures. [ref. Allchin, Keller, Wiederhold 80]
3. High Level Data Models
Research in the KBMS project at Stanford has developed a high level
structural model which permits queries of relational generality to
be processed, while permitting the physical design to take advantage
of specified semantic constraints. This permits efficient binding
to be utilized in the implmentation, so that the generality does
not reduce access efficiency, or vice versa. A current database
system seletion process at Stanford has reinforced the impression
that todays systems, that are concerned with efficiency, impose
horrible limits and constraints upon the users [ref. John Sack's
report to Ed Shaw, forthcoming]
!
D. Method of Approach
We would plan to acquire a system consisting of a modest processor,
fairly large disk capacity, an ethernet connection, and some
device as tape or an ATL connection for backup storage. We expect
the processor to have a large virtual address space, automatic
paging, adequate bandwidth to handle the expected devices and network
services, and good reliability aspects.
The aquisition might proceed in two phases. A prototype would use the
facilities of the SU/IBM joint study in Distributed Computing. Before
service could commence however, a service-oriented system would be
obtained to avoid impact on the research activities of the project.
The availability of the project machine would continue to enhance
development and provide some backup capability.
During the acquisition process formal interface specification will be
developed and circulated for comments. We expect that the interface
will have the appearance of a relational query and update language,
with an additional specification of structural integrity constraints.
All data will be referred to for retrieval by name, that is in symbolic
form. Names are assigned by the user, and may be interpreted
by the user as numbers. Names will exist for
Tables or relations
Lines or records (including pages, i.e., fixed length records)
Fields of records (optional)
Connections between relations (optional)
A schema will be kept in the processor in order to establish symbolic
to physical correspondence, record the data format, note the
structural constraints, and catalogue redundant access structures as
indexes, hash tables, etc. The data format specification can help
with selection, sharing, and compression of data, but an uninterpretable
binary format should also be provided. Data would be kept normally
in the owner's format, but authorized users could request translation
of character-set, integer, and real formats during retrieval. Update
using converted formats might be severely constrained to avoid the
introduction of errors.
The users may require that access structures be created or not be
created; in the unspecified case the storage system is free to decide
to create these structures, trading increased update cost for more
rapid retrieval according to its internal cost functions. The schema
itself will be kept in virtual memory to take advantage of
processor cache or paging technology.
!
We expect the system to retrieve blocks of data from the mass storage
device of system optimal size, select the units desired, if needed
combine them with related data, and ship them to the users as sets
of records, with a cardinality of 1 to very many. The optimal size
for internal manipulation and transmission may be a size near to the
most common current block size, say 2000 characters.
The records may be placed in compressed form on storage. Decompression
may take place in the mass storage system or, by mutual agreement,
on the users side. The latter can significantly reduce communication
bandwidth for certain types of data. An investigation whether user
limits can help the compression algorithms is of interest.
Backup of data may be accomplished by infrequent full or incremental
dumps, and an extensive logging service for updates. Since updates
can be specified by field, rather than by record or block, the volume
to be logged might be considerably less. Periodically the logged
data might be consolidated with data from the physical areas updated,
to make reapplication of voluminous logs more efficient.
The system would be brought into operation serving a few specific
projects with a close interaction to develop and improve the
approaches. The design of the system will not constrain the
attachment of larger or different storage devices than those now
foreseen, and should also handle multiple network connections to
increase bandwidth.
The reliability of the GIANT system will be of great concern and
duplicate hardware, sharing the mass storage devices and networks
should be possible, and perhaps part of the initial implementation.
Associated efforts could involve replacement of user file servers
with identical services to be provided wholly through GIANT,
transforming those systems to truly computational engines. Such
systems would have a much lower cost. Even the Stanford-Livermore
S-1 system however is a candidate for such an operation.
!
E. Resources
People
Budget
Equipment
!Appendix A
Comparision of Systems
---------------------------------------------------------------------------
device | GIANT | proto-GIANT | CIT service | DEC | Syst.Con. |
---------------------------------------------------------------------------
| | | | | |
CPU | ? | IBM 4331-2 | IBM 3033 | ? | DEC-11 |
| |(joint study)| | | |
| | | | | |
Channel | ? | upgrade ? | incl. | ? | incl. |
| | | | | |
Contr'r | IBM 3880 | adap. 3201 | incl. | incl. in | incl. |
cost | $65.450 | $3.780 | | first dsk| |
| | | | | |
Disks | IBM 3880 | IBM 3370 | IBM 3350 | ? | incl. |
first | | | | (2 disks)| |
cap. | 2.500 Mb | 819 Mb | incl. | 929 Mb | 2.600 Mb |
cost | $97.650 | $38.850 | | $120.000 | total |
#lim.| 2/contr.| ? | | ? | |
more | | | | (2 disks)| |
cap. | 2.500 Mb | 819 Mb | incl. | 967 Mb| |
cost | $81.000 | $24.570 | | $51.000 | |
#lim | 3/first | 3/first | | 3? | |
| | | | | |
Totals ------------------------------------------------------------------
Min.sys| | | | | |
cap. | 2.500 Mb | 819 Mb | (incr. | 929 Mb| 2.600 Mb|
cost | $163.100 | $42.630 | rate) | $120.000 | $125.000 |
$/Mb | $66/Mb | $56/Mb | $40/Mb*month| $130/Mb | $48/Mb |
Max.sys| | | | | |
cap. | 20.000 Mb | 3.270 Mb | (bulk | 3.850 Mb| ? |
cost | $746.790 | $116.340 | rate) | $273.000 | |
$/Mb | $38/Mb | $36/Mb | $9/Mb*month | $71/Mb | |
| | | | | |
--------------------------------------------------------------------------
Notes | wo CPU, ch | wo CPU cost | incl. people| estimates | prelim. |
| | | | | info. |
--------------------------------------------------------------------------
!Appendix B
Comparison with Commercial Products.
System Concepts and MASSTOR are commercial ventures into this area.
A brochure of MASSTOR made it clear that their aproach differs considerably
in scope from the GIANT proposal.
MASSTOR is a new systems company in Sunnyvale proposing to provide
a shared remote large storage system (VSS). The devices proposed
are
3350 disks - 630 MB each
3850 IBM mass storage (vido tape technolog)
ATL Automatic tape reel handlers.
Communication hardware is coax transmission, the protocol is not
specified but does not seem to be Ethernet.
They propose to execute the low level I/O commands issued by channels
of the served systems.
This avoids changes to user interfaces but puts extreme demands on
the network since the commands cannot be understood at a high level,
and hence have to be executed literally, without any data reduction
by selection of data in the VSS.
Their principal software effort is towards effective staging from
massive tapes to fast disks.
-------
Gio - I'm impressed with the start you have made on a proposal.
However, I have two reservations about the present approach.
First, the initial emphasis on Stanford's need may
jeopardize getting money for it as a research proposal. Second,
I have some doubts about whether we want to be involved with C.I.T.
I have talked to Ed, and he thinks that CSD might be able to swing
a ten gigabyte system at a cost that I estimated at $700K (and
Ralph regards as reasonable) without involving C.I.T. The meeting
left me with continued doubts about their interest and capacity
to avoid excessive costs. I hope to phone you later this weekend
when I have had a chance to think a bit more about it.
∂09-Aug-80 1132 Wiederhold at SUMEX-AIM (Response to message)
Date: 9 Aug 1980 1125-PDT
From: Wiederhold at SUMEX-AIM
Subject: (Response to message)
To: JMC at SU-AI
cc: csd.feigenbaum at SCORE
In response to your message sent 09 Aug 1980 0944-PDT
I agree with you reservations, and know we should focus it on one
or the other, partially depending wher we
will look for resources. If an excessive emphasis is given to
rapid service the research components will be neccessarily
compromised.
I would like not exclude CIT completely, I think it is not
healthy for the university as an instituition to operate a
service facility that lives completly seperate from academic
concerns. The same notion holds for the SLAC operation.
Perhaps a notion could be that when things are working ,
and have no longer a research component, that then it
could become a service run by and for CIT. The time would be quite
unspecified perhaps, and would depend on CSD having new and
better resources.
Personnally my strongest interests are of course in the
system interface specification, for me a critical research issue
in the building of all large systems.
I may not be at home this afternoon, taking the kids boating,
but will be at home ( 851-8363 ) all Sunday and prob. Monday.
Gio
-------
∂09-Aug-80 1255 REG
To: BIS, JMC, TOB, DCL
I'm willing to have CSD-CF support Dick Gabriel for three or four months
for the purpose of getting MACLISP updated and stablized for SAIL and for
TOPS-20.
∂09-Aug-80 1759 LWE
Have message (Sat PM) - will call ASAP
∂10-Aug-80 0216 RPG
∂10-Aug-80 0129 JMC
Is there a way within STEP of evaluating an expression on the side?
What I do is go into a αB break loop at whatever point is interesting
(just type αB) at step. Then you're in a read-eval-print loop.
<alt>p continues in the step. It will prompt with ? as usual.
You can then type αP to print the current expression that is being
asked about, or αβP to pretty-print it.
-rpg-
Thanks. Actually I am presently interested in seeing more of the
value of the previous expression that it has just printed to
a set depth and length. Does this expression have a name?
By the way BIS, I assume that BIS's proposal to pay you to clean up
MACLISP and which was accepted by REG, meets with your approval.
∂10-Aug-80 0235 RPG Value
The value of the last thing should be in %value, as the current expression
in in %ce.
I assume that the offer is acceptable, contingent on the salary not being
ridiculously low. I expect Oct 15 is a good starting date, and I've left
room in my proposed time schedule to entertain improvements and frills
suggestions.
-rpg-
Thanks.
BIS is the pusher, REG will be the offerer. I will add persuasion if asked,
as I have, more or less unasked.
∂10-Aug-80 2010 RPG Offer
To: REG
CC: BIS, JMC, DCL, TOB, RPG
I accept your offer in principle, contingent only on a satisfactory
salary decision. A starting date of Oct 15 would be acceptable.
I expect that given a 3-4 month time period there will be plenty of
time available for more projects than have been discussed to date. I
will be entertaining suggestions during the next 2 months for useful
tasks to do.
Currently I am considering a screen editor to enhance the current
in-core editor, an extension to FAIL allowing it to produce FASL files
(as MIDAS does now), smarter hisegment management, and BPS management
routines (which everyone writes independently).
-rpg-
∂11-Aug-80 1438 FFL
To: JMC, FFL
Please call Doug Hofstadter at 326-67ll. He will be here for a few days.
∂11-Aug-80 2201 LLW . . .The Tents Of The Wicked
To: JMC
CC: LLW
∂07-Aug-80 2304 JMC
Congratulations.
[Thanks, John. Having learned to live with rather harsh and sustained
adversity, the challenge now facing us seems to be how to survive major
success. Lowell]
∂12-Aug-80 0213 LWE
Trying to call you - will try again tues.
∂12-Aug-80 0828 FFL
To: JMC, FFL
Please call Jane Akre at 505-265-766l.
∂12-Aug-80 1000 JMC*
connie about notes
∂12-Aug-80 1718 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Date: 12 Aug 1980 1716-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
To: JMC at SU-AI
cc: csd.feigenbaum at SU-SCORE, csd.friedl at SU-SCORE
In-Reply-To: Your message of 12-Aug-80 1651-PDT
I just talked with Paul Friedl. We set up at tentative meeting time
for next monday at 4pm. Is that OK on your schedule?
Ed
-------
The time is ok; your office, I presume?
∂12-Aug-80 1926 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) meeting next Monday
Date: 12 Aug 1980 1921-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: meeting next Monday
To: jmc at SU-AI, csd.friedl at SU-SCORE
cc: csd.feigenbaum at SU-SCORE
reconfirming the time: 4pm
Let's do it in my conference room.
Ed
-------
∂12-Aug-80 2001 REG
Ok, I've added Professor Erbring to the people data base. We'll
monitor his usage for awhile to see if he deserves to buy an aliquot
∂12-Aug-80 2243 LWE saving digests
thanks for all the help!!
i followed your suggestion and entered to $request to ns (works fine).
question: is there a waits feature that would allow me to substitute
a file for my tyy, to receive ns response to a search request (if
so, i could save digests back to aug 1 as well - can't be done via
automatic feature since start date already passed)? if not, that's
fine (it's only a couple of weeks difference); i am truly delighted
at having a chance to access your n e a t facility! (by the way,
lwe is still "unknown" to waits, it seems). thanks again.
∂13-Aug-80 0843 JRA SAIL USAGE
JOHN,
I JUST GOT A MESSAGE FROM DENNY BROWN PRECIDTING MY REMOVAL FROM THE MACHINE
SINCE RUTH AND I WILL BE LOSING MONEY ON THIS CONFERENCE, HOW ABOUT BLESSING
ME AS A USER AT LEAST 'TIL THE END OF THIS YEAR?
JOHN
I'll see what I can do about regarding helping the LISP conference as a
legitimate overhead expense justifying an overhead account till the end of
the year.
∂13-Aug-80 1459 100 : REM via SU-TIP PCNET/DIALNET
To: MRC at SU-AI, JMC at SU-AI
I now have a computer with s-100 bus, and have also an auto-everything
modem (300 baud). Within the next week I expect to start work on a
full implementation of PCNET. DIALNET is rather fuzzy because the
computer doesn't have interrupts (as far as I know) and DIALNET might
be rather difficult to implement without them. Thus I am not at
this time formally requesting a 300-baud port into SU-AI DIALNET, although
when I get interrupts I may re-open the question.
∂13-Aug-80 1524 FFL
To: JMC, FFL
Lois Shapiro of the Canadian Broadcastin Corp. wishes to speak with you
about the panel at the conference on Wednesday in which you are participating.
Do you wish to call her, 415 664 1348, or is there someone else I may refer
her to?
I called Lois Shapiro.
∂14-Aug-80 1221 CG Kreisel's address
Kresel's addres in Italy is:
c/o Dr. F. del Franco
80122 Napoli
Via Giorgio Arcolea 19
Italy
∂14-Aug-80 1458 RWW working
Yes I think we ought to talk. how about early next
week after the conference.
roichard
How about Monday 3pm here?
∂14-Aug-80 1927 RWW
sounds great I will see you then and also this weekend
∂15-Aug-80 0245 LWE
thanks again! (you're always one step ahead of me)...
∂15-Aug-80 0900 JMC*
royko.xgp
∂15-Aug-80 1105 100 : REM via AMES-TIP Fountain of canon balls for levitating a skytower
To: RWG
CC: JMC
(Vacuum chamber containing up&down streams of bowling balls used to
support very thin tower as it is built toward the sky.)
First approximation indicates although there is vertical force
to keep the tower up, horizontal forces of reaction
of balls against contour of tower will be unstable. Consider what happens
if there is a very slight kink (bend) in the tower just for a brief moment.
Balls traving thru the kink must change directin to remain within
the confines of the vacuum chamber. As they change directin, thus change
momentum, there is a reaction against the walls of the chamber,
in the direction that increases the amount of deviation from straightness.
Thus instead of remaining vertical the tower will kink in two at the
slightest deviation from vertical. If the tower is large in diameter
(say 10 cannonball diameters) it may be possible to stabilize by diverting
balls in the "wrong" direction in the vicinity of the kink to force
the kink back to straight, but if at any time the kink becomes big
enough for balls to strike side of chamber, this method breaks
down and the kink suddenly goes into positive-feedback mode as above.
Has anybody analyzed this mechanism for stability more thoroughally?
I believe Lowell Wood and Rod Hyde have analyzed stability.
∂15-Aug-80 1119 FFL
To: JMC, FFL
I have more pendalex folders if you need them. I do not have in stock the
tabs for them. Will order them today. If you want to do ahead and do your
sorting, just put the folders in the file with a slip of paper in the front
of each one with the file name you wish. As soon as I get the tabs, I will
go thru them and type and insert the tabs for you.
∂15-Aug-80 1358 FFL
To: JMC, FFL
Jussi Ketonen's appointment on ARPA expired 8-3l-80. Betty Scott asks if
it is to be continued and if so, at the same level? He is presently paid
61700 monthly.
It is to be continued with a raise. Suppes's office has details of
when raise starts, but I suppose it is Sept. 1.
∂15-Aug-80 1502 CJS classnotes for the Fall quarter
To: JEF at SU-AI, RWF at SU-AI, JMC at SU-AI, DPB at SU-AI
CC: CET at SU-AI, CJS at SU-AI
Okay, I've checked with both the Bookstore and the printer, and here's
where we stand. If everybody can get the master copies for their various
classnotes to me by September 9th (the 10th at the very latest), we can
still get them to the Bookstore before classes start. Will anyone have
a problem meeting this deadline?
Connie
∂15-Aug-80 1654 OWE Visitors from Norway
To: "@SEM.DIS[SEM,VER]" at SU-AI
Assist.prof. Stein Gjessing, from univ of Oslo, and Kay Olsen ,from univ
of Tromso, will visit the verification group at Stanford during the week
25 to 29 august. Their main interests are parallel programs (design,
scheduling, specification & verification of monitors and processes)
and micro computers. They would like to talk to people with similar
interests. For more information, or suggestions, contact me( I am a
visiting scholar in the verif. group - also from Norway) , owe at SU-AI.
I will be back on 25.august to answer any mail.
Olaf Owe
∂16-Aug-80 1148 Kanerva at SUMEX-AIM Menus
Date: 16 Aug 1980 0912-PDT
From: Kanerva at SUMEX-AIM
Subject: Menus
To: JMC at SAIL
cc: Siegman
Your comment about menus in your Criteria for Usefulness of Computers
in Offices (1980)
Menus are bad, because as soon as a user gains the
slightest experience, he hates having the screen cluttered
up with changing menus.
brings to mind the following passage:
From TRAVELS INTO SEVERAL REMOTE NATIONS OF THE WORLD BY LEMUEL
GULLIVER, by Jonathan Swift (1726).
(In Part III, A Voyage to Laputa, Balnibarbi, Luggnagg, Glubbdubdrib,
and Japan, Chapter V on the Grand Academy of Lagado, about a page from
the end of the chapter.)
We next went to the school of languages, where the
professors sat in consultation upon improving that of
their own country.
The first project was to shorten discourse by cutting
polysyllables into one, and leaving out verbs and
participles, because in reality all things imaginable
are but nouns.
The other, was a scheme for entirely abolishing all
words whatsoever: and this was urged as a great advantage
in point of health as well as brevity. For, it is plain,
that every word we speak is in some degree a diminution
of our lungs by corrosion; and consequently contributes
to the shortening of our lives. An expedient was there-
fore offered, that since words are only names for THINGS,
it would be more convenient for all men to carry about
them, such THINGS as were necessary to express the
particular business they are to discourse on. And this
invention would certainly have taken place, to the great
ease as well as health of the subject, if the women
in conjunction with the vulgar and illiterate had not
threatened to raise a rebellion, unless they might
be allowed the liberty to speak with their tongues,
after the manner of their forefathers: such constant
irreconcileable enemies to science are the common people.
However, many of the most learned and wise adhere to the
new scheme of expressing themselves by THINGS; which hath
only this inconvenience attending it; that if a man's
business be very great, and of various kinds, he must be
obliged in proportion to carry a greater bundle of THINGS
upon his back, unless he can afford one or two strong
servants to attend him. I have often beheld two of those
sages almost sinking under the weight of their packs,
like pedlars among us; who, when they met in the street
would lay down their loads, open their sacks, and hold
conversation for an hour together; then put up their
implements, help each other to resume their burthens,
and take their leave.
But for short conversations, a man may carry
implements in his pockets and under his arms, enough
to supply him, and in his house he cannot be at a loss;
therefore the room where company meet who practice this
art, is full of all things ready at hand, requisite to
furnish matter for this kind of artificial converse.
-------
∂17-Aug-80 0050 ME DM3025 questionaire
To: BIS, DCL, FWH, LES, TOB, WP, ZEN
CC: RPG, JMC
We have had some trouble on certain of the DM3025s with horizontal
wavering of the screen, apparently because of power supply trouble.
If there are any more such of these terminals with this problem,
we may be able to have this problem fixed while the terminals are
undergoing the final upgrade. Please let me know if your DM 3025
shows symptoms of the text wobbling or wavering, especially in the
horizontal direction, but even in the vertical.
Please also let me know very soon when you can bring in the DM 3025
you have for the upgrade. This must be done soon.
My terminal is in the shop. ZEN is long gone, and the terminal he
was using is one of those belonging to CASBS which we are buying
and which Hersche knows about.
∂17-Aug-80 1606 MINSKY at MIT-AI (Marvin Minsky)
Date: 17 AUG 1980 1905-EDT
From: MINSKY at MIT-AI (Marvin Minsky)
To: jmc at SU-AI
Susie was trying to find you. If you read this, call her.
∂17-Aug-80 2059 RWG
To: REM, JMC
intuitively (which i surmise is how rem is operating, too),
it ought to be possible to concentrate the momentum transfer toward the
tip, so as to create lots of upward tension, from which one should be able
to create stability.
∂17-Aug-80 2104 RWG
i haven't done much with finding out about california environmental
impact statements. you are doubtless aware that legal firms which prepare
such things are in the yellow pages under "environmental and ecological
services", e.g. Ecological Impact Studies Inc, and about four others. i haven't
the chutzpah to actually call any of them up. i have tried friends of the earth,
peninsula conservation center, and p.a. office of conservation, where i spoke
with an engineer named Mike (telephonically). i think that there is no specific
state counterpart to EPA, and therefore no general state requirement for EIS.
rather, various state and local agencies have various requirements depending on
the activity, e.g. PUC, State Energy Commision, Cal OSHA, and Bay Area Air
Quality Management District, <water quality homologue>, and the city agencies
that want building permits, etc. Palo Alto, it turns out, does require an
Environmental Impact Assessment for certain activities. There's an 800
number for bay area air pollution complaints. (there's also a statewide
Solar Energy Information Hotline (800), which i should call).
It turns out that the low interest loans for attic insulation are "voluntarily"
provided by PG&E, although the PUC is supposed planning to force all utilities
to assist with solar hot water heating, so Mike thinks. Also(?), he says that
a provision of the National Energy Act of 1978(?) calls for billions to be
spent by US utilities along similar lines, but it is still grinding through the
bureaucracies.
my guess is that the "horrors of insulation" have quietly but thoroughly
permeated the agencies preparing these programs, since the perpetrators are
locally nonstupid.
the strategy may have to be local vs state or federal vs federal (in particular,
the Residential Conservation Service, whatever that is or becomes), instead
of state vs state. anyway, i don't seem to be very good at this, but will
persist if you don't already have better data than the above.
Many thanks. Please persist.
∂17-Aug-80 2218 RWG
denny brown says i evaporate from waits next fortnight. i currently
have 300k filled with life hacks, fractal graphics, math paper pox files, a few
experimental programs, and miscellany. i have barely touched any of it since
last november, indicating that it is all pretty dispensable to me. what would
be painful to lose, however, would be home access to MIT-MC, e.g. when i set it
one last computation while at parc, drive home, and see if it converged or ran
out of core. is there any sense in which CSD could find this activity
sufficiently in its interests to permit its continuation?
I'll see what I can do about your continued access.
∂17-Aug-80 2354 PMF Eschewing Kinky Tower Designs
To: REM at SU-AI
CC: LLW at SU-AI, JMC at SU-AI, minsky at MIT-AI, RAH at SU-AI
∂17-Aug-80 1728 REM via SU-TIP
JMC says maybe you analyzed lateral stability of vacuum column that
permits the cannon-ball fountain to function for levitating a tower
to space. Is your analysis written up somewhere (online or elsewhere)?
[Only in Rod's manuscript equations and (LLL) computer output. The analysis
indicates that the tower is subject to the mechanical analog of the
magnetic kink instability, moreover with a regrettably short growth time.
Unable to devise a tower variant to deal with this purely passively, we've
altered basic aspects of the tower design to obviate the problem with an
active control system. We continue to grind away on the various analyses
needed, the design itself, and even the documentation thereof; the latter
*really* will appear sometime this side of the indefinite future. Lowell]
∂18-Aug-80 0949 Feigenbaum at SUMEX-AIM visit of important person from ONR
Date: 18 Aug 1980 0948-PDT
From: Feigenbaum at SUMEX-AIM
Subject: visit of important person from ONR
To: buchanan, jmc at SAIL, lenat, TW at SAIL,
To: CSD.GENESERETH at SU-SCORE, shortliffe
cc: csd.genesereth at SCORE
The Chief of Naval Operations, under impetus from the Chief of Naval Research,
has convened a panel on the state of AI and its ultimate usefulnes to the
Navy. This is the highest level "review" of AI that the Navy has ever done.
The chairman of that panel is Dr. Albert Brandenstein of the Center for
naval Analyses. He is visiting AI centers to acquaint himself with the field
before he gets the show rolling (prima facie a sensible attitude!). He will
visit us on September 4.
I will be gone by that time. I would like to ask Bruce to coordinate the
arrangements for his visit. My view is that this guy is worth time and effort,
since if the panel gives a strong positive statement, good solid ONR monies
will flow for years into the future.
Terry and John, who could use more money soon, should attend carefully to this.
All the other people listed have some sort of ONR money too. Binford
should be alerted.
Brandenstein will show up in my office or Bruce's office at 9am on Sept. 4.
His background is engineering.
Ed
p.s. let's get together a packet of material that Bruce can send in advance
with a cover letter to prepare him for the visit (e.g. the new HPP brochure;
some scientific papers; department brochure, etc.)
His address is:
2000 N. Beauregard St.
Alexandria, Virginia 22311
-------
I have rediscovered and reread Feigenbaum message about Brandenstein. I
forgot because of all the conferences. 9am is still very early for me, but
I do want to meet him.
∂18-Aug-80 1140 LGC PUB Problem
Your file COUNTE[E77,JMC]/2p/33l gives rise to an "UNDEFINED IDENTIFIER A2"
error when I attempt to PUB the file. Is this something that you can easily
fix?
Your presentation on counterfactuals at the workshop yesterday has led me
to look again at what you have written on the subject, and this will
probably lead to my writing up some comments on that material during the
next few days.
I have fixed that bug, but the file is just a start on a draft, and
it has never been pubbed.
∂18-Aug-80 1848 Wiederhold at SUMEX-AIM For your info - data collected about our computing for NRC .
Date: 18 Aug 1980 1846-PDT
From: Wiederhold at SUMEX-AIM
Subject: For your info - data collected about our computing for NRC .
To: jmc at SAIL
Computing Capability for Computer Science Research at Stanford.
Estimated for the National Research Council Study on the
Conduct of Basic Research and its Effect on Industrial Development.
Summer 1980. Gio Wiederhold and Forest Baskett.
Equipment Person # CPU @MIPS Main MEMORY # Disks Capacity/disk
x.x/sec Kbyte Mbyte
----------------------------------------------------------------------
Sail KL | REG | 1 | 1.3 |10,600(2.25Mw)| 8 | 200 |
KA | REG | 1 | .3 | none | none | |
----------------------------------------------------------------------
CIS VAX1 | REG | 1 | 1.0 | 2,500 | 1 | 200 |
VAX2 | REG | 1 | 1.0 | 1,000 | 1 | 200 |
----------------------------------------------------------------------
SCORE | REG | 1 | 1.3 | 4.710 (1Mw) | 4 | 200 |
----------------------------------------------------------------------
SUMEX 10 | Rindf.| 2 | .6ea.| 2,305 (512Kw)| 4 | 200 |
20 | Rindf.| 1 | .3 | 2,305 (512Kw)| 1 | 200 |
----------------------------------------------------------------------
ALTOS | REG | 16 | .3ea.| 256 each | 2 each | 2.5 |
FS | REG | 1 | .3 | 256 | 3 | 2x300 + 80 |
----------------------------------------------------------------------
IBM4331 | Noah | 1 | .2 | 1,000 | 4 | 65 |
ser.1B| Noah | 1 | .3 | 128 | 1 | 9 |
ser.1DF| Noah | 2 | .3 | 256 | 2 each| 58 +9 |
ser.1E| Noah | 1 | .3 | 256 | 1 | 9 |
----------------------------------------------------------------------
Total CS + CSL | 30 | 12.9 |29,668 | 62 | 4972 |
======================================================================
Other CS + CSL usage
----------------------------------------------------------------------
CIT3033 | 1 % | 2 | 5ea. | 4,000+2,000 | | 400 |
| | | | | | 200 |
----------------------------------------------------------------------
SLAC | 1 % | 3 | 3 ea.| 4,000 each | gobs | |
----------------------------------------------------------------------
LLL S-1mI| 20% | 1 | 3. |36,000 (8Mw) | 1 | 300 |
----------------------------------------------------------------------
======================================================================
!
---------
I am still trying to get CIT and SLAC data.
Gio
-------
∂18-Aug-80 2106 REM at MIT-MC (Robert Elton Maas) Female person you asked about putting up
Date: 19 AUG 1980 0000-EDT
From: REM at MIT-MC (Robert Elton Maas)
Subject: Female person you asked about putting up
To: JMC at SU-AI
CC: FFM at MIT-MC
With about $3500 of electronic equipment here, not one penney insured,
I'm very reluctant to let anyone I don't fully trust stay here unattended.
FFM (Steve Kudlak) on the other hand has almost nothing worth stealing
in his apartment, so if this person isn't arrogant or violent or dishonest
he might be willing to put her up. But he wants to meet her first to be
sure there's nothing horrible about her personality. Please contact
FFM at MIT-MC (SKU at SU-AI but he's not there much) if you think it's
worth trying (he gets along with down and out people better than I do,
so it seems to me that it's worth trying).
∂19-Aug-80 1134 FFL Phone call from Paul Raeburn of Associated Press
To: JMC, FFL
Would like to see you for a few minutes on Wednesday. Is it possible inthe
morning or in early afternoon or after 5 p.m.? He will be attending some
sessions of the conference. If you can call him, his number in SF is
55 62l 7432. If he does not hear from you, he wil be checking back with me
to see if you can speak with him.
I have a breakfast tomorrow at 7:30am, which must be over by the time
sessions start at 9. I can meet the AP man at 9:15 here. I also
have a meeting of the III board of directors at 1:30 tomorrow in L.A.
and need a PSA reservation to get there on time from San Jose returning
as soon after 5:15 as possible. Please call the AP man and leave message
as to result.
∂19-Aug-80 1219 TOB VAX
To: FB at SU-AI, EAF at SU-AI, JMC at SU-AI
Col Larry Druffel would like to meet for a couple of minutes
to arrange access on VAX for Image Understanding. He has
Bob Kahn's word to go ahead.
Tom
If you declare a meeting time, I'll try to come, but it would seem that my
presence is not essential. I'll be gone most of tomorrow, though 10am
would probably be feasible. Thanks for keeping me informed with the
option of intervening.
∂19-Aug-80 1425 FFL
To: JMC, FFL
Do you want a car in Los Angeles tomorrow?
∂19-Aug-80 1432 FFL
To: JMC, FFL
Your reservation is made for lo:45 from S.Jose (could be ll a.m. from SFO))
and 5:30 p.m. return on Wednesday, August 20, SJO-LAX and return, charged
to your American Express card. I will pick up the ticket this afternoon.
∂19-Aug-80 1441 FFL Two parties connected with alum association activities next week
To: JMC, FFL
The Association office called to ask if you wish to attend with one or
two others:
l. Barbecue, Sunday, August 24, 6 p.m.
2. Dinner, Thursday, August 28, 6 p.m.
To reply to Shirley, 7-2021, or I will call her for you.
Barbeque on 24th yes, by myself. Dinner on 28th no.
∂19-Aug-80 1448 FFL
To: JMC, FFL
You say you can meet the AP man "here" at 9:l5 a.m. Where is here?
In your office or elsewhere?
"Here" meant my office.
The AP man called again, and we made the appointment.
∂20-Aug-80 0037 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: VAX
Date: 20 Aug 1980 0035-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: Re: VAX
To: TOB at SU-AI, FB at SU-AI, EAF at SU-AI, JMC at SU-AI
In-Reply-To: Your message of 19-Aug-80 1219-PDT
Tom, you can catch me around 9am or around noon or around 1:30
around the Tressider cafeteria; or around Annenberg Auditorium
during the AAAI business meeting.
Ed
-------
∂20-Aug-80 0127 RMS at MIT-AI (Richard M. Stallman)
Date: 20 AUG 1980 0416-EDT
From: RMS at MIT-AI (Richard M. Stallman)
To: JMC at MIT-AI
The funding for my trip has been arranged
but I still haven't been able to find anywhere
to live once I get there.
Can you tell me who I can contact to ask for help?
∂20-Aug-80 1307 DRUFFEL at USC-ISI VAX cycles
Date: 20 Aug 1980 1255-PDT
Sender: DRUFFEL at USC-ISI
Subject: VAX cycles
From: DRUFFEL at USC-ISI
To: JMC at SU-AI, FB at SU-AI, Feigenbaum at SCORE
Cc: TOB at SU-AI, Kahn
Message-ID: <[USC-ISI]20-Aug-80 12:55:47.DRUFFEL>
Gents,
Tom Binford's iu group is preparing to convert the ACRONYM system
to the VAX for tranfer to th IU testbed and needs VAX cycles.
Bob Kahn and I have discussed this some time ago and would like
him to have access to the machines which DARPA is providing to
support other research at Stanford provided that does not have an
adverse effect on those other efforts. M best information
indicates that the image work could be comfortably accomodated.
Do any of you have an objection wto working out a cooperative
arrangement?
Regards, Larry
∂20-Aug-80 1726 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Wednesday, August 27
SPEAKER: Stein G. Jessing, University of Oslo, Norway
TITLE: Monitors with Associated Processors
ABSTRACT:
A monitor can conveniently be implemented as a process.
The monitor proedures (operations) are then executed by the processor
of the monitor, not by the calling process. The talk will concentrate
on comparing this monitor concept with the concept of distributed processes
and the concept of communicating sequential processes. Some scheduling
techniques will be presented, and specification and verification of such
monitors will be discussed.
∂20-Aug-80 2356 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: VAX cycles
Date: 20 Aug 1980 2352-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: Re: VAX cycles
To: DRUFFEL at USC-ISI, JMC at SU-AI, FB at SU-AI, Feigenbaum at SU-SCORE
cc: TOB at SU-AI, Kahn at USC-ISI, rindfleisch at SUMEX-AIM, engelmore at USC-ISI
In-Reply-To: Your message of 20-Aug-80 1255-PDT
This is the gist of what I said to Larry at dinner this evening, plus some:
1. There are only two VAXes under Baskett's and Feigenbaum's control, not four.
2. The spirit of the proposal is exactly right, and I will do everything I can
to be facilitative.
3. Tom Binford's request has not yet been made quantitative. How much of the
VAX? Time-sharing or single-user mode? Until what time? The VAX is not all that
high powered a machine; and image understanding researchers can generally use
every computer cycle available in the universe. Also a user tends to get
dependent on a particular type of machine; then when our load builds up we
can't get them off. I look forward to receiving a quantitative proposal from
Tom, so that Baskett, Ullman, and I can evaluate its impact on our VLSI work.
4. Bob Balzer needs VAX cycles to develop INTERLISP, and nothing I can think of
on the VAX in my project could possibly have higher priority than helping ISI
get that done. That will consume cycles in this fiscal year.
5. I would like to see Tom not hard-wired to the VAX but connected to it
via ETHERNET (a more general solution).
Tom Rindfleisch can carry out my part of the discussions in my absence.
I'm sure everything can be made to work out smoothly.
Ed
-------
∂21-Aug-80 0739 JRA
do you know yet if my account goes away sept 1??
No, it doesn't.
∂21-Aug-80 0814 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Date: 21 Aug 1980 0811-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
To: JMC at SU-AI
In-Reply-To: Your message of 20-Aug-80 2359-PDT
The session is at 11am in Memorial Auditorium....Ed
-------
∂21-Aug-80 1005 TOB
To: csd.feigenbaum at SU-SCORE, FB at SU-AI, JMC at SU-AI,
druffel at USC-ISI, rindfleisch at SUMEX-AIM
We seek access to VAX in time-sharing, and ask for 20% share for 18 months.
Our use would build up slowly over one year; we would use less than 10% averaged
over one year. My best evaluation is that we need a 20% share at the end of a
year. After 18 months we should evaluate again and consult with ARPA.
Now vision uses about 10% of the KL10. The VAX has about half that
performance. Because of arrangements for graphics, image displays,
terminals, and software use would build up slowly. We would still have a
large share of our computation on the KL10.
I would like connection to VAX to be left free to decide as a technical
matter, ie by whatever effective and expedient means. I am strongly in
favor of the principle of using Ethernet. My estimates of when complete
dependence on Ethernet will be possible are not encouraging.
Ethernet interfaces and software do not exist now; putting a condition
on use of non-existing technology would introduce unpredictable (probably
large) delays.
∂20-Aug-80 2356 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) Re: VAX cycles
Date: 20 Aug 1980 2352-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: Re: VAX cycles
To: DRUFFEL at USC-ISI, JMC at SU-AI, FB at SU-AI, Feigenbaum at SU-SCORE
cc: TOB at SU-AI, Kahn at USC-ISI, rindfleisch at SUMEX-AIM, engelmore at USC-ISI
In-Reply-To: Your message of 20-Aug-80 1255-PDT
This is the gist of what I said to Larry at dinner this evening, plus some:
1. There are only two VAXes under Baskett's and Feigenbaum's control, not four.
2. The spirit of the proposal is exactly right, and I will do everything I can
to be facilitative.
3. Tom Binford's request has not yet been made quantitative. How much of the
VAX? Time-sharing or single-user mode? Until what time? The VAX is not all that
high powered a machine; and image understanding researchers can generally use
every computer cycle available in the universe. Also a user tends to get
dependent on a particular type of machine; then when our load builds up we
can't get them off. I look forward to receiving a quantitative proposal from
Tom, so that Baskett, Ullman, and I can evaluate its impact on our VLSI work.
4. Bob Balzer needs VAX cycles to develop INTERLISP, and nothing I can think of
on the VAX in my project could possibly have higher priority than helping ISI
get that done. That will consume cycles in this fiscal year.
5. I would like to see Tom not hard-wired to the VAX but connected to it
via ETHERNET (a more general solution).
Tom Rindfleisch can carry out my part of the discussions in my absence.
I'm sure everything can be made to work out smoothly.
Ed
-------
∂21-Aug-80 1144 FFL Professor Ketonen's apppointment
To: JMC, FFL
In coordinating the paperwork with Joyce Firstenberger in Prof. Suppes'
office I learned that she understood his salary was to be $30,000 per year.
This past year his salary was $20,400. Since we are to pay one half, it
will be an additional demand on Dr. Weyhrauch's account of $4,800. Was
that your understanding?
Yes, I agreed with Suppes on the increase.
∂21-Aug-80 1438 Denny Brown <CSD.DBROWN at SU-SCORE> Scheduling of the AI courses.
Date: 21 Aug 1980 1435-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: Scheduling of the AI courses.
To: csd.lenat at SU-SCORE
cc: csd.tajnai at SU-SCORE, csd.feigenbaum at SU-SCORE,
csd.genesereth at SU-SCORE, buchanan at SUMEX-AIM, tob at SU-AI,
jmc at SU-AI, tw at SU-AI
Doug, I just got wind of your plans to teach 222 and 224 during Winter
Quarter. When we organized the curriculum last winter, I thought
we agreed that 222 would be in the fall and 224 in the spring (with
223 in winter.) Changing that to do 222 and 224 in winter would mean
doing all of 222, 223, and 224 during winter; this would be a disaster.
In particular, MSAI students would have to take all three at the
same time, and no other AI courses in their first year. It is less
drastic, but no less inapropriate for PhD students interested
in AI.
It is normally our practice to be very flexible with regard to
teaching schedules, letting faculty members do what they want.
I don't think we can do so in this case.
What do you want to do about the problem?
(I'm copying the other AI faculty members to let everybody know
what's going on, and in case I misunderstood the plans.)
I'll send a copy of the Courses and Degrees blurb, and a sample
CSAI schedule in another message. (So people can read it or not.)
-Denny
-------
∂21-Aug-80 1448 Denny Brown <CSD.DBROWN at SU-SCORE> The courses and degrees blurb re CSAI
Date: 21 Aug 1980 1445-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: The courses and degrees blurb re CSAI
To: csd.lenat at SU-SCORE
cc: csd.tajnai at SU-SCORE, csd.feigenbaum at SU-SCORE,
csd.genesereth at SU-SCORE, buchanan at SUMEX-AIM, tob at SU-AI,
jmc at SU-AI, tw at SU-AI
This is what's in the bulletin about CSAI. It exists on-line in the files
<CSD.FILES>CSAI and <CSD.FILES>BULLETIN. Sending it now in case you'd prefer
it this way. -Denny
MASTER OF SCIENCE IN
ARTIFICIAL INTELLIGENCE
The degree of "Master of Science in Computer
Science:Artificial Intelligence" may be conferred
upon students who wish to develop a competence in the
design of substantial knowledge based AI applications. The
degree will be administered by the Committee for Applied
Artificial Intelligence, composed of faculty and
research staff of the Computer Science Department.
Present members include Bruce Buchanan,
Chairman, Doug Lenat, Mike Genesereth, and Tom Binford.
The CS:AI program will begin in autumn quarter each
year. Normally, a student will spend two years in
the program. Each quarter the student will register
half-time (9 units) and serve as a research or
teaching assistant half-time (20 hours per week).
The first year will involve acquiring the fundamental
concepts and tools through course work and project
involvement. During the second year, the student will
implement and document a substantial application.
A student should indicate preference for this
degree at the time of applying for admission.
Admission to the CS:AI program will be limited by the
amount of financial support available, and by the
amount of research supervision we can provide. To be
considered for this program, an application should
reach us by January 15.
The degree of "Master of Science in Computer
Science:Artificial Intelligence" is intended as a terminal
professional degree. Students completing this program will
have no advantage over other Ph.D. applicants; admission to
MS/CS:AI may negatively affect a subsequent Ph.D. application.
Students planning to obtain the Ph.D. degree are strongly
advised to apply directly for admission to the Ph.D. program.
PROGRAM REQUIREMENTS
Programs of at least 54 quarter units that meet
the following guidelines will normally be approved:
1. Core AI. At least four AI courses.
Required are C.S. 206 or C.S. 222, C.S. 223, and
C.S. 224. The fourth course may be chosen from C.S.
275, C.S. 276, C.S. 226, or C.S. 227.
2. Classical hardware and software. Choose 2
courses from the following: C.S. 211, C.S. 212, C.S.
311, E.E. 282, C.S. 142, C.S. 143, C.S. 144A, C.S.
145, C.S. 246A.
3. Theoretical computer science. Choose 1
course from: C.S. 154, C.S. 155, C.S. 156, or C.S.
206.
4. Practicum. 9 units of C.S. 225. A substantial
A.I. system is implemented and documented.
5. Practicum related electives. 9 units in
courses related to the C.S. 225 project. Acceptable
courses will be determined by the project supervisor,
depending upon the application area of the project.
Examples include courses in Physical Science, Social
Science, Computer Science, or Mathematics.
6. Electives. Additional courses (usually five)
to bring the total to 54 or more quarter units. These
courses may be in departments other than Computer
Science.
Courses taken to satisfy guidelines 1 through 5 will
normally be taken for a letter grade. As in other
MS programs in the Computer Science Department, a
3.00 grade point average must be maintained in these
courses. Electives taken under guideline 6 may be
taken Pass/No credit; if taken for grade a 2.50 grade
point average in these courses must be attained.
CS:AI programs that deviate from one or more of the
above guidelines in order to meet the valid
objectives of individual students will be considered
by the CS:AI Committee on an individual basis. In
particular, students are not expected to take courses
when they have had the equivalent subject matter
previously. The student should submit a written
statement of individual objectives and how the
program and previous preparation meet these
objectives.
A successful experience in this program is likely to
require an undergraduate education in the sciences, with
at least a moderate exposure to computing concepts
and practice.
-------
∂21-Aug-80 1453 Denny Brown <CSD.DBROWN at SU-SCORE> Sample CSAI program
Date: 21 Aug 1980 1451-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: Sample CSAI program
To: csd.lenat at SU-SCORE
cc: csd.tajnai at SU-SCORE, csd.feigenbaum at SU-SCORE,
csd.genesereth at SU-SCORE, tw at SU-AI, jmc at SU-AI, tob at SU-AI,
buchanan at SUMEX-AIM
This is the sample CSAI program which can be found in the file
<CSD.FILES>CSAI. -Denny
This page contains a sample program.
An Example Program in CS:AI
First Year Second Year
Fall Winter Spring Fall Winter Spring
---------------------- ----------------------
222(1) 223(1) 224(1) AI v ? AI v ? AI v ?
144A(2) 143(2) 154(3) PRE(5) PRE(5) PRE(5)
AI v ? AI v ? AI v ? 225(4) 225(4) 225(4)
Where AI means the fourth AI course, ? means a free elective,
and PRE means a "practicum-related elective." (And v is "or").
The numbers in parens refer to the guidelines which the course
satisfies.
Alternatives for the first year to get the "breadth" requirements
are: 142 given F and W. 156 given F. 246A given W.
The "AI v ?" from fall quarter may move to Spring to get more
breadth done right away.
-------
Denny - It seems as though you're right, but I prefer to say nothing
unless it seems required.
∂21-Aug-80 1528 Denny Brown <CSD.DBROWN at SU-SCORE> Course schedule issue resolved
Date: 21 Aug 1980 1524-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: Course schedule issue resolved
To: csd.tajnai at SU-SCORE
cc: csd.lenat at SU-SCORE, csd.genesereth at SU-SCORE,
csd.feigenbaum at SU-SCORE, buchanan at SUMEX-AIM, tob at SU-AI,
jmc at SU-AI, tw at SU-AI
Solution is: CS222 will be given in Winter along with CS223.
(Mike and Doug agree that this is reasonable.)
CS224 will be given in Spring. (This avoids the jumble in Winter.)
Since CS206 is available in Fall, and is an appropriate first course for
AIers (and others), there's no longer a problem.
Thanks to Doug and Mike for quick response and solution. -Denny
-------
∂ extract from Human-nets with ITT ad from Sept. Sci Am
Which of these things do you want your future telephone to do?
Here are some telephone services that sooner or later could be
offered to telephone customers ... through ITT's new System 12
telephone exchanges. Which services would you choose?
o An electronic telephone directory -- so your phone will look up
the numbers you want.
o Shopping from home -- a TV screen on your phone will show you a
selection, you'll order by pushbutton.
o Electronic mail -- you'll send and receive letters by a printing
device in your phone.
o A telephone information bank -- to let you call up data from
central files by pushbutton.
o Banking by phone -- your phone will flash your checking-account
balance, and let you pay bills from your chair.
o Newspapers by phone -- a machine in the phone will receive and
print the morning paper while you're sleeping.
o Long-distance business conferences -- you'll "meet" with business
associates over a phone that carries your picture as well as your
voice.
o Instant travel service -- your phone will flash updated train and
airline schedules on a computer screen (and then book seats for
you).
o Telephone alarm system -- will automatically signal police or
fire department when your home alarm goes off.
Imaginary? Not at all.
Many of these things are already available (or soon will be)
through ITT technology.
And as newer and newer customer features come along, they'll all
be easily accommodated by our System 12 telephone exchanges.
These exchanges are totally new, totally digital. And so flexible
that advanced customer features can be added without any interruption
of your telephone service.
Inevitably, as the world's demand for voice and data communication
grows, we're moving toward a total information delivery network --
built around services like these.
It's what we at ITT call Network 2000 (TM). A fully digital
network, with fully distributed control, easy to expand and change
as customer needs change.
There's only one exchange system in the world that meets all the
future requirements of Network 2000.
ITT's new System 12.
("signed") Compagnie Generale de Constructions Telephoniques, a French
associate of ITT
∂22-Aug-80 1000 JMC*
mail hurwitz stuff
∂22-Aug-80 2053 RPG Last Attempt
To: DPB at SU-AI, JMC at SU-AI
CC: RPG at SU-AI
This is to remind you that I would really love to have Dave Waltz
flown out for my orals at Stanford expense. He is willing to give a talk
on the morning of my orals on some of his research at Illinois.
If you decide that this is not possible, I will do it at my own expense,
and I will pass the hat, literally, beforehand to find help with
the funding.
-rpg-
I am now not quite as broke as I was and am willing to pay half of the
cost of RPG's requested trip for Dave Waltz. How about the Department
paying the other half?
∂22-Aug-80 2113 Rindfleisch at SUMEX-AIM FYI RE FILE SERVER DISKS
Mail-from: ARPANET site SUMEX-AIM rcvd at 22-Aug-80 2110-PDT
Date: 22 Aug 1980 2102-PDT
From: Rindfleisch at SUMEX-AIM
Subject: FYI RE FILE SERVER DISKS
To: WIEDERHOLD, JMCCARTHY, ADMIN.GORIN at SU-SCORE
Date: 22 Aug 1980 1049-PDT
From: Tucker
Subject: Large disk and tape systems
To: Rindfleisch
The following is an update (as of August 1980) of information about
large scale disk and tape systems.
I. SYSTEMS INDUSTRIES
Steve Hammerslag
56-732-1650
DISKS
These folks have added a 600 megabyte, non-removable, drive to their
line which previously had the 300 megabyte removable as the biggest item.
Actually the 600 (a CDC 9775) has an unformatted size of 675 megabytes.
Perhaps the 675 figure is the best one to use in calculating costs relative
to other companies.
Cost:
Controller and first drive $33,210
additional drives 25,650
(8 drives max on one controller)
This comes out to about $45 per megabyte for a 4 drive system which
has about the capacity (2.7 gegabytes) of the smallest 3380 configuration.
This drops to about $39 for an eight drive (5.4 gegabyte) system. One does
not get down to these price ranges with the 3380 until one has a four storage
unit (10 gegabyte) system (half the max size).
The above System Industries costs are for a Unibus connection with
about 1 megabyte transfer rate. An PDP 11/70 Cash Bus interface costs
about $7500 additional and a VAX SBI Bus interface costs and extra $11,000.
Thus in these last two cases the per megabyte cost goes up by $4 or so.
The transfer rates for these last two are about 1.2 megabytes.
If one sees an advantage in having the lower capacity removable packs
of about 300 megabytes each the cost is
Controller and first drive $21,510
additional drives (8 max) 15,660
So the costs are higher by at least 1/3. And the same additional
costs apply the the 11/70 and VAX options.
TAPES
Tri density 800/1600/6250, 125 ips STC drives via System Industries
Costs: Controller and first drive $28,890
additional drives 15,840
The above is for the Unibus, 11/70 Cash Bus add $4000. These tapes
are not yet available for the VAX.
All of the above System Industries prices include a 10% Stanford Good Guys
discount.
SYSTEM CONCEPTS
Mike Levitt
55-442-1500
I have placed three calls recently (in the past two weeks) and have not
yet been put through to Levitt. I get the impression from those answering
the telephone that they need some time to think about our inquiry. Their
position about 6 months ago was a dual 675 megabyte drive system including
an 11/34 and lots of talk about networking software for $125,000. The
latest IBM equipment and the other available 675 megabyte systems now
probably have them in a bind which they don't wish to talk about.
They claim they will get back to me next week.
-------
-------
∂22-Aug-80 2332 MINSKY at MIT-ML (Marvin Minsky) tower stability
Date: 23 AUG 1980 0234-EDT
From: MINSKY at MIT-ML (Marvin Minsky)
Subject: tower stability
To: MINSKY at MIT-ML, llw at SU-AI, jmc at SU-AI, rah at SU-AI
To: rem at SU-AI
About stability. On reflection, it seems to me that we can make
a truss structure as follows. Begin by building four (or three)
towers, beginning at the same place on earth,
but diverging. The diverging towers are lashed together by Kevlar
ropes. My intuition says it will be sufficient to do this at
altitudes of increasing spacing -- lets say, at 1 km, 2 km, 4 km, up to
32,000 km, which is exponential. (I realize that that is not
particularly reasonable, because it should
be related to inverse squares or somethng, but it makes nice similar triangles.
Now, what about the weight of the cross-link tension elements?
If the towers diverged by 1/60 radian, then they would be too heavy,
more or less, and would break -- that is, if made of untapered Kevlar.
And tapering wouldn't really help much. However, if
the divergence is 1/10 of that, then there is no pronlem.
(Incidentally, the most critical cross-strut is at an altitude of
one Earth Radius. Below that, the struts are too short to be too heavy, and above that, their weight decreases faster than their increasing length; besides, they don't have to pull so hard there, I think.
Now, what do we gain with the diverging towers? Mainly, North-South
stability. If Rod is right about using Coriolis for equatorial plane
stability, then we need only two towers, leaning apart, North and South.
The two different angles now permit global, rather than local control over the
polar-plane stability, now controlled not by beam deflection, but by changing the
particle flow in the two directions.
Then, the conjecture is that the cross-tension cables will permit adequate
lateral stabilization (with active control). The two towers
diverging by 1/10 degree angle give the Kevlar a
safety factor of 10. (Back of airline ticket calculation, needs
checking.) If the Coriolis thing is too flaky -- e.g., causes
too messy transfer into the longitudinal modes, then the 3 or 4-tower system
only costs a little more.
In this conception,
each "tower" is a full, two-way mass flow. I think separating the flows
causes impossible long-distance energy transfer problems.
I also did not consider having the towers curve away from each other.
Very likely, that is a somewhat better configuration, becuase the
fall-off of gravity makes the lateral members better able to
handle larger forces at higher altitudes, because they
don't have to bear so much of their own weight. I suspect that an
initial divergence ofjust a few millidegrees would, probably, be quite
enough, and my 1/10 degree is overkill, causing unnecessary
stresses.
∂23-Aug-80 1711 Wiederhold at SUMEX-AIM Re: FYI RE FILE SERVER DISKS
Mail-from: ARPANET site SUMEX-AIM rcvd at 23-Aug-80 1708-PDT
Date: 23 Aug 1980 1655-PDT
From: Wiederhold at SUMEX-AIM
Subject: Re: FYI RE FILE SERVER DISKS
To: Rindfleisch, JMCCARTHY, ADMIN.GORIN at SU-SCORE,
To: feigenbaum
In response to the message sent 22 Aug 1980 2102-PDT from Rindfleisch
Thanks for the info. My sense of the meeting with Friedl is
that therevis no great anxiousness to shovel significant
percentages of Megabucks into this project. My general
IBM scuttlebutt says that IBM will be supporting joint projects
stronger, so that is not yet out. Some progress on what
we have now would be helpful to convince them that
further investment is worthwhile.
I will continue to collect numbers and ideas, the
cope and methods for such a project are becoming clearer to me and I may
be able to state them more convincingly as time
goes bye.
Have fun Gio
-------
∂25-Aug-80 1426 FFL Call from Bennett Karmin, I.B.M. 408 256 5307
To: JMC, FFL
Understands you may have or have access to a film, video tape or written
print on the Lighthill Report, which he would like to borrow. He wants to
have an article on it in THINK, the house organ. Can I call him back for
you?
Nils: Do I remember correctly that you have the Lighthill tape?
∂26-Aug-80 0851 Nilsson at SRI-KL (Response to message)
Date: 26 Aug 1980 0849-PDT
From: Nilsson at SRI-KL
Subject: (Response to message)
To: JMC at SU-AI
cc: NILSSON
Yes, I have two cassettes (part 1 and part 2) of the Lighthill
debate (Called "Controversy: Robots" by the BBC). I can either
send them over to you or can send them directly to Karmin
(I will need his address). I forgot how I came into possession
of these tapes, but SRI will be glad to loan them free of charge
to whomever wants to borrow them. -Nils
-------
Fran - Please tell Karmin to call Nilsson at SRI.
∂26-Aug-80 1640 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum) [John McCarthy <JMC at SU-AI>:]
Date: 26 Aug 1980 1637-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Subject: [John McCarthy <JMC at SU-AI>:]
To: jmc at SU-AI
YES, THAT LOOKS RIGHT. I WONT BE ABLE TO INTERACT MUCH ABOUT
IT UNTIL I GET BACK FROM FRANCE IN JANUARY....ED
----------------------------------------------------------
---------------
Mail-from: ARPANET site SU-AI rcvd at 21-Aug-80 0148-PDT
Date: 21 Aug 1980 0148-PDT
From: John McCarthy <JMC at SU-AI>
To: csd.feigenbaum at SU-SCORE
Ed: Does the following agree with your recollection of what we
promised to do?
Feigenbaum and McCarthy promised the editorial board to propose
a scheme for using part of the journal's money for computerization.
Ideas considered included
subsidizing lecture
commissioning survey articles
on-line form of the journal
on-line bibliography
We have a year to come up with proposals.
---------------
-------
∂26-Aug-80 1730 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Date: 26 Aug 1980 1725-PDT
From: CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
To: TOB at SU-AI, FB at SU-AI, JMC at SU-AI, druffel at USC-ISI,
rindfleisch at SUMEX-AIM
cc: buchanan at SUMEX-AIM, engelmore at USC-ISI
In-Reply-To: Your message of 21-Aug-80 1005-PDT
Tom, here's the situation as I see it on the VAX (the one under my control):
1. The INTERLISP group at ISI is saying that a run-time package of INTERLISP
for VAX will be available in the Spring.
2. Our use will begin at that time and "ramp up" sharply to a point
later in the year when we will be heavily dependent on VAX.
3. Unfortunately your projected use "ramps up" at the same time, potentially
producing a heavy use conflict in the period 7/81 to 12/81.
4. There are great uncertainties abouthow much computing~
ARPA will be placing in our CSD environment and when. Maybe
there are no worries about compute power at that time. I don't
know.
5. There is absolutely no problem with your use now. But if there
is a use conflict later, there is no graceful way to get you to
reduce your use later if that should become necessary (though
of course it may never become necessary). The 20% use level
you propose is certainly reasonable
for the next 10-12 months.
6. As you can tell, I am agonizing about this. I am inclined
to say yes for now for reasons of "community spirit" within
the ARPA-supported community; and because the machine is
lightly loaded now. But:
a. I propose that this use be renegotiated in July of 1981
when we can more accurately assess the computing circumstances
of the time.
b. That we agree NOW , before there is any crunch, that
there be a definite cutoff date of 12/31/81, unless otherwise
negotiated because of favorable circumstances.
7. Attaching "smoothly" via Ethernet is essential. Attaching
lots of devices in an ad-hoc fashion means a lot of down-time,
and ours is not that kind of machine. I.e. ours is not a
"lab machine" but a "service" machine. I believe that you
have an ethernet interface for your PDP-11, and the VAX
is on the ethernet, so I can not forsee great problems. In any
event, this will serve to focus the attention of your software
people on the (inevitable) ethernet developments.
In summary: the answer is "Yes" (hesitatingly, and with caution)
with caveats about renegotiation, definite time limit, and
request tht you don't disrupt the service too much.
Day-to-day coordination should go through Tom Rindfleisch, who
will be directing the activities of this VAX center for me
(Forest is directing the activities of his own VAX center).
Ed
-------
∂27-Aug-80 1020 FFL
To: JMC, FFL
I have tried repeatedly at different hours to reach John Haugeland but
have been unsuccessful. Do you have suggestions?
Please send a note to his office address.
∂27-Aug-80 1645 REM via SU-TIP Deaf access to SU Comp.Sci. ?
There is a hard-of-hearing person in San Carlos who wants to study
Computer Science when he attends college. I suggested Stanford, MIT,
or CMU. He prefers Stanford because it's nearby. The problem is, how
can he get in contact with people in Comp.Sci. dept? (talking on phone
via voice is almost impossible for him) He has a Bell 103 modem (30 cps),
not a Weitbrecht modem, thus even if SU has a deaf info number it wouldn't
help him because deaf numbers are all Weitbrecht usually. Would it be
reasonable to give him the SU-AI dialup number so he could link to
you or other SU-AI/COMPSCI people and discuss his plans and SU's features?
He has a manual-select answer-mode modem, so you could callhim directly
from a 300 baud terminal at 593-9891 if you wanted.
∂27-Aug-80 2353 JK disk allocation
by doing a dsksize I have just found out that my disk allocation
has been decreased from 250 (which already was too small) to 120!
What should I do?
I decided to make extensive changes to the underlying ekl code -
should have a core image ready by end of this week.
Ketonen's (JK) disk allocation has been reduced from 250K to 120K.
It should be 250K.
∂28-Aug-80 0111 Boyer at SRI-KL (Bob) Meeting
Date: 28 Aug 1980 0111-PDT
From: Boyer at SRI-KL (Bob)
Subject: Meeting
To: JMC at SU-AI
As I understand it, I'll come visit you next Tuesday around
2pm at Jacks hall to discuss an example you have (involving
primes of the form 4n+1) that you speculate may require more
than elementary number theory but less than set theory.
I would also like to mention to you my opinions on the
matter of finite quantification within LISP.
Thanks for your remarks to AAAIC about Moore's and my work.
Our main forum for expressing our exceedingly high opinion
of your work is our work. In the sense of Thomas Kuhn,
we're working in your paradigm.
-------
Next Tuesday at 2.
∂28-Aug-80 1652 DES
John,
I'll be leaving this Sunday to start (assistant) professing at Columbia,
and probably won't have a chance for a proper "farewell" before then, but
wanted to say goodbye and invite you to visit whenever you're in New York.
(Address and telephone data is in NUDATA[1,DES].)
When my dissertation is printed, you'll find yourself in the list of those
acknowledged, but I wanted to express my gratitude to you directly as well,
not only for our direct contact, but for the influence your ideas have
had on my thesis work, as well.
I'll try to stop by to visit when in town. Best wishes in the interim.
David
∂28-Aug-80 2315 LLW Raining Momentum On The Just And The Unjust
To: JMC
CC: LLW
∂28-Aug-80 1426 JMC Tahoe bomb
What is the effect of refrigerating a bomb with liquid nitrogen?
At some temperature detonators might become inoperative.
[John, regrettably most detonators feature a short section of resistance
wire embedded in a very sensitive explosive (e.g., mercury fulminate); the
firing unit usually dumps enough charge through the wire to bring it to
white heat in a microsecond or less, so that it both superheats and shocks
the surrounding explosive rather vigorously, flinging a healthy detonation
wave into it. This sensitive explosive usually becomes more sensitive as
the temperature drops, as it crystallizes more thoroughly and is usually
piezoelectric to start with; the thermomechanical stressing it receives
during supercooling sometimes will set it off `spontaneously.' The
surrounding `booster charge' and the main body of the explosive also
usually have such thermal characteristics. There's no really nice way to
`turn off' a chemical explosive that I know of, but if you have time to
work around it as they did at Tahoe, there's a lot that can be done to
minimize the damage that it'll do when it goes off. Lowell]
∂28-Aug-80 2354 GG leaving
We are now going back to Italy. We would like to thank you for all you did
to make our stay here possible and useful.
Pina and Maria Gini
Have a good trip, and come back again sometime.
∂29-Aug-80 0954 FFL Visit of Prof. Iwata on Sept. 25 and 26
To: JMC, FFL
Yes, Tom Binford will be available and will make arrangements.
∂29-Aug-80 1002 CSD.BARB at SU-SCORE Visitor from ONR
Date: 29 Aug 1980 0957-PDT
From: CSD.BARB at SU-SCORE
Subject: Visitor from ONR
To: jmc at SU-AI
John, I need a few reports or papers you have written that you would
like me to submit with all the other department information and papers
that are going to the visitor from ONR. Also I was asked to make sure
that 9:00 a.m. is o.k. with you for the meeting on 9/4. Your prompt
response will be most appreciated, as this package will have to leave
Stanford today in order to be there on time. Many thanks, Barb
-------
9am is almost never ok with me for meetings. Also no-one has even
told me the name of this ONR person. You can collect recent papers
from Fran.
Should Barbara Roberts ask you for papers from me, please give her
the AI Journal Circumscription reprint, the Israel speech, "Concepts
as objects " and "Recursive programs as objects ...
∂29-Aug-80 1055 FFL Ketonen, Doyle, Goad, Talcott office arrangements
To: JMC, FFL
I have informed Betty Scott of your requests for them. She tells me that
each person automatically receives a desk, bookcase, and file cabinet. Those
will be taken care of as quickly as possible. Some will have to be ordered
so there may be some delay in getting settled.
∂29-Aug-80 1304 rwg at MIT-AI (Bill Gosper)
Date: 29 AUG 1980 1518-EDT
From: rwg at MIT-AI (Bill Gosper)
To: MINSKY at MIT-AI, jmc at SU-AI, llw at SU-AI
CC: DANNY at MIT-AI, RWG at MIT-AI, rod at SU-AI, pmf at SU-AI
CC: hpm at SU-AI, rem at MIT-MC, pourne at MIT-MC, forwrd at MIT-MC
Say, have you guys already thought to apply your momentum rigidification
principle to vacuum balloons? In effect, you could make a high pressure,
low density gas out of cannon balls ... those Franklin Institute types
with their kilometers-wide ballon proposal would be interested. I'd be
anxious to hear Rod Hyde's assessments of energy budget, choice of circular,
polygonal, or shuttling trajectories, and whatever.
∂29-Aug-80 1324 FFL REMINDER
To: JMC, FFL
Do you wish to send request for hotel reservations for Sigma Xi?
Yes, for me alone for whole meeting. I guess it would be Friday
and Saturday nights.
∂29-Aug-80 1325 FFL REMINDER
To: JMC, FFL
Please note your parking sticker number so I may apply for your new one.
∂29-Aug-80 1450 CSD.BARB at SU-SCORE ONR Person
Date: 29 Aug 1980 1446-PDT
From: CSD.BARB at SU-SCORE
Subject: ONR Person
To: jmc at SU-AI
I'm sorry I thought Feigenbaum mentioned it in his message, anyway
the name is Dr. Albert Brandenstein. I will see what I can do
about the time. Thanks, Barb
-------
∂29-Aug-80 1952 LLW Vacuum Balloonery
To: rwg at MIT-AI
CC: LLW at SU-AI, JMC at SU-AI, minsky at MIT-AI, RAH at SU-AI,
danny at MIT-AI, ROD at SU-AI, PMF at SU-AI, HPM at SU-AI,
rem at MIT-MC, pourne at MIT-MC, forwrd at MIT-MC
∂29-Aug-80 1304 rwg at MIT-AI (Bill Gosper)
Date: 29 AUG 1980 1518-EDT
From: rwg at MIT-AI (Bill Gosper)
To: MINSKY at MIT-AI, jmc at SU-AI, llw at SU-AI
CC: DANNY at MIT-AI, RWG at MIT-AI, rod at SU-AI, pmf at SU-AI
CC: hpm at SU-AI, rem at MIT-MC, pourne at MIT-MC, forwrd at MIT-MC
Say, have you guys already thought to apply your momentum rigidification
principle to vacuum balloons? In effect, you could make a high pressure,
low density gas out of cannon balls ... those Franklin Institute types
with their kilometers-wide ballon proposal would be interested. I'd be
anxious to hear Rod Hyde's assessments of energy budget, choice of circular,
polygonal, or shuttling trajectories, and whatever.
[Bill, it's too easy to make vacuum balloons in more conventional
fashions. When Rod was worrying a couple of years ago about how he was
going to deliver his multi-km-scale nickel-iron asteroids down to
Earth-surface in a fashion that people were willing to pay for (rather
than pay to *not* have delivered, which is the more natural state of
affairs for deliveries of ≥10 billion ton masses from orbit), we came up
with a nice scheme one nite for running said asteroids through a suitably
scaled hot-wire cheeese slicer to dice it up into slabs a few cm thick,
and then cutting the slabs into some regular polygon or another with the
same technique. You then use conventional deep-welding techniques to
assemble these slabs into a suitable polyhedron, and chuck enough of the
scraps inside to bring the mean density of the entire beast to the
*average* density of the atmosphere a carefully selected few km distance
above the ground (the remainder of its interior being filled with good low
density vacuum, and its strength being quite ample to keep it from
imploding even during reentry). You then de-orbit it with a gentle kick
into the upper atmosphere at the appropriate location, and its comes to
rest (subject to breezes, etc.) at the spot designated by your customer,
its bottom bobbing gently a hundred meters or so above the local surface.
You have thereby put a dozen years of planetary consumption of
already-smelted iron on some lucky fellow's doorstep, where he can tow
it to a cut-up area, and then `pop its spoilers' by letting the air into it.
We didn't pursue this idea because Rod decided that, rather than have one
guy pay to have the nickel-iron asteroid processed into a balloon and
delivered to him, it would be more cost-effective to have *everybody* pay
to *not* have have the *unprocessed* asteroid delivered to them.
Re your suggestion, gases have the nasty problem that, even if their
temperature in one direction is initially very much higher than in all
others, this ceases to be the case after very few collisions of the gas
species with itself, and it's very tricky to avoid such collisions for an
interestingly long period of time. What is usually of interest is
momemtum per unit mass of gas and, for this, it's hard to beat hydrogen at
room temperature, which darts around at km/sec specific momentum
densities. . .
Lowell]
∂31-Aug-80 0242 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #60
Date: 31 AUG 1980 0527-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #60
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Sunday, 31 August 1980 Volume 2 : Issue 60
Today's Topics: Security - Control of Cryptography Research,
Electronic Mail - Impact on PO Workers,
Telephone Rate Structures - User Equipment Charges,
WORLDnet - Development
----------------------------------------------------------------------
JBARRE@MIT-AI 08/30/80 13:53:21 Re: cryptography
There is an article in the September, 1980 Omni, page 84, called
"Unbreakable Code," which discusses some of the current work going on
in the area of cryptography.
------------------------------
DLW@MIT-AI 08/30/80 03:54:08 Re: postal workers
The post office is still useful for physical transfer of parcels and
documents. Also, I like to send personal letters by U.S. Mail rather
than net mail as a matter of style (the legality issue is something
else). If it must be phased out or contracted, that can happen
gradually as the price of electronic mail for the masses slowly
declines from its initial high (I assume) rates...
------------------------------
LLOYD@MIT-AI 08/30/80 13:37:34
Re: Displaced manual laborers (I.E. Postal workers)
First we need to take a step backward from the problem and try to
view it globally. First of all, the problem is not one of displaced
workers but rather one of a cultural revolution.
Mankind has progressed through the hunting/fishing economy, the
agrarian economy, the industrial revolution, and now we are entering
the information revolution. If we view the demographics we are going
to see a shift from manufacturing/producing to information management
and related services. Those of us here on HUMAN-NETS are uniquely
prepared to make the transition.
Unfortunately, most of the people in the US are NOT prepared. Granted,
the Postal Workers Union will protect the postal workers for awhile
but in the end the result will be the same; They WILL be displaced. We
are already beginning to see the handwriting on the wall. In Detroit,
we are feeling the effects of the impact of automation. The American
auto manufacturer generates about 20 autos/year*person. In Japan the
figure is up to 50 autos/year*person. Is there anyone in the group who
does not agree that the Japanese automobile is a more cost-effective
product?
The key to the problem is education (we've heard this one before).
We have to coerce the mainstream away from blue-collar jobs into
technical/professional careers. This will require the revamping of
the educational system in the United States. We need teachers who
can make science and mathematics interesting. We need to stress the
importance of an enlightened society. We need to build people who
are flexible (I don't know what kind of education you received in
High School but mine was LOUSY. If it weren't for two or three
teachers, I would not have been able to survive college).
Now comes the rub;
H O W D O W E A C C O M P L I S H T H I S ?!?!?!?
I don't know the answer to that one. Perhaps what we can do is to
help change the attitudes of society at the grass-roots level. We
need to encourage the children to hero-worship intelectuals as well
as athletes and TV stars. I don't know what your life was like
in grammar and high school but I was ridiculed as "Mr. Brain" all
throughout my tenure there (until I got to college). I hope to
impress everybody with the idea that we all have to do something.
Just discussing it in HUMAN-NETS is insufficient. Why don't you
(collectively and/or individually) go to a lower school and try to
interest at least ONE student in something other than Mork and Mindy
(not that that doesn't have it's place). Most kids are inquisitive
enough to listen if your subject is made understandable.
Well I have pontificated enough (and rambled some too). Do you think
that I am anywhere near the mark?
Brian Lloyd
------------------------------
Date: 30 Aug 1980 0329-PDT (Saturday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: technology
Nobody who knows me personally would ever call me an "anti-technology"
type (a glance around my apartment quickly verifies that I am deeply
immersed in a variety of technological subjects.) On the other hand,
I am not blindly pro-technology -- I cannot view the world through
rose-tinted glasses and accept every pro-technology argument that
comes my way.
One of the classic lines I've heard over the years runs something
like: "Automation will free the masses from drudgery and allow them
to be completely happy and fulfilled..." This may well be true in
the long run, but in the short run, what I see is increasing numbers
of technologically displaced persons with nothing but hate toward all
things smacking of high technology. One of the problems, of course, is
that it is impossible to instantly take millions of "obsolete" people
and retrain them to fit into the "society of the future". Yet these
people exist. What are we to do with them? Are they LESS human than
we are, ergo they have fewer "rights" than the rest of us?
If we blindly proceed along pretending that all these people can be
simply and easily transferred away from their old jobs into meaningful
new ones, we are fooling ourselves.
Yes, the postal service is definitely a situation where automation
could be usefully employed, and it should be. Yes, "the unions"
have too much power and tend to be reactionary about such matters.
But why is this the case? Basically because the unions have one
primary concern -- the immediate welfare of their members. They are
not concerned with the wonders that technology will bring in terms of
the "better life" in 15 or 20 or 30 years. They are more concerned
with making sure that their members will be in good shape tomorrow
and the next day. Yes, many union wage scales are crazy, I certainly
would not deny that! But the important thing is that we can never,
ever expect to gain the support of the general population for the
overall concept of "automation", unless it can be demonstrated that
people's immediate lives will not be disrupted or destroyed by the
transition effects of high-technology phase-in. Gee, that sounds
awfully intellectual, let's put it a simpler way:
Just because something CAN be done, doesn't mean that it IMMEDIATELY
SHOULD be done. After all, the whole point of all we do is supposedly
aimed at achieving happy, secure, fulfilled lives. In some cases, this
may well mean balancing the long-term advantages of some technology
against the short-term and medium-term negative effects that the
technology may have on PEOPLE. This does not mean stop progress
-- it simply means to consider more than JUST technology when
dealing with these issues.
End of sermon.
--Lauren--
------------------------------
Date: 30 Aug 1980 1322-PDT
From: John McCarthy <JMC at SU-AI>
Technological Unemployment
In spite of all obstacles their unions can raise, postal workers
will eventually be displaced by electronic mail. This will be good,
because they will then be employed doing other things society needs.
They are entitled to bargain about the terms of this displacement.
For example, it is negotiable demand to protect the job of a 50 year
old postal clerk with 15 years to go till retirement. Severance pay
is negotiable and is part of the cost of introducing the new
technology.
The worry (at least 100 years old) of technology causing permanent
unemployment has so far proved groundless. There is no correlation
between increases in productivity and increases in unemployment. In
the last couple years in America, there has been a REDUCTION in
productivity and an increase in unemployment. On the whole countries
with low technology have more unemployment than countries with higher
technology (e.g. China and India have much unemployment), but these
phenomena probably have a common cause (lack of entrepreneurship)
rather than one causing the other.
For the last 30 years, some people have been saying that America
is just about to run out of jobs because of automation. If we were
anywhere near self-reproducing robots with human intellectual ability,
there would be something substantial to adjust to. At present,
however, the problem is the reverse. Low productivity, coupled with
the increasing cost of social and environmental programs (whether or
not the increases are justified) are causing people to feel poor and
demand increases in income that cannot be met without excessive
inflation. Also, we have neglected to develop synthetic fuels and
have stalled nuclear energy. One reason we feel poorer is that our
work is going to build palaces and mosques and giant air terminals
in Saudi Arabia?
Many people hastily assume that "mindless" (do they mean manual?)
work is unnecessary. Perhaps it justifies getting out of doing any
and relieves the conscience of people who regard their own (perhaps
well paid) work as useless. Richard Lamson gave examples of assembly
line work as mindless and caring for the sick as necessary. This is
stacking the deck, because both are necessary and both are often
boring.
------------------------------
LLOYD@MIT-AI 08/30/80 11:54:44
Re: What does an additional telephone instrument cost . . .
REM@MC was asking what it costs to operate an extension telephone.
First, I don't think it is power consumption that TPC is worried
about. With the phone on-hook there should be no power drain (unless
the bell coupling capacitor is leaky). The only time an instrument
should represent a DC load is when the instrument is off-hook. Each
instrument does, however, represent an AC load when the ring voltage
is placed on the line (hence the ringer equivalence value on user
supplied hardware).
In actuality you are paying for a service contract on the equipment
and line. TPC is effectively guaranteeing you continued service NO
MATTER WHAT! That is expensive service indeed (I know that they fall
short of their goal much of the time, but you have to admit that phone
service in the United States as a whole is pretty decent).
TPC cannot deal with us on an individual level. They have to deal
with their subscribers on the basis of a statistical universe. If Joe
Blow destroys an instrument a month (for whatever reason), we all have
to pay for it. Also, think back to the last instrument you returned
to TPC: Was it in good shape? Was it good enough to go out again?
How old was it? It probably wasn't worth keeping. All these costs
are represented in your monthly rental fee.
I didn't intend to write in support of TPC but it just turned out that
way.
Brian Lloyd
------------------------------
REM@MIT-MC 08/29/80 06:38:56
Re: Telephone network vs. de novo computer network
InfoWorld for September 1 (page thirteen) has an article by Jim Edlin
(who is he?) asking the question which is the appropriate network for
mass-market telecommunications between microcomputers, and answering
TPC. Quoting Fortune magazine's estimate of present value for the AT&T
telephone network (not counting equipment owned by other companies
in the Bell network) he says that anyone attempting to create an
alternative network of equivalent bandwidth will have to invest
about the same amount ($111 billion) just to get all the equipment.
That amount won't be available to spend on nice equipment (terminals,
high-resolution displays, processing power) and software etc. He
suggests strongly that we should use the existing telephone network
efficiently, using efficient modems (9600 baud or whatever the line
will suport), data-compression, instead of spending all our capital
attempting to build an alternate comunications network. Spend our
capital adding to TPC facilities rather than trying to replace them.
In a short time there will be LSI-modems using digital filtering to
achieve optimum data transmission over voice-grade TPC lines, and if
there is a large enough market the prices will drop to something we
can all afford. The only problem I find is with local loops: if TPC
charges us through the nose for the link from our terminal to the
local concentrator, we won't be able to keep the 9600 baud modems
occupied enough to make that local connection cost-effective (long
distance hops are no problem, we multiplex many users so that the
modems and long distance lines are sending 9600 bits every second
of every day that they are in use; since they aren't leased, we
can add or delete circuits dynamically as bandwidth demands). Does
anybody who has read the full article care to rebut the author and
my agreement with most of what he says? If we go that route for
everything except local home-to-CO (central office) hops, does
anybody have a good solution for the local hops? (Maybe each
neighborhood could have a sort of Aloha-net, tying into the
neighborhood concentrator; Thus use VHF or UHF radio for local
hops, TPC for everything else?)
------------------------------
End of HUMAN-NETS Digest
************************
∂31-Aug-80 1655 JK ekl
tonight - am just now testing all the rules.
will have a compiled version ready then; the current core image is
quite large.
Please mail me a message then. I'm looking forward to trying it.
∂31-Aug-80 2308 JK
a version of ekl is ready - alias under ekl,jk and type: run ekl
the file tst1.ppr[ekl,jk] contains a "pretty printed" (some of the
parameters have not yet been set quite right) testproof in ekl.
pages 22 to 24 give a brief description of the forms of current commands
that can be given to the "top level" (this exists solely for debugging
purposes: the question of the correct language and formats of lisp
functions to use I have so far paid no attention to. The only thing I am
confident about is that the internal lisp functions that manipulate proofs
and the data structures are sound. I think that I will continue to avoid these
questions by designing first the tautology, decide and sort algorithms.
The user interface and the programming language for ekl
will be easy to implement once one has decided how they should look like,
which is the only truly deep question left.)
the example in tst1.ppr is somewhat misleading since it is far more convenient
to use the "fexpr" versions of the commands: thus a typical proof could start like:
(start-proof tst)
(fdecl p constant truthval universal)
(fdecl q constant truthval universal)
(fnamerange dd 1 2)
(fassume p dd)
(fassume q dd)
(f∧i 4 5)
(∧e 6 1) etc.
∂01-Sep-80 0000 JMC*
colmer.le1.
∂01-Sep-80 0640 SOL via MIT-TIP ap digests
at one time I saved digests and NYT pageones on dftp. When DFTP was
taken down, I had little time to retrieve all my files, and put on tape
about 40-50 megabytes of news related to communications. The rest
I left there. If there is some way to convince DCA to reopen the
Ampex DFTP terrabit tapes, I still have my directory so I can
find the rest of the stuff. I suspect they haven't erased anything.
richard
∂01-Sep-80 0742 SOL via MIT-TIP news saving
I would have continued to save news, except for the
pain of transferring files over the net and then the expense and
delay of having tapes mounted here. If this could be simplified so I would
not have toworry about its reliability, I would conitnue to save alot of
news ontape. In fact,this isone of the areas I intend to work on next
richard
∂01-Sep-80 0802 RPG
∂31-Aug-80 1706 JMC pattern directed computing
Are there any general purpose pattern matching programs available
in maclisp or accessible packages. Production systems are also of
interest. I'm currently working on a chapter for our book.
Yes, there are two related ones for MacLisp, one of which is
a standard part of SAIL MacLisp. This one is on MATCH.>[AID,RPG]
and is a general tree-structure matcher with both ? and * variables
(? matches a single object; * any number of objects). ?<var> and
*<var> bind. Also, you can put predicates on any ? or * variable:
with * variables, you can put the predicates on the entire list
that matches or on each individual element.
The other one is on UMATCH.>[AID,RPG], and allows pattern variables
on both sides (pattern and data), with the option, then, of the
predicates either applying or not to pattern variables. I'm not
sure that this version is quite a unifier on trees, but it's close.
Otherwise, it is just the two-way version of the above.
In each case correct backtracking is done to ensure consistency of
match (using a continuation style), and each is tail-recursive for
speed. The latter is about 2500 lines of code (single function), so
it might be had to read. I'd be happy to talk about how they
work with you, so feel free to ask.
-rpg-
∂01-Sep-80 1733 RPG Pattern matcher inverse
If you are going to do something like use the pattern matcher to
essentially set some variables and then instantiate based on that,
there is an inverse function, called %instantiate of all things,
which is now in match.>[aid,rpg]. It used to be somewhere else,
but I decided it belonged with the matcher. In any case, you can then
write something like:
(cond ((%match '(some *losing *pattern ?this ?is) data)
(%instantiate '(?this ?is a very *losing pattern))))
-rpg-
Thanks for the inverse. It is something I wanted.
A simple simplifier for sums and products sits in match.lsp[e80,jmc].
Among other things it takes (plus 2 x 3 y) into (plus 5 x y). The
match part uses %match ok, but maybe instantiate requires extension
to handle the instantiate part. The relevant case is noted in the
file.
∂02-Sep-80 0229 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #62
Date: 2 SEP 1980 0516-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #62
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Tuesday, 2 Sept 1980 Volume 2 : Issue 62
Today's Topics: Security - Control of Cryptography Research,
Electronic Mail - Impact on PO Workers,
WORLDnet - Development, Telephone System - Maintenance,
Telephone Rate Structures - Long Distance
----------------------------------------------------------------------
Date: 1 September 1980 1057-edt
From: Richard Jay Solomon <Solomon at MIT-Multics>
Subject: cryptography
While I agree with John McCarthy about the general thrust of U.S.
cryptographic research, I want to emphasize that our security
establishment has been much more open than most other countries
-- otherwise we wouldn't know anything at all about their history.
Several interesting points:
1) The history of N.S.A. does go back to World War I when
the Army Signal Corps did a little codebreaking and began to get
sophisticated about enciphering military messages with the aid of
AT&T, however SECRECY about their activities really didn't begin in
earnest until just before World War II. In fact, some of the basic
scientific principles about cryptography were published because the
work was done outside of government auspices by private research labs
for corporate uses during the early 1920's! Fortunately for us, the
private researchers were hired by the Army during the Depression (it
was the only job they could get) and broke the Japanese Purple code
in time to save hundreds of thousands of lives in WWII.
2) About all we know about other modern nations' attempts at
codebreaking and codemaking is that the British cracked the Nazi
codes enciphered on the Enigma machine using an early version
of a computer, and that that effort led to the development of a
vacuum-tube stored-program machine called "Colossus" by Alan Turing
in 1943. Apparently the British invented the modern digital computer
first, and yet they won't give any useful details to establish
absolute historical priority because of their paranoia about crypto-
graphic secrecy, almost forty years after the event! Contrast that
with the relative freedom to do historical research about U.S.
security agencies. (I have been reading unclassified NSA, OSS and
CIA files at the National Archives about WWII and afterwards on
cryptography and computers. Only names and actual codes or still
valid code-breaking techniques are whited out. Imagine doing that
in the U.S.S.R.)
3) Interception of electrical communications has been a fact of
life for the security of nations at least since the introduction of
trans-national telegraphic communications -- say from before the U.S.
Civil War. The British oceanic cable companies dominated worldwide
telegraphs starting about 1880, and with the aid of Marconi's
invention they had an early lead on radio communications as well. For
the present argument, I will assume the British had had an early lead
on decryption and interception as well; and hence their early lead in
machine computation. David Kahn's book, "The Codebreakers", (1965;
based mostly on U.S. archives) argues that the U.S. was somewhat
innocent of these matters (as noted in 2 above) until the late 1930's.
Be that as it may, the point is WE have been able to find out what OUR
government is up to, more or less to the present, and our principle
clues to what others have been doing in this vital area primarily
come from American sources. Seeing cryptography in an historical
perspective sheds some light on the problem of worldwide computer
(telegraphic)-communications for the future.
4) Try doing historical research (as I have been engaged during
the past few years) in this area anywhere but in the U.S. Despite all
the paranoia we hear about our intelligence agencies, they are much
more open about the past than that of our allies or enemies. The
National Archives has many documents yet to be perused by historians
on the role of computers and mathematicians in winning WW II. Except
for blanked out details of codebreaking techniques, or the name of
some individual who may still be serving in intelligence, one can
read of real missions on which the tides of war may have hinged; I
find it more engrossing than the phony macho stuff on TV and in the
movies about the war. Unfortunately, most historians, like policy
makers, haven't the technical background to understand these
documents, so gloss over their importance. If we are going to
envision future high-tech communications systems, we shouldn't
gloss over the role of intelligence either.
5) If I have whetted your appetite for more info, I'm sorry to
say the list of published material (for obvious reasons) is scanty.
Besides Kahn's 1965 book, his new one on "Hitler's Spies" (1978)
gives a good insight into the Nazi-Soviet intelligence debacles of
WWII. Books on the cracking of Enigma and such like "The Wizard War"
are so heavily censored by the Foreign Office, that they are worse
than useless for policy making. (Kahn submitted his book voluntarily
to DOD -- before Freedom of Info Act days -- and I suspect that's
why his book has little to say of post-1944 computer technology).
PLUG: I am looking for more info, leads, anecdotes and friendly
patter about the role of these things in the history of computers and
communications. Some day I hope the British will stand up and take
credit for bringing us into this age, even if it came from an
unexpected arena.
Richard Solomon
------------------------------
Date: 1 Sep 1980 1618-PDT (Monday)
From: Edwards at UCLA-SECURITY (David Edwards)
Subject: postal worker displacement
Regarding the recent discussion of displacement of postal
workers by technology advances, I think several good points have
been made so far. However, I have what I think is a somewhat
different view of the situation. First, I can't agree with Brian
Lloyd that we have to "build" a different type of person starting
at an early age. I don't think that's possible (although the
current system certainly has negative effects on some people, I
don't think ALL people could be convinced that they actually like
doing technical work). A good example of this is the USSR, where
they have been trying for something like 60 years to "re-educate"
people to have no self-interest and strive only for the collective
good. This clearly hasn't worked.
The situation we are looking at is another example of the trend
toward greater efficiency in providing the goods and services that we
require and want. It's clear that by automating the postal service
we can get the same services (probably better) for a smaller cost in
human labor. This means that, on average, a person has to put in a
little less work each day to provide himself with the things he wants.
This is because the amount of money he has to pay for his mail should
decrease. Therefore, the wealth of all the non-postal workers has
slightly increased, and the wealth of the few postal workers has
greatly decreased, due to them losing their jobs. If each person was
capable of doing any job, one good solution would be to decrease the
length of the work week, since people could sustain their same level
of living with less work. This would provide jobs for the ex-postal
workers.
That, however, is unfortunately not the case. But I have a
somewhat optimistic theory about what actually happens. First,
it doesn't work to just have the people with jobs provide for the
unskilled people who have been displaced. It is human nature that
people need some work that is a challenge to keep them happy. I
simply don't believe it would be utopia to lie around and have to
do nothing all day in order to live a comfortable life. People even
might eventually get tired of TV. It seems to me that increased
leisure time will cause an increased demand for leisure activities
and services, thereby creating more jobs. It's sort of an equilibrium
situation. Moreover most of those leisure-time services tend to be
low- or unskilled type work. Therefore, we have nothing to worry
about other than the overhead of displacing the workers and placing
them into the new jobs. Theory has it that capitalism automatically
promotes that, but the labor unions sure aren't helping matters.
Any comments?
Dave Edwards
P.S. I know this is long, but since I contribute about once every
few months, it evens out. Right?
------------------------------
REM@MIT-MC 09/01/80 07:00:37
Regarding old vs. new equipment... A naive analysis says that the cost
of something is different depending on whether you already own it or
not. If you don't own it, you must either pay rent or pay interest
plus principal on a loan (or equivalently pay cash and prorate it
somehow for its useful lifetime). But if you do own it, all you have
to pay is taxes and maintenance (which includes energy it consumes).
Now suppose there is something old you own already and something
newer and better you don't own yet (such as an old crossbar switch
and a new ESS). We assume that the new thing is more efficient than
the old one, thus if you could own either you'd pick the new one. If
The Phone Company could wave a magic and and own ESS's instead of
crossbars, they'd do it. But so long as merely maintaining the old
thing is cheaper than buying the new thing, it is optimal to keep
the old thing, and use the new thing only for (1) replacement of
broken old things and (2) increasing your total number of things
when you'd have to buy one or the other anyway.
So the answer to "why doesn't TPC replace their whole network with
digital stuff?" is that it costs more to buy all new stuff than to
simply keep maintaining old inefficient stuff. Of course when merely
maintaining old stuff costs more than buying new stuff, there'll be
a mad dash to buy the new stuff (but as the dash approaches madness,
the law of supply and demand will keep the price up, or create long
delays from ordering to delivery, so the madness will be gradual).
So the higher price of touch-tone phones might just be that it costs
more to buy new touch-tone phones than to merely maintain old
pulse-dial phones. (If this price difference is for connection of
a user-owned phone to network, rather than purchase or rental of a
touch-tone headset, and if the lower price for a pulse-dial connection
is to an ESS that can handle touch-tone, and if there isn't a shortage
of touch-tone dialer units in the ESS, then maybe those prices are
suspicious after all.) As for all-digital TPC, with A/D conversion for
voice, I suspect the cost of new digital units is MUCH MORE EXPENSIVE
currently than merely maintaining analog units (local CO's, regional
switching nodes, etc.). Remember that it's impossible to replace
the whole net overnight, and when you replace one CO or node with
all-digital you have to install A/D conversion at all interfaces
with neighboring analog equipment. I suspect all-digital equipment
will start in major cities, include some cross-country links, and
lastly filter into the smaller cities and rural areas. But I yield
to Lauren and other knowledgeable people for factual inside info...
------------------------------
LLOYD@MIT-AI 09/01/80 11:26:06 Re: Telco Service
Telco indeed services their equipment. I was refering to the entire
system (lines, CO's, etc.). Also, when your phone breaks, they do
indeed come to your home to fix it. As such, it costs in terms of
manpower and overhead to provide maintenance service to individual
subscribers.
Considering the complexity of the voice network, Telco's field
service beats the hell out of DEC's. (a good portion of the time
when DEC field service drops by for PM, they break your computer.
TPC is better than that.)
Brian Lloyd
------------------------------
Date: 1 Sep 1980 1505-PDT (Monday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: rates up ...
Long-distance rates are going to go DOWN, eh? Not the way I read
current tariff proposals at both the intrastate and interstate level.
Rates are headed UP substantially. The primary argument is that the
"competition" at the business level provided by alternative networks
is making it harder to recover the designated rate of return from toll
service, and thusly rates have to go up. This situation is bound to
get worse, not better. And, by the way, the proposals for increase
are being granted by and large.
--Lauren--
------------------------------
End of HUMAN-NETS Digest
************************
∂02-Sep-80 0858 FFL Sigma Xi
To: JMC, FFL
I called Mrs. Forslund about the glossies. Also made reservations for you
;on Friday and Saturday, October l0 and ll, Town and Country Motel, San Diego.
Do you want me to make air reservations? What time do you wish to leave and
return? You speak at l:l5 on Saturday. Do you want a car while there?
∂02-Sep-80 1056 FFL re papers to give to Barbara Roberts
To: JMC, FFL
Is the Israel speech you refer to ISRAEL.Spe[W80,jmc]?
Is the Recursive programs... FIRST[W79,jmc]
Is the Circumscription CIRCUM.W80[W80,jmc]? I know it is a reprint that
I can get from Connie Stanley, but she is on vacation and I don't think anyone
else can help.
I did not find a file name for Concepts as objects which suited me. Is it
the Concepts work you did in Fall of l979?
I may be anticipating difficulty. Barbara Roberts has not yet
asked me for any of these.
All the papers are in printed form in the file cabinet on right as
you come into my office.
∂02-Sep-80 1102 FFL
To: JMC, FFL
I am getting require let pub, etc. on top of everything that I start with ET.
Is that good?
I'm puzzled, and I'll be in shortly.
∂02-Sep-80 1300 JMC*
What to discuss with Boyer.
∂02-Sep-80 1308 teitelman at PARC-MAXC Re: FLIP
Date: 2 Sep 1980 13:01 PDT
From: teitelman at PARC-MAXC
Subject: Re: FLIP
In-reply-to: Your message of 2 Sept. 1980 9:09 am PDT (Tuesday)
To: John McCarthy <JMC at SU-AI>
FLIP is currently not in use inside of the Interlisp System, nor to my knowledge,
by any other applications program. If you wish to experiment with it, you simply
load the relevant files into an INTERLISP system and go. If you wish to pursue
this, let me know and I will make the files available to you
warren
My interest isn't in FLIP per se, but in pattern directed computing
in general. Is some other system for pattern dircted computing
in general use among INTERLISP users? I am writing the chapter
on pattern directed computing for McCarthy and Talcott "LISP: Programming
and Proving". Larry Masinter just expressed the opinion that
pattern matching compilers (and perhaps he meant pattern matching
systems in general) aren't useful. Is that your current opinion also,
and if so why?
∂02-Sep-80 1353 Masinter at PARC-MAXC Re: pattern matching
Date: 2 Sep 1980 13:49 PDT
From: Masinter at PARC-MAXC
Subject: Re: pattern matching
In-reply-to: Your message of 31 Aug 1980 1936-PDT
To: John McCarthy <JMC at SU-AI>
cc: masinter
The normal way that I hear that people are still using a package is that I get
complaints about it -- and I haven't heard too many complaints lately. I know
that personally I have found very few instances where the PMC (or any PMC)
would be useful....
I doubt that there are very many people at SCORE who use it frequently enough
to be of much help -- I'll be glad to offer any assistance that I can. As for the
"=>" not working, I suggest that you avoid the infix ":" notation, and use the
more conventional macro notation:
DWIMIFY((match X with ('A ! $2 $1) => <$1 $2>)).
should display the translation.
Larry
Thanks, Larry. I would be interested in your opinions as to why pattern
matching compilers (or is it pattern matching systems in general) are
unlikely to be useful.
∂02-Sep-80 1608 FFL
To: JMC, FFL
I have an inquiry about requirements for entering the A.I. program.
Does A.I. have a program for the M.S. degree?
If a students wants a combination A.I.-E.E. degree, with which department
does he start? Again is there a Master's program?
If interested in the PhD, does the student write to you, what information do
you need, should he be encouraged. I guess I am asking who is the counselor.
One of the callers seemed vague about just what AI is.
There isn't an AI program separate from the general program of the
Computer Science Department. The department has both
masters and PhD programs, and inquiries should be made to the Department
Admissions Secretary. Dot Dale is the secretary still, I think, and I
think Gene Golub is still admissions chairman.
∂02-Sep-80 1712 Masinter at PARC-MAXC
Date: 2 Sep 1980 17:10 PDT
From: Masinter at PARC-MAXC
In-reply-to: Your message of 02 Sep 1980 1358-PDT
To: John McCarthy <JMC at SU-AI>
cc: masinter
(If you think you could use them, I am willing to generate a couple of
paragraphs on the design of the Interlisp Pattern Match Compiler. I also have a
pattern-match compiler which generates directly the "ByteLap" which is the
machine-independent output of the byte compiler... which generally is
substantially better code than translating directly to Lisp. )
Why a pattern-match compiler is not a frequently used tool for most
programmers:
(It isn't so much that one is unlikely to be useful-- just that it is unlikely to be
useful frequently.)
It seemed that there are few situations were the discrimination is complex enough
to require the general mechanism of pattern matching -- most of the time that I
want to match something, I can tell if it matches by looking at the CAR. It might
be that, for novice programmers, the PMC would know about some built-in
routines that the programmer might not, even for simple patterns. In those cases
where the pattern matching is complex, the results of (at least my) pattern match
compiler were too locally optimized: there was no way to share or cross optimize
the pattern matches (was it the Lisp70 system which did?).
Although the intent of a conditional is clearer when it is expressed as a pattern
match rather than an AND of EQs, the pattern match language has its own
peculiar syntax to learn, so that the net improvement in code legibility is small
when using the PMC.My most typical use of the PMC was to generate the initial
lisp code, which I would then edit myself, i.e., define the function initially
using the pattern match, and then, in the editor, use the GETD command while
sitting on the match expression to expand the expression into the generated lisp
code, which I could then edit.
I see it is mainly a function of the kinds of program people write. In
writing an algebraic simplifier, complex patterns like a sum, some
of whose terms are sums, arise all the time. Thanks for the message.
∂02-Sep-80 2023 JK ekl
made a new and somewhat more compact core image with a couple of
bugs fixed and the base changed to ten. By the way, if you want to
play with the real lisp functions that operate on the proofs, they
can be found in lcmds.lsp. It would be useful for me to find out
whether you would really prefer to do this anyway inspite of all
the variables involved.
I think I can solve the problem of introducing goal structuring and
meta theory in one fell swoop: construe proofs as operators on
other proofs.
To tell the truth, I was a bit discouraged by difficulties in
following the writeup, and I think I'll need a demonstration
lecture. The idea of construing proofs as operators on proofs
seems promising.
∂02-Sep-80 2239 JK ekl
ok; what would be a good time to talk ?
∂03-Sep-80 1050 Denny Brown <CSD.DBROWN at SU-SCORE> Re: Department funds for Waltz' expenses
Date: 3 Sep 1980 1046-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: Re: Department funds for Waltz' expenses
To: RPG at SU-AI
cc: jmc at SU-AI, tob at SU-AI
In-Reply-To: Your message of 3-Sep-80 0032-PDT
Mail-from: ARPANET site SU-AI rcvd at 3-Sep-80 0033-PDT
Date: 03 Sep 1980 0032-PDT
From: Dick Gabriel <RPG at SU-AI>
Subject: Reply
To: DPB at SU-AI
I have not heard from you regarding the new financial situation
wrt Waltz and JMC's generous offer of 50% of the amount.
I am hitting TOB up for some amount since HHB is one of Waltz's
advisees and will be conferring with Waltz when he is out here.
-rpg-
I raised the issue at a staff meeting last Friday, but forgot to
get you the reply.
It was decided that the central department funds
should not be used for what is essentially research activity.
It is
appropriate for projects to support such when they can afford to. But
the department is relatively poor (broke at this point.) Readers are
not required to attend the orals, so (at least for the near future) we
won't be allocating department funds for expenses of reading committee
members. -Denny
-------
It seems to me that the decision not to finance Waltz's trip if made
on the grounds that
"It was decided that the central department funds
should not be used for what is essentially research activity." is a
mistake. Obtaining readers for a thesis is essentially an
educational activity. I am unfamiliar with the circumstances
under which Waltz became a reader of Gabriel's thesis, and I
don't even know who is Gabriel's adviser - perhaps Cordell Green.
However, being able to get external readers for thesis projects
for which an outsider is more appropriate than additional
internal readers is a big advantage, and it is a favor to us
when someone from another university undertakes this task.
It also gives our students more freedom in choice of thesis
topic and protects the student against changes in faculty
interests or personnel during the preparation of his thesis.
Moreover, having external readers promotes inter-university
comparison of dissertation and oral committee standards. In
many British universities external readers are required and
their transportation is an accepted element of the cost of
having PhD students.
I think the next Department meeting should discuss the question
of external readers.
∂03-Sep-80 1344 FFL
To: JMC, FFL
I found that name of the company from which the department ordered its "special
order" bookcases. and called to inquire availability and price for you.
SPECIAL ORDER: Ash wood with dark oak finish, same design as in office.
SPECIAL ORDER: Ash wood with dark oak finish, same design as office cases.
72" $191
84" 210 This is the height of the office cases.
Have to be special ordered and will take some weeks.
STANDARD: Ash wood with light oak finish. Design differs in that it has
toe space in the front.
72" $l75
84" l93
Probably 30 day deliver if ordered reasonably soon.
All of these are estimates. Tax would be added.
Boling Chair Co., Wholesale Office Equipment Co., 986-6972 (Ray)
84TB is the standard with walnut finish, price =
walnut, dark oak, light ash at Seeley Mudd
∂03-Sep-80 1308 FFL
To: JMC, FFL
Please call Susan McCarthy. Is at work now, but will be home later.
Work: 77l49 Home: 366 0221
∂03-Sep-80 1215 TOB
∂26-Aug-80 1730 CSD.FEIGENBAUM at SU-SCORE (Edward Feigenbaum)
Ed:
I accept your generous offer.
Tom
∂03-Sep-80 1615 RPG My Reading Committee:
Dave Waltz was my thesis advisor at the University of Illinois
when I was a grad student there. He and had had worked extensively
on the initial specification of the PLANES system, which is an early
data base system with a natural language interface. When scouring
Stanford for readers, Terry Winograd, because of the language part
of my thesis, became my advisor. Since Cordell Green wanted a natural
language generation program for PSI, hee became my second reader.
The only other people in the department in AI at the time was
EAF and JMC, neither of whom seemed as appropriate as Waltz in
terms of language issues and acquaintanvce with me as a student.
-rpg-
∂03-Sep-80 1644 CSD.ULLMAN at SU-SCORE
Date: 3 Sep 1980 1642-PDT
From: CSD.ULLMAN at SU-SCORE
To: JMC at SU-AI
In-Reply-To: Your message of 3-Sep-80 1411-PDT
I agree that the reading of a thesis is an educational function, not
a research one. The point I have trouble with is the notion that
Waltz's presence is essential to the educational process. For example,
three of my Princeton students had their final exam last June. I attended
two at my own expense; Princeton wouldn't have paid, and I wouldn't ask
them to. The third I did not attend at all.
However, to be fair, I must listen to arguments that say the Gabriel case
is special, and if you or he have a reason why Waltz must be brought here,
I'll see that funds are made available.
Further, I'll be happy to put a discussion of the policy on the menu
for the next faculty meeting, Sept. 30. OK?
-------
Don't try to be too fair; I don't want to use you up as department
chairman. I don't see Gabriel as a special case. I think Princeton
should have paid, and I will argue that we should encourage outside
readers and other forms of external participation in our educational
process. Sept. 30 will be fine.
∂03-Sep-80 2242 RPG I don't want to appear impertinent, but...
if you put the following at the beginning of your file you
will obtain better results. The syntax I use for NCOMPLR is
loser.1(kt)
-rpg-
(declare
(special *u *v *w ?x ?y *z)
(*expr %instantiate)
(*lexpr %match))
(defun if macro (l)(cons 'cond (ifxxx1 (cdr l))))
(declare
(defun ifxxx1 (u) (cond ((null u) nil) ((null (cdr u)) (list (list t (car u))))
(t (cons (list (car u) (cadr u)) (ifxxx1 (cddr u)))))))
Well that worked much better. I take it loser.1(kt) refers to the
syntax for getting the desired directory as well as file name.
The first time I tried omitting the match.fas← and it clobbered my match.lsp,
but there were so many things wrong that ...
What is the syntax to make load take a directory other than the one
it thinks is the default?
∂03-Sep-80 2249 TOB request for display purchase
Andy Bechtolsheim has offered to build a 1024x1024
display with 8 bits per pixel combined with a
high performance display (1000 lines without interlace)
for $20k. I would like to have it for the Horton box
to make some beautiful graphics. Can we buy it from
ARPA? Do you favor it?
Tom
Is this a private venture for Bechtolsheim or is it an internal Stanford
cost? If Bechtolsheim is available to do it, I would like to persuade him
(within Stanford or otherwise) to build the large department wide display
system.
Also I thought you were broke.
John
∂03-Sep-80 2330 TOB
This is a private venture for him. As far as the department system,
that might be too large for him (a real manufacturing effort). I personally
hope that he doesn't become overcommitted before I get a display.
The NSF is broke. ARPA is a question mark. I did not make clear that
I was asking if you were feeling both rich enough and especially generous
to get it for us out of other than ARPA IU funds. Otherwise I will find
a way in time.
Also I thought you were broke.
I am feeling neither rich nor generous. However, the design effort on
the departmental system may not be so great, and Ralph has more money
than brains for the task. I think I'll tell Ralph about it, and he
can explore the matter with Andy if he has no better scheme. The
Data Disc may die soon, and Schnurle is tied up on Ethernet until
at least January according to Ralph.
John
To: REG
The following should be self-explanatory.
John
∂04-Sep-80 0033 RPG Load
(load '((dsk (lo ser)) foo bar)) is it, I think.
-rpg-
∂04-Sep-80 0114 ME DMs from CASBS getting new modems
To: REG, JMC
We're getting UDS modems installed in the four DMs recently obtained
from CASBS. The modems will cost 400 or 405 plus tax and need to be
charged to some account. I assume they should be charged to the same
account that paid for the DMs.
That's correct assuming that's the total and not the cost per modem.
If it's the cost per modem, then it doesn't seem in accordance with
what Les negotiated with Product Associates in connection with the
inadequacy of the Product Associates modems.
∂04-Sep-80 0141 ME modems for DMs from CASBS
To: JMC, REG, LES
∂04-Sep-80 0121 JMC
To: ME, REG
That's correct assuming that's the total and not the cost per modem.
If it's the cost per modem, then it doesn't seem in accordance with
what Les negotiated with Product Associates in connection with the
inadequacy of the Product Associates modems.
ME - According to the letter from CASBS, their purchase price for the DMs
was reduced by 400 per terminal because of faulty modems. Our price to
CASBS was 2/3 of CASBS's reduced price. The 400 (which is mentioned in
the letter from CASBS) is the full price per modem because we are getting
brand new, working, UDS modems. Actually, 405 is what we are paying
Product Associates per modem for the upgrades of our year-old DMs to
replace the losing P.A. modems with UDS modems; that's the amount we
withheld in our payment to P.A. for the DMs, since the modems were
unreliable. I'm not aware of any other price Les negotiated with P.A.
∂04-Sep-80 0856 TOB Happy birthday
Happy birthday, John.
Best regards.
Tom
∂04-Sep-80 0912 CSD.ULLMAN at SU-SCORE Bibliographies
Date: 4 Sep 1980 0911-PDT
From: CSD.ULLMAN at SU-SCORE
Subject: Bibliographies
To: CSD-Faculty: ;
Early this year, I plan to file several grant proposals for money to
support our ongoing and future computer needs. These will require from
each of us the usual curriculum vita and bibliography, and a statement
of a few hundred words on our current research activities.
In at least one case, we may be able to include CSL people in the proposal,
so material from people whose affiliation is wholly or largely CSL would
be appreciated.
I would like these materials by Sept. 15. I would really really like them
by Sept. 22, and I NEED them by Sept. 30. Would you send hardcopy or the
name of an appropriate file to Jeanie Lagerwerff (csd.jeanie).
-------
∂04-Sep-80 0918 FFL
To: JMC, FFL
Yes, I left Carolyn's chair in your office temporarily until the
great furniture shifting takes place for both offices so that it would
not by mishap disappear.
∂04-Sep-80 1100 JMC*
check on Maclisp manuals.
∂04-Sep-80 1631 CSD.FRIEDL at SU-SCORE FILE SERVER PROPOSED PROJECT
Date: 4 Sep 1980 1629-PDT
From: CSD.FRIEDL at SU-SCORE
Subject: FILE SERVER PROPOSED PROJECT
To: MCCARTHY at SU-SCORE
I JUST WANTED TO LET YOU KNOW THAT I HAVEN'T DIED, AND AM STILL FOLLOWING
UP ON OUR LAST MEETING WITH ED AND GEO. I HAVE HAD ONE GOOD MEETING WITH
PEOPLE FROM THE IBM SAN JOSE LAB AND THEY APPEAR TO BE INTERESTED IN
COLLABORATING ON SUCH A PROJECT. I ALSO FOUND OUT THAT CIT MIGHT ALSO BE
INTERESTED, AND I WILL BE FINDING OUT MORE IN THE NEXT TWO WEEKS
-------
Thanks Paul.
∂04-Sep-80 2039 MRC
Happy Birthday
∂04-Sep-80 2258 LES
To: JMC, REG, ME
modem upgrade
Yes, $405 is the price that was agreed upon for the upgrade of the
modems in the SAIL Datamedias.
∂05-Sep-80 0046 POURNE at MIT-MC (Jerry E. Pournelle)
Date: 5 SEP 1980 0344-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
To: JMC at SU-AI
Just got back from Boston (World SF Convention). Your right of
reply is intriguing. More on subject when I recover from jet lag.
Jerry
∂05-Sep-80 0056 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #65
Date: 5 SEP 1980 0342-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #65
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Friday, 5 Sept 1980 Volume 2 : Issue 65
Today's Topics: Security - Control of cryptography research,
Automation - Dealing with its impact,
WORLDnet - Development,
Telephone Rate Structures - Long distance
----------------------------------------------------------------------
CBF@MIT-MC 09/04/80 07:10:56 Re: NSA & Civil Liberties
Bill Lewis asserts that in a conflict between academic freedom and
issues of national security, he would ALWAYS choose academic free-
dom. I find that a most remarkable statement in view of some of the
tradeoffs involved. It has already been asserted that allied code
breaking efforts during WWII probably saved hundreds of thousands
of lives. I would say based on material unclassified/uncovered in
the last few years the total could be substantially greater than
that and perhaps the course of the war itself. But what does this
counting matter? Its rather clear that whatever the costs might
be, if it stands in the way of Bill Lewis's academic freedom, then
its got to be paid.
In any event, as far as I remember things, this business with the
NSA took place about 2 years ago, and essentially concluded with
the NSA backing down. Have I missed a major piece of news recently?
Lastly, to answer Bill Lewis's question about what the use of code
breaking is since the Rivest algorithm is published: aside from
possible mathematical "bugs" in the algorithm (unlikely?), it is
also not the case that every foreign nation in the world will start
using these techniques tomorrow. In fact, with LSI/microprocessor
technology almost exclusively the domain of the west (you know, the
"us" side, as opposed to the "them" side), I would expect it could
be rather cumbersome and expensive to outfit most communications
gear with Rivest encoders decoders. There is also great possibility
for the usual implementation bugs/carelessness etc. to creep in.
------------------------------
Date: 4 September 1980 1531-EDT
From: Jerry.Agin at CMU-10A
Subject: Technological Unemployment
The job security of postal workers relates to questions of techno-
logical unemployment in general. It is particularly important that
we who are contributing to the technology understand the implications
of what we are doing. It's my guess that stories of technological
layoffs are largely exaggerated, but my guesses and rhetoric aren't
worth anything without some hard facts to back them up. Can anyone
produce a documented case of any worker who was without a job because
s/he was replaced by a machine?
------------------------------
Date: 04 Sep 1980 1419-PDT
From: John McCarthy <JMC at SU-AI>
"While the situation probably will remain one of increasing
employment for awhile, it must inevitably end up with fewer
jobs, once the factories that build automatic factories are
built. I don't have the answer to this problem, but it
definitely calls for more than just saying, 'Well, it won't
happen for awhile, and it might turn out okay.'"
There is good reason to take precisely the attitude criticized above.
The attitude of doing something now about eventual job loss has led
in England to bitter struggles against new technology that have put
England far behind France and Germany in standard of living and with
no less unemployment. The trouble is that over-sensitivity to distant
threats, whether military or environmental or social or unemployment,
leads to bureaucracies that don't have to justify their activities by
current achievement. They can simply claim success by showing that
the disaster they purport to be averting hasn't happened. In the case
of technological unemployment, the burden of proof is on the advocates
of action to show that the situation is different from that of ten and
twenty and forty years ago when immediate action against the same
danger was also being demanded.
------------------------------
Date: 3 Sep 1980 2147-PDT (Wednesday)
From: Edwards at UCLA-SECURITY (David Edwards)
Subject: Converting people to technical jobs
Brian Lloyd brings up a good point which has fallen out of the
discussion about displaced postal workers. With the imminent shift in
the job market from manual to technical jobs caused by advances such
as WORLD-NET, it's worth considering the social ramifications of the
increased emphasis on training technical people. My earlier point
was that there are certain people who by nature are not and probably
cannot be interested in doing technical work of certain types. I
admit I'm going out on a limb here, but it's been my experience that
some people, apparently very early on, develop an orientation for or
against certain types of mental activities. I feel that it would be
very difficult, if not impossible to change these people.
However, Brian makes a very valid point that most technically-
oriented people are subject to negative influences such as peer
pressure in the high-school age range. I'm going to go out on
another limb and claim that most of those who ostracize "super-brains"
or "eggheads" do so out of jealousy, not hatred of the subject matter,
such as science or math. I witnessed the class buff in history, for
example, receiving the same treatment when I was in high school. This
is often compounded when the brains start congregating with others of
their own kind, and limiting their free-time activities to those that
don't interest the bulk of the people. The brains then start looking
like "others" with a whole set of their own activities and friends,
and things they can do that others can't. And we all know that people
can be very cruel to those they consider "others".
Sadly, I think that would be very difficult to change, and it
will certainly continue to hurt the conversion to a more technical
society. What I think might be more worthwhile looking at is how
many people THINK they wouldn't like doing technical work, when in
fact they would. The classic example of this is how girls in the
past have very often been MADE to think they wouldn't like technical
work (or wouldn't be capable of it), just because they are female.
So my resulting point is that the key thing we should be
concerned with is convincing people to at least give technical work a
try, since it just might not turn out to be a mystical art after all.
I'm constantly amazed at how many people have that attitude, without
actually looking at it to see for themselves. Those are the people
who could be added to the ranks of technical workers...and Brian is
right that this type of encouragement must be done early in a child's
development. It is the responsibility of educators not to limit the
horizons of children by instilling in them irrational reasons why
they shouldn't be able to do something. This is what we must guard
against.
Any other ideas on this?
Dave Edwards
------------------------------
KRAUSS@MIT-MC 09/04/80 09:03:21 Re: cable bus networks
It is true that there are no cable bus networks (such as Ethernet)
operating on CATV cables right now. Later this year, my company
(M/A-COM) will be testing this concept on the Viacom system in San
Francisco and the Manhattan Cable system in lower New York. This
experimental project will also involve common carriers Tymnet and
SBS, and will use 10 GHz digital microwave radio as well as CATV
cable for local distribution of digital signals. The cable access
units will include RF, since the 1.5 Mb/s signal will displace a
slot normally used for video. Both the 10 GHz radio and the cable
bus will use TDM out from a central node to subscribers, and TDMA
from subscribers back to the central node. Transmission tests over
these two cable TV systems showed very low bit error rates, so we
anticipate a successful program.
---Jeff Krauss---
------------------------------
Date: 4 September 1980 11:33-EDT
From: Robert W. Kerns <RWK at MIT-MC>
Subject: "Contention nets not suitable"
MALIS@BBNE raises the point that Ethernet-style contention nets
are limited in the number of subscribers that can be put on them
since the available bandwidth descreases as the number of contending
stations increases. But this unremarkable fact is not related to it
being a contention network, but is rather a fact of life in ANY net.
The available bandwidth is inversely proportional to the traffic. Of
course, the Ethernet experience is with high-bandwidth applications
such as connecting personal computers with a remote filesystem. It's
not clear just how many subscribers you could fit on a cable because I
don't think we know just yet how much bandwidth the average subscriber
will use AVERAGED OVER TIME.
Right now, the Telephone dial network is very inefficient in it's
use of copper local loops. This is because historically it has been
cheaper to string wires than to build switching equipment. The
economics have changed; now it makes excellent sense to have lots
of smaller networks connected by SMALL computers to handle "gateway"
functions between small networks. This is very similar in style to
the way the telephone net is currently arranged; what is different
is the optimum size of the local "exchange".
It's also not clear that there is any reason at all to want to put
traffic from the computer on "completely different channels". There
are many ways to design your net such that no single host can block
transmissions for long. If you're talking about a high-bandwidth
cable, you might have to wait a few microseconds before you transmit.
So what? Even a twisted pair network can provide delays on the order
of a few milliseconds under reasonable load.
The big problems are not how to physically move the data (although the
MOST ECONOMICAL solution is still a matter of lively debate, as is how
to get there from here) but how to ADDRESS AND ROUTE the data.
------------------------------
Date: 4 Sep 1980 1154-MDT
From: THOMAS at UTAH-20 (Spencer)
Subject: Telco rates/deregulation
I am afraid (living in a basically remote area) that if long distance
rates are deregulated to "what the market will bear" that we will see
the same situation we currently have with airplane fares. If you
happen to live in NY or LA and want to get to the other, this is
great. However, if you happen to live in Salt Lake City and want to
get to NY, this is not so good (it costs at least twice as much to go
from SLC to NY as from LA to NY - even taking into account discount
fares). This is, of course, due to competition - there are a lot more
people who want to go from LA to NY than from SLC to NY - thus those
few of us who do must pay more for the 'privilege'.
-Spencer
------------------------------
Date: 04 Sep 1980 0951-PDT
From: Paul Martin <PAM at SU-AI>
Just a small correction to erroneous TPC statements...
Reading the recent rural vs. urban service claims in HN, I have
heard repeatedly stated the oversimplified notion that "TPC must
give service at a standard cost to any user, despite the higher
line costs associated with radically lower densities in rural
areas..."
I personally know of two cases where TPC has not extended service
to rural customers due to the higher costs associated with them.
The La Honda CA phone exchange has only 100 lines (all phone numbers
are the same except the last two digits), and for at least 6 years has
told prospective customers that they must wait through a queue until
enough people move or die to free a line. The cost of a new trunk over
the Santa Cruz mountain range is cited as prohibitively expensive. La
Honda has grown dramatically due to the general population pressure of
the SF Bay Area, but the phone service hasn't.
The other case is in such remote areas as the outlying regions near
Las Cruces NM and Whitethorn CA. In both areas, rural customers are
offered phone service only if they will pay thousands of dollars to
clear a right of way and install cables.
So while I don't know the rules in effect, apparently there is some
practical limit on the investment to serve rural customers required
for TPC....
------------------------------
Date: 5 September 1980 00:09 edt
From: SSteinberg.SoftArts at MIT-Multics (SAS at SAI-Prime)
Subject: gee, only ten cents worth of parts
Good grief! A nice looking Van Gogh costs a few hundred thousand
bucks and it only cost him about ten or so for the paint plus the
canvas and brushes. You can make a store quality apple pie for <50
cents worth of apples (a good one runs about a dollar). Gee, a floppy
diskette costs only $2.50 in quantity and it only takes 5 minutes to
back up, obviously Radio Shack is ripping me off completely when I pay
$100 for SCRIPSIT. Make your own beer. A 75 cent head of lettuce is
made from .2 cents worth of seeds (actually less).
Come on, you can build anything cheaper if you don't pay yourself for
it. That was why batch processing was so popular, so what if it takes
me twice as long it's 30% cheaper and us professionals are only worth
$2.50 an hour. When someone sells a modem for $200 they are probably
making surprisingly little money off of it. Someone has to buy the
parts and keep them in stock, someone has to put it together, someone
has to see if seems to work, someone has to shove it into a plastic
bag, someone had to decide to sell it, someone had to put an ad in a
magazine which charged them for running it [the ad], and so on. Not
only that but they have to pay rent, taxes, electric bills, phone
bills, insurance bills and buy things like desks, chairs, tables,
pens and so on. It adds up pretty fast.
On the other hand it costs $20,000 to build a $6000 car from parts.
------------------------------
End of HUMAN-NETS Digest
************************
∂05-Sep-80 0438 Robert Elton Maas <REM at MIT-MC>
Date: 5 September 1980 07:35-EDT
From: Robert Elton Maas <REM at MIT-MC>
To: HUMAN-NETS at MIT-MC, JMC at SU-AI
(Re: distant threats and bureaucracies to worry about them)
On the other hand, if we wait until a disaster happens before we start
worrying about it and planning how to avoid it (nuclear war for example),
it'll be too late when it has already happened and the bureaucracy now
has hard evidence to justify the expense of studying the disaster. I
think each person must judge what is worth worrying about and what isn't.
When we have a plausable mechanism by which a disaster could happen, but
it hasn't yet happened evenonce, we can't say a priori that it is or isn't
worth worrying about. I happen to think nuclear war is worth fretting
about to the point of maybe banning nuclear weapons, but that unemployment
due to technology is worth taking pains to minimize but likely to be
solvable without stopping advance of technology. Since some people fear
mass unemployment as much as I fear nuclear war, there is no a priori way
to say one disaster is worse than the other, we have to judge each
thing-to-worry-about individually and not make flat statements that
we shouldn't establish bureaucracies to worry about things in the future.
On the other hand, GOVERNMENT bureaucracies sometimes seem to spend a lot
of money to get nowhere, REGARDLESS of whether the problem is immediate
or long range, and the more there are the harder it is to keep track of
them all. These voluntary scientific-elite mailing-list discussions
(HUMAN-NETS, ENERGY, etc.) are an interesting alternative...
Now that we've been at it about 8 months in HUMAN-NETS, would anybody
like to compare the effectiveness of (1) government bureaucracies
(2) scientific organizations via meetings (3) scientific journals via
articles and letters to editor (4) network mailing lists?
∂05-Sep-80 0439 REM at MIT-MC (Robert Elton Maas)
Date: 5 SEP 1980 0737-EDT
From: REM at MIT-MC (Robert Elton Maas)
To: JMC at SU-AI
P.s. only the first (large) paragraph of the preceding HUMNET msg was
specifically to you (rebuttal), second (small) paragraph was merely to
HUMNET general audience (i.e. pretend you didn't see the second paragraph
until it apears in next digest)
I'm not really satisfied with what I said in my last HN. Maybe
the point is more like this. There is a tendency to regard people in the
future the way we regard our children - requiring our guidance. They
don't need our guidance, and it's just as irrelevant as the ideas of the
people 100 years ago on what our problems will be. Moreover, if we act on
a problem about which action can be postponed, we use up our resources -
mainly our extremely limited political ability to act sensibly in several
areas at once. Since, our social knowledge is non-quantitative, and
social movements to encourage action have positive feedback from success,
society is likely to overreact and go into oscillation.
Suppose, for example, that a movement against assembly robots in
the automobile industry acquired a charismatic leader, successfully forced
GM to abandon a major project and defeated a major politician. This would
cause anti-robotism to become a successful cause, and it would attract
adherents in the U.S. and other countries. The success of the cause over
a ten or twenty year period might be entirely independent of whether it
saved any jobs. Eventually, there would be a reaction, and its success
might be independent of whether it produced economic growth.
The main cure would be to develop a successful scientific
social science. Pending that, it is best to put off almost anything
that can be put off. Saving jobs in the next ten or twenty years is
a valid goal - if not an all consuming goal. Preventing a nuclear
war in the next ten to twenty years (though maybe we can't really see
ahead more than ten) is worth almost any action that can be shown
to do any good. However, a claim that some action of no use today
will prevent nuclear war even thirty years from now needs to be scrutinized
extremely carefully.
Supposing someone had claimed six months ago that worker
discontent in Poland over the next ten years would cause a big strikes in
August 1980 and that this would cause a revolt in East Germany in 1981,
which would cause the older generation of Soviet hardliners to start a
nuclear war. This analyst might have further claimed that the best action
was to have the CIA assassinate the Lech Walesa, an obscure Polish worker,
which would postpone the Polish strikes by three years, at which time the
Politburo situation would be different and permit real liberalization
without war. The hypothesis would be conceivable, but no-one could
possibly have good enough reasons for believing it to justify an
assassination.
This is the case for concentrating on seeing clearly what is in
front of our nose. How does it justify developing science and potentially
useful but long range technology? That requires betting that our
successors are likely to have at least as clear a picture of their
interests as we have of ours and have a good chance of benefiting from the
greater variety of options provided by improved knowledge and technology.
∂05-Sep-80 0923 FFL
To: JMC, FFL
Please call Barbara Klein in Washsington re reviewing an article for
SCIENCE magazine. Dial 1 to get Operator who will give you
800 424 9730.
∂05-Sep-80 1237 TOB trip
I plan to go to MIT for a week on Sept 14, on ARPA
travel. Is this ok?
Tom
Have a good trip.
∂05-Sep-80 1311 RPG `Research activity'
To: DPB at SU-AI
CC: JMC at SU-AI
For the record, I agree with JMC that the reasoning behind your `expenses
for readers' decision is hollow, but I will leave that battle for another day.
-rpg-
∂05-Sep-80 1319 Nilsson at SRI-KL Book
Date: 5 Sep 1980 1320-PDT
From: Nilsson at SRI-KL
Subject: Book
To: JMC at SU-SCORE
cc: Nilsson
Continuing our discussion during AAAI, Tioga would be interested
in publishing McCarthy papers on AI. Can you send me a list of
those that you would like to have included. In the meantime, I can
talk with Knuth about the possibility of TEX-ing them and getting
output on the alpha-type machine. Then we can discuss more
details. -Nils
-------
∂05-Sep-80 1815 REM via SU-TIP
That sounds reasonable, you should maybe submit it to humnet, even though
it's a bit esoteric (but then so was my message). On the other hand, the
example of polish unions is a bit off subject. If you could replace it
with something closer to computers/automation/jobloss/networking anything
of interest to humnet, I'd say definitely submit to humnet.
<<Diversion, there's an old radio show from the days of radio, a science
fiction program about a society with time machines who goes back to
assasinate Einstein in 1900 so that he won't later develop the theory
of relativity that brought on nuclear weapons in the 1940's. It backfired
because somebody else invented the same theory and conditions happened
to be worse, the person with the time machine was arrested for possessing
knowledge that was top secret....>>
I probably overreacted to your original message, and when you find a good
statement of what you really meant to say it would be good to submit it.
∂06-Sep-80 0316 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #66
Date: 6 SEP 1980 0559-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #66
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Saturday, 6 Sept 1980 Volume 2 : Issue 66
Today's Topics: WORLDnet - Development,
Telephone System - Echo Suppression,
Telephone Rate Structures - Long Distance,
Automation - Dealing with its impact
----------------------------------------------------------------------
REM@MIT-MC 09/05/80 07:48:19 Re: cable bus networks
I think the problem with direct A-->B (instead of
A-->centralrepeater-->B) communication on cable-TV systems is that
they are designed for a top-down deliver with reverse channel for
replying to the master, not in a ring like Ethernet coax where
messages cycle around. In a cable-tv system, unless nodes A and B
happened to be on the same branch they couldn't talk to each other
without going up to some forking node anyway, and rather than
install switching at every forking node it's simpler to just put
a giant repeater (echoback node) at the very top and require all
communications to go through it. This also simplifies the equipment
since only that top repeater has to worry about contention between
two nodes that both try to transmit at the same time.
I'm not an expert in this, so don't ask me for grungy details.
By the way, David Harris, not on Arpanet hence not on HUMAN-NETS, is
presently working on installing some sort of Ethernet on a cable-TV
system soon to be installed in the Palo Alto (Calif.) area. I think
his design will have a repeater node at the top as I suggested, with
all other nodes able to talk to each other only through the repeater.
------------------------------
Date: 5 Sep 1980 01:04:20-PDT
From: ESVAX.jimbo at Berkeley
Subject: Echo Suppressors
Would someone out there mind explaining how echo suppressors work on
long lines? I find them to be a rather large nuisance that makes it
almost impossible to carry on an intelligent conversation. The symptom
is that I can't hear the other party when I am talking. (This may be
something different?) Is there any way to disable it? (I seem to
recall something about a tone which cuts out the echo suppressor.)
Jim Kleckner (esvax.jimbo@berkeley)
------------------------------
Date: 5 Sep 1980 0455-PDT (Friday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: rural telco rates
Yeah, it is true. There are definitely some pathological cases where
telco's are exempted from the "equal servicing" of rural customers,
with the blessing of the state PUC involved. But by and large, there
are not too many of these exemptions, relatively speaking.
--Lauren--
------------------------------
Date: 5 Sep 1980 10:42 PDT
From: Hamilton.ES at PARC-MAXC
Subject: Rural telephones
Most people don't realize that there still exist phones with a hand
crank instead of a dial or keypad. It's even more fun trying to dial
IN to such stations. (Any other Deep Springers out there? Ever spend
ten minutes trying to convince an operator and three supervisors that
"Deep Springs toll station #2, out of Bishop, CA" is a number that
exists and can only be routed manually?)
--Bruce
------------------------------
Date: 5 Sep 1980 1421-MDT
From: THOMAS at UTAH-20 (Spencer)
Subject: Telephones in remote areas
If you own all the land over which the lines will be strung (or if
you can make an arrangement with the land owner), you can obtain a
"Telephone Installer's License" from the phone company and string
your own lines and install your own phone. I'm not sure of the
economics of this (what it costs to get the "license"), but I know
that my father installed a phone at our summer place this way, and
didn't say anything about outrageous cost - it was definitely cheaper
for them than having the phone company install the lines and phone.
-S
------------------------------
Date: 4 Sep 1980 1108-PDT
From: ROODE at SRI-KL (David Roode)
To all you people who favored continued subsidy of local service
by long distance: I'll agree if we change to one price long
distance calling. I.e., just like Telenet and Tymnet, it will make
no difference to the rate if you are communicating Reno to Sparks
or Los Angeles to Boston. It's the only way to be truly fair to
the people who happen to live in rural areas away from the action.
Actually, if you think about it, it is unlikely that the long distance
service subsidy to basic service makes that much difference in the
urban vs. rural controversy. Why not just say the city dweller's
basic service rate is subsidizing the country resident's? Actually
it probably makes the most difference when considering installation
costs. In fact all people who keep the same service for above average
periods of time are subidizing installation service for others. I'll
bet this counteracts the higher cost of rural service. A few farm-
houses on some remote telephone line spur may well pay for the cost of
installation through 40 years of uninterrupted and unchanged service.
City turnover rates must be higher than those in the country. I've
been told Telco claims the actual cost of installation averages around
$100 or $150. This subsidy is of greater impact than the urban/rural
one I imagine.
------------------------------
Date: 5 September 1980 1158-EDT (Friday)
From: Robert.Frederking at CMU-10A (C410RF60)
Subject: Reply to John McCarthy's reply
I certainly didn't mean to imply that I want the Federal Govt. to
regulate automation. What I want is similar to what Jerry Agin wants:
I want to KNOW that we're going to be okay, not just guess. I'd be
satisfied with a thorough study by someone (Natl. Academy of Sciences?
Rand?) about what would probably happen, and how to prepare for it, if
we need to prepare. There may be one, but I haven't heard about it.
The Natl. Academy of Sciences did a study on "Databanks in a Free
Society" a few years ago, and it came up with a very carefully thought
out book with some re-assuring results that didn't seem to be a white-
wash (they decided things were just as bad with pre-computer record
keeping, or even worse. Unfortunately, Congress didn't pay too much
attention to them due to pressure from business lobbyists).
As regards other government regulation of "distant" threats, I
don't find any of the problems listed as distant: the Russians are
unfriendly and imperialistic, there were no songbirds in my hometown
until 10 years after DDT was banned, and if you or I were black, we
would find it extremely difficult to be where we are today given the
lousy education we would almost certainly have gotten as children.
------------------------------
Date: 5 September 1980 07:35-EDT
From: Robert Elton Maas <REM at MIT-MC>
Subject: distant threats and bureaucracies to worry about them
On the other hand, if we wait until a disaster happens before we
start worrying about it and planning how to avoid it (nuclear war
for example), it'll be too late when it has already happened and
the bureaucracy now has hard evidence to justify the expense of
studying the disaster. I think each person must judge what is
worth worrying about and what isn't. When we have a plausible
mechanism by which a disaster could happen, but it hasn't yet
happened even once, we can't say a priori that it is or isn't
worth worrying about. I happen to think nuclear war is worth
fretting about to the point of maybe banning nuclear weapons,
but that unemployment due to technology is worth taking pains
to minimize but likely to be solvable without stopping advance
of technology. Since some people fear mass unemployment as
much ask I fear nuclear war, there is no a priori way to say
one disaster is worse than the other, we have to judge each
thing-to-worry-about individually and not make flat statements
that we shouldn't establish bureaucracies to worry about things
in the future.
On the other hand, GOVERNMENT bureaucracies sometimes seem to
spend a lot of money to get nowhere, REGARDLESS of whether the
problem is immediate or long range, and the more there are the
harder it is to keep track of them all. These voluntary
scientific-elite mailing-list discussions (HUMAN-NETS, ENERGY,
etc.) are an interesting alternative... Now that we've been at
it about 8 months in HUMAN-NETS, would anybody like to compare
the effectiveness of
(1) government bureaucracies
(2) scientific organizations via meetings
(3) scientific journals via articles and letters to editor
(4) network mailing lists?
------------------------------
Date: 4 SEP 1980 1105-PDT
From: RAHE at USC-ECL
Subject: Displaced workers
I don't think the middle class were screaming about the poor not
having to suffer because they were poor. A good example was recently
a mother of 5 children (no husband and 20 - yes twenty - years old)
threatened to camp out in front of the housing authority in the city
near me unless she was given a house to live in. The letters to the
local paper complained not that she should not get the house but as
one man put it he works 2 sometimes 3 jobs at 16 hours a day 6 days a
week, has to pay his own bills and noone is giving him a house.
Another suggestion was to give her a hysterectomy, not a house.
Interesting problem along the lines of costs for new technology: Can
we use the same arguments about the pill which is currently having a
rather dramatic effect on the teaching profession? There currently
seems to be a shortage of children to teach. I bring this up lest we
forget that some pretty bizarre things can fit the current arguments.
------------------------------
Date: 5 September 1980 2218-EDT
From: Lee.Cooprider at CMU-10A
Subject: displaced workers and related topics
a) Some of the discussion about automation displacing workers has been
too simplistic. Worker displacement does not come about only along
the lines of "we installed a machine to do your job, goodbye". It
also comes through non-hiring (postponing the displacement for 1/2
generation), routinization (permitting cheaper labor for the same
task) and stagnation of work roles.
b) The effects in the PO can be different than in other examples of
automation. The primary reason for automation in a corporation
is not to provide better service to more people but to make more
money within the short to medium term. Therefore the decision to
install computing and communication equipment will come about when
the additional control, speed and capability can be converted into
cash. The increased profit often results from the ability to use
less skilled people to get the work done (e.g. NY Times type-
setting, machining, word-processing). In the case of the more
political PO, it is theoretically possible to install equipment
because "it is a good idea". However, government auditors can
indirectly force the same phenomena to occur there.
c) We should stop trashing the unions. Labor unions should be strong.
In our economic system the notions of private property have been
so construed as to make people who own lots of things the defacto
economic oligarchy. They get to decide issues that affect people's
entire lives (e.g. whether to relocate textile manufacturing
operations to Alabama or Taiwan, how much to spend on training
future workers). The labor organizations have correctly attempted
to gain some control over these decisions; they have not, however,
had perfect foresight nor good planning information. Our strategy
should be to develop labor as an ally, to assist unions in making
technologically wise decisions, help them defend against exploitive
displacements (e.g. promote the use of phrases such as "the cost of
change" and "responsibility for change"). The unions can be very
effective and computing people, among others, could well use their
support on other issues.
d) The US economy is not, overall, suffering a decrease in
productivity. This myth is being used to justify all sorts of
regressive policies. The productivity measures are insensitive
to many time-variables and misrepresent many important issues. For
example, productivity measures show that a teacher with 40 pupils
is more productive than one with 25. Services cannot be so simply
evaluated. There are many other problems with these statistics.
We may believe that additional technological human-expanders are
desirable, but this is not because the US is currently in decline
in this area (despite US News claims).
e) Work is interesting. People want to work, to cooperate, to contri-
bute to the common effort. We do not save people by deleting their
jobs either in the short run (during which they must buy food, etc)
nor in the long run in which they want to do useful and perhaps
significant things. There are many things that could be done today
by people now unemployed, and given the projected adequacy of the
economy to provide all with essentials (which is not quite yet
true), such jobs make perfect sense. However, the economic deciders
do not have a mechanism for non-capital forming work. I believe
this problem to be chronic, and the continuing "crises" evidence
for that.
------------------------------
End of HUMAN-NETS Digest
************************
∂07-Sep-80 0047 ME ns
∂07-Sep-80 0025 JMC ns forgets
I just finished reading all the news stories, and left NS with the e command.
Immediately, saying NS again, I try #, and it has forgotten half the stories.
20 were remembered and 21 forgotten.
ME - I think the correct statement is "NS remembers". It remembers the
stories you've already seen and doesn't show you them again unless you
ask for them. If you thought you'd seen all of the stories, perhaps you
had actually only looked at one wire's worth, or just only looked at
half. If you say # and you only get a small number, it should say
"not counting 21 stories already seen" or so.
It did say it but I ran through all it showed me, and then exited again,
again said NS #, and it had forgotten all of them. I can demonstrate
the effect if you wish to peek.
∂07-Sep-80 0114 ME NS remembers...
...only 64 stories already seen, to limit its use of TMPCOR file space.
You had seen over 64 stories, although some had come in before 9pm,
when the wires go to a new day. After midnight, NS defaults to looking
back only to 9pm (before midnight it looks back to 9pm the previous day).
So it started up and read the TMPCOR file listing 64 stories already seen,
but of those the # command only retrieved 22. I copied your NS tmpcor
file before you logged out so I could reproduce the problem, and indeed
when I said #/FROM YESTERDAY, instead of just #, NS said "... (not counting
64 stories already seen)".
OK, thanks.
∂07-Sep-80 0248 RWG at MIT-MC (Bill Gosper) evening juggling of differential eqns
Date: 7 SEP 1980 0545-EDT
From: RWG at MIT-MC (Bill Gosper)
Subject: evening juggling of differential eqns
To: jmc at SU-AI
ok, well the lispmachine has some advantages and disadvantages over
macsyma@mc, depending mostly on me moving some diffeq code to a
parc file server. i should be able to manage that tomorrow pm, but
i may hit a snag, in which case we might just use mc (and fight
the core bound). barring lossage, tonite (sun) should be the first of
several ok nites for me. i'll try to reach you before supper.
Tonight is good for me.
∂08-Sep-80 0404 REM via SU-TIP Your ideas on concentrating on immediate instead of sheer-guesswork
(HUman-nets, polish strikes, ...)
Have you written up your ideas in an essay? It runs contrary to the
"an ounce of prevention is worth a pound of cure" which drives the
anti-nuke anti-technology anti-computer movements as well as the
more sensible health-prevention etc. I'm not sure where the optimum
lies, somewhere between your extreme and the ounce-of-prevention extreme.
Your side needs to be presented so we can consider both and find the optimum.
∂08-Sep-80 1100 FFL
To: JMC, FFL
I am not feeling well and am going home. Expct to be in tomorrow.
∂08-Sep-80 1415 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Re: Faculty interests file
Date: 8 Sep 1980 1409-PDT
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: Re: Faculty interests file
To: JMC at SU-AI
In-Reply-To: Your message of 29-Aug-80 1443-PDT
Thank you for the suggestion. Incidentally, the file is
<CSD.Files>Fac-Int.TEX.
I'm putting up a new one for the research associates and affiliated
faculty:
<CSD.Files>Resassoc.Int.Tex.
Carolyn
-------
∂08-Sep-80 2353 ROB AP vs NYT
∂08-Sep-80 2350 JMC
To: ROB, ME
NS is losing lots of characters.
--------------------
Hmm. I haven't been using NS much. Can you tell me if
both AP and NYT are losing, or is one losing more than t'other?
- Rob
∂09-Sep-80 1612 RWG
sorry, i didn't read my sail mail yesterday, will mail eqs tonite.
Sorry! I had supposed you would have your mail automatically
forwarded to wherever you read it. Is MC the best place to mail it?
∂09-Sep-80 1634 RWG at MIT-MC (Bill Gosper)
Date: 9 SEP 1980 1931-EDT
From: RWG at MIT-MC (Bill Gosper)
To: jmc at SU-AI
Here's a writefile of the eqs and time-invariant sol'ns to 1st order in h.
(note equivalence of c1 with %c1, etc.)
(C5) MAP(PART,EQS,[3,2,2,1]);
dR1 dR2 dW1 dW2
(D5) [---, ---, ---, ---]
dT dT dT dT
(C6) MAP("=",%,%-D3);
dR1 dR2
(D6) [--- = - B1 R1 W1 + E1 H R2 + A1 R1, --- = - B2 R2 W2 + A2 R2 + E2 H R1,
dT dT
dW1 dW2
--- = F1 H W2 + D1 R1 W1 - C1 W1, --- = D2 R2 W2 - C2 W2 + F2 H W1]
dT dT
(D7) DONE
(C8) ORD1;
%C1 A2 B1 F1 H %C2 %D1 E1 H A1 %C2 A1 B2 F2 H
(D8) [R1 = --- - ----------, W1 = ------------ + --, R2 = --- - ----------,
%D1 %D1 A1 B2 %C1 %D2 B1 B1 %D2 %D2 A2 B1
%C1 %D2 E2 H A2
W2 = ------------ + --]
%C2 %D1 B2 B2
(C9) CLOSEFILE(JMC,MAIL);
The actual macsyma internal s-expressions are in mc:rwg;jmc >, with all the
Cn's and Dn's %ified.
∂09-Sep-80 1748 DPB Offices and desks in MJ
To: JMC, AVB
A detailed count of offices (being used as offices now) and desks
(as currently allocated) gives:
Offices: 89
Slots in these offices: 185
Giant offices are included in the above count, namely:
352 - 8 slots. (Should be conference room; is being used as office.)
416 - 7 slots. (Used as group office for otherwise un-housed PhD
students.)
450 - 12 slots. (Is and will remain student offices.)
022 - 6 slots. (Relatively poor catacomb space in basement. TOB
students now.)
All other offices have at most 4 (student) slots.
Public terminal rooms which don't have people assigned are not included.
These are:
225 (HPP terminal room) Could be a 2 person (3 student) office.
Now has (planned) 2 Altos, 4 terminals,
HPP Reports Library.
320 (NA terminal room) Probably stays for SLAC connection, else is
a conference room. Office only under duress.
Now has 4 or 5 terminals.
460 (Public term. room) Jungle offices, or could continue as
public terminal room for un-housed.
Now has Altos, DDs, Score terms. (I don't know
how many off hand.)
408 (Public term. room) Jungle offices, or continue as
public terminal room for un-housed.
Now has Altos, DDs, Score terms. (I don't know
how many off hand.)
XSP MAP.XGP[PER,CSD] for a map of the building with occupants as given
by the PEOPLE file (PEOPLE.DAT[PER,CSD]). Since there are some open slots
today, you won't get exactly the same figures by counting heads. I counted
planned slots; and I have the plans in my head.
-Denny
∂09-Sep-80 2232 kirkley at sri-unix Hello... but...
Date: 9 Sep 1980 at 2222-PDT
From: kirkley at sri-unix
Subject: Hello... but...
To: jmc at sail
cc: kirkley at dnsri, kirkley
I can't seem to make sail do a link or send or whatever.... are you getting my
messages and/or will you reply.
CAlled Wirth recently... he's fine.
Regards,
Charles Kirkley (from 1965 etc)
Vituperative Associates
PO Box 3694
Stanford, Calif 94305
415-328-323$1
Electronic mail : kirkley@dnsri (Deaf network at SRI)
Please send at least one "nil" and I'll leave you alone or
whatever.
-------
I wasn't looking at my terminal when you sent me the message.
If you're still there you can send me messages, and I'll reply.
∂10-Sep-80 0353 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #70
Date: 10 SEP 1980 0644-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #70
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Wednesday, 10 Sept 1980 Volume 2 : Issue 70
Today's Topics: Databases and Publishing - Copyrights and HNT archive,
Automation - Dealing with its impact
----------------------------------------------------------------------
Date: 9 September 1980 1126-EDT (Tuesday)
From: Robert.Frederking at CMU-10A (C410RF60)
Subject: Copyright and publishing
I'm not absolutely certain, but I believe that once you
distribute written or oral material without a copyright or
non-disclosure agreement, you have surrendered any rights you
had to it. If anyone has said anything they don't want published,
they're in trouble. I find the preceding hard to believe, since
they have voluntarily splattered their opinions all over the
country, in some cases onto relatively public media like the
CMUA Bboard.
------------------------------
REM@MIT-MC 09/09/80 10:29:55
I don't think there is any absolute sense in which some task is
necessary. Necessity is based on some goal. If you have no goal
whatsoever, not even to stay alive, nothing is necessary. If you
merely want to stay alive, you have to eat and drink and get some
sleep. These might be considered the essential industries, providing
food and water for cities where there isn't enough space for everyone
to grow or hunt food and scoop up water from streams. Beyond that
it depends very much on what you want to enjoy. Who would think in
1920 that computer manufacturing and programming would in 1980 become
almost an essential industry to maintain industry and government? I
think as it becomes possible to do more and more things that weren't
feasible before (vacations on Mars or Titan, pocket Chess computer
that makes instant moves of world-championship quality, ...), people
will demand these things, and industries will exist to satisfy this
demand, and jobs associated with these industries will be considered
"essential" to people who couldn't imagine having to live without
modern conveniences like supermarkets, pocket-master-chess, running
water, indoor toilets, daily flights to New York and Titanville,
daily HUMAN-NETS digest ... Although unemployment might be high
(20 percent at times), I doubt we'll ever be satisfied with just
the essentials of 50 years ago and not the contemporary "essentials"
we've grown used to. Thus we'll never be satisfied getting what the
machines and the 2% of population who maintain them can provide, and
not getting also what the other 78% of the population can provide.
Thus we'll never put 98% of the population totally out of useful
things to do. (On the other hand, "never" is a long time. Let me
say not in the next 50 years.)
------------------------------
Date: 9 Sep 1980 (Tuesday) 1921-EDT
From: WESTFW at WHARTON (William Westfield)
Subject: What to do with displaced workers
The obvious thing to do with 90% of the people not needed to work
is to start a new manifest destiny and using them to colonize other
worlds, etc. Of course, this would require cheap space travel,
which isn't likely. I have noticed that SF-LOVERS and HUMAN-NETS
are currently running very parallel in this area. Maybe we need
a list TECHNOLOGY-IMPACT ?
Bill W
------------------------------
End of HUMAN-NETS Digest
************************
Subject: What will displaced workers do
"What do do with displaced workers" is not quite the right way of putting
it. Suppose a union thinks it can get another dollar an hour out of
an employer or industry, and the workers have to decide whether to take
it in direct pay or in a job security plan whereby the employer's right
to fire redundant workers is restricted or removed. How a worker will
vote in the union bargaining committee will depend on how self-confident
he is about the economy and his own personal ability to get another job.
Union officials have as an additional consideration the size of the
union which affects their personal positions. Historically, American
workers have mostly chosen higher pay, and this has paid off.
Perhaps I have digressed, but it has been more important what workers
have done for themselves in getting new jobs than what anyone, employers
or government, have done for them.
There is an apparent contradiction in the notion that displaced workers might
take part in colonizing space if it weren't too expensive. The expense
consists precisely in that it requires a lot of work.
Let me again claim that present new technology is making little progress
in displacing most jobs. If only construction technology were getting
cheaper rather than more expensive, we could rebuild our houses
and cities. If only bureaucratic technology were getting cheaper the
fraction of their income universities spend on administration could fall
rather than rise. If only automotive technology were getting cheaper
more rather than fewer people could afford new cars. Micro-processors,
given present well developed ideas for using them, seem to be a peripheral
technology, less capable than the automobile and trucks of displacing
labor. Of course, when robots of servant level capability can be built,
everything will change, but that will require new AI ideas.
∂11-Sep-80 0034 POURNE at MIT-MC (Jerry E. Pournelle) deadlines
Date: 11 SEP 1980 0332-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
Subject: deadlines
To: POURNELLE at MIT-MC, JMC at SU-AI
Deadlie rushes up for your contribution to THE SURVIVAL OF
FREEDOM anthology.
End of month at latest?
JERRY
Can do.
∂11-Sep-80 0835 FFL
To: JMC, FFL
The incoming PhD student whom I mentioned yesterday will be in to see
you this afternoon at 3 p.m.
Pat Worley
∂11-Sep-80 1650 JRA
price is $20+postage to lisp company, po box 487 redwood estates ca,95044
Please state a postpaid price, so people can buy without further
dickering.
∂12-Sep-80 1121 FFL
To: JMC, FFL
Wm. Stockton of the NY Times, Science Editor, called. Doing a Sunday piece
on AI. Wishes to come out on either September 26 or October l6 to talk with
people here and would like to see you. Can you call him back today or Monday
about this? 212 556 7l4l.
∂13-Sep-80 0132 POURNE@MIT-MC
From: POURNE@MIT-MC
Date: 09/13/80 04:30:06
POURNE@MIT-MC 09/13/80 04:30:06
To: JMC at MIT-MC
Just read yours to human-nets. Trying to be sensible with that
crowd is a valiant thing. Live long and prosper.
I doubt you'll ever get robots who'll be better house servants
than are human beings; it's hard to find a robot that CARES. You
may recall our housekeeper. However--Luz LIKES to work, feels
horrible if there is nothing to do, and MAKES work when there's
not some (as an example while I was back east she removed all the
books from my book shelves and dusted them as well as shelves; I
wasn't here to dirty the floor and she trhnks she ought to put in one hour a day in my office...
But finally--our problem in the US is that most people have been
taught that it is somehow demeaning to clean other people'{s
houses. Our housekeeper, being from an oldr tradition, is proud
to work for an author and a professor... an attitude I am sure
is going away fast.
JEP
∂13-Sep-80 0722 HUMAN-NETS@MIT-AI HUMAN-NETS AM Digest V2 #73
Date: 13 SEP 1980 1014-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS AM Digest V2 #73
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Saturday, 13 Sept 1980 Volume 2 : Issue 73
Today's Topics: Automation - Unions & ALPA/DC9
& Social/Sexual Biases & Dealing with its Impact,
Home Information Retrieval - Classified Ads & Terminals,
Telephone Rate Structures - Long Distance,
On propriety - Deletion Control Message
----------------------------------------------------------------------
Date: 12 Sep 1980 1017-EDT
From: Steven J. Zeve <CCIS.ZEVE at RUTGERS>
Subject: Labor unions and wage price controls.
This is in reply to a comment of a few issues back. [see HNT V2 #69]
Under voluntary wage price controls labor unions in general did not
stick to the 7% target. They went far over it. The only people who
claimed the unions stuck to 7% were Carter's advisors. It was a
little strange since the union members apparently thought they were
getting more like 10+%. No one I know figured out how the government
could make 10% be 7%.
steve z.
------------------------------
POURNE@MIT-MC 09/13/80 04:05:46
Somehow it comes as no surprise that a federal bureaucracy (FAA)
when doing spot check on flight crew makes the (very costly)
third crewmember become a passenger and perform no flight duties
--although the airlines contractually have to have the chap there.
Why doesn't that surprise me? Am I finally becoming infected with
the benign logic of Washington?
------------------------------
POURNE@MIT-MC 09/13/80 04:44:49 Re: do you believe Darwin?
The theory of natural selection states that adaptations take
place over VERY long time periods; Darwin, indeed, was concerned
that the age of the Earth (as then understood) was insufficient to
allow his mechanisms to work.
Now if you believe Darwin's theory, you will note that for
something like 1.5 million years the human race toodled along with
very definite sex-specific roles, and indeed seemed to select for
sexual dimorphism more extreme than that of many mammals.
We are now told that 5000 years of cultural molding is more
important than the immensity of pre-civilizational years.
Perhaps. But I doubt Darwin would have accepted the idea; nor
do I think it very persuasive except for our emotional desires to
prove a political point.
Equality is a legal concept, enforceable by courts and hangmen,
and perhaps ought to be enforced by the armed police; but equivalence
is a supposedly scientific assertion and ought to be subject to
scientific inquiry -- when in fact, any inquiry into the "equivalence
of the sexes" principle results not in rational debate, but epithets.
As witness the ones I will now receive.
JEP
------------------------------
Date: Fri, 12 Sep 80 05:53-EDT
From: bmstrazz at DNGC
Subject: workforce displacement
Why all the talk from alarmists about possible total displacement of
existing work force? Are robots going to do the work of garbagemen,
lavatory janitors, agricultural migrants, policemen, soldiers,
carpenters, recreational area workers, firemen, auto mechanics, etc.
I think not, even if robots can be devised to do some of these
menial tasks.
Barry Strassler
------------------------------
Date: 12 Sep 1980 at 2009-CDT
From: wilcox at UTEXAS
Subject: ATT Elect. Info Service
Dateline: Austin, Texas
From: Austin American Statesman
"ATT has chosen Austin for a trial run of its Electronic Information
Service in which a select group of customers will be able to call
up phone listings and grocery and department store ads on a small
computer terminal in their homes and businesses.
...740 customers will be selected for the trial which begins in mid
1981 and continue for 14 months. ... Customers chosen for the Austin
experiment will get a typewriter-like keyboard attached to a display
screen similar to a television. This terminal will be connected by
telephone lines to a Southwestern Bell computer.
Customers will be able to call up white, yellow, and blue (Government)
page listings and ads placed by department and grocery stores. ...
Austin was chosen because it is an upscale market -- high educational
levels, income levels.
In a similar Albany test only telephone listings were available. ...
A more elaborate trial is underway at Coral Gables where ATT provides
equipment and Knight-Ridder chain is supplying the computer stored
data base.
There customers can look up ads and order merchandise from Sears and
JCPenney catalogs, do their banking, get airline schedules, read AP
newswire, check ads and order from grocery stores, order books, check
weather, and keep track of how their elected representatives voted on
recent issues.
...We havent even come close to figuring out what the charges would
be if customers paid for the devices). ... Bell said EIS could be
available to customers by the end of the 1980s."
--Jim
------------------------------
Date: 12 Sep 1980 1544-PDT (Friday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: yellow pages
While I can certainly imagine a "threat" to white page type listings
from electronic systems, I have far more trouble imagining serious
threats to newspaper or yellow page advertising, at least in the
short term. For many purposes, newspapers are simpler to use
(you can rip out the ad, circle items of interest, etc.) True,
sophisticated systems could provide electronic analogues of these
functions, but current Viewdata/Teletext systems do not have such
niceties. Another point: The electronic systems provide only the
crudest form of "display" ads (pictures, fancy fonts, etc.), which
many advertisers feel are VERY important and would be unwilling to
give up.
--Lauren--
------------------------------
REM@MIT-MC 09/13/80 04:51:03
If TPC-USA gives out free terminals instead of printed yellow pages,
I hope they give out Bell-103 ASCII terminals so that those lucky
customers who first get them can use them for non-TPC services as
well (CBBS and similar stuff, and whatever else becomes available
via Bell-103 ASCII terminal). Meanwhile, does anybody know what
kind of terminal TPC-France is planning to give out?
A neat idea would be to install a usage-meter on each such terminal.
The difference between the reading of the meter on the terminal and
the accounted time connected to TPC automated directory would be time
used for non-TPC, which might be charged as rental. That wouldn't be
too unreasonable. TPC wouldn't be providing free terminals for access
to non-TPC services, but on the other hand customers wouldn't have to
pay the full monthly rental on a terminal if they were using it only
a little for non-TPC services. If TPC didn't find a way to charge for
non-TPC usage, it might put terminal rental companies out of business
so the whole free-terminal thing would probably be declared unfair
competition. But protecting against non-TPC use by using non-ASCII
(such as by wiring encryption into the terminals) would be really
nasty and selfish. Any TPC employees on this mailing list who want
to express feelings or provide information? Any FTC or FCC people
want to agree or disagree with my amateur legal opinion?
------------------------------
Date: 11 Sep 1980 0934-PDT
From: Geoff Goodfellow <GFF at SU-AI>
Subject: News Story on the alternate Long Distance phone networks.
Ma Bell Gets More Competition
By FLOYD NORRIS AP Business Writer
NEW YORK (AP) - If you live in one of more than 100 metropolitan
areas, the American Telephone & Telegraph Co. is no longer the only
game in town for long distance calls. And Ma Bell, the nation's top
telephone company, isn't happy about the competition.
International Telephone and Telegraph Corp. on Wednesday became
the third major company to announce a cut-rate, long-distance service
for residential customers.
''Anyone now spending over $20 a month on interstate
long-distance calls can save substantially with this service,'' said
Patrick J. Ryan, executive vice president and general manager of U.S.
Transmission Systems, the ITT subsidiary handling the service.
The burgeoning service has so far made only a small dent in AT&T's
long-distance revenues, which last year were $18.23 billion.
But MCI Communications and Southern Pacific Communications are
latching on to many customers since they began promoting the services
that had been available to businesses for several years.
In June, a federal judge ordered AT&T to pay $1.8 billion in
damages to MCI after a jury found the communications giant violated
federal antitrust laws in denying intracity phone collections to MCI.
MCI says 60,000 residential customers have signed on since it
entered he market in March, and SCP estimates it has about 8,000
since beginning residential service last October.
''It's obviously a potentially serious drain on revenues,'' said
AT&T spokesman Jay Gross. AT&T doesn't object to the competition,
he said, but contends the rules should be changed to force the other
companies to charge higher rates.
The three competing residential services cover calls only among
the major metropolitan areas and only after the regular business day.
All charge a monthly fee, either $5 or $10, plus a per-call
charge that they say is much less than Ma Bell's. ITT says its
discounts range from 15 percent to 80 percent for each call.
Customers dial a local number, paying a local charge, if there
is any, and are then connected to the microwave system that hooks
them into the distant city, where local lines again are used. An
identification number is dialed by the customer.
AT&T says its fees are higher because long-distance charges
subsidize local service and because its competitors serve only
lucrative markets, ignoring small towns.
It wants the Federal Communcations Commission and Congress to
require the competitors to pay more for local connections and to
use some of their revenues to subsidize local service provided by
phone companies.
ITT said it serves more than 100 metropolitan areas, compared to
fewer than 80 for MCI and SPC. Included are seven cities in Florida,
a potentially lucrative market since many retired people live there.
ITT will offer two types of residential service between 8 P.M.
and 8 A.M. on weekdays, plus all day on holidays and weekends. The
others provide service starting at 5 P.M.
One, called City-Call I, costs $5 a month and five to 16 cents
a minute per call, depending on time of night and distance. City-Call
II costs $10 a month, $5 for the first two hours of calls, plus $2.50
for each extra half-hour. There is a one-time connection fee of $10
for both types of service.
ITT also will let residential customers buy its regular business
service, which covers calls any time at a monthly charge of $10 plus
varying per-call fees at least 15 percent below Bell rates. The
connection fee is $30. MCI and SCP also allow residential customers
to buy their normal business service, but say few do so.
On a 1-minute call from Charlotte, N.C., to New York, ITT said it
will charge 13 cents between 8 and 11 P.M., and six cents thereafter.
It said AT&T would charge 34 or 21 cents depending on the time of
night.
To make a fair comparison, AT&T spokesman Grossman said, the
charges a customer pays for local service should be added to the
ITT figure, along with a portion of the monthly charge.
------------------------------
Date: 12 Sep 1980 1551-PDT (Friday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: (deleted)
For Pete's sake! (Whoever HE is.) Do some people ever have HARD
noses. The little blurb I quoted was from a widely circulated
"journal", and quoted a publicly available document. The point
THEY (and I) were trying to make is that sometimes the shreds
that are left after all the classified material is removed can
be rather humorous. At least, **I** thought it was humorous!
My most sincere apologies to the multitudes whom I must have
offended with that TASTELESS item. But I gotta tell ya',
sometimes I (deleted) and just can't (deleted) people.
--Lauren--
P.S. To those of you who sent me direct mail appreciating
that entry, you should be ashamed of yourselves!
--LW--
------------------------------
End of HUMAN-NETS Digest
************************
∂13-Sep-80 0718 REM via SU-TIP Progress on XGP-->PRESS program
To: JMC, REG
Tonite I wrote the first part of the program, at MIT-MC, then FTP'd it here
and checked to be sure it worked here too, it does.
Review: I decided a long time ago on MacLISP as the language to use for
all these document-compiler and translation programs. For this task,
I decided to first write a function that reads in a PRESS file and creates
internally an s-expression representing its structure, using existing
PRESS files as test data to debug it; then write the inverse function
which creates a PRESS file given its structure in memory, using the first
function to verify it works; then do same two things for XGP file, finally
write a function that maps the XGP-file s-expression into the PRESS-file
s-expression. Applying three functions in sequence, XGPREAD,
XGP-TO-PRESS, and PRESSPRINT, we have the desired XGP --> PRESS
translator. I expect to support both MIT and SU-AI XGP files,
and also for input plain text files. (RWK expressed an interest in
the MIT XGP compatibility, maybe even eventually replacing the existing
MIT program that is written in MIDAS which does a subset of what I
plan to do plus a bunch of stuff like spool-queue and EFTP; that is
replacing the relevant format-conversion subset, and more, but not
the spooling and EFTP stuff, at least not soon.)
To play with what I have now... alias to [1,REM] then say R PLISP
and wait for my default MacLISP environment to start up, then say
(REMLOAD 'PRESS) and then say (PRESSREAD '|some su-ai filename|)
for example (PRESSREAD '|PREFMT.IMG[ETH,REG]|).
∂13-Sep-80 0900 JMC*
cleaning
∂13-Sep-80 1533 JK
Some items concerning ekl - comments welcome.
I have sketched out my proposal for a top level and the proof checking language
for ekl on p. 19 of doc.ekl[ekl,jk]/19p
on [ekl,jk] - the ekl interpreter will be the same as the maclisp interpreter
with the exception that ekl data types be treated as atoms to preserve their
integrity.
On p. 18 of doc.ekl[ekl,jk]/18p
you can find my view on goal structuring and the use of
proofs as operators.
The file doc.new[ekl,jk]
contains brief descriptions for the underlying algorithms
for the ekl decision procedures (tautology, simplication, decide etc.)
I will start the coding for the decision procedures and the goal structuring
tomorrow.
I'll look. Incidentally if you give references to page 19
of doc.ekl on [ekl,jk] in the form doc.ekl[ekl,jk]/19p, the reader
can switch to that file just by giving the αxpo command.
I still think I need a lesson to try things out.
∂14-Sep-80 0331 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #74
Date: 14 SEP 1980 0624-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #74
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Sunday, 14 Sept 1980 Volume 2 : Issue 74
Today's Topics: Home Information Retrieval-Classified Ads & Terminals,
Telephone Services - Repair Experience,
Automation - ALPA/DC9 & Dealing with its Impact,
CBBS - Announcement about Boston/Cambridge/Maynard
----------------------------------------------------------------------
Date: 13 September 1980 11:36-EDT
From: Frank J. Wancho <FJW at MIT-MC>
Subject: Electronic Yellow Pages
Here in El Paso, Mountain Bell is dividing the yellow pages into
"regional" editions WITHIN the city and the businesses are really
upset about this - to the point of picket lines, etc. Their
objection is that they will have to pay extra to have their ads
in other regional editions to get the same exposure they have now.
Do you really think they will go along with an electronic yellow
pages in lieu of hardcopy yellow pages? Will the phone company
then be able to justify placing a yellow pages terminal in each
household without a hardcopy? Judging from the current furor,
I think not. It would be interpreted by the businesses as just
another way to raise still more revenue by the phone company -
just like regional editions...
--Frank
------------------------------
Date: 13 Sep 1980 1435-PDT (Saturday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: free terminals
The day ATT or GTE gives out "free" terminals to anyone other
than very narrow groups (like the deaf), on a mass scale, A LOT
of people will be surprised, including me. I just cannot see it.
Most of the current test systems are giving out Viewdata type
terminals (this includes France) since they have (crude) color
graphics capability. None of the Viewdata specs I have seen to
date were really 103/ASCII compatible, except in the most limited
sense. Most Viewdata systems have limited keyboards, and operated
1200/75 split or something similar.
--Lauren--
------------------------------
POURNE@MIT-MC 09/13/80 04:00:54
One thing about informational elites: pity those at the bottom of a
300 baud information well!
(This first thought of by PDL, but it's a valid concept. If you're
down at 300 baud you simply don't have the acces to nets that the
9600 and above aristocrats do...)
------------------------------
Date: 12 Sep 1980 2318-EDT
From: ROBG at MIT-DMS (Rob F. Griffiths)
Subject: [FLAME!]
We are having an addition to our house done, and the construction men
came out for the first time on monday, and promptly managed to cut
our phone lines. So we called TPC, from a neighbor's house, and they
said they would send a man 'right over' to fix it. Well, today, at
12:00 PM, the man finally arrived. We were without phone service for
all of this week. Has any one else had any problems with having lines
repaired?? I am greatly upset at the rates that TPC charges for our
local service, and them takes a whole week to show up. Is my case
just an exception??? Or have other people out there had this problem?
-rob
------------------------------
LLOYD@MIT-AI 09/13/80 11:17:52 Re: FAA flight crew spot checks.
In reply to Jerry's (POURNE@MC) comments about FAA spot checks of
flight crew performance, I have to come to the defense of the FAA.
The DC-9 is rated for a flight crew of TWO. A crew of two must be
able to perform ALL flight tasks. Just because ALPA, United, and
Western have decided that on those airlines DC-9's must carry a
crew of three doesn't mean that the FAA should change their rules
for those two airlines.
If the regs say two pilots can fly it then, damnit, two pilots
better be able to fly it.
I also hold that, whenever possible, the cockpit workload on the pilot
should be reduced. When I went for my instrument check ride, the regs
required me to perform ALL cockpit duties (navigation, communication,
answer dumb questions, etc.) myself. That doesn't mean that I don't
welcome qualified assistance when on an instrument flight (besides,
when I have help, I am a safer pilot).
<<disable flame mode>> Brian Lloyd
------------------------------
REM@MIT-MC 09/13/80 17:30:11
[I wanted to reply to this person directly instead of to HUMAN-NETS,
but the mailer claims DNGC isn't a server site on net and thus
doesn't think it can send mail there, hence this...]
Date: Fri, 12 Sep 80 05:53-EDT
From: bmstrazz at DNGC
Subject: workforce displacement
Why all the talk from alarmists about possible total displacement
of existing work force? Are robots going to do the work of
garbagemen, lavatory janitors, agricultural migrants, policemen,
soldiers, carpenters, recreational area workers, firemen, auto
mechanics, etc.
I think not, even if robots can be devised to do some of these
menial tasks.
Barry Strassler
[in some sense <robots> are already doing work of garbagemen somewhat.
We put our apartment garbage in a large bin, the garbage truck comes
in and lifts it up and dumps it in the truck innards. A human drives
the truck and performs rendesvous and docking with the garbage bin,
but once we get automated highways this could be one of the first
unpleasant jobs to be automated (a micro-transponder on each bin
might help, and would only cost as much as the labor cost of a couple
week's human driving). As for lavatory janitors, policemen, soldiers,
firemen -- I'd volunteer to do some of this via telepresence as soon
as the technology becomes available. The other three would be harder
to automate I suppose.]
------------------------------
MITTON@MIT-AI 09/14/80 05:16:02
CBBS/Boston (in Randolph, Mass), 617-963-8310, has closed up shop.
CBBS/Cambridge (which has been down since it's equipment was stolen
in Feb) is now back on the air at 617-864-3819. I am running this
system using the former CBBS/Boston equipment and database.
CBBS/Boston is now *OFF-THE-AIR* please erase the 963-8310 number
and pass the word! Mitch sold his equipment to the New England
Computer Society to replace our stolen system and so that he could
retire gracefully from the CBBS scene. CBBS/Cambridge will continue
to offer 24 hour CBBS services to the Boston area.
Also; please pass the word that CBBS/Maynard, formerly 897-0346,
which has been down since last November, should be considered
permenently off-the-air. DEC is moving me and the DECnet group
to the sticks of Tewksbury, Mass in a few months anyways.
Dave Mitton
------------------------------
End of HUMAN-NETS Digest
************************
∂14-Sep-80 1532 VRP dialnet
Do you know SU-AI's process ID for dialnet?
No. I'm afraid my level of knowledge about Dialnet is quite low.
I confess vagueness as to what a process ID is. However, I think
MRC and Tim Eldredge would know and wouldn't mind being phoned even
today if it's urgent.
∂14-Sep-80 1616 Buchanan at SUMEX-AIM Natl Rsch Council position paper
Date: 14 Sep 1980 1616-PDT
From: Buchanan at SUMEX-AIM
Subject: Natl Rsch Council position paper
To: jmc at SU-AI
John,
After you declined to write a position paper for the NRC Panel on Basic
& Applied Rsch in Comp. Sci about AI, the request filtered through
Feigenbaum to me. I am preparing such a draft now and would
appreciate some time with you tomorrow (Monday), if possible,
to talk to you about it.
There are several topics mentioned in the minutes of the meetings,
but I'm confining my thinking now to just the major ones mentioned
as guidelines for the reports. The outline I'm wrking from is
appended.
Comments by msg would be almost as helpful as talking, if that's
more convenient.
thanks,
Bruce
DRAFT FOR NATL RSCH COUNCIL PANEL ON BASIC AND APPLIED RSCH IN COMP.SCI.
B.G. Buchanan
What is AI research?
Theoretical Rsch
Experimental Rsch
Where is it done?
Universities
Industrial Research Labs
Non-Profit Research Labs
Why is AI impt for industry?
Product development
Services
In-house pgms for increasing productivity
Strengths & Weaknesses in Current Rsch Posture
Strengths:
critical mass of rschers in major centers
clearly identified publication media
sustained funding for major centers from DARPA
mix of basic & applied rsch
cooperation among scientists across geographical lines
(fostered by ARPANET)
Problems:
personnel
computing cycles / memory
insufficient basic research ideas
proprietary nature of indus. rsch
insufficient attention to past work
no tradition of controlled experiments
mixed quality of published work
technology transfer
(rewriting pgms; developmt effort for I/O & other details;
"marketing")
funding outside of major centers
(AI rsch is big science)
Recommendations
Government & Industry
provide fellowships for promising young minds (post-docs)
support more basic research
underwrite journal subscriptions & conf. proc. for libraries
encourage govt & indus. rschers to teach
provide subsidies for new hardware
Universities
shorten time for PhD
enlarge MS programs
sharpen criteria of quality in AI
-------
I have one comment on your outline. Most of the other reports I have seen
and the instructions are rather heavy on history. They emphasize what
developments in the field have reached practice and were they done
in universities or industry. You are probably as well qualified
as anyone to do this for AI. If you want to talk now, I am at
7-4430.
Further comments on recommendations: Since one of our major expenses
is software, re-equipping with software is as important as re-equipping
with hardware. For example, support for the maintenance of MACLISP and
INTERLISP would be very helpful.
I doubt that a recommendation that would apply to all sciences is
especially helpful unless some contrast can be pointed out. For example,
there is no more obvious need to underwrite journal subscriptions in
AI than in any other field. On the other hand, there are sources of
funds for postdocs in other fields and nothing in AI or even in computer
science generally.
∂14-Sep-80 2100 JMC*
reiter letter
∂15-Sep-80 0846 REM via SU-TIP Status of XGP/PRESS program
To: JMC, REG
I've finished the font-directory parser, and written the EL parser.
It doesn't know all the EL opcodes, but if it encounters one it doesn't
know it asks the user how many bytes (see transcript), stores that info
in an ALIST, and proceeds with parse. It now parses everything in the
PRESS file except DL and ignored chunks of file. I'm working on parsing
the DL (via pointers from the EL) next.
See PRESS.LST[1,REM] for a PTYJOB transcript of this morning's run of program.
Note how easy it is in PLISP (MacLISP with full BIBOP allocation except for
arrays, which I'm not using) to reallocate in the middle of a program that
runs out of space. The parsed things are currently put on globals. When
it is finished, it will put everything together into one giant S-EXPR and
flush the globals before RETURNing. Then I'll be ready to start on the
inverse function which will create a PRESS file de novo from the S-EXPR
representing its parse.
∂15-Sep-80 1102 CSL.LAB.SALLY COmputer Forum Memo
Date: 15 Sep 1980 1008-PDT
From: CSL.LAB.SALLY
Subject: COmputer Forum Memo
To: cSD.Dbrown, CSL.JLH
cc: csl.sally at SU-SCORE
Remailed-date: 15 Sep 1980 1059-PDT
Remailed-from: Denny Brown <CSD.DBROWN at SU-SCORE>
Remailed-to: CSD-Faculty: ;
Would you be good enough, Denny, to send this on to the CSD
faculty, and would you, John, please send this on to the CSL faculty:
To: CSD/CSL Faculty
From: Sally Burns, Assistant Director
Subj: New Policy concerning Maintainers' Fees
Stanford Computer Forum
In Mike Flynn's memo of 21 May l979, the concept of maintainers' fees to
encourage CSD/CSL faculty to participateactively in warm relations with
our Computer Forum members was established. Specifically, a fee of
$l,500/compnay/membership year was to be depositedinto the unrestricted
funds of the faculty member volunteering to foster this relationship.
At its meeting at the end of Spring Quarter of this year, the
Computer Forum Committee examined the results of this format. The
Committee concluded that liaison activity has either not increased or not
been documented as increasing. They, therefore, wish to establish a
modification to this policy.
The portion of the membership fee designated for the maintainer
will be withheld from deposit into the unrestricted funds of that faculty
member until some evidence of acitvity between that faclty member ad that
affiliated company can be substantiated. The most acceptable method of
documenting such activity will be in the preparatin of a brief "trip
report" in HARDCOPY FORM to be kept in the central files in my office.
This report should include news of a site visit to a particular facility
of a company, incorporating such details as date visited, name of faculty
member visiting, topics discussed or title of seminar given, and names of
industrial colleagues contacted. Alternatively, brief reports filed
concerning visits to the campus by industrial scientists of our Forum
companies (including hosted activities, subjects discussed, etc.) will
likewise indicate that some liaison is in effect.
Upon receipt of this brief report, the $l,500 maintainer's fee
will be deposited to the designated maintainer for that membership year.
Please note that it IS possible for the official maintainer to
arrange for some other faculty member to pay the site visit. Terms of
payment to this surrogate maintainer will be negotiated between the two
parties involved.
This policy in no way affects the $l,500 for the finder's fee.
The finder also gets to be the maintainer for the first three years, but
the $l,500 for maintenance again must be documented as in the above format
before it is deposited.
cc: D. Brown
B. Scott
-------
∂15-Sep-80 1112 Denny Brown <CSD.DBROWN at SU-SCORE> Richard's office
Date: 15 Sep 1980 1109-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: Richard's office
To: jmc at SU-AI
cc: rww at SU-AI
I'm waiting for confirmation from you (and RWW) that it's OK to move
Anne Gardner and Paul Martin into RWW's office for fall quarter. Richard,
though on leave, is expected to use the office too. Paul plans to start at
SRI Jan 1. By then we will decide whether Anne will be RWW's official
roomie, or we'll replan. The only potential problem we see is that both
Paul and Anne are confirmed smokers. Even if there is no hope for Anne
being Richard's partner, I'd like to take advantage of the fall quarter
only option. Both are now officed in the hall, so quick response would
be appreciated by me and them. -Denny
-------
I talked to Richard. It is agreeable to him provided he gets to keep his
desk. Maybe another table could be moved in. He doesn't mind the fact
that they smoke.
∂15-Sep-80 1115 FFL
To: JMC, FFL
Your reservations to Vancouver and return made.
Leave SFO 8:30 p.m., Western Flt 469. Arrive Vancouver 10.25 p.m.
Leave Vancouver Saturday, 8:15 a.m., Western Flt. 460. Ar. SFO 10:15 a.m.
Friday nite possibility: On Pac-Western. Leave Vancouver 5:25 p.m. and
arrive Seattle 6:45 p.m. Leave Seattle on Western at 7:40 p.m. and arrive
in San Francisco at 9:2l p.m. Please note that these flights are presently
filled - you would have to check to see if space were then available.
∂15-Sep-80 1421 PHY McCarthy book
To: Nilsson at SRI-KL
CC: JMC at SU-AI
There are two ways to go:
(1) TEX John's papers.
(2) Write a program that will translate the XGP files produced by PUB into
our DVI (device independent!) encoding. (We produce Alphatype files
starting from this DVI format.)
This second path was already attempted by Les Earnest and he almost got it.
(Some glitches remain in his program.)
I think it might be worthwhile to try to TEX things again (you may find
nicer layouts, etc.).
Finally (and firstly): Yes, the Alphatype is available. Charges are
$10 per output sheet (about 6 pages of the Art of Computer Programming
per sheet). Wish you the best decisions.
Luis Trabb Pardo
Research Associate
LTP at SU-AI
∂15-Sep-80 1616 kirkley at sri-unix tnx for the reply.
Date: 13 Sep 1980 at 0350-PDT
From: kirkley at sri-unix
Subject: tnx for the reply.
To: jmc at su-ai
cc: kirkley at dnsri, kirkley
rsvp: kirkley@dnsri (deaf-net at sri)
file: namadr:john mccarthy <jmc at su-ai>
thank you for the reply to my message.
i pick up mail about every three days.
dnsri is sharing a small unix machine... brevity is next to godliness.
i am volunteering work for the deaf network... hence my access to dnsri.
i am holding a once in a lifetime gathering of the "vituperative
associates" tonight (sat 14 sept 80) at my house at 159 churchill
ave (cross street alma) 6 to 10 pm. i would be honorored by your
attendence if possible. other guests... raphael, rulifson, duda,
ingalls, engelbart, rosen,... and many in wheel chairs, and some
deaf (from my other volunteer work).
please call 415-328-3231 for details or to leave message with answering
machine.
thank you for your kind support and help over the years.
your student,
charles r. kirkley (kirkley@dnsri)
**
Your message came on MOnday at 4pm,, so I can't attend gathering.
DNSRI is not recognized as a host by our ARPAnet program. However,
SRI-UNIX is. It seems to me that some non-Stanford person also
complained that DNSRI wasn't recognized, so maybe it's a general
phenomenon.
∂16-Sep-80 0328 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #76
Date: 16 SEP 1980 0619-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #76
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Tuesday, 16 Sept 1980 Volume 2 : Issue 76
Today's Topics: Electronic Banking - Auto-teller security,
Home Information Retrieval- Terminals & Classified Ads,
Telephone Services - GenTel vs PacTel,
Automation - Social/Sexual Biases
----------------------------------------------------------------------
Date: 15 Sep 1980 1724-PDT
From: Billy <BRACKENRIDGE at USC-ISIB>
Subject: [Dwight E. Cass <CASS at USC-ISIE>: Beware of UCB]
For you non-Californians UCB is one of our smaller California banks,
but one of the earlier to go to the ATM. The larger banks will be
doing so soon. Dwight promises to keep us informed as the drama
unfolds.....
---------------
Date: 15 Sep 1980 1201-PDT
From: Dwight E. Cass <CASS at USC-ISIE>
Subject: Beware of UCB
People with UCB Day/Night Teller Cards - BEWARE!
I have just had $10 withdrawn from my savings account by the machine,
WITHOUT MY KNOWLEDGE. After talking with UCB about this, they inform
me that I am responsible for all transactions which the machine's
records under my card number, whether or not I actually did it.
I am still talking with UCB about this problem, but I felt everyone
at the Institute should be warned!
---------------
[ For our earlier discussion of security problems with auto-tellers,
see ELECTRONIC BANKING - AUTO-TELLER SECURITY in [HNT V2 #3-13,17].
-- RDD ]
------------------------------
KWH@MIT-AI 09/15/80 22:41:18
I used AI for several months at 300 baud through a TIP, and now
that I am in Cambridge, with local high-baud terminals, the difference
is very sharply felt. It was the 300 baud data rate that led me to
initially propose a "skimmer" program, and such a device would make
many tasks an order of magnitude easier at 300 baud. The biggest
problem was when I had to "read" anything, be it mail or online
documentation - you cannot scan a document like you can turn pages
or advance a 1200 or 9600 baud terminal. If 300 baud terminals are
provided for WorldNet, the central utilities should provide a lot
of good on-line utilities, such as EMACS or skimmers to the users.
-Ken
(The things on skimming are mentioned in early summer HNT's)
[ For our earlier discussion of electronic analogs of skimming, see
DATABASES AND PUBLISHING - SKIMMING, [HNT V1 #101-107,109,111].
-- RDD ]
------------------------------
Date: Mon, 15 Sep 80 16:42-EDT
From: bmstrazz at DNGC
Subject: free machines - an illusion
Lauren speaks of his skepticism of free machines (if any) given out
by ATT, GTE or some other carriers. I wish to add to this discussion
in that the machines we, the deaf receive are the obsolete types that
the carriers are only so anxious to get rid of. If we, the deaf did
not grab these machines then these would either be dumped in the ocean
(not a joke as it occured some time ago much to the wrath of the deaf
community) or somewhat "sold" at the highest bid to radio hams and
computer hobbiests. When we obtain these machines, obsolete that
is, then we are a generation or two behind the latest teletypewriter
machine on the scene. Witness model 15/19's given out when model 28
were popular; model 28 given out in favor of model 32; model 32 then
given out in favor of model 42, and so on.
A blessing or a curse, depending on how one looks at it. We bless the
28's and 32's we get but curse the inadequacy of these machines when
looking at the features of the 42's.
Any comments?
Barry Strassler (BMStrazz at DNGC)
------------------------------
Date: 15 Sep 1980 0128-PDT (Monday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: yellow pages
Hah! You think businesses are screaming about small "regional"
yellow page books! They did that here in L.A. years ago. To get
all the yellow pages for the full San Fernando Valley, for example,
takes about six separate books.
Now they've hit upon some new ones around here. First of all, they
are pushing the "neighborhood" phone book even more -- meaning even
SMALLER yellow pages books. And a totally new idea: they have split
the L.A. central yellow pages into TWO books. Now, last year, for
the first time, it was so big they put it in two volumes (A-L and
M-Z). Note that this is JUST very central L.A., none of the rest of
the city and suburbs. This year, it is also two books, but one is
the "business to business" book an the other is the "business to
consumer" book. That is, one is delivered to businesses and the
other to homes. The concept promoted is that this makes it easier
for you to find what you want. Of course, the reality is that now
you have to advertise in BOTH in many cases. My understanding is
that this same idea will be applied to other regions in L.A. and
around the country shortly.
By the way, a residence user CAN get a copy of the business book,
but you must ask for it specially. It takes a pile of telephone
books about fifteen feet high, at least, to cover all of the L.A.
region (area code 213).
--Lauren--
------------------------------
Date: 15 Sep 1980 0159-PDT (Monday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: PacTel VS. GenTel
Yeah, there sure is a difference between the two. But I have to
come to GTE's defense in one respect, they are generally easier to
deal with concerning "odd" requests (like private circuits and such).
Bell has detailed regs on everything including the color of pencils
operators use (or so it seems). Sometimes GTE seems to make up the
rules as it goes along...which can sometimes be good. Well, sometimes
not so good either. Above the installer level, however, I have found
GTE to be more responsive when it comes to esoteric sorts of circuits.
They DO have some smart people, but their typical installer is, well,
shall we say, "undertrained"?
Mike's rendition of the strange GTE data installer is true. What
actually happened is that I was having a touch-tone modem installed
for a voice response system. I went into UCLA on a Saturday to
supervise the installation. I had a hard time getting anyone to
admit the modem existed in the first place, by the way. So the guy
puts it in, and then starts to pull over his serial I/O testing box.
I say, "Oh, don't bother with that, it's only for ASCII data, this
is a touch-tone modem." He stared at me, and then at his tester,
and then at the modem. Then he said (increduously), "You mean that
thing actually receives touch-tone and converts it to data?" "Yep,
that's right", I replied. "GOLLY! That's NEAT!", he exclaimed...
--Lauren--
------------------------------
Date: 15 Sep 1980 07:28 PDT
From: Chapman.ES at PARC-MAXC
Subject: do you believe Darwin? [HNT V2 #73]
I agree with Jerry's point about evolution: Nature has been selecting
for sexual dimorphism for much longer than Man has been trying to
reduce it. However, I remember an article in Science several years
ago about a tribe of primitive hunter-gatherers in Africa. In this
tribe, the women and the men performed equal tasks in just about
everything, including staying home and playing with the kids, and
the women had just as much say in governing tribal affairs as the
men. It was suggested that this tribe was typical of prehistoric
hunter-gatherer societies. It may be that this is a more realistic
picture of the way Man spent 1.5 million years developing, and the
sex-specific roles were only a (relatively) recent development of
cultural molding. The current struggle to equalize male and female
responsibilities and opportunities may be a "return to nature",
except we are so far from our roots that we don't recognize it as
such.
------------------------------
Date: 15 September 1980 1157-EDT (Monday)
From: Robert.Frederking at CMU-10A (C410RF60)
Subject: Human adaptation
There are at least two kinds of adaptation people can do. One is
to physically evolve to fit a new environment. This could even allow
eventually living in a low oxygen atmosphere, I would guess. The
problem is that it is incredibly slow (there is STILL a problem with
the 5 billion year age of the earth being a little too short for what
has happened). Also, the human ability for an individual to adapt his
lifestyle may very well have removed natural selection as a force in
mankind's development (except to make people more sexually attractive
to each other and less likely to use contraceptives). There is also
the nasty tendency for evolution to work through mass death.
As far as changing one's lifestyle to fit the environment, this
would certainly solve most of the problems. But why should people be
forced to make changes in their lives that the majority don't want?
It is more reasonable to change the environment, if possible. A good
example is that it would be quite easy to survive on a planet that
has a very high level of particulate pollution, just by building dust
masks for everyone and eating vegetables. But who wants to do that?
Instead, you invent a way to force large, profit-oriented organiza-
tions to cut back on their polluting. You pay a price, but you get
an atmosphere that is pleasant. The thing that we must do is to
decide which way we wish to deal with our new opportunities. Do
we evolve to fit a new society, or do we find a way to make that
new society comfortable for the kind of human beings that already
exist? The choice must be made by the majority (in our society)
with whatever information the specialists can provide on the
available alternatives.
This last statement is a bit too idealistic, I admit. But I
think that it is the underlying mechanism behind what policies a
democratic country follows. My main point is that we shouldn't
think that we MUST go ahead with a particular course of development,
even if it means re-engineering mankind. We should keep an open
mind about the alternatives available and support the one that
will most greatly benefit the greatest number of people.
End of diatribe.
------------------------------
End of HUMAN-NETS Digest
************************
∂16-Sep-80 0656 REM via SU-TIP Progress
To: JMC, REG
I have PRESSREAD parsing everything correctly, am doing final cleanup
before bringing it to SU-AI, unfortunately MIT-MC crashed. After one
last little cleanup, will test it here, then start work on PRESSPRINT
the inverse function sometime soon.
∂16-Sep-80 1311 FFL
To: JMC, FFL
please call Peter Hart, 9 323 3616, today.
I was talking to Peter Hart today, and the question of their (Fairchild's)
need for high quality terminals arose. I suggested that he might
to well to lend an engineer if he had one to the SUN terminal
project, and he is willing to consider it, but he won't even be
in his office for two weeks. He suggested that you call him after
this time if you are interested in pursuing the possibility.
His number (apparently a home number) is 323-3616.
∂16-Sep-80 1343 FFL
To: JMC, FFL
Pleasecall Alphonse Juilland, 9 325 0949.
∂16-Sep-80 1418 Denny Brown <CSD.DBROWN at SU-SCORE> New PhD student summary info.
Date: 16 Sep 1980 1411-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: New PhD student summary info.
To: CSD-Faculty: ;
cc: CSD-Research-Associates: ;, csd.tajnai at SU-SCORE, csd.dale at SU-SCORE,
csd.bscott at SU-SCORE
All of the faculty and most of the research associates in the
department have been sent summary information on the new PhD
students. (If you didn't get this, and wish you did, ask
Dot Dale to send you a copy.) Some of the information is
confidential, so don't distribute these publicly. You should
use these summaries to identify students who might be interesting
to you for research support, academic advisee, etc.
When students arrive, Carolyn and the student advisors provide
some orientation. I try to talk to all of the PhD students
to determine how their current interests relate to our situation.
(The information on the cover sheet you got comes from the
applications; it is nearly a year old.) I point them
to appropriate faculty members for possible research support.
They are encouraged to make appointments with those faculty
members to determine if a confluence of interests and needs
exists. (Have your secretaries ready to schedule new PhD
students according to your desires. I tell them to specify
that they are PhD students; you probably want to postpone talking
to new Masters students until the quarter starts.)
When a PhD student/research advisor match has been determined,
you should let Carolyn know. She'll keep track of who has
been placed and who hasn't, and pass on the info to Betty etc.
It seems like there is more research support available than
PhD students to fill the slots. It is worth your while to
talk to the students if you want to support new ones, else
they'll be grabbed by somebody else.
After research support has been determined, I'll assign each
student an academic advisor. In most cases, the academic advisor
and the research advisor will be different. This ensures that
the student has more than one initial contact in the department.
If you don't want to see any new students, let me know and I won't
suggest they see you. More complex filters should be handled by
your own secretaries.
-Denny
-------
∂16-Sep-80 1421 CSD.ULLMAN at SU-SCORE Travel for thesis committee
Date: 16 Sep 1980 1420-PDT
From: CSD.ULLMAN at SU-SCORE
Subject: Travel for thesis committee
To: jmc at SU-AI
I'm in the process of preparing backup material for the meeting on Sept. 30.
Do you wish to supply any written statement regarding your motion?
If so, may I have it by Friday.
-------
The issue recently arose of the Department paying the
travel expenses of David Waltz to attend Dick Gabriel's thesis
oral, Waltz being one of Gabriel's readers. The Department
declined to pay, and Tom Binford and I arranged for payment,
although neither of us is involved with Gabriel's work. I think
the Department should be willing for outside reader's to attend
thesis orals, when there is no research source involved. There is
altogether too little interaction between CS faculty in different
universities as is, and interaction at the thesis committee level
is very valuable in developing a common computer science public
opinion about what constitutes a good thesis and a good state
of knowledge for PhDs in computer science.
∂16-Sep-80 1506 CSD.ULLMAN at SU-SCORE
Date: 16 Sep 1980 1500-PDT
From: CSD.ULLMAN at SU-SCORE
To: JMC at SU-AI
In-Reply-To: Your message of 16-Sep-80 1430-PDT
I have made a listing of your comments. Do you want them circulated?
-------
Yes, please circulate.
∂16-Sep-80 1520 ME macros upon login
∂16-Sep-80 1047 JMC
Can read macros be automatically loaded on login?
ME - Not yet, but soon.
∂16-Sep-80 1741 Buchanan at SUMEX-AIM Natl Rsch Council Panel
Date: 16 Sep 1980 1740-PDT
From: Buchanan at SUMEX-AIM
Subject: Natl Rsch Council Panel
To: jmc at SU-AI
John,
I'll have a much better idea after the panel meeting about the kind
of document they would like to have. I'm taking an extended outline,
which I'll put in the next msg to you, and have incorporated your
suggestions. Thanks for your help -- I'll appreciate more when I
get back.
bgb
-------
Have a good trip. I think the outline is basically what is wanted.
Please give my regards to Jack Schwartz if he's there.
∂16-Sep-80 1742 Buchanan at SUMEX-AIM NRC draft (4 pages)
Date: 16 Sep 1980 1741-PDT
From: Buchanan at SUMEX-AIM
Subject: NRC draft (4 pages)
To: jmc at SU-AI
DRAFT FOR NATL RSCH COUNCIL PANEL ON BASIC AND APPLIED RSCH IN COMP.SCI.
Bruce G. Buchanan
Stanford University
Sept. 18, 1980
1. What is AI research?
Artificial Intelligence (AI) is that branch of computer science
dealing with symbolic, non-algorithmic methods of problem solving.
One long range goal is to develop a theory of intelligence. A short-term
goal is to develop computer programs that exhibit some aspects of
intelligence. Understanding both the theoretical and practical aspects
of intelligent problem solving is important in either case.
2. Where is it done?
[Universities, Industrial Research Labs, Non-Profit Research Labs]
With only a few exceptions, all AI research to date has been
carried out in major universities. The most notable exceptions are
RAND, SRI-International, Xerox-PARC, and BBN (Bolt Beranek & Newman).
At the 1956 Dartmouth Conference, which many take to mark the formal
birth of AI, three major centers were represented by a handful of
individuals:
MIT, Carnegie-Tech (now CMU), IBM
A recent tabulation [SIGART, Nov.,1979] of universities offering
educational and research opportunities in AI listed 68 places
around the world, 47 in the U.S. Two and a half years ago, a
tabulation of corporate AI locations [SIGART, April, 1978] listed
20 industrial AI research groups in the U.S., with the number much
larger today.
Funding for AI research has been concentrated in the last two
decades at three major universities:
MIT, CMU, Stanford
Although this is still true, there is enough support for significant
research at another dozen universities as well, including:
Illinois, Maryland, Massachusetts, Rutgers,
Rochester, Texas, Yale
3. Why is AI important for industry?
[Product development, Services, In-house programs for increasing
productivity]
Industrial interest in AI is growing rapidly. At the 1980 meeting
of the American Assoc. for AI (AAAI), a tutorial entitled "Technology
Watch for R&D Managers" drew 109 paid attendees from 27 large companies
and 31 small ones (and 8 Federal agencies). Several large companies have
established AI research groups, including:
Schlumberger, Fairchild (now part of Schlumberger), Xerox,
Texas Instruments, General Motors, Hughes Aircraft.
Why?
Using the concepts and methods in the AI literature, it is possible
to develop new product lines and new services. TI's toy, Speak & Spell,
is an example from past work on speech synthesis. IBM's (and others')
word processing systems are examples from past work on interactive text
editing. Telenet is a commercial computer network growing out of the
work of the AI community on the ARPANET.
Equally important, but not so flashy, benefits can be derived from
companies using AI techniques in-house. Firms engaged in genetic
engineering, for example, see the potential of AI-based programs (MOLGEN)
in keeping track of and reasoning with the rapidly changing knowledge
in that field. Among Schlumberger's goals are to use AI programs to
aid in the interpretation of oil well data and to aid engineers in
the maintenance of complicated equipment.
4. Strengths & Weaknesses in Current Research Posture
Successful AI research has already led to new products and services
and better in-house operations. Largely this is because of the transfer
of ideas, though, rather than a transfer of programs. The products
and services one points to are by-products of AI research, and not
the prorams in which the ideas were initially demonstrated.
The reasons for and impediments to success are listed below.
Because the experimental side of AI research requires considerable
computing resources, sustained funding at a high level is a critical
component of success, as with MACSYMA and DENDRAL which were developed
over a period of ten years. A shortage of qualified persons with
ideas and skills is the most serious impediment.
A. Strengths:
sustained funding for major centers from DARPA
critical mass of rschers in major centers
cooperation among scientists across geographical lines
(fostered by ARPANET)
complex software & large computer systems available via network
clearly identified publication media
mix of basic & applied rsch
B. Problems:
personnel shortage
insufficient computing cycles & memory
continued maintenance of AI languages (Interlisp, Maclisp)
insufficient basic research ideas
proprietary nature of industrial research
insufficient attention to past work
no tradition of controlled experiments
mixed quality of published work
technology transfer
(rewriting programs; development effort for I/O & other details;
"marketing"; size of target computer)
5. Recommendations
A. Government & Industry
provide fellowships for promising young minds (post-docs)
subsidize new hardware to replacing aging computer
subsidize maintenance of AI languages (Interlisp, Maclisp)
B. Universities
shorten time for PhD
enlarge MS programs
sharpen criteria of quality in AI
!
6. CASE STUDY: DENDRAL
[material to be taken from Lindsay, et al., Mc-Graw-Hill, 1980]
History
Current Status
Factors in Its Success
Technology Transfer
Dissemination
Working relations between University & Industries
-------
∂16-Sep-80 1842 PLS POP2
I don't know about "expert", but I've certainly used it. I don't
think I have a manual here, but I can probably find out anything about
it I don't know fairly easily.
∂16-Sep-80 2000 JMC*
Sarah letter
∂16-Sep-80 2107 CSL.FB at SU-SCORE
Date: 16 Sep 1980 2103-PDT
From: CSL.FB at SU-SCORE
To: JMC at SU-AI
In-Reply-To: Your message of 16-Sep-80 1514-PDT
Thanks. If they are willing to do that without claiming a propriatary
interest in the result...
-------
As you probably know Peter Hart left being head of the AI Center at
SRI to head a new AI group at Fairchild but with the sponsorship
of Schlumberger, which I think bought Fairchild or something like
that. His interest is in getting good display facilities for his
group not in getting a product, so I think he will be reasonable.
My idea was merely that they would share the effort by donating
the services of an engineer. I think it isn't very highly probable,
because maybe they don't have the person, but it may be worth pursuing,
since things (from my point of view anyway) seem to be going slowly.
Probably you understand all that, but since I was in a hurry this
afternoon and therefore concise, I want to be explicit now.
∂17-Sep-80 0023 kirkley at sri-unix
Date: 17 Sep 1980 at 0005-PDT
From: kirkley at sri-unix
To: JMC at SU-AI
cc: kirkley, kirkley at sri-tsc, kirkley at dnsri, kirkley at kl
sorry the invitation to the gathering was delivered late.
≥dnsri (deafnet sri) lives on a small unix...also known as sri-tsc for
telecommunication science center.
i am now receiving the computer sciend dept list of publications.
please recommend a general survey paper that describes your present system and
immediate goals.
nice to chat with you again after such a long hiatus.
I'm not sure what you mean by "present system and immediate goals".
Do you mean our computer system or my research goals or the AI Lab
or AI generally or the computer science department?
∂17-Sep-80 0722 CSL.FB at SU-SCORE
Date: 17 Sep 1980 0720-PDT
From: CSL.FB at SU-SCORE
To: JMC at SU-AI
In-Reply-To: Your message of 16-Sep-80 2303-PDT
I'm completely with you and, as I apparently succeeded in saying Saturday,
we are quite short of engineering talent.
-------
∂17-Sep-80 0852 FFL
To: JMC, FFL
Please call collect by 6 p.m. Eastern time Alice Digillio of Book World of
the Washington Post re doing a couple of book reviews for them.
∂17-Sep-80 1604 CSD.JEANIE at SU-SCORE Faculty Meeting
Date: 17 Sep 1980 1559-PDT
From: CSD.JEANIE at SU-SCORE
Subject: Faculty Meeting
To: buchanan at SUMEX-AIM, rwf at SU-AI, or.dantzig at SU-SCORE, ghg at SU-AI,
jgj at SU-AI, dek at SU-AI, jmc at SU-AI, ejm at SU-AI, zm at SU-AI,
csd.ullman at SU-SCORE
cc: csd.bscott at SU-SCORE, csd.jeanie at SU-SCORE
Professor Ullman has scheduled a special meeting of the full professors
to discuss the promotions of Tom Binford and Robert Tarjan. This meeting will
begin at 4:00 directly after the regularly scheduled faculty meeting Sept. 30.
The files on Binford and Tarjan are available in Betty Scott's office.
Jeanie
-------
∂17-Sep-80 1643 CSD.GENESERETH at SU-SCORE Siglunch, Friday the 19th
Date: 17 Sep 1980 1634-PDT
From: CSD.GENESERETH at SU-SCORE
Subject: Siglunch, Friday the 19th
To: Siglunch-crowd: ;
cc: : ;
This week Jon Doyle has agreed to talk to us about selected topics from
his thesis. I include the abstract from that document below. We'll meet
in the gazebo as usual, at 12:05 sharp.
A Model for Deliberation, Action, and Introspection
by Jon Doyle (doctoral research done at M.I.T.)
This thesis investigates the problem of controlling or directing the
reasoning and actions of a computer program. The basic approach explored
is to view reasoning as a species of action, so that a program might
apply its reasoning powers to the task of deciding what inferences to make
as well as deciding what other actions to take. A design for the architecture
of reasoning programs is proposed. This architecture involves self-
consciousness, intentional actions, deliberate adaptations, and a form of
decision-making based on dialectical argumentation. A program based on this
architecture inspects itself, describes aspects of itself to itself, and uses
this self-reference and these self-descriptions in making decisions and
taking actions. The program's mental life includes awareness of its own
concepts, beliefs, desires, intentions, inferences, actions, and skills.
All of these are represented by self-descriptions in a single sort of language,
so that the program has access to all of these aspects of itself, and can reason
about them in the same terms.
-------
∂17-Sep-80 2343 100 : kspace via AMES-TIP
Want to see the business plan for the Text Network Program on Monday?
∂17-Sep-80 2347 100 : kspace via AMES-TIP
If you're looking for a mailbox you won't find one.
∂18-Sep-80 0336 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #78
Date: 18 SEP 1980 0618-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #78
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Thursday, 18 Sept 1980 Volume 2 : Issue 78
Today's Topics:
Going to COMPCON?,
Electronic Banking - Apple prototype & Auto-teller security,
Computers and the Deaf - CA Bill on Free TDD's
----------------------------------------------------------------------
Date: 17 September 1980 1334-EDT
From: Paul Hilfinger at CMU-10A
Subject: COMPCON panel
For the benefit of all us stay-at-homes, would some munificent COMPCON
attendee take notes on the "Cartesian programmer vs. hacker" panel
debate? Formally published accounts of such things, when they exist,
never quite capture the flavor (not to mention many of the important
points) of the real thing. Many thanks.
------------------------------
Date: 17 Sep 1980 0158-PDT (Wednesday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: electronic banking
I wonder what real use (other than to check balances) direct dial-up
access into your bank's system would really provide? Sounds like
another great avenue for "unauthorized transactions" that the banks
can legally force you to make good, even when you know nothing about
them. Direct access even neutralizes the minimal protection that a
physical bank card provides. If more banks provided direct funds
transfer to/from savings and loans, there would be some minor value,
but since your home system can't accept deposits or dispense cash (at
least mine can't!), useful applications still require presence at the
public teller.
------------------------------
Date: 16 Sep 1980 2355-PDT
From: ROODE at SRI-KL (David Roode)
Subject: Electronic Teller
I suggest complaining in writing and requiring a written explanation
of the unauthorized charges. Some federal law recently took effect
making all electronic funds transfer (EFT) subject to special require-
ments for disclosure and Error Resolution procedures and Unauthorized
Charges. The I.D. procedures are so loose I can't see how they can
operate other than the phone company does with all the fake 3rd party
billing ("bill-to") calls -- in the favor of the complainer. I believe
the Federal Reserve Board is the Agency with jurisdiction and a
complaint to them should help. The machines must save the banks big
money in terms of human tellers, so they can afford to self-insure.
UCB isn't that small a bank either -- they are statewide, and have 8
offices listed in the Palo Alto area phone book alone. Their ATM
service is statewide. I believe Crocker Bank's ATM network is
somewhat bigger. Bank of America also has a large ATM network in
Santa Clara County with some 30 machines, and Wells Fargo has a Bay
Area network with 44 locations. I have been told that ATM's lagged
in California because so much of the banking is via Banks with many
(100+) branches so diffuse geographically.
------------------------------
KLH@MIT-AI 09/17/80 21:30:07
Re: CA PUC (OII 70) - Text of OII 70, hearing schedule
Today I'm including today most of the PUC's OII 70 announcement.
Again, this is a bit lengthy, and somewhat stuffy to boot, but is
good background. Skim it, noting IN PARTICULAR the list of issues.
Note also the procedural details, since they have been an obstacle
to some people who wanted to send input.
----------------------------
BEFORE THE PUBLIC UTLITIES COMMISSION OF THE STATE OF CALIFORNIA
Investigation on the Commission's )
own motion to provide for a pro- )
gram for the furnishing of tele- ) OII No. 70
communications devices to the )
deaf and severely hearing impaired )
to be implemented by each California )
telephone utility. )
----------------------------------------
ORDER INSTITUTING INVESTIGATION
In 1979, the California Legislature enacted SB 597 (signed by
the Governor on September 28, 1979), which included among other
provisions a new Section 2831 of the Public Utilities Code providing
for a program to furnish deaf and severely hearing impaired telephone
subscribers with a device for communications. Section 2831, which
became effective on January 1, 1980, reads as follows:
2831. The commission shall design and implement a program
whereby each telephone corporation shall provide a telecom-
munications device capable of servicing the needs of the deaf
or severely hearing impaired, together with a single party
line, at no charge additional to the basic exchange rate, to
any subscriber who is certified as deaf or severely hearing
impaired by a licensed physician, audiologist, or a qualified
state agency. The commission shall phase-in this program, on
a geographical basis, over a four-year period ending on
January 1, 1984. The commission shall establish a rate
recovery mechanism to allow telephone corporations to
recover costs as they are incurred under this section.
To carry out the mandate of the Legislature, as set forth in
Section 2831 of the Public Utilities Code, the Commission hereby
opens this Order Instituting Investigation to determine what services
and equipment are required to meet the needs of the deaf community,
what rules and conditions should be applied to the provision of such
services and equipment, what programs of implementation should be
established, and what rate recovery mechanism should be authorized
to recover costs of the program. Each telephone utility in the state
is named a respondent herein, as set forth in Appendix A. Other
interested parties to be sent copies of this order are listed in
Appendix B.
Prior to the issuance of this order of investigation, the
Commission staff has elicited the views of the telephone utilities
and organizations of the deaf as to their views on how the program
should be implemented. Based upon responses by those parties and
upon the staff's review of the problems, we have developed the
following list of issues which must be addressed in this proceeding.
ISSUES:
1. What type of equipment is required to provide a telecommuni-
cations device capable of servicing the needs of the deaf or severely
hearing impaired?
2. What will be the cost of providing such equipment?
3. How should the program of providing such equipment be financed?
4. How should the investment in nonrevenue producing equipment be
treated in the utility rate bases?
5. What method of rate recovery should be adopted to allow tele-
phone corporations to recover costs as they are incurred as required
under section 2831 of the Public Utilities Code?
6. What are the best estimates of the deaf population requiring
special telecommunications devices?
7. What is the availability of equipment to meet the needs of
the deaf?
8. In pursuing the program to meet the needs of the deaf, will
there be equipment available to meet the needs of the non-deaf in
communicating with the deaf?
9. What types of basic equipment should be provided without
additional charge to the certified deaf?
10. What types of optional equipment or features should be
furnished to the deaf at an additional charge?
11. How should rates and charges for optional equipment or
features be determined?
12. Should a deposit or other security provision be required to
protect the investment in equipment furnished the deaf?
13. Does the statutory requirement of a single party at no charge
in addition to the basic exchange rate contemplate that the basic
exchange rate includes suburban mileage charges outside a base rate
area?
14. How will the requirement for single party service be met in
rural areas where only party line service is available?
15. What are the requirements for deaf telecommunications devices
in the areas served by independent telephone companies, and how should
such requirements be determined?
16. How can a deaf equipment program be provided in the service
areas of the smaller independent telephone companies?
17. Should an equipment and revenue recovery pooling arrangement
be developed for the smaller independent telephone companies?
18. Should any telephone companies be exempted from the program?
19. What are the requirements for customer instruction in the use
of the deaf telecommunications devices?
20. Should basic telecommunications devices be furnished to the
certified speech impaired on the same basis as furnished to the
certified deaf?
21. Should a pilot project be undertaken at an early date to test
distribution plans?
22. What does the statutory provision for phasing in the program
on a geographical basis contemplate?
23. What priorities should be assigned in determining schedule of
geographical area implementation, and should priorities be assigned
to categories of persons within geographical areas?
24. What will be the continuing additional requirement for
providing devices in the future after all needs are initially met
in the basic four-year program?
25. What will be the costs of maintenance of basic telecommuni-
cations devices for the deaf, and what are the rate recovery methods
to offset such costs?
26. What is the estimated useful life of equipment for the deaf,
and what provisions for replacement at the end of the useful life
should be made?
Evidentiary hearings in this proceeding will be scheduled at
an early date to consider the above issues. Based upon the record
therein, the Commission will issue such order or orders as required
to implement a program for the provision of telecommunications devices
to the deaf in accordance with the requirements of Public Utilities
Code Section 2831. Parties desiring to address the issues herein
should submit to the Communications Division, Legal Division, and
Administrative Law Judge Thomas E. Daly, a statement within thirty
days after the issuance date of this order, listing the issues they
wish to address, a brief summary of their position on each such
issue, and a list of the witnesses they plan to present. Parties
having substantive testimony will be required to present their
testimony in written form to the Commission and to interested
parties at least ten days in advance of scheduled hearings.
As a basis for resolving the economic issues of this proceeding,
the Commission staff has prepared a proposed plan for a Deaf Equipment
Acquisition Fund which is attached hereto as Appendix C. In their
direct testimony, parties are invited to comment upon the staff's
proposal. Where parties disagree with the staff's proposal, they
should present such alternate proposals as they deem appropriate.
In accordance with the foregoing and good cause appearing
therefore,
IT IS ORDERED that:
1. An investigation is hereby instituted on the Commission's own
motion to determine the requirements for telecommunications devices
for the deaf and hearing impaired and to provide for a program whereby
telephone corporations will provide such devices to the certified deaf
in accordance with the provisions of Section 2831 of the California
Public Utilities Code.
2. Each of the telephone utilities listed on the attached Appendix
A is hereby made a respondent in this proceeding.
3. A prehearing conference in this proceeding will be held before
Administrative Law Judge Thomas E. Daly and/or Commissioner Gravelle
on Friday, June 20, 1980, in the Commission's Courtroom, 350
McAllister Street, San Francisco. At that prehearing conference
procedural issues will be taken up, appearances will be received
from parties desiring to participate in this investigation, and
further hearing dates will be scheduled and announced. Those
wishing to participate and learn of the hearing schedule should
attend the prehearing conference.
4. The Executive Director is directed to cause a certified copy of
this order to be served forthwith upon each of the respondents herein
listed in Appendix A, and a copy of this order to be mailed to each of
the parties listed in Appendix B, and thereafter to cause appropriate
notice of hearings to be mailed at least ten (10) days before such
hearing.
Dated May 6, 1980, at San Francisco, California.
John E. Bryson
President
Vernon L. Sturgeon
Richard D. Gravelle
Claire T. Dedrick
Leonard M. Grimes, Jr.
Commissioners
----------------------------
Appendix A:
A list of 25 (yes, 25!) telephone companies in California.
Appendix B:
A list of 65 "interested parties" including lawyers,
deaf/handicapped service organizations, manufacturers, associations,
and apparent randoms. This isn't as inclusive as it seems, since for
one thing all are in California ONLY, and many of the most interested
parties were neither listed nor notified (e.g. HEW, FCC, SRI, TDI,
NAD, non-CA manufacturers, etc). Some managed to get wind of it,
others didn't. However, it's a tough (perhaps even hopeless) job
to make such a list.
Appendix C:
The PUC staff's proposal for a "Deaf Equipment Acquisition Fund".
[If you notice the acronym, you might well suspect the PUC has a sense
of humor...] It runs to 6 pages. I'm only including the surcharge
table presented; if people want the gory details I could probably send
them.
Computation of Monthly Telephone Surcharge
for DEAF
Item Estimated quantity
1. Number of devices for certified deaf, end 1983 200,000
2. Cost per device $600
3. Total cost of devices (1) x (2) $120,000,000
4. Statewide number of main telephones 13,200,000
5. Dollars required per telephone (3) / (4) $9.09
6. Dollars per year (5) / 3 years $3.03
7. Dollars per month (6) / 12 months .2525
8. Monthly surcharge per main or equivalent main station USE .25
The basic idea is for the costs to be directly supported by
the subscriber base. In particular, the $.25 or whatever surcharge
would be listed separately on the monthly phone bill and identified
specifically as the "Deaf Equipment Acquisition Fund surcharge".
(Understandably this bothers many deaf people, who are worried about
a consumer backlash directed at them -- they want the identification
removed, unless the phone company agrees to identify all the OTHER
components of the basic phone bill!)
I'm getting ahead of myself... to continue, here's the upshot of the
prehearing:
--------------------------------------------------
ADMINISTRATIVE LAW JUDGE'S RULING
As a result of the prehearing conference held in San Francisco on
June 20, 1980, the following procedure will be followed in the above
proceeding:
(1) Public hearings will be held at San Francisco on August 25, 26,
27, 28, 29, and September 2, 3, 4, and 5, 1980, for the purpose
of hearing from those individuals and organizations who wish to
make comments and suggestions on the proposed program. Those
wishing to make a presentation will be required to submit
prepared statements and to serve copies thereof, along with
copies of all exhibits to be offered, upon all parties that
made appearances at the prehearing conference. Such service
shall be made at least ten days prior to the August 25th
hearing. The names and addresses of the parties are set forth
in the attached certificate of service.
To facilitate this phase of the hearing and as an accomodation
to those making a presentation a time schedule will be prepared
and mailed to all parties prior to the August 25th hearing.
(2) Following the initial hearing the staff will prepare a proposed
program incorporating a synthesis of the recommendations and
suggestions offered along with recommendations of its own.
Further public hearings will be held at San Francisco on October
6 and 7, 1980, for the purpose of receiving the staff's presen-
tation. The staff will serve copies of its presentation upon all
parties of record at least ten days prior to the October 6th
hearing.
(3) Those wishing to comment upon the staff's proposed program will
be afforded the opportunity at hearings to be held on October
16 and 17, 1980, at San Francisco. Such comments may be made
orally. It will not be necessary to have them in the form of
prepared statements.
Dated June 26, 1980, at San Francisco, California.
Thomas E. Daly
Administrative Law Judge
------------------------------
I don't currently plan to transcribe any more documents; the worst
is over, and future messages will be shorter and meatier. I'll be
describing the UNWRITTEN assumptions the PUC is making, some of
the issues that the prepared statements focus on, and what sort of
testimony was given at the hearings just concluded. I should warn
you, though, that some of it will doubtless elevate your blood
pressure.
--Ken
------------------------------
End of HUMAN-NETS Digest
************************
∂18-Sep-80 0900 JMC*
print freedo
∂18-Sep-80 0920 CSD.GOLUB at SU-SCORE Re: Faculty Meeting
Date: 18 Sep 1980 0918-PDT
From: CSD.GOLUB at SU-SCORE
Subject: Re: Faculty Meeting
To: CSD.JEANIE at SU-SCORE, buchanan at SUMEX-AIM, rwf at SU-AI,
or.dantzig at SU-SCORE, ghg at SU-AI, jgj at SU-AI, dek at SU-AI,
jmc at SU-AI, ejm at SU-AI, zm at SU-AI, csd.ullman at SU-SCORE
cc: csd.bscott at SU-SCORE, CSD.GOLUB at SU-SCORE
In-Reply-To: Your message of 17-Sep-80 1559-PDT
Jeanie: Gene is again on sabbatical as you no doubt know. He will
return to Stanford on December 19.
Janet
-------
∂18-Sep-80 1141 ROB Pies and Wires
∂17-Sep-80 1612 JMC
Bob Floyd noticed some of the famous pie on a stair rail at floor 3.5.
--------------------
Oop. I better clean it up! Thanks for Bringing It To My Attention.
- Rob
PS: If a news wire line seems dead, you may be able to resurrect it by
typing "TTY EXIST nn", where nn is 16 for AP wire, and 17 for NYT.
∂18-Sep-80 1630 FFL
To: JMC, FFL
A gentleman by the name of Om Vites, who said he is a visitor in CSD
wishes to see you Monday afternoon. He will stop by to see if you are
available.
∂19-Sep-80 0014 100 : John Kelley (kspace) via AMES-TIP
I understand that you'd like to see a copy of the business plan for the
Text Network Program. It's being revised, but I'll try to bring a copy
by on Monday. I'd like to ask you some more questions about how the
program will be coordinated witnh the CSD wand how research in connection
with the program could be organized.
Thanks
John Kelley
∂19-Sep-80 0103 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #79
Date: 19 SEP 1980 0354-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #79
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Friday, 19 Sept 1980 Volume 2 : Issue 79
Today's Topics: ACM '80 Panel, Electronic Banking - Possible Services,
Home Info Retrieval - Classified Ads & Yellow Pages,
Telephone Services - GenTel vs PacTel,
Computers and the Deaf - CA Bill on Free TDD's
----------------------------------------------------------------------
Date: 18 September 1980 1352-EDT (Thursday)
From: Paul.Hilfinger at CMU-10A (C410PH01)
Subject: Panel at ACM'80
According to my preliminary schedule, there will be a panel discussion
entitled "The Computerized Destruction of Western Civilization" at the
ACM'80 National Conference on Wednesday morning, 29 October. SIGCAS
is the sponsor.
One would naturally suppose that "The Computerized Destruction of Western
Civilization" panel was a spoof, but considering that it's SIGCAS (Special
Interest Group on Computers and Society (of ACM)), it
probably isn't. SIGCAS impresses me as one of the most blindly technologically
reactionary conglomeration of know-nothings, I have ever encountered,
based on all the issues of their newletter available a few months ago
in the Stanford library. Their specialty is horrific "what if" scenarios
about data banks, electronic funds transfer and other computer applications.
Whether it is from lack of imagination or merely ideology, they don't see
any of the positive potential of computerization and the enormous costs in
delay implementing it. For example, if we can get rid of cash, we can get
rid of crime for cash. A robber can't hold you up and demand that you
transfer money to his account - not unless he can prevent you from talking
and conceal the fact of your demise. By helping delay EFT, SIGCAS
assumes responsibility for the deaths of perhaps several thousand people
per year. Another example is that the Freedom of Information Act could
be made effective by requiring that all such information be on computer
file. Indeed the discussion of telephone company practices in HUMAN-NETS
would be much better informed if the Public Utilities Commissions were
required to keep their decisions and the tariffs they allow on file. (I
apologize having the same urge to impose a duty on government agencies
that others have to impose duties on the telephone companies).
Is anyone out there interested in forming an organization that would
devise and press for useful applications of computer technology in
society.
------------------------------
Date: 18 Sep 1980 1009-PDT
From: Zellich at OFFICE-1
Subject: ATM Lag
The note on ATM lag in California is interesting. Here in Missouri we
have an even greater lag: Branch banks are illegal in MO, and remote
ATM's have been declared to be branches!
--Rich
------------------------------
Date: 18 Sep 1980 1812-PDT
From: Jim McGrath <JPM at SU-AI>
Subject: electronic banking
In response to Lauren's message: direct dialing means that you can
conduct any activity you normally could have done except those that
require physical contact. Obviously you will never be able to get
money (cash) from your account, nor deposit cash or physical checks.
But you can authorize electronic transfers of funds.
I would love being able to pay my phone bills via computer. Most
utility payments could be managed this way. Eventually, all payments
to companies that can afford to be hooked up can be managed that way.
One can also transfer money between accounts. I could transfer
money from savings to checking to cover a new check, or to my charge
accounts to cover a new purchase.
Three problems: to get users interested in using it (having things
done by a home computer-network hookup beats the telephone, and
probably is sufficient), to get the companies interested (starting
with utilities, since they have the largest number of customers and
are fairly centralized), and security.
Security is, as always, the big technical problem. But right now a
bank will transfer money for you over the phone with no more identi-
fication than your birthday and mother's maiden name. So I do not
think security is a big problem, unless the criminal attacks the bank
records for large customers. But that simply means that people with
large accounts will pay (in a monthly service charge) more for the
implementation of tighter security for THEIR accounts.
Details depend upon the system used (are transactions conducted
through a net? direct dial up? how are the records organized?,
etc...)
Jim
------------------------------
Date: 17 Sep 1980 0158-PDT (Wednesday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: electronic bill paying service
I recently signed up for an electronic bill paying service. While
the paperwork is still being processed, here is how it works. I
have established an account at a local savings and loan (Allstate
Savings here in L.A.) This special "bill pay" account draws 5%
interest. Allstate has a (long) published list of firms that have
agreed to accept payments via the "bill pay" system (and the list
is constantly growing). It includes most of my "regular" bills,
particularly utilities and VISA. I have specified a list of
merchants whom I want to pay via this system. Whenever I have a
bill from one of these merchants/entities, I dialup (on an INWATS
number) their voice-response system. I can then very simply
touch-tone in merchant numbers and amounts, and the checks will
be mailed within 24 hours. Much to my amazement, they seem to
have done just about everything right, and I feel fairly secure
with the arrangement. The interesting point is that in addition
to drawing 5% on the money (slightly less than a conventional
account), the cost is very minimal: for accounts over $1000,
there is no charge for use of the system. For accounts under
$1000, there is a 10 cent fee per bill. This still is less
than the cost of a stamp, and overall I should do alot better
than having the same money in a non-interest-paying checking
account. Since I have pre-specified the allowable merchants
to pay, there are virtually no opportunities for fraud even
if some random person discovered my account/passnumber pair.
Any overpayments to these merchants would quickly show up on
my next bill. Of course, the bank sends out a monthly statement
detailing all transactions.
--Lauren--
------------------------------
Date: 17 September 1980 04:29-EDT
From: Robert W. Kerns <RWK at MIT-MC>
Subject: Yellow pages terminals
On the other hand, properly marketed, businesses might just love
electronic yellow pages due to possibly WIDER distribution. They
could be listed under MORE catagories, and people from a WIDER AREA
could examine their listing (long distance yellow page directory...).
They could include MORE information with their listing, and update it
MORE frequently. The problem with splitting directories is that the
were getting 1/2 as much exposure for the same cost. I don't blame
them for being upset. As a subscriber, I'd be upset to have 1/2 the
selection in my yellow pages!
------------------------------
Date: 16 Sep 1980 0936-PDT
From: Weinstock at SRI-KL (Charles B. Weinstock)
Subject: Multiple Yellow Pages
Chicago has had the twin yellow pages (Commercial and Consumer) for
at least 4 years, and I understand is also putting out regional books
now too. In the case of Chicago, the Red Book (as they call the yellow
pages because the cover is red) a single volume was unwieldly (the
average apartment dweller would have trouble storing it).
------------------------------
Date: 17 September 1980 06:17-EDT
From: Frank J. Wancho <FJW at MIT-MC>
Subject: yellow pages
Then there's Ft. Lauderdale with no yellow pages (at least that I
found), but a separate section of the white pages, in the same
format as residential listing, i.e., a residential only white
pages and a business only white pages - not a bad idea...
Houston has had A-L, M-Z yellow pages for several years, now, but
I don't live there any more, thank goodness!
--Frank
------------------------------
Date: 16 Sep 1980 at 1036-PDT
From: obrien at Rand-Unix (Mike O'Brien)
Subject: Different phone companies
I wish to point out that the messages Lauren and I have been
putting forth about PacTel vs. GenTel are not random storytelling, but
make a valid point: regulation of something as big as the phone system
is essentially a natural process, and not subject to rule-making. It
is much closer to the SEC or the Pentagon than to the workings of a
normal legislative body in that there is essentially no free choice
about what goes on at the lowest levels. It just happens.
Supposedly the PUC ensures that everybody gets the same level of
phone service. This mechanism replaces the notion of competition, as
it obtains in other industries (like cars). It doesn't work. At the
lowest level GenTel is so notorious that there is actually some level
of consumer revolt. It can't work of course, the remarkable thing is
that it is there. Who would ever dream of taking on the phone company?
These folks would...
Between the glitter of Santa Monica and the ghetto of Venice
lies the community of Ocean Park, which is best described as funky.
It resembles a university town without the university. Recently
the Ocean Park Community Organization (OP is actually part of Santa
Monica but doesn't act like it) began an organized revolt against
GenTel, and has put several VP's through the grinder at community
meetings. The last I heard, they were actually considering leasing
outside lines into a PBX which would then act as the community's
wholesale replacement for GenTel. Of course GenTel would then
(presumably) rip out phones, poles, lines, CO, and all, but of
course the regulatory picture is so murky that goodness knows
what would actually happen.
Doubtless this is merely a scare tactic to generate some
unfavorable publicity for GenTel, but it does indicate that these
people are DESPERATE! So much for PUC's...
------------------------------
DGSHAP@MIT-AI 09/17/80 14:49:39
Re: Krieger Krieger Boggin and Krieger
The fact that personal benefit is such a strong motivation for
legislation is rude enough, but I find the avarice with which
this particular plot was carried out to be truly objectionable.
And I am sure that this type of situation is a mainstay (a
necessary corollary?) of the legislative mechanism. Ie, every
time a bill is passed, someone (or some corporation(s)) who
stood to gain substantial financial benefit from the bill
lobbied for it, and some legislator advanced his or her own
personal causes (by gaining clout, reputation, or personal
satisfaction) by backing the bill.
This is all logical enough. What is bothersome here is the
conspicuous absence of the the purer motivations (altruism, the
desire for social benefit, truth, justice and the american way).
What we do see is a demonstration of original sin; the unopposed
application of somewhat seedy political tricks in pursuit of the
satisfaction of greed.
Dan
------------------------------
Date: Wed, 17 Sep 80 10:46-EDT
From: Breunig at DNGC
The article from Oakland Tribune (Gannet news service) about
the legislation benefitting SSI is quite interesting. It
brings to mind an editorial I wrote in the TDI publication
in November, 1972 in another context..........
Editorial; Snakes in the Garden of Eden
Teletypewriters are one of the most exciting and successful programs
for deaf people, their families and their organizations in many
years. Not only are deaf people able to communicate with each other
unassisted over the telephone lines but a group of deaf agents and
servicemen have found a happy opportunity to serve the fellow members
of their deaf community through reconditioning and installing
equipment.
It is with a feeling of deep regret then, that we have received a
number of letters complaing about mudslinging in the "war of the
coupler makers." Competition is the name of the game, and it is
acceptable practice for manufacturers to praise and promote their
own products. But this "war" appears to have gone beyond the
bounds of gentlemanly courtesy. There is far too much badmouthing
of a competitor's product.
Personally, we no longer pay too much attention to letters mailed out
to agents and members by one side or the other, letters which seem to
contain innuendos and half-truths, if not outright falsehoods. From
our point of view the letters accomplish nothing but ill-will and,
furthermore, seem to us to be a terrible waste of time and energy
which could more profitably be used to develop the growth of the
network. It is high time that such unethical and un-businesslike
tactics be stopped forthwith.
H. Latham Breunig
So it seems that the snakes are still in the Garden of Eden. I
have never approved of the company in discussion nor its personnel
from day one. Its not surprising that PACTEL and GTE have rejected
the product. The complaints about it have been legion.
Latham Breunig
------------------------------
Date: 18 Sep 1980 0333-PDT (Thursday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: deaf equipment surcharge
It seems to me that the $.25 surcharge per subscriber is really a tax,
much the same as the 911 implementation tax in California or the other
federal and city taxes. As such, I feel it should be listed on all
phone bills the same as any other tax. Trying to prevent a consumer
backlash by "hiding" the cost is not a good idea.
Something about this plan bothers me, but I cannot put my finger on
it. Somehow I get the impression that there is a vaguely unfair aspect
to saddling the state's telephone subscribers with what is essentially
a special interest group cost, regardless of how deserving that group
may be. I realize I risk the wrath of the hearing impaired community
that reads HUMAN-NETS, but I generally speak my mind on such issues,
and this is no exception.
By the way, do they plan to add this charge to ALL phone bills
including private leased circuits?
--Lauren--
------------------------------
End of HUMAN-NETS Digest
************************
∂19-Sep-80 0819 Kanerva at SUMEX-AIM DEC-20 business plan
Date: 19 Sep 1980 0818-PDT
From: Kanerva at SUMEX-AIM
Subject: DEC-20 business plan
To: JMC at SAIL
Jon, Jeannie, and I are putting together a plan, a draft of which
Ralph has seen. Ralph asked for a copy for you. We are trying to
get an updated draft out today, and I would like to bring it to you
and talk with you about it. The plan is on the agenda for a meeting
of Ed Shaw with the vice-presidents on Monday the 29th.
We should get our draft together this afternoon. When and where
might I find you. If I hear nothing from you, I will leave the
draft in your mail slot in Margaret Jacks. - Pentti Kanerva
-------
∂19-Sep-80 1123 FFL
To: JMC, FFL
Dr. Clara Torda stopped by to see you. She will try again on Monday
afternoon or another time if you have a suggestion for a better apppointment.
Her number is 7-3774.
∂19-Sep-80 1531 FFL
To: JMC, FFL
mail jmc,ffl
John Kelley would like to speak with you Monday afternoon, or some other
convenient time, about the text network program. 9 328 6077.
∂19-Sep-80 2108 RMS at MIT-AI (Richard M. Stallman) buyer for emacs
Date: 20 SEP 1980 0007-EDT
From: RMS at MIT-AI (Richard M. Stallman)
Subject: buyer for emacs
To: JMC at SU-AI
Date: 17 Sep 1980 1558-PDT
From: John McCarthy <JMC at SU-AI>
Dick Cogan 216 371 1306 was told to call you about buying the right
to use EMACS on a Dec-20 at I forget what university.
How did you suggest that he call me?
At the moment, I have no way to call him without quite a
handful of change.
He was told by someone at M.I.T. that he could reach you at my number
and that you had something to do with it. Why not call him collect?
∂20-Sep-80 0146 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #80
Date: 20 SEP 1980 0435-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #80
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Saturday, 20 Sept 1980 Volume 2 : Issue 80
Today's Topics: Computers and the Deaf - CA Bill on Free TDD's
----------------------------------------------------------------------
Date: Fri, 19 Sep 80 06:38-EDT
From: BMStrazz at DNGC
Subject: surcharges for the deaf
I wish to reply to Lauren's surcharges comments. If it bothers
him grievously that twenty five cent surcharge is being proposed
in California, then please rest assured that it bothers grievously,
the deaf Marylanders to have a ten cent surcharge on our monthly
telephone bill to subsidize the 911 number. This 911 number is a
voice number, and is utterly useless to the deaf unless they are
lucky enough to have hearing children around in state of emergency
to call up this number.
If tax payers protest using their tax money to finance special
groups, and if telephone users protest paying surcharges to
finance special groups, then who is going to pay for it?
BMStrazz
------------------------------
Date: 19 Sep 1980 0827-PDT
From: LEWIS at SRI-KL (Bil Lewis)
I think that I know what Lauren is refering to WRT TTYs for the
deaf. Being provided a TTY for no cost above normal telephone rates
is a privilege. To have society as a whole underwrite the costs of a
minority of any sort is a privilege, and not a right as it has been
often regarded as. I believe that this latter feeling has generated
some ill-will from those who don't think of themselves as being so
privileged (?).
People like Ken, who is deaf, and myself, who is blind as a
bat, would have simply perished in a more primitive society. I
think that we are fortunate to have been able to survive to become
the highly skilled computer personnel that we are. Further, the
society is fortunate to have us survive, and the cost of providing
special services is certainly returned in full.
My conclusion is that special sevices are a privilege for
those who receive them, but that they are a necessary economy for
the society to provide. I would have the TTY costs show on the
bill, along with everything else including rural subsidies, or
not at all.
regards, bil
------------------------------
Date: 19 SEP 1980 0451-PDT
From: RAHE at USC-ECL
Subject: D-E-A-F tax
I would certainly agree with Lauren that the 'surcharge' is a tax
and should appear on the bill as such, along with any other taxes.
The earlier notes had said that it shouldn't be included unless all
costs that that make up the bill are itemized. I would point out
that currently no costs are itemized on bills. The addition of the
'surcharge' as an item in the bill is the listing of a TAX, not a
COST.
------------------------------
KLH@MIT-AI 09/18/80 23:55:45 Re: CA PUC (3) - Hearing details
Several "interested parties" submitted written statements and
appeared at the Aug 25-29, Sep 2-5 hearings. One thing people should
understand (and this surprised me at first) was that the hearings were
VERY much like a courtroom, and officially, only material which has
been formally submitted as evidence, or heard in court as testimony,
is allowed to be put "on record" for consideration by the PUC.
In other words, unless you have a lawyer, you may not have much
of a voice. I was really amazed. Just like in the movies, you had
these smiley-smiley pinstripes jumping up and shouting "Objection"
and whispering with their clients and asking the most inane cross-
examination questions. They weren't all like that, of course; some
were very helpful. The main reason for their presence seemed to be
that they knew the procedural ropes, and knew what they could get
away with, and thus could advise their clients on how to best
present their evidence.
The judge simply stepped through one "witness" at a time. Some
parties had two or more witnesses; others had only one. Some called
in outside people, others represented themselves. But in order to
have anyone on the witness stand, you had to register your intentions
in advance -- I'm not sure how far in advance, but they certainly
seemed to prefer something like 10 days.
The procedure for each "witness" was the same. After the
swearing-in, the judge would ask "is this written testimony yours? Is
it accurate to the best of your knowledge?" etc. to which the witness
would reply yes-yes-yes-etc. Then the judge asks "Is there anything
you wish to add?". This is the kicker: if there was anything that you
DIDN'T include in your written testimony (distributed several days
ago, remember, before you have any idea what other parties are going
to say), this is the ONLY chance you will EVER have to say it! That's
right -- no second chance. After you finish answering this (and in
most cases, people just said "no"), the judge declares the witness
available for cross-examination. This means that each interested
party, in a set order, is given the opportunity of asking questions.
Two tricks here: to be allowed to ask questions, it's not enough to
just walk in the room, as a member of the great unwashed public. You
have to register as an "interested party" beforehand, otherwise you
keep your trap shut. Furthermore, asking questions is all you can
do; if you disagree with something the witness claimed, you can't
just say so, as in a debate -- you have to play Perry Mason and
butcher the rules of clear speaking and basically talk out of the
side of your mouth in order to make a point.
The reason I describe these procedures in detail is to explain
why there is some information that the PUC will not see "on record",
and why the hearings are not very accessible to the public. Most of
the interesting things I learned came from people in the halls, not
in the courtroom. That's true for many other tribal functions of
course, but dammit, it can be frustrating. Doubtless it ensures that
antagonistic parties interact in a nice orderly manner, but it seems
to suppress many datums which otherwise would been brought to light.
Remember, you have lawyers in there who are more interested in
silencing the testimony of opposing parties than they are in
discovering the true facts (particularly if the facts are
technical in nature, since not many understand them).
By the way, to follow the hearings requires a bunch of money.
If you can't attend physically (take time off from work, fly to SF,
etc.) then you can, of course, ask for transcripts from the court
recorder. However, they run $.90 per page, and the average hour's
worth of testimony is estimated at $20 to $22. That's over a hundred
bucks a day, for two weeks. You pays your money and takes your
chances.
Enough of that.
------------------------------
Here are the parties who were scheduled for the hearing, in
the order of appearance:
(1) The Pacific Telephone and Telegraph Company [Ma Bell!]
(2) Continental Telephone Company of California
(3) Citizens Utilities Company of California
(4) California Independent Telephone Association
(5) The Siskiyou Telephone Company, et al. [??]
(6) Dept of Rehabilitation, State of California
(7) DCARA and Bay Area Center for Law and the Deaf
(8) California Association of the Deaf, Inc.
(9) Deaf Informed Communications Resource Center, Inc.
(10) Krown Research [TDD mfr]
(11) Micon Industries [TDD mfr]
(12) Northern Communications
(13) Northern Telecom Limited
(14) General Telephone Company of California
(15) Novation, Inc. [potential TDD mfr]
(16) Specialized Systems, Inc. [TDD mfr]
(17) SRI International
(18) Plantronics, Inc. [potential TDD mfr]
(19) Telecommunications for the Deaf, Inc.
[TDD means Telecommunications Device for the Deaf - a more general
term than TTY.]
This doesn't represent the exact schedule since there were a few
last-minute shuffles and additions. Unfortunately I only have direct
knowledge of the last 2 days, and am still waiting for second-hand
transcripts of the earlier sessions. For some reason, despite the
judge's ruling, several of the parties either did not prepare any
statement or did not distribute them. That's a shame since not only
are the statements "free", they're very informative and usually get
the goods across in a coherent (non-cross-exam) fashion.
Now, about the unwritten PUC assumptions. When we (at SRI)
talked with a PUC "Utilities Engineer" over the phone, he explained
in more detail what the PUC was trying to do. They want to provide
the MINIMUM possible; their wording is "basic black phone equivalent".
That is, they don't want the equipment to furnish any other services
than would normally be available to a hearing person with a typical
black dial-type telephone; the idea is to provide EQUAL access, and
not a smidgin more. Anyway, one of the corollaries is that (hold on
to your seat now) ANYTHING involving computers is a LUXURY.
There have been really incredible arguments going on about
what constitutes "equal access" and the "minimum"; I would say
that a majority of debate has simply centered on the issue of "what
equipment constitutes 'basic service'"? In fact, the manufacturers
almost turned the hearings into a trade show, each of them smoothly
promoting their own products while rabidly tearing into those of
others. Two of the most significant issues were whether ASCII and/or
hard-copy capability should be provided; interestingly enough, the
manufacturers and consumers almost uniformly opposed each other on
these two -- the mfrs say "no", the users say "yes".
Other factors such as rate recovery plans are mostly between the
PUC and the phone companies and haven't generated much response from
any side; at least not that I know of. The statement from General
Telephone was one of the most reasonable I saw; in particular they
said:
"General believes that the mandate of the Legislature as expressed
in Section 2831 of the Public Utilities Code may best be carried
out by the establishment of a separate body or committee made up
of the telephone companies and/or organizations whose purpose is
to work with the deaf and severely hearing impaired."
They claim that this would result in more efficient and effective
distribution, and their arguments make sense. Pacific Telephone, of
course, simply assumed that they'd be doing everything themselves,
with a cold nod acknowledging the existence of "smaller independents".
They produced by far the most detailed breakdown of costs, with an
interesting peek into their contract/lease cost planning system. For
example, there are about 32 lines worth of catgeories and totals, with
a few pages of explanation, that demonstrate how they arrive at final
total costs for the TDD equipment they currently provide. It's kind
of hard to summarize this stuff in any reasonable way, and I don't
know if people are interested in those aspects, but I'll pluck &
present whatever info seems needed.
Oops, am running late and want to go catch Shogun tonight; have
to cut this short. Tomorrow will continue with the equipment speci-
fication issue, demographics, and anything else that pops up. In the
meantime, think about it... if you couldn't use your voice or ears
with the telephone, what device or system would you consider minimal
for interpersonal telephonic communication? Don't forget, you're
paying for it one way or another.
--Ken
------------------------------
Date: Wed, 17 Sep 80 08:10-EDT
From: BMStrazz at DNGC
Subject: Between the Lines
I wish to add to KLH's Calif PUC piece. This is a very sensitive
issue, and thus I hope the facts being presented are as objective
as can be. The Specialized Systems, Inc story has made the front
pages of the two major, national deaf publications, so to any
well-read deaf people, this episode is no secret.
What is the issue is not the dealings supposedly behind the scene, but
the assertion of some of the California deaf leaders to the effect
that the deaf "do not need" a TDD of a hard copy version. Citing equal
accessibility rights of the deaf, the reasoning is that if the hearing
people (non-deaf) do not need hard copy versions of their voice calls,
then the deaf also do not need this hard copy privilege. This is a
crucial issue since the corporation (SSI) mentioned in the newspapers
has had phased out the hard copy printer production, according to
statements made in the SEC prospectus. The SSI products are chiefly
soft copy as contrasted to the hard copy equipment turned out by
Porta Printer, which is one of SSI competitors.
It was the concern of the TDI that if the PUC goes by the line that
the deaf do not need copy of their TTY conversations, then non-hard
copy companies could reap hugely at the expense of hard copy
companies.
Statistical studies indicate that a great majority of the deaf people
prefer hard copy, and the alleged statement of one California deaf
leader contradicting the wishes of a great majority of the deaf is
serious.
And this is why TDI sent Joe Slotnick the vice president (of Marina
Del Rey, California) to Sacramento hearings to restate the hard copy
sentiments of the deaf.
This, in essence, is the gist.
Barry Strassler
------------------------------
End of HUMAN-NETS Digest
************************
∂20-Sep-80 1133 Kanerva at SUMEX-AIM Business plan for a text system
Date: 20 Sep 1980 1133-PDT
From: Kanerva at SUMEX-AIM
To: JMC at SAIL
cc: REG at SAIL, Sandelin, Siegman
Subject: Business plan for a text system
I have left for you in Margaret Jacks the drafts of a DEC20
business plan and a proposal to DEC that Jon, Jeannie, and I have
been working on in cooperation with Ralph. There is also the agenda
for Ed Shaw's meeting with Kennedy and the vice-presidents on Monday,
September 29.
I have a big favor to ask: that you give your criticism of the
business plan to us before we prepare a copy for the vice-presidents
this Wednesday. I would like to come to talk with you about the plan
Monday or Tuesday. Monday from 2 to 4 is the only time that I could
not make it.
We hope to submit a plan that you would like to see implemented.
As the September 29 meeting could be critical, please consider talking
about the plan to at least some of the vice-presidents before the
meeting. I believe that your doing so would influence the outcome
of the meeting.
- Pentti Kanerva
-------
∂21-Sep-80 0813 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #81
Date: 21 SEP 1980 1056-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #81
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Sunday, 21 Sept 1980 Volume 2 : Issue 81
Today's Topics: Computers and the Deaf - CA Bill on Free TDD's
----------------------------------------------------------------------
Administrivia: Free TDD discussion and pending mail
Recently, I have distributed 4, very long background messages from
Ken Harrenstein <KLH at MIT-AI> discussing the CA bill on free TDD's.
We discussed several ways of distributing the material. We chose to
distribute the material directly because the material needed to go out
as quickly as possible, because each message focusses on a different
aspect of the situation, and because the material is directly relevant
to several past HNT discussions in addition to our recent discussion
of Computers and the Deaf.
Unfortunately, this has meant that a number of messages on a variety
of topics have been delayed. These messages have been stored and will
appear over the next several days along with further discussion of the
CA TDD bill and other new material.
-- Roger <Duffey at MIT-AI>
------------------------------
KLH@MIT-AI 09/21/80 02:20:00 Re: CA PUC (4)
This should be the last background message necessary. The first
one dug up some history of the bill itself; the second contained the
exact PUC documents listing issues and considerations; and the third
described the general atmosphere and restrictions of the hearings
just concluded. In this, the fourth, I will concentrate mostly on
one specific issue of the hearings: the type of equipment required,
with particular respect to the question of ASCII capability.
------------------------------
To understand why this is so important, consider the demographics
for a moment. The PUC asked for estimates on the size of the deaf
population requiring TDD's, and received figures ranging from 43,532
(Pac Tel, 1980 data) to 280,142 (DCARA, 1978 data). Most people
seemed to accept the 200,000 figure the PUC used. At any rate, these
numbers are ONLY relevant to the number of units provided "free";
remember that for every "free" unit, you'll almost certainly have a
market for several additional units -- family, friends, relatives,
businesses, government agencies, institutions... The latter entities
are particularly significant owing to the provisions of recent "equal
access for the handicapped" bills, which for the deaf are usually
taken to mean that TDD access must be possible (plus an interpreter,
if needed). To belabor the point, SSI has directed much of its sales
push to such institutions, and consequently has sold more TDD's in
its few years of existence than all the other TDD manufacturers put
together have EVER sold.
How much is that? Well, estimates for the number of TDD's now in
the US range from 40,000 (low) to 70,000 (high). This means that the
California bill has opened up a guaranteed market which presently can
far exceed that of the entire country, without half trying. And just
wait until other states consider similar legislation -- unless the
California example becomes a disaster, you can bet the legislators
will make it easy for themselves by pretty much emulating the west
coast. No wonder the manufacturers are scrambling to get their
teeth in.
It should be apparent now that whatever particular type of
equipment is selected, it will have great economic impact. The
most important factor, of course, is the STANDARDS which are
selected. Many of us have seen some standards battles and can see
the effects of good and bad standards decisions... ASCII, RS-232C,
etc. And in standards as it is elsewhere, possession is nine tenths
of the law -- it's very difficult to mandate any changes in huge
economic investments.
And this leads to what I consider the most important issue:
whether the equipment should be ASCII compatible. I think it should.
Since that may seem obvious to most Human-Nets people, let me explain
why it is NOT obvious to most of the parties at the PUC hearing.
Earlier Human-Nets messages from Strassler and Marotta have
explained that TDD's use the 5-bit Baudot character code, with the
Weitbrecht modem. This represents an existing investment and a de
facto "standard"; with few exceptions, the manufacturers, phone
companies, and PUC itself were highly critical of the need for
adding ASCII/Bell-103 capability. (Note I said "add", not "replace
with"). The reasons for this are varied. The actual STATED reasons
went something like this:
* The cost of a "free" TDD should be kept at a minimum.
* Adding extra "frills" like ASCII/103 would increase
the cost by $100 or more.
* Nobody can type at 300 baud.
* ASCII is a "luxury".
* Giving data processing equipment to the deaf would discriminate
against the hearing; access should be equal, not superior.
* Market surveys "prove" the deaf don't want it.
* TDD's can talk to a computer too, you don't need ASCII for that.
There are other reasons, of course. But let me first lay out
some case for ASCII.
There are about 4 million ASCII terminals in the US, and this
figure is increasing rapidly (several hundred thousand produced per
year, but some are just replacements) And I've seen estimates that
there are about 40,000 to 50,000 personal computers in northern
California ALONE. How many of these units can now, or will be,
used to talk with the deaf? Damn few; they're all based on ASCII,
and those which have modems usually have the standard Bell 103 type.
Ironically, all of the currently produced TDD's (as opposed to surplus
teletypewriters) are built with ASCII-based components; the "C-Phone"
TDD is just a Micro-Term ACT-4 with a Weitbrecht modem appended! And
don't forget the steadily cheaper (V)LSI hardware that has ASCII built
into it.
What I'm afraid of is that a decision by the PUC which results
in distributing only Baudot/W terminals will result in continued
isolation of the deaf community, as well as a GREATER cost to the
public. When public data communication services become pervasive,
you can just bet they'll be ASCII. Existing networks like Telenet
and Tymnet require ASCII terminals for access; the ARPAnet is even
removing EBCDIC support from the TIPs, if not already gone. The
deaf will be left behind again, locked by archaic standards into
talking with themselves rather than the world-net at large. At
SRI we've sweat blood trying to get those damn TDD's to interact
reasonably with a simple electronic mail service; I just can't see
a similar effort being spent by the organizations and individuals
who will bring about the "network nation" that Human-Nets discusses.
To do this would incur great cost; forcing people to invest in
parallel systems doesn't make much sense. "What do you mean, I
have to buy another terminal?"
Now as to UNSTATED reasons why some parties want to avoid ASCII.
Since availability of equipment was an issue, the manufacturers were
frantic to convince the PUC that their current product lines satisfied
the requirements of the bill; after all, if the PUC didn't like any of
them, it would be a whole new ball game. And as long as the deaf use
Baudot/W, the manufacturers have a CAPTIVE market; if it were to
include ASCII equipment, the industrial behemoths might start paying
attention and wipe them off the map. Incidentally, the judge
specifically said that the PUC would not, repeat not, recommend any
particular product -- that actual selection of equipment meeting the
standards would be up to the telephone companies. This didn't stop
the trade show.
For some reason, any mention of c*mp*t*rs would freak out the
PUC. No one mentioned the fact that the little microprocessors
sitting inside most TDD's could be used for doing other interesting
things, and the idea of being able to talk with a computer over the
phone was heretical. It's hard to explain; you probably won't find
it on the record, but any time you were talking in the hall with a
lawyer, or a PUC person, or whatever, and tried to explain what
wonderful things could be done with terminals, they would throw up
their hands and say in horror, "God, don't tell them computers talk
ASCII!". Apparently, they want to be sure that the devices can ONLY
be used for talking to the deaf, and anything additional will have
the public up in arms whether it costs them anything or not.
Seemingly, even if it's cheaper.
Which brings me to the point that, frankly, there is only one
objection to a "dual" terminal (with both ASCII/103 and Baudot/W
capability) that I can seriously consider -- money. I think
everyone now realizes that it's not a question of the CODE, since
converting from ASCII to Baudot, and vice versa, is trivial for
the micros; even the keyboard (full 4-row 128-char ASCII, vs.
3-row Baudot) makes no difference, especially since some TDD's
are actually using ASCII keyboards! What makes the cost higher
is the MODEM.
Unfortunately, many people think of a dual modem as being two
separate couplers stuck together. With this model in mind, a price
increase of $100 isn't surprising. However, if the modem is designed
from the start for both uses, the extra cost becomes much less --
for the SRI-built dual modems, it's about $16. Of course the actual
retail price will be something like 3 times this, depending on the
manufacturer, but for current technology this isn't too bad. It's
when technological trends are taken into account that you can see
this price differential approach zero. In fact, one of the most
surprising things learned from the testimony is that the crossover
point is already here!
Novation, Inc. was a surprise party, with a nice present. They
demonstrated their recently developed "Personal Communication Device"
which has:
* Built-in telephone capability
* Light controller [i.e. Line Signal Indicator]
* Minimum 32-40 character alpanumeric display [LCD, dot matrix]
* 1000 character memory with offline editing capability
* Full four row ASCII compatible keyboard
* Multinetwork direct connect modem
- compatible with existing TDD network (45.5 baud)
- 103 answer and originate compatible (300 baud)
- 202 compatibility (1200 baud)
* Auxiliary printer port for printer option
Notice the triple modem capability. If your eyes haven't bugged out
yet, the cost is quoted at $325 in quantities of several thousand.
It is a very compact and portable unit; the keyboard feel is bad,
but that certainly isn't too hard to fix.
Options include:
* Battery pack
* Acoustic adaptor and carrying case
* Printer [and a neat one it is, too]
* Additional light controllers
* Emergency synthesized voice [yup, a voice synthesizer]
* Additional memory
It is extremely impressive. In private talks with the Novation
people (including the president, B. F. Kessler), they explained some
additional things which impressed us even more, but which are still
under wraps. I think it's safe to confirm that one of the keys to
this device is their development of a totally LSI custom modem chip.
I am not sure, but it makes sense to deduce that it is capable of
handling any FSK signalling scheme in programmable fashion.
Novation, by the way, was the ONLY manufacturer or technically
competent party, apart from SRI, to state as its position "That
currently postulated acoustic Baudot devices do not represent current
technology and will further limit and segregate deaf people from the
mainstream of present and future personal communication." This led
some people to think SRI was in cahoots with Novation, although we
were as surprised as anyone, and probably happier. I should also
mention that another company (Ultratec, headed by Rob Engelke) is
right now marketing a TDD called the "Superphone" with many of the
same features, including ASCII/103 capability and voice synthesizer,
for a slightly higher price; unfortunately they are based in Wisconsin
and were not present at the hearings to answer Pacific Telephone's
claim that they are not for real and don't have the requisite
production capacity. Sigh.
And now for Plantronics. They were the last party to testify
at the hearings, and they'll be the last party in this message.
Ummm. I don't know where to start, quite... basically, Plantronics
is very depressing. In 1980 they sold $97 million worth of electronic
products & systems to telecommunications & "productivity" industries,
and have a very tight relationship with telephone companies; most of
their equipment is designed to phone company needs and specs. They
have mounted a very slick, very professional campaign to get their
TDD entry, the "Vu-Phone", selected by phone companies for their
TDD needs. In fact, they have no plans to market the VuPhone to
individual consumers at all -- either the phone company buys it,
or they forget it.
Those people angry at SSI would do well to turn their attention
to Plantronics instead. It's my personal belief, and that of other
knowledgeable people, that unless the PUC requires ASCII/103 or
hardcopy capability, Plantronics has the game sewed up. This is
depressing both because the VuPhone isn't exactly the best device
I've seen, and because of their attitude.
The device: it has no hardcopy; they claim it will be available
as an option, but it hasn't been demonstrated, and the picture they
showed looked like a very expensive type of printer. What really
burns me up, though, is the constant claim of their salesmen that
"yes, they have ASCII". There is, in fact, a little switch on the
panel with two positions marked "ASCII" and "Baudot", and on the
witness stand the production manager asserted that the VuPhone had
a Bell 103 modem... HOWEVER, the keyboard itself is 3-row Baudot
(you want to send atsign? Tough luck) and only after persistent
questioning will they admit that they don't in fact plan to include
Bell 103 capability. Their user's manual even says the switch "Will
not work now -- it is for a feature to be added later". Sure, sure.
The attitudes: were incredible. If you are looking for a
company singularly unmotivated by concern for the deaf, Plantronics is
as good a candidate as any. Their lawyer was the veritable epitome of
evasiveness; he was literally the most objectionable, logging several
objections per witness. Exactly the sort of person you can trust to
shake undying friendship with one hand, while his other hand slips a
knife in your back. I realize one apple does not a bushel make, but
let's have a sample from the cross-examination of the production
manager:
<After establishing that the VuPhone lacks 103>
Q: Why isn't it included?
A: <Words to the effect that giving the deaf a data terminal
would be too much, computer access isn't needed...>
... I don't have an ASCII terminal at my house!
Q: Do you have a Baudot terminal at your house?
A: Uh, no.
Q: Are you aware that voice-only computer access is possible?
<OBJECTION!! from PUC lawyer (a second ahead of Plantronics
lawyer) Judge upholds objection, terminating that
line of questioning...>
The point was, of course, to try to establish that right now
you can interact with computers by voice, hence computer access per
se is not a "luxury". As it turns out, Plantronics has produced a
"Roadrunner" telephone management network with a "Speakeasy" feature,
"translating the voice into digital data for the Roadrunner computers,
and synthesizing a spoken reply by translating digitial data into
spoken words." This is from their annual report, and they obviously
are aware of the technology, whether or not they admit it.
Don't go away just yet, this one is interesting. After the
hearings were closed, we managed to catch some Plantronics people
away from their lawyer, and they dropped a bombshell -- ostensibly
the REAL reason why they weren't going to provide ASCII:
"If we do that, IBM will sue AT&T and tie up everything
for several years..."
They mumbled something about the FCC Computer Inquiry decision;
apparently the phone company isn't supposed to be in the business
of providing "data processing equipment", under any guise. I'm not
sure I believe that, since I can't figure out any reason to keep that
information private. If it is in fact a legitimate concern, certainly
the PUC ought to know about it, as well as us consumers.
Perhaps some of you have thoughts about that, beyond the question
of social taxation. I believe in fact that in two years the phone
co. is required to divest itself of such equipment by forming an
(unregulated) subsidiary, but can't seem to find my reference just
now.
------------------------------
That's all. This particular message has of course been slanted
to reflect what I feel is the critical issue at stake; there are
others which I've left out or just skimmed, such as that of hardcopy
vs. softcopy. Perhaps I should send another message about that point,
since it certainly inspired a lot of rhetoric at the hearings, and is
tied in with notions of "equality". At any rate, I'm interested in
knowing what people think about:
(1) the bill's intent (assuming the best) and possible alternatives;
(2) the definition of "minimum required" and "equal access";
(3) the technological tradeoffs w.r.t. equipment;
(4) the future effects (eg on public electronic mail);
(5) how to best exert influence on the PUC/appropriate bodies,
and what to recommend.
Ideas or feedback might well prove valuable; there's still
time. I'll be sending updates as things develop -- the preliminary
PUC recommendation is due on the 26th, for example. If anyone is
curious about some aspect that I've slighted, I'll be glad to
extract and send any relevant information I have.
--Ken
------------------------------
End of HUMAN-NETS Digest
************************
∂22-Sep-80 0201 LWE debate
can't tell yet - looks like more comments focus on "who won" (again)
than on content. actually i was trying to see if either wire carried
verbatim transcript that could be used for machine content analysis
(not so far, it seems - will take another look tomorrow)...
by the way, should i mail a tape to dump digest file(s) on before they
get too big? please advise at your convenience.
Unfortunately, transcripts are carried only on the high speed wires
(Datastream) to which we can't afford to subscribe. They usually
mention it on the regular wire when they do it, so I suppose they
didn't in this case. I don't know about machine analysis, but the
comments on who won say quite a bit about the newsman, especially
concerning the intensity of their feelings. Would you agree that
Reagan is getting substantial advantage from the fact that many
people who disagree with him can't work up either anger or fear,
and may therefore express themselves by voting for Anderson or
just stay home. I'll look into the tape question, but we have no
regular operators, so it isn't obvious how to get it done. It may
cost you extra. Actually you should inquire directly by MAILing
a message to REG.
∂22-Sep-80 0540 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #82
Date: 22 SEP 1980 0823-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #82
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Monday, 22 Sept 1980 Volume 2 : Issue 82
Today's Topics: ACM '80 Panel,Electronic Banking - Possible Services,
Telephone Rate Structures - Long Distance,
Computers and the Deaf - CA D.E.A.F. Surcharge
& TDD Design, Technology - Effects of Regulation
---------------------------------------------------------------------
Date: 21 Sep 1980 0043-PDT
From: John McCarthy <JMC at SU-AI>
One would naturally suppose that "The Computerized Destruction of
Western Civilization" panel was a spoof, but considering that it's
SIGCAS (Special Interest Group on Computers and Society (of ACM)),
it probably isn't. SIGCAS impresses me as one of the most blindly,
technologically reactionary conglomeration of know-nothings, I have
ever encountered, based on all the issues of their newletter available
a few months ago in the Stanford library. Their specialty is horrific
"what if" scenarios about data banks, electronic funds transfer and
other computer applications. Whether it is from lack of imagination
or merely ideology, they don't see any of the positive potential of
computerization and the enormous costs in delay implementing it. For
example, if we can get rid of cash, we can get rid of crime for cash.
A robber can't hold you up and demand that you transfer money to his
account - not unless he can prevent you from talking and conceal the
fact of your demise. By helping delay EFT, SIGCAS assumes responsi-
bility for the deaths of perhaps several thousand people per year.
Another example is that the Freedom of Information Act could be made
effective by requiring that all such information be on computer file.
Indeed the discussion of telephone company practices in HUMAN-NETS
would be much better informed if the Public Utilities Commissions
were required to keep their decisions and the tariffs they allow
on file. (I apologize having the same urge to impose a duty on
government agencies that others have to impose duties on the
telephone companies).
Is anyone out there interested in forming an organization that would
devise and press for useful applications of computer technology in
society.
------------------------------
Date: 20 Sep 1980 (Saturday) 2308-EDT
From: MORGAN at WHARTON (Howard Morgan)
Subject: Bank by phone - the other side of the coin
Bank by phone service has been available in the Philadelphia area
for several years now, and, like any good hacker I signed up for
it a while ago. In general, it is quite convenient, working about
the way Lauren's bank in CA does, with preasssigned merchants etc.
But their problem handling isnt nearly as good. On May 28th, I
dialed up and told "George" (Girard Bank's name for their ATMs
and Bank by phone) to pay a major credit card bill. On June 6th
the credit card co. sent me their standard dunning letter asking
for payment. I called them, explained that I had payed via BBF 9
days earlier, and they said they would check their records. Three
days later they called and said they still had no money and were
canceling my account. I asked them to call the bank, giving them
the reference codes supplied by the friendly synthesized voice at
the end of each call. They called the bank, who confirmed that I
had given payment instructions, but there was still no money around.
Finally, on June 17th, 20 days after I called (and had the money
removed from my account) the payment showed up at the credit card
co. We are still negotiating on the reopening of the account, and
the bank has been more helpful after I put some high level pressure
on them, but I no longer feel as secure (in the non technical sense)
about payments going out on time.
------------------------------
Date: 22 Sep 1980 2355-PDT
From: ROODE at SRI-KL (David Roode)
Subject: WATS lines
AT&T has petitioned the FCC for approval to raise WATS rates for the
big users by 75%. From an AP news article reporting on the petition:
"The impact of this will be disastrous for many businesses and
consumers," said Mike Hess, an attorney for the Committe of
Corporate Telephone Users. "The impact is unevenly skewed upon
LARGE USERS [emphasis added --rdr], the people who use these
lines efficiently, and we will fight it."
An Avis Rent-A-Car V.P. went on to say this would increase his WATS
bill $2 or $3 million (from $5 million). He threatened to increase
car rental by $.50 or $1 per day.
Personally, I think many of these 800 numbers are a waste and
encourage inefficiency in a way similar to that of people using
411 instead of the phone directory. Furthermore, it is a way that
business escapes subsidizing local service with its long distance,
unlike ordinary individuals. The article states that the increases
will affect 10% of WATS users and 90% will see slight decreases but
it observes that AT&T still must justify the rate advantages afforded
WATS users as opposed to regular long-distance.
------------------------------
REM@MIT-MC 09/20/80 04:38:59
Rebuttal to claim that nothing is itemized on telephone bill
(except of course the long distance charges and recently
message unit charges):
Pacific Telephone (Calif.) itemizes on each monthy bill:
Taxes US: .42 911: .06
Property tax adjustment .81 CR (due to Jarvis-Gann bill)
Credit for intrastate toll CPUC decision 90316 .43 CR
(I think everything except 911 tax, and maybe the property tax
adjustment, is a function of how much I've used my phone, so
the numbers may be different for other people's Pac.Tel. bills.)
------------------------------
Date: 20 September 1980 1107-EDT (Saturday)
From: Richard H. Gumpertz <Rick.Gumpertz at CMU-10A>
Subject: Deaf Equipment Surcharge
One of the things which bother me about the $0.25 surcharge (and
may be what is bothering Lauren) is that it is effectively a tax
for supporting a social program but the tax is imposed only on
telephone users. Why not support it out of the general fund of
the state government? There are already many programs which
benefit a limited number of people which are paid for out of
broad-based taxes; why should this one be paid for via a
limited tax?
Rick Gumpertz
------------------------------
Date: 19 Sep 1980 1417-PDT
From: ROODE at SRI-KL (David Roode)
Subject: D.E.A.F. surcharge
There are reasons why the $.25 per month D.E.A.F. surcharge can be
considered other than a tax. Do we consider it a tax when we provide
wheelchair access in a building? Do we consider it a tax when we
spend extra money in recruiting so as to potentially attract women
and minorities? Personally I think it is OK to subsidize it but am
opposed to anything completely free -- perhaps there should be a
$1/month charge levied on this extra equipment. And $600/unit seems
ridiculously high. With quantity production, one would think they
could get a modem for $50 and a single-line LED display for $250.
------------------------------
KLH@MIT-AI 09/21/80 19:20:20 Re: Surcharge = tax?
With respect to the messages about terming the D-E-A-F a "tax",
here is some testimony from Pac Tel:
"The surcharge would be separately identified on each customer bill
as 'DEAF', 'SB597', 'HCAP', or similar description. Pacific plans
to display the surcharge on the tax line of the bill. The amount
of the surcharge must be included in the amount subject to 911,
city utility users, and Federal Excise taxes. "
"Since non-certified [i.e. non-deaf] subscribers are apt to view
the surcharge as either a tax or a non-conventional additional
subscriber fee, refusals to pay the surcharge could result. We
do not view the surcharge as a tax. It is an amount paid for
the provision of telephone service to certified customers. When
customers refuse to pay approved rates for service provided, the
Company takes appropriate action to collect the amount due. In
contrast, the small percentage of customers who refuse to pay
the "911", city utility users, and Federal excise taxes are
reported to the taxing agencies, and they take steps to collect
those taxes. Pacific Telephone Company acts only as a conduit
to collect these taxes and remits them to the taxing agencies."
I don't know whether the surcharge would apply to ALL bills, such as
private leased lines. The wording of the PUC from Appendix C is
"Each participating utility shall bill the specified surcharge
monthly to each main station or equivalent main station that it
serves. Surcharges shall be placed on foreign exchange services
by the utility normally rendering the bill to such services."
...
So it depends on the definition of "main station". Also, the .25 is
an AVERAGE, and
"The distribution of this surcharge by class and grade of service
will be determined by the Commission."
------------------------------
REM@MIT-MC 09/20/80 05:07:32
Re: Barry Strassler's remarks on hard vs soft copy.
I'm afraid I must agree that ordinary people aren't given free
answering service or free tape recorder, by which to collect calls
while not at home or by which to record calls when attended, thus
deaf people shouldn't be given free hardcopy which would also permit
those luxuries the rest of us don't have. If a deaf person wants to
keep a record of a call, (s)he can attach a tape recorder to the line
(after notifying the other party) and record the modem sounds, which
can be played back into the modem at any later time to recreate the
whole conversation just like a hearing person can record the voice
transcript. I therefore believe that to provide the cheapest available
video terminal that can display enough lines to maintain coherence (8
lines of at least 40 characters should be enough to hold one or two
sentences while the person is reading them) is all that is really
required by law. Anything beyond that which deaf people "prefer"
is like other things that people can buy if they really want them
enough.
I would like to see all handicaps exactly compensated so nobody loses
or rightly feels cheated.
(I just wish medicine could reach the point where blind people could
be equals in society like deaf people with terminals nearly can be.)
------------------------------
Date: 21 Sep 1980 at 1342-CDT
From: david at UTEXAS
Subject: Response to "ideal" TDD question
First, thank you, Ken, for an interesting and well written series of
reports on CA TDD developments.
[54 lines on public TDDs, free enterprise, and widespread computers.]
I have great difficulty trying to decide how I would "ideally" set
things up to provide for worthwhile public needs, such as providing
equal access for the handicapped to public services. But I do feel
opposed to the idea of government GIVING anything away, whether it
be to individuals or corporations. In my mind there are a couple of
problems: unearned or undeserved, even unsought-after gifts are too
little valued; and, this directly mixes the government in what should
be a free market decision.
My advice to the California PUC would be this: the responsibility
of the government to provide equal access to public services should
require that all operators for the public phone companies and other
public services (fire, police, hospitals, etc.) be provided with
telecommunication devices for the deaf. Because in this setting
it would be important to serve the widest possible audience, they
should select devices with the widest possible communication range.
Individuals and businesses should be free to select what ever device
they think they need, and then deduct some amount from their income
tax if the purchase was for the purpose of communicating with the
deaf. The needy deaf should be provided for. If the government
does, it should select the most cost effective device. Since they
should already have all capabilities at public places, it is likely
that ascii, 103 or 212 would be the choice.
This seems to have several advantages. Instead of raising everyone's
taxes to support a praiseworthy yet socialistic cause, this would
reduce taxes. Users will have complete freedom of choice in the
devices they select, though certainly they can be guided by the
public information gathered for the public TDD selections. In such
a free market, producers will be driven by competition to produce
the best or the cheapest device.
I'll bet we could have the greatest impact in this matter by
approaching various terminal manufacturers and convincing them of
the really vast market that could be developed for their product
by bringing this issue to the public, say with advertisements in
the popular media. As Ken points out, the market is more than the
strictly deaf, it includes everyone that wants to communicate with
them. And in fact, it is more than that. My parents, with a small
business, are starting to think about what a computer could do for
them. Well, if they had a computer "on the phone" I know our
communications would be better, strange as that may sound.
If we could convince manufacturers (...what sort of convincing should
we have to do? It's in their own interest.) to really "popularize"
computers, then the price could become very low, and the devices
would be very widespread.
Is this all just so much hot air? If it is as obvious as it seems to
me, why hasn't someone done it? Has our national drive waned to such
a point that we need the government to create and influence what
should in any case be a lucrative market?
david m. phillips (david@UTEXAS)
------------------------------
Date: 19 Sep 1980 1449-PDT
From: Steve Saunders <SAUNDERS at USC-ISIB>
Subject: Laws
"Any sufficiently promising technology must be regulated,
or it will succeed."
- from The Official Explanations, Paul Dickson (Delacorte 1980)
as quoted in Machine Design
------------------------------
End of HUMAN-NETS Digest
************************
∂22-Sep-80 1102 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Honeywell wants a speaker in Phoenix
Date: 22 Sep 1980 1100-PDT
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: Honeywell wants a speaker in Phoenix
To: CSD-Faculty:
cc: csd.tajnai at SU-SCORE
I received a call this morning from Barbara Kuklewitz of Honeywell in
Phoenix. They are planning a two-day meeting of approximately 200
managers. They would like to have someone from our department give
a talk on "Future of Computers." Also, share some research results.
They would like for the speaker to speak twice (once on each day) but
would be willing to get a different speaker for the second day.
Honeywell is not on my list of Computer Forum members. I'll forward
this msg to Sally Burns. It is possible they could be recruited.
If you are interested in being their speaker, please send me a message,
or call Ms. Kuklewitz at 602/866-2730.
Carolyn Tajnai
-------
I would be willing if the time works out.
∂22-Sep-80 1050 FFL
To: JMC, FFL
Call from Sigma Xi (Jill Hammons, 7l4 452 3624). They would like to have
an advanced copy of your speech, abstract or brief review sent as soon as
possible. Mrs. Janke who is in charge of redia releases then will call you
to interview you over the phone. Do your have something?
I don't have anything right now. Maybe something short by Friday.
∂22-Sep-80 2042 Kanerva at SUMEX-AIM DEC-20 configuration
Date: 22 Sep 1980 2041-PDT
From: Kanerva at SUMEX-AIM
To: JMC at SAIL
Subject: DEC-20 configuration
I have an early version (August 13) of the configuration plus
recollections on how the presently proposed one differs from it.
Ralph has the up-to-date picture, which is about as follows:
2060 CPU with cashe
1.5M words MOS memory
4 massbus controllers
2 176Mbyte disk drives (RP06), controller, & two channels
1 967Mbyte disk drive (RP20), controller, & channel
2 6250 BPI, 125 ips tape drives, controller, & channel
32 asynchronous terminal ports on the front-end 11, later
to be supplemented by EtherTIP ports.
In addition, there will be Diablo and possibly Printronix and Canon
printers plus terminals that correspond to the Datamedia Elite 2500.
These will not be from DEC.
The budget has been based on this configuration, with a 40% discount
from DEC, which we were told to be possible. Since the fast tapes are
very expensive ($167,400 for two plus controller before discout), it
might make sense to have slower drives (like the ones at LOTS and SCORE)
and a second 967Mbyte disk instead.
- Pentti Kanerva
-------
Thanks, Pentti.
∂23-Sep-80 0337 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #83
Date: 23 SEP 1980 0623-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #83
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Tuesday, 23 Sept 1980 Volume 2 : Issue 83
Today's Topics: COMPCON Panel Cancelled,
WORLDnet - Fiber optics/Telematics Companies Query,
Computers and the Deaf - TDD Design & Doing Something,
ACM '80 Panel
Electronic Banking - Possible Services & Auto-teller Ref
----------------------------------------------------------------------
Date: 22 Sep 1980 (Monday) 1337-EST
From: DOLO at NBS-10
Subject: PLEASE NOTE!!!!!
Cartesian Programmer Debate Tuesday evening cancelled -- paper will
be presented at the short notes session 2:00-3:30pm same day, Tuesday,
September 23.
------------------------------
Date: 22 Sep 1980 1023-PDT
From: Chesley at SRI-KL
Subject: Fiber optics companies
I'd be interested in finding out which are the major companies
involved in fiber optics. Who's providing all that fiber optic cable
that the phone companies are beginning to install?
And as long as I'm at it, what about other companies involved in
communication? In other words, who will be building the Worldnet?
Send replies to me, and I'll summarize for the list.
--Harry...
------------------------------
DGSHAP@MIT-AI 09/22/80 14:49:44 Re: hobble me, hobble my brother
One of the most interesting themes in these hearings is the
luxury vs. compatible service conflict. I see the hearings
making a bee-line towards giving the deaf comfortably
inconvenient equipment. By that I mean equipment which is
just frustrating and antiquated enough so that the hearing
public can be sure that the deaf community has not been given
a leg up. I have this scenario in my head where the political
pressures (to legislate in favor of the deaf, but to provide
for minimal phone service only) force the selection of a
terminal which is less capable than other alternatives, but
more expensive at the same time. This is sort of a bizarre
interaction between the value of spending a political dollar,
and the negative value of getting too much for your money.
Dan
------------------------------
LLOYD@MIT-AI 09/22/80 22:03:07 Re: DEAF system
Since it costs just as much to develop special hardware to make sure
that the deaf dont get any "undue advantage", why not let them have
what they want? A cheap hardcopy terminal (ala Teletype 43) doesn't
cost significantly more than a decently built CRT.
Brian Lloyd
------------------------------
Date: Monday, 22 September 1980 14:33-EDT
From: John A. Pershing Jr. <JPERSHING at BBNA>
Subject: The Myth of Equal Access
I guess it comes as no surprise that the exponents in the CA PUC
debate over TDD's want to make sure that they provide EQUAL access
to the deaf, but not one iota MORE access. However, don't these
people realize that there just isn't such a thing as equality??
For heavens sake, deaf people CANNOT have access which is equal to
non-deaf people, unless we eliminate phones and make everybody use
teletypes. SO WHAT if a deaf person gets a subsidized ASCII terminal
and can use it to log into the computer from home (in addition to
communicating with other deaf friends)? This "advantage" can hardly
make up for other disadvantages that the deaf have to put up with
from day to day.
This issue of providing exactly equal access brings to mind a story
which I encountered quite a few years ago (was it a Woody Allen
movie??). The story depicted a future society in which true equality
had been legislated. This meant, for instance, that people who were
fast runners were forced to wear heavy sandbags all of the time in
order to slow them down. Admittedly, this is a bit extreme; however,
the moral is that "equality" is a relative concept. Given two frobs
A and B that are "equal", one can always find inequalities if one
looks closely enough.
The relevant question isn't one of equal access; rather, it is the
age-old question of economic tradeoff. After reading Ken's messages
on the CA PUC hearings, it seems that, indeed, most of the discussion
is really centered on economic issues (actually, beating around the
bush of economic issues). In fact, the "equal but not greater-than
access" argument is merely corner-cutting in disguise. This is all
well and fine -- if we halve the cost of TDDs, then twice as many
people will get them.
However, the travesty of this whole process is the fact that the needs
of the deaf are merely being used as a vehicle for the economic gain
of businesses (what does a terminal manufacturer really know about the
needs of the deaf, anyway?). These pro/con arguments aren't focusing
on the real needs of the deaf; they are focusing on the problem of
the Baudot manufacturers getting driven out of the marketplace by IBM.
Now, there's an example of a minority, special-interest group which
wants a subsidy! Golly gee, what's going to happen to those poor,
disadvantaged entrepreneurs if ASCII TDDs become the standard? Where
is their "equal access" to the market place?
Perhaps the PUC hearings should be held in two parts. The first
part would be an investigation of the needs of the deaf: essential
features, bells and whistles which make life a LOT nicer, and frobs
which are nice but not really essential. No names would be used in
this part of the hearing (such as "ASCII" or "Baudot"), although the
optional feature list might include "capability to communicate with
computer". Also, this part of the hearing would be mostly populated
by direct representatives of the deaf -- none of this trade-show
crapola. The second part of the hearing would then concentrate on
economics: given the list generated in part one, where is the point
of diminishing returns. This part of the hearing would inevitably
include many trade-showish aspects [can/should this really be
avoided?], but should concentrate most of its effort on the abstract
issue of cost/benefit tradeoffs. Independent consultants should
be retained, to try and introduce a bit of objectivity into the
proceedings, too. Note, in contrast to part one, this part would
be mostly populated by [non-deaf] managerial and engineering types.
(Perhaps we should include a part three, in which the lawyers stage
one of their dramatic productions. The transcripts of this part
would be ignored by the PUC; however, they would be released to
the press to provide something sensational to report.)
-jp
------------------------------
Date: 22 Sep 1980 1826-PDT
From: Craig W. Reynolds (at III via Rand) <REYNOLDS at RAND-AI>
Subject: political action on free TDDs
This is a question for KLH, but I thought all of us laid-back
California hackers might want to know:
Given the current stage of this governmental drama, what is the
best way of expressing opinions on the whole issue? Letters to
Human-Nets presummably do not get counted in PUC votes. To whom
do we write to insure that the forces of darkness don't win
another one (such as non-ascii TDD)?
------------------------------
Date: 22 Sep 1980 1506-PDT
From: ADPSC at USC-ISI (Attn: Don)
Subject: Applications of Computer Technology
Subject: (and associated ramifications)
I must disagree with Mr. McCarthy's analysis of what will happen
to us when we fully computerize (HN, V2 #82, 9/22/80). While I
know little of SIGCAS, I have always held that a bit of healthy
scepticism never hurt.
The idea of completely eliminating cash (and thereby eliminating
crimes for cash) is, when taken to its logical extreme, a bit
farfetched. Within the next hundred years or so the requirements
for some amount of fiat money per person will not be eliminated
(based on a study by the Fed.) If we make ourselves into a
"cashless" society are we eliminating "crime for cash" or are we
merely removing the immediate opportunity for armed robbery? I
find the prospect of electronic chicanery equally disconcerting.
With a gun or with a TTY, either way you're out some bucks.
I will agree with you that Freedom of Information Act requests could
be made easier. However, putting great masses of information on-line
ignores the other side of this two-edged sword, the Privacy Act. The
fact that you have had some form of criminal action taken against you
is no reason for my being able to learn everything about it if I was
not involved. While we all have a right to know, we also have a right
to be left alone. (My feelings on this matter are based on being a
former Justice Department employee.)
Public utilities commissions are required to keep their tariffs on
file and available for public inspection during normal business hours.
In addition, in most states I am aware of, the utilities themselves
are required to do the same. Are we to blame technology because the
gas company doesn't carry front page ads in local newspapers telling
what they are charging? Go up and ask them. Write a letter to your
representative(s).
I agree that improvements in technology should be implemented when
and where possible. But let's not jump into these things and change
simply for the sake of improved technology. We must ask ourselves if
what we are replacing it with is an improvement. A lot of pregnant
women in the 50's felt great while taking thalidomide.
Don
------------------------------
Date: 19 Sep 1980 1642-PDT
From: Lee W. Cooprider <COOPRIDER at USC-ISIB>
Subject: Phone banking
In February of 1976, I established an account like that described by
Lauren with the Pittsburgh S&L called Dollar Savings Bank. Merchants,
visa, inter-bank transfers, insurance, utilities and randoms were all
payable by touch-tone codes 24 hours a day. Transfers to other banks
were guaranteed to be delivered with the 9:00 courier and out-of-town
deliveries mailed by that time. They soon combined this with a draft
account (read checking account) which automatically transferred money
from the interest bearing account to the draft account when necessary
in the exact amount of the draft. Phone transactions cost $.10,
two-thirds of the price of a stamp; drafts were free. Interest paid
at 5% or 5-1/4% depending on options.
I liked the service very much and in fact retained my account there
even after I moved to California two years ago. I was of course
very disappointed that there was no comparable service here (until
recently) and amazed by the backwardness of California's banking
system (as mentioned by others). I have been told that innovation
is stifled in California because B of A controls the state banking
commission and is exceeding conservative w.r.t. banking policies.
By the way, paying bills via the Dollar SB system took about 1/3 the
time of writing checks, perhaps even less. That alone makes direct
interaction with the bank a win.
------------------------------
Date: 20 Sep 1980 2039-PDT
From: Geoff Goodfellow <GFF at SU-AI>
Subject: AP News Story on ATM's.
Consumer Scorecard: Banking by Machine
By LOUISE COOK Associated Press Writer
Millions of Americans are doing their banking by machine these
days, and people who use the automated tellers have to take extra
steps to protect themselves and their money.
There are an estimated 14,000 machines across the country in
locations ranging from bank lobbies to apartment complexes, from
shopping centers to airports. The terminals dispense cash, accept
deposits and let you check your balance - 24 hours a day, 365 days
a year.
Before you start pushing buttons, you'll have to learn some
new words and phrases.
Begin with "electronic funds transfer," usually abbreviated
as EFT. The phrase covers the whole field of automatic banking
and payments. It not only includes the increasingly familiar
bank-operated terminals, but also applies to the relatively new
"debit cards," which look like credit cards, but are used instead
of cash or checks to make purchases.
Next, there's "access device." That's the card you use in an
ATM - an automated teller machine. When you insert your access
device in an ATM, you'll also have to use your PIN - your personal
identification number, a secret code, without which the machine
won't work.
Users of the EFT system are protected by "Reg E," a regulation
issued by the Federal Reserve Board to implement legislation passed
by Congress in 1978.
"Reg E" limits the amount of money you can lose if your card is
used without your authorization. As long as you notify the issuing
bank, savings and loan association, etc., within two business days,
you cannot be charged more than $50 - no matter how many transactions
are made with your card. The notification can take the form of a
personal visit, a letter or a telephone call. The longer you wait
to report a lost or stolen card, the greater your liability.
Citibank, the nation's second largest commercial bank, has some
advice for customers who want to give themselves added protection.
Among the suggestions:
- Choose a personal identification number that's easy to remember
so you won't have to write it down. Avoid a number based on your name,
telephone or Social Security number or address. A thief who steals
your wallet will probably have access to those numbers and try them
first.
- Do not share your code. "More often than you'd expect," says
Citibank, "youngsters with access to a parent's EFT card have dribbled
out funds to supplement an allowance. So have trusted secretaries or
other employees who've been asked to make transactions for the EFT
card owner."
- Always wait to collect your receipt - no matter how rushed you
are. You'll need the receipts to check your monthly statement and in
case of disputes.
- If you make a mistake - or if the machine does - call the bank
immediately. There should be a customer service telephone nearby.
------------------------------
End of HUMAN-NETS Digest
************************
∂23-Sep-80 1020 FFL
To: JMC, FFL
Zohar asks that kyou meet at 12:45 instead of 12 noon today.
∂23-Sep-80 1258 RPG Damages
To: JMC, TOB
Dave Waltz has sent the receipts and other admissions of expenses,
coming to $426.80. I await your instructions.
-rpg-
Give receipts, etc. to Frances Larson.
∂23-Sep-80 1439 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Date: 23 Sep 1980 1436-PDT
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
To: JMC at SU-AI
In-Reply-To: Your message of 22-Sep-80 1105-PDT
John,
I learned from Sally Burns that Honeywell does belong to the Forum and
Zohar is their liaison. I forwarded the msg to him and asked him to
follow up. I told him you were interested.
I will leave it with you and Zohar unless I hear further.
Carolyn
-------
∂23-Sep-80 1449 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) Evans & Sutherland
Date: 23 Sep 1980 1448-PDT
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: Evans & Sutherland
To: JMC at SU-AI, FB at SU-AI
cc: CSD.Dbrown at SU-SCORE
EAF received a letter from Teri Adams, Employment Supervisor of Evans
and Sutherland. She will be visiting Stanford on Friday, Nov. 7, for
college recruitment and "to meet with professors to get a better
understanding..." "Your name was recommended tto me by our President,
Dr. David C. Evans, as a person who could provide such information and may
possibly know some students who are interested in Computer Graphics."
She sent an annual report.
Since Evans and Sutherland does not belong to the Forum, this type of
courtesy is not usually afforded. However, because Evans was a member of
the Advisory Committee and Ivan Sutherland is a colleague, it is perhaps
more politic to go along once.
I will mail her a brochure on the forum and put up an announcement that
she is coming.
Do you want to see her? Recruit the company for the forum?
Carolyn
-------
I have some good will left for Evans and Sutherland, but I really don't
know anything now that is likely to be of use to them. Therefore, someone
else would be better to tell them something of use and recruit them
to the forum.
∂23-Sep-80 1524 PAT lib.lst
To: JMC
CC: FFL, ROB
well, i've searched every directory i can think of and then some and can't
find it. the last dump date was Oct 25,1979, a very suspicious date as
that was around the time th machine went down for the move. i had thought
that i had moved it to one of JMC's areas. the best thing seems to be to
restore it. the tape is P1604, LIB.LST[PUB,PAT]. sorry i can't be of more
help. Patte
Thanks, Patte. We'll restore from that dump.
∂23-Sep-80 1536 Buchanan at SUMEX-AIM Natl Rsch Council mtg
Date: 23 Sep 1980 1536-PDT
From: Buchanan at SUMEX-AIM
Subject: Natl Rsch Council mtg
To: jmc at SU-AI
It was interesting. The fellows from Bell Labs and IBM seemed downright
hostile to AI and were not able to contain their snide comments.
That part was disappointing. I think I can contribute something along
the lines of the outline I wrote and will send you a revision
in the next week or so for comments. It made everyone feel better that
you were adding your input to this paper --- would you care to co-author it?
You mentioned a few weeks ago that you might have a spare terminal
upstairs to replace this old TI I've been using. At the time I thought
I'd be getting a Heath, but that's held up now while we try to
get them to make good on repairs for the ones we have. Do you still
have a terminal I could borrow?
thanks,
Bruce
-------
∂23-Sep-80 1900 JMC*
abstract for sigma xi
∂24-Sep-80 0321 REM at MIT-MC (Robert Elton Maas)
Date: 24 SEP 1980 0618-EDT
From: REM at MIT-MC (Robert Elton Maas)
To: JMC at SU-AI
That's a briliant idea, assigning SIGCAS blame for muggings etc. including
resulting deaths. I agree that SIGCAS should find imaginative scenerios
on both sides instead of just one side. I suggest you write up some
scenerios, including the one of "EFT delayed 10 years, 10,000 people die
as a result of muggings and armed robbery that wouldn't have happened if
big-computer kept track of all 'money'", and send them as letters to
editor of SIGCAS, if they permit such things. If not, complain to ACM to
have SIGCAS forced to print letters to editor or droped from affiliation.
Yes, I'll join organization to find appropriate uses for computers and to
publish scenerios of how much better life will be with than without. But
I thought we were already doing such things via Arpanet mailing groups?
If you want to collect messages on that subject, and take it upon yourself
to edit for public consumption (flushing AI jargon and all references to
personal use of Arpanet) and then release them to the press for public
distribution (you're famous enough to be interviewed on TV news if you
have something interesting to say on a topic they're investigating), go to
it!
∂24-Sep-80 0359 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #84
Date: 24 SEP 1980 0647-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #84
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Wednesday, 24 Sept 1980 Volume 2 : Issue 84
Today's Topics:
Electronic Banking - Cashless Society & Possible Services,
Computers and the Deaf - TDD Design & Doing Something
& Story Reference
----------------------------------------------------------------------
Date: 24 September 1980 0432-EDT (Wednesday)
From: The Twilight Zone <Doug.Philips at CMU-10A>
Subject: McCarthy's note
I doubt if this society will become cashless in the near future and
even in the 'medium' future. I seem to remember a story on 60 minutes
not too long ago (read less than a year) about how many people there
were that dealt substantially in cash, and as a result paid less/no
federal income tax. As I remember, there were a LOT of these people.
Simply the fact that they exist and that they probably like things the
way they are means you have a substantial block of people dealing in
cash and that they will probably resist efforts to do otherwise. As
I see it, a 'cashless' society is a long way off.
Doug
------------------------------
Date: 23 Sep 1980 (Tuesday) 0516-EST
From: GERMAN at HARV-10
I want to express my admiration for Professor McCarthy's remarkable
argument that opponents of EFT are responsible for muggings, but I
think he has pushed things too far. The logic is unassailable --
if nobody carried cash, people would not try to rob it from you.
One hears the phrase "cashless society" in connection with EFT, but
I always thought it was just a slogan -- not to be taken literally.
EFT cards are just a replacement for credit cards (which are a
replacement for checks), and credit cards have not eliminated all
use for cash.
The main benefit of EFT goes to creditors who will be able to
collect payments more quickly and avoid bad debts.
Steve German
------------------------------
Date: 23 September 1980 2042-EDT (Tuesday)
From: Joe.Newcomer at CMU-10A
Armed robbery is only one way to obtain cash. Selling swamps in
Florida, envelope-stuffing businesses, numbers, drugs and sex are
other ways of relieving people of their wealth. And if you don't
want the transaction recorded (as the cashless society implies),
and there is no scrip, then SOME form of scrip will be invented.
Gambling casinos, mining towns, and at one point not too many
years ago, local banks, all did it. I would expect to hear
someone offering "two Big Louies on 762" or a crisis because the
"Frankies were devalued after the shootout with the East End gang
showed he was losing control".
------------------------------
Date: 23 Sep 1980 (Tuesday) 0653-EDT
From: MOSESJ at WHARTON (Jack Moses)
Subject: living in a time warp....Illinois
Whatever you do, *DON'T* move to Illinois.
I am a resident of the Chicago area and I find it most amusing that
discussions of ATMs and EFTs contain references to 'back in 1976'
or 'a couple of years ago'.
The McFadden act is a|ive and well and living in Illinois. The
state politicians have never repealed or amended the act so we
have no branch banking. To be sure, there are a few ATMs around,
but they stand like idle robots in dark hallways threatening all
who pass. EFT stands for 'Extended Free Transit' which means
for another 10 cents you can ride around the Loop twice (subway).
And the body politic of Chicago is still debating the existence
of Cable TV so we have *NONE*.
So while I read of the virtues and difficulties some of you have
encountered with EFTs and ATMs, I long for the day when I may
have the privilege of realizing same.
oops, gotta run, my stagecoach is coming. ta ta
Jack
------------------------------
Date: 23 Sep 1980 1142-PDT
From: MJMARCUS at USC-ISI
Subject: KLH's TDD note in [HNT V2 #81]
Plantronics seems to have a very simple model of the FCC's 2nd
Computer Inquiry Decision. (You may remember that I solicited
requests for copies in this space several months ago and sent
out 80 copies before we ran out. If you would like to read the
actual text seepage 31319 of the May 13, 1980 Federal Register.)
While the decision might be paraphrased in part as saying TPC
isn't supposed to be in the data processing business, it actually
says that BELL and GTE must form separate subsidiaries to provide
customers with any customer premises equipment (CPE) be it black
phones, TDD's, PBX's, or computers. AT&T recently took the first
steps to set up such a separate subsidiary. It is interesting to
speculate ho this might affect the California problem. Since PT&T
is forbidden to supply any customer with any CPE post-1982, the
TDD's may have to come from the new subsidiary. Is there a legal
way for it to take money from PT&T's regular subscriber's? There
may be questions of state vs. federal jurisdiction. The whole
thing could keep a lot of lawyers well employed.
You might be interested in knowing that production ASCII/Baudot
terminals are sold for ham use and regularly advertised in QST
and like journals. However, these do not include modems.
Mike
------------------------------
Date: Tue, 23 Sep 80 12:07-EDT
From: ljs at DNGC
Subject: Soft or Hard TTD Device ?
First of all, Ken (klh) did a good job to inform such facts we
need to know. I would call him "Jack Anderson II" in the deaf
world. (Jack Anderson is the syndicated newspaper columnist
exposing inside facts).
Also, such comments shared by other HNT readers are very acceptable
and informative. I agree that the PUC in Calif should have copies of
all reports and comments. Ken or Ron, can you see that the copies be
send to the PUC as the courtesy of HNT ?
Regarding the kind of TDD (soft-copy or hard-copy), please be aware
that there is only one kind of black simple telephone available to
the hearing community regardless of tape recorder or not. In the
deaf world, there is no such simple TDD comparable to the simple
telephone. There is only one kind of voice that can be coming out
of the phone speaker, but there are various printouts that can be
coming out of various TDD devices. You have no choice on what voice
you would like to hear but you have the choice on what printout you
would like to read: LED, CRT Display, Braille, Black Characters on
Paper, etc.
For those Deaf-Blind Telephone Customers, the Telephone Company
would have to provide expensive TDD device that can print out in
Braille, since the law says it must be accessible to all equally.
On the question whether LED and Hard Copy are equally accessible, you
have to be aware that anything that can be readable is accessible, but
not the luxury necessity.
As the matter of fact, if the hearing person listens to the phone,
he has free hands to write down important information. But if the
deaf person uses his eyes to read the soft-copy printout, especially,
the one-line display, he has the hardship to write down the similar
information. And that is not accessible. Therefore, each deaf
person is different from other deaf person, while the hearing
person does not.
Louis J. Schwarz
Board Director, Region III (Md, Va, DC, Pa, De, and WV)
Telecommunication for the Deaf, INc.
------------------------------
From: bmstrazz at DNGC
Date: Tue, 23 Sep 80 10:27-EDT
Subject: spate of comments
The following are my views in response to a spate of messages on the
Calif TDD-PUC thing:
1. To REM at MIT-MC: the reading level of the average adult deaf
person lies between third and sixth grade. coupling this with
the LED character display count of some 10-36 digits (depending
on particular brand of TDDs being used), this is the reason why
the average deaf prefer hard copy.
2. To REM at MIT-MC: there are two companies that manufacture CRT-type
TDDs. Unfortunately both are on shaky grounds and I wouldn't be
surprised if both go under within a short time.
3. To David M. Phillips: perhaps bell 103 and 212 modems are the ideal
for the deaf. Putting all hot air aside, I have not yet seen a
demonstrable 103 or even 212 modem that can "talk" like a regular
weitbrecht device without having to fuss around too much with half
and full duplexes and with originate and answer modes. more on
this later.
4. To KLH: his views certainly merit validity. I only wish to
elaborate further. Seriously existing these days are standards
because there are three possible modems for deaf TDDs-
a) the Baudot Weitbrecht modem
b) the Ascii Weitbrecht modem
c) the Bell 103 modem (assuming future technology renders the
feasibility of this device for deaf talk-tdd purposes)
We hear all the hot air rhetoric about deaf going ASCII; fine but
which ASCII modem???? Some are talking about ASCII Weitbrecht
modem while others talk about the Bell 103 modem. The deaf should
go ASCII, no question about it, but I personally and in my position
as spokes- person for the TDD industry would discourage this move
unless standardizations agreements can be thrashed out. Asides from
Novation, there are a couple more companies pondering entering the
TDD market, but they are all waiting until this such standardizations
agreement have been effected.
As for Plantronics, my immediate criticism is that they would market
their TDD only thru Telcos. The deaf should have an equal opportunity
to pick out the TDD they want. I live in Maryland, under auspices of
the C & P Tel Co, and I understand they are disinclined towards vu
phone. So assuming I would want vu-phone, I would have to rely on
the black market to get myself this such a device.
The vu-phone people incidentally demonstrated their device at a deaf
convention at Cincinnati last July and they were dismayed to receive
such a heavy negative responses from deaf conventioneers about the
inavailability of this device.
To DGSHAP at MIT-AI: hope what is not lost among the hearing people
is the heavy infighting among telephone and communications people
on what is the best for the deaf. The average deaf is totally and
utterly unaware of the infighting going on over and above their heads
by the "gods" higher ups. I would never repeat never put myself in
the position of deciding what is the best for the deaf (even though I
am deaf) without consulting the average deaf man on the street on how
they feel re various TDD devices. Remember if the deaf consumer does
not buy, all of the ascii manufacturers will go under; and the 75,000
odd baudot devices on the market is too formidable a statistic to be
swept aside by these rhetoricians.
To John A. Pershing Jr: you are quite erroneous in saying that TDD
manufacturers, especially the hearing manufacturers, have the deaf
captive market all to themselves. There are eleven TDD companies
on the scene. One is from Canada, so that leaves us with ten. Of
the exalted ten, three are deaf-operated, so presumably some of
them might have a good slice of the California pie.
To Craig W. Reynolds: in case you are unaware, the TDI is the voice
of the Deaf Telecommunications Industry. We have a significant role
in influencing thoughts of people involved with telecommunications.
Right now we are trying to establish standardizations. If you have
views, kindly direct these to me.
Last of all, an irony. A company which should have known better
aroused my ire last week. They called me up on TDD, and requested
that I switch to voice. I told them I am deaf. They asked me to
find a hearing secretary. I told them my secretary is also deaf.
They persisted by asking me to grab a hearing person around to
interpret voice for me. I got very angry and asked them "What
is the purpose of your new TDD that you are planning to market
nationwide very soon if you insist on using voice????"
Moral there, huh?
Barry Strassler
------------------------------
Date: 23 September 1980 2042-EDT (Tuesday)
From: Joe.Newcomer at CMU-10A
On the whole terminals-for-the-deaf controversy, I am surprised at
the number of references to the hearings which implied that all of
the people at the hearings were in fact capable of that (hearing).
Evidence seemed to show that "polls" were taken of the deaf popu-
lation. Were there /any/ technical people with hearing impairment
who were called on to testify? Did the group being represented
indeed have ANYONE who was a member of the group as an active
participant? I know that if I were deaf I would damned well not
want to be restricted to some obsolete technology foisted upon me
by some well-meaning but ill-advised people who themselves thought
they were doing me a huge favor. Did anyone ask any deaf people
to compare the performance of a slow baudot terminal with a 1200
BPS CRT in terms of their own effectiveness at communication?
With 1200 baud, more than just point-to-point typing is possible;
a whole gamut of services not readily available could be made
available...for example, the latest weather information when you
need it (not depending upon what was true when the newspaper was
printed, or even having to go out and buy one). Compare the
information bandwidth of the audio system, and 45.5 baud...I can
probably (opinion, but should be verified) absorb information by
listening much faster than the same information can be typed at
45.5 baud, so if you talk "equality", shouldn't a deaf person be
entitled to get information AT THE SAME RATE as someone who can
hear? 5 chars/sec is NOT equality.
Yes, I'd like to help convince these turkeys that 45.5 baud baudot
is blatant /de facto/ segregation, and that current ASCII based
technology is already cheaper and getting more so. How can I help?
Joe Newcomer
------------------------------
Date: 24 September 1980 0220-EDT
From: The Moderator
The story about equalization is Kurt Vonnegut's "Harrison Bergeron",
which I have listed as having appeared in Edmund Crispin's /Best SF 7/
and Tom Boardman's /An ABC of Science Fiction/. I suspect a Vonnegut
collection of short stories would also include it.
-- <Joe.Newcomer at CMU-10A>
The story that JPERSHING mentioned is from the collection "Welcome to
the Monkey House" by Kurt Vonnegut. The story goes on to detail how
all classes of people are assured of equal service. Smart people wear
earphones that deliver ear-piercing blasts at random intervals (so
that they cannot take undo advantage of their brains), strong people
wear weights, and attractive people are forced to wear disfiguring
masks. The protagonists are two extremely gifted individuals who
doff their shackles and stage a soul lifting dance,.... before they
are shot down by the attorney general.
Now in my mind, the problem was that the sand-bags were too light.
-- Dan <DGSHAP at MIT-AI>
------------------------------
End of HUMAN-NETS Digest
************************
∂24-Sep-80 0806 FFL
To: JMC, FFL
Of course they wanted the whole speech, but I don:t thing the length really
matters. It is for their media relations person who, after receiving your
material, wants to interview you on the telephone and then she will decide
wkhat to release to the media. I suppose she's looking for something provo-
cative/newsworthy.
∂24-Sep-80 0810 FFL
To: JMC, FFL
Do you remember that a Professor Iwata is coming on the 25th and 26th to
see you. You asked Tom Binford to make arrangements for him to visit in
the department.\
∂24-Sep-80 0908 Baskett at PARC-MAXC Re: Evans & Sutherland
Date: 24 Sep 1980 09:05 PDT
From: Baskett at PARC-MAXC
Subject: Re: Evans & Sutherland
In-reply-to: CSD.TAJNAI's message of 23 Sep 1980 1448-PDT
To: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
cc: JMC at SU-AI, FB at SU-AI, CSD.Dbrown at SU-SCORE
I received the same letter. I know Dave Evans and he is a great fellow. But I
don't know Teri Adams and I am not especially interested in meeting with her,
especially since E & S is not a member of the computer forum. I'll bet that they
would join if she asked Dave to have them join. So keep up the pressure.
Forest
∂24-Sep-80 1032 FFL
To: JMC, FFL
A Paul Horwitz who was referreed to you by Prof. Hans Bethe called. He
works in area of aapplication of computer science to nuclear safety and
would like to speak with you. He will call back about 3 p.m. (6 p.m
his time), hoping that you will be available then.
∂24-Sep-80 2252 JK
one more comment on the takeuchi function:
I can prove that the tak(x y z) converges for all reals iff
tak(x y z) converges for all x y z s.t. one of x-y, y-z, x-z
is an integer.
This might be of use since the latter condition is an invariant of
the inductive definition. Of course, if two of the differences are
integers then we already know that tak(x y z) converges.
I have put a copy of Nozaki's paper on your door clip.
∂25-Sep-80 0131 RWG at MIT-MC (Bill Gosper)
Date: 25 SEP 1980 0431-EDT
From: RWG at MIT-MC (Bill Gosper)
To: jmc at SU-AI
Any new thoughts on Volterritoriality? If you come over
again, you might try to bring a copy of Bernard Cohen's (?) plutonium
snorting challenge, which i'd like to xerox and (digitally) mail to energy@mc,
unless you feel otherwise. Also, did you see this in 30Aug TimesTrib?
NOTICE OF DRAFT ENVIRONMENTAL IMPACT REPORT
A Draft Environmental Impact Report for the Menlo Park/East Palo Alto and
Districts Sphere of Influence Study is available for public review in the
office of the San Mateo Local Agency Formation Comission, 590 Hamilton
Ave., Redwood City, CA. Reviewers of the EIR should explain the basis for
their comments and wherever possible, should submit data or references in
support of such comments. It is suggested that the staff of the Formation
Commission be supplied the name of a contact person who is available for later
consultation should this prove necessary.
The review period has been set by the Executive officer at 45 days from the
date of distribution. The last day for the submittal of comments shall
therefore be FRIDAY, OCTOBER 10, 1980.
Huh? If you need an EIR for a STUDY... The real environmental threat is the
idea of an Agency Formation Commission--the bureaucracy has grown genitals!
(I'll try to call them when I get in phase.)
It is Bernard Cohen, and I'll try to bring one. I have several things I
have to do before returning to the Volterra problem. I'm puzzled by the
EIR. I assume it's Federal, but what Federal action is involved?
∂25-Sep-80 0539 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #85
Date: 25 SEP 1980 0820-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-AI
Subject: HUMAN-NETS Digest V2 #85
To: (@FILE [DSK:DUFFEY;DUFFEY HNTXX6]) at MIT-AI
HUMAN-NETS AM Digest Thursday, 25 Sept 1980 Volume 2 : Issue 85
Today's Topics: IBM General Robotics Division,
Electronic Banking - Cashless Society & Possible Services,
Computers and the Deaf - TPC Standard Query & TDD Design,
Home Information Retrieval - Radio Shack Videotex
----------------------------------------------------------------------
Administrivia: A reminder
Recently, several suggestions have been made for distributing copies
of the HUMAN-NETS discussions to various people outside this group,
especially with respect to the COMPUTERS AND THE DEAF discussion.
The primary problem with distributing copies of these discussions
to an outside agency, such as the CA PUC, is that it will draw
attention to us and more dangerously to the ARPAnet itself.
Another problem, which we cannot casually dismiss, is the rights of
the people who wrote those comments. While HUMAN-NETS is a type of
public forum, I question whether anyone has the right to distribute
those comments outside this forum without obtaining the permission
of the originators.
The value of HUMAN-NETS is that it brings together a wide range
of knowledgeable people to discuss the issues and to share their
expertise. These discussions can lead to constructive action in
several ways: by leading to discussions at conferences or public
meetings; by pointing out where information, resources, and
people are available; and by simply fostering greater awareness
of these issues. However, we cannot seek to distribute these
discussions outside this informal network.
-- RDD
------------------------------
From: Nomdenet at Rand-Unix
Date: 24 Sep 1980 at 1121-PDT
Subject: IBM and Robots
From the "Look Ahead" Section of Datamation for September (p.13):
A new IBM division may soon be unveiled out of GSD [General
Systems Division] in Boca Raton. Project White Cloud, IBM's
internal development group for industrial robots of the sort
used to assemble CRTs in Raleigh, has reportedly been reorga-
nized as the General Robotics Division. Scuttlebutt is the
new division marks a major commitment by IBM in an emerging
marketplace. For now, a small GRD sales force will sell only
internally to IBM.
A. R. White
------------------------------
LLOYD@MIT-AI 09/25/80 00:12:52 Re: Cashless society
I like cash. It's almost untraceable.
Brian Lloyd
------------------------------
Date: 25 Sep 1980 (Thursday) 0055-EST
From: GERMAN at HARV-10
Subject: more on EFT
In my opinion, the thing that makes widespread EFT unattractive is
not the possibility for fraud, which is certainly there, but the
likelihood of aggravating errors. Mixups do occasionally happen
with checks, but important pieces of paper do not often get lost,
duplicated, or garbled in storage and transmission. Just for the
sake of argument, suppose there was a large national EFT network
and that it and its hosts were about as reliable as the ARPA net.
Now you might say that they would try to do better than that, but
assuming that it is built in the near future and considering the
usual standards in commercial data processing, I would have my
doubts. It helps to think of it as putting your money under
the control of a huge automatic billing system with hundreds
of thousands of terminals. They are likely to adopt rather
low standards for all the usual reasons ("We have to save
costs or pass them on to the consumer...").
The ARPA net is wonderful, but you have to be careful with it.
Messages and files go wrong often enough that I would not use it
to do my banking. And when something does go wrong, you can fix
it. I would suggest that one of the reasons the ARPA net is usable
is that it is not operated by a large bureaucracy. If you dont
like the way it is working on a given day, you can check things
for yourself. Do you really expect Bank of America's computer to
warn you, "System is Flaky -- Watch Out?" And if something does
go wrong, you may not find out for a few weeks. Then, without
written records you will have a hell of a time proving your case.
You may like the idea of saving a few cents on postage, but if
you value your time the benefit can be wiped out if there is even
one mixup a year. Customers should also beware of paying for EFT
indirectly, for example, through a high required minimum balance.
To tie up $1000 at only 5% costs you maybe $25 a year assuming
that you would normally keep only a few hundred in a checking
account (in New England, checking accounts can pay interest)
and the rest in something paying a higher interest rate.
------------------------------
Date: Wed, 24 Sep 80 08:22-EDT
From: bmstrazz at DNGC
Subject: a question and a few comments
Putting aside temporarily the heated ASCII-BAUDOT talks, I need
advice and perhaps one of you can help.
Yesterday I was visited by an elderly deaf couple of Hague, Virginia.
This is a small town that I had never heard of until it was mentioned
to me. The telephone company serving this town is the Continental
Telco of Virginia.
They bought Phonics TDD and were dismayed to see that it did not
work. They made the rounds of telephone companies and even television
repair shops around, and for all of their efforts, these people just
scratched their heads and said sorry. Frustrated, they drove all
the way up to Washington DC area just to see me (a trip of some 450
total miles) and explained their problem. Their unit was tested on
the premises and there was nothing wrong with it. I am no engineer,
or even worse am no technocrat, but I suspected the problem did not
lie with their TDD but with the differing standarizations of that
Continental Telco of VA which rendered incompatible that TDD if
attempted to bring messages across towards another Telco territory???
I understand USA has 1,500 different Telcos and many of them have
their own standards, and which is not bad enough for voice calls
but bad enough for TDD-created calls.
Any advice on this?????
now for my comments:
To Joe Newcomber - why are we Baudot and not ASCII was not a
decision of our choice but of necessity; Robert Weitbrecht, the
inventor of the first deaf TDD (in 1964) chose Baudot for two
reasons:
1. ASCII data transmission was in its infancy and not much
was known about it at that time
2. Baudot machines at that time were plentiful and could
be had for the asking and thus much easier on the deaf
pocketbook to go Baudot instead of ASCII.
To DGSHAP at MIT-AI - We, the deaf are not race horses demanding
that better bred horses (the hearies) be burdened with weights so
as to handicap them down to our level. By the same token, you cannot
train a crippled man to be an olympic sprint champion. We, the deaf
are not asking for equality in the sense that the hearies be burdened
with some equalizing handicap. We only ask for equality in terms of
accessibility and to us these are TDDS, and interpreters among other
things. Do not even forget these simple fundamental needs of our
people.
Barry Strassler
------------------------------
Date: 24 Sep 1980 1423-EDT
From: BERN at MIT-DMS (Bern Niamir)
Subject: deaf access equality?
It is amusing to hear (actually read) that CPUC and the TDD lobby
are advocating equal access when this is practically impossible and
access by deaf people has always been and will still be less than
that of hearing people. Even the most sophisticated TDD equipment
will not be able to communicate with all the organizations and
residences on the telephone network. Would the deaf person with
his ASCII terminal and multi-standard modem be able to call up
the bank, his broker, the local drug store, or a hearing friend?
Probably not for the foreseeable future. Is the equality seeking
PUC willing to put a TDD in every home and office?
The only way to partially compensate for this inherent inequality
is to allow access to computers and computer networks. This is
most important too because the average home and the average office
will gradually and eventually establish access to computers for
the reason of the economies involved and not because of the need
to talk to the deaf.
------------------------------
Date: 24 Sep 1980 1830-PDT
From: Craig W. Reynolds (at III via Rand) <REYNOLDS at RAND-AI>
Subject: TDD voucher plan
One way of accommodating both sides of the what-type-of-TDD-should-
be-provided-by-the-state debate is to allow the bureaucrats to
select a (presumably lo-tech undesirable and old fashion) TDD
as the "standard"; then rather then buying them for the deaf, a
voucher for the expected purchase price would be given (with no
strings) to the individual deaf.
Since (as has been mentioned here in HN) the "offical" choice may
well turn out to be not only slow and non-ascii, but more expensive
to boot - this should provide a large enough allowance (via the
voucher) to pick up a nice, fast, ASCII device. The main drawback
I see with this plan (there must be others) is that for the "naive"
deaf (ie not computer hackers) it may be too much trouble to select
a device off the open market, but they can fall back on the offical
recommendation.
On that story by Vonnegut (Harrison Bergeron) - as I remember it was
not the "attorney general" who was the heavy - rather the office was
called the "handicapper general".
------------------------------
Date: 24 Sep 1980 0922-PDT
From: ROODE at SRI-KL (David Roode)
Subject: Question
Does anyone know how many Baudot Modems, and how many Bell 103 Modems
and how many Vadic triple-threat Modems there are in SERVICE in the
world? in the US?
------------------------------
Date: 25 September 1980 01:35 edt
From: Frankston at MIT-Multics (BOB at SAI-Prime)
Subject: Free enterprise?
In answer to david@UTEXAS,
The basic problem with the TDD issue is that of local optimization
versus global optimization. With our technical expertise we can
see global issues that the users do not necessarily see when making
purchase decisions. What's worse is that a PUC may be required by
statute to make local optimizations to minimize immediate cost.
My consolation is that when computing is available to the masses, the
mass market technology might be so cheap that the existing machines
won't be an issue. Rather than petitioning the PUC, perhaps we should
try to get Radio Shack to include a dual modem in a version of their
Videotex for the deaf. (Hmm, wonder why they weren't at the hearings)
------------------------------
Date: 23 Sep 1980 1842-PDT
From: ROBERTS at USC-ISI
Subject: videotext and tdd's
The discussion of the communication needs of the deaf coincided with a
recent visit I happened to make to my neighborhood Radio Shack store,
where I discovered an upcoming product that could have a significant
impact on the development of computer networks for use by the general
public, including the deaf. it seems that in November Radio Shack
will be making the debut of its TRS-80 videotex (sic) terminal. A
few quotes from their sales blurb follow:
"A TRULY AFFORDABLE TWO-WAY INFORMATION RETRIEVAL SYSTEM --- The
TRS-80 Videotex terminal works with any telephone to turn ANY TV set
into a 2-way information retrieval system. It features an alpha-
numeric keyboard and 4K memory that allows you to store information
for later viewing -- cutting 'on-line' time to a minimum. The
videotex terminal provides fully interactive communication, so you
can answer questions and exchange ideas, or even play computer games
with other videotex users. Videotex software is built-in, and we
include one free hour on the CompuServe information service with
purchase....It's easy to install -- simply plug in a telephone cord,
plug into an AC power outlet, and connect to the antenna terminals
of your TV set. Available November 30 at over 6000 Radio Shack
stores nationwide..."
"The TRS-80 videotex concept. Imagine having a public library,
that never closes, right in your home or office! Or volumes of
stock market data at your fingertips! Picture a continuously
updated newspaper you don't have to fetch from your doorstep!
Or a low-cost electronic mail service!! ... Radio Shack's TRS-80
videotex format -- consisting of sixteen 32-character lines
PLUS color graphics -- makes it all possible. Videotex makes
interactive information retrieval a reality ANYONE can take
advantage of--today!..."
They then go on to talk about the many services which are or will
be available using this terminal. They also talk about the deal they
have worked out with CompuServe, and all the wonderful services and
info you can get from them, including electronic mail. They also talk
about a 300 baud interface, so I assume it runs at 30 cps. The price
for all this, including the modem? $399.00.
This device, assuming that it is introduced and performs as
described, poses some pretty interesting policy questions, it seems
to me. Radio Shack claims that 90% of the US population lives within
5 miles of one of their stores, and since I've got 3 or 4 within that
distance, I'm inclined to believe them. Given this distribution
network, and also given that they've sold over 200,000 of the TRS-80
computers, my initial feeling is that Radio Shack may well set the
standard for this type of terminal, simply because of their marketing
power, independent of the merits of the terminal. This could well
mean that work now under way by various bodies trying to determine
generic videotext and teletext standards could well be preempted by
this machine. 16 lines of 32 characters may be the public standard
for a data page from now on.
Undoubtedly the price of this machine will fall in the future,
if the price history of the TRS-80 is used as a guide. This may
make it an attractive terminal for the needs of the deaf, specially
considering the built in 4k of memory and the ability to display
16 lines at once, as opposed to just one. A couple of thoughts
occur to me, however. Undoubtedly, this terminal works only in
an originate-only mode, so how do you have direct terminal to
terminal conversations? On the other hand, if the deaf had wide-
spread access to an electronic mail service, how much real-time
interactive dialogue would there be? (Perhaps someone from
DEAFnet can supply some data on this)
In a somewhat related vein, I would think that one of the
problems that deaf people would (do) have is having to put up with
high telephone bills when they call long distance, because it takes
longer to type out what you want to say than to actually speak it.
In this connection, using a packet-switched network like Telenet
might be considerably cheaper, since your cost would be based on
the volume of characters sent, plus your local connect time. This
would seem to be another argument for the deaf having access to an
ASCII compatible terminal.
All in all, this latest Radio Shack development is thought
provoking. This is true on many levels, including such trivia as
why doesn't it have a control key, when the keyboard looks like a
regular ASCII model, more or less? You can find a picture of it
in the TRS-80 computer catalog, p.35.
Ollie Observer
------------------------------
End of HUMAN-NETS Digest
************************
∂25-Sep-80 0900 JMC*
call bank
∂25-Sep-80 0956 PJH via ROCHESTER
John, I tried to see you before leaving California, but I couldnt
find you in. I just wanted to thank you for getting me invited
to one of the most interesting years I have ever , or probably
will ever, spen(d/t). I have a sense of regret that you and I
didnt spend more time talking about the issues that are in the
center of our work, and I wanted to say that I look forward to future
chances to get together on technical matters, and lets
keep in touch. We are both terrible correspondents, but there is always
the net; and lets stay on one anothers mailing lists for memoranda
and working papers, etc..
bets wishes to Carolyn and yourself from Jackie and me
pat
∂25-Sep-80 1107 CSD.JEANIE at SU-SCORE Faculty Meeting Agenda
Date: 25 Sep 1980 1101-PDT
From: CSD.JEANIE at SU-SCORE
Subject: Faculty Meeting Agenda
To: CSD-Faculty: ;
cc: csd.bscott at SU-SCORE
Agenda for Faculty Meeting September 30:
AGENDA ITEMS
Reports:
1. Space (Brown)
2. Teaching Loads (Brown)
3. CIS (Ullman)
4. Degree Candidates (Tajnai)
5. Math Sciences Committee (Schreiber & Herriot) (A)
Old Business:
1. Policy on Report Distribution (B)
New Business:
1. Licensing Policy (C)
2. Travel for Thesis Readers (D)
3. Split of Visiting Committee
4. Use of IBM Grant
A,B,C,D refer to materials being sent through the mail regarding the
corresponding agenda item.
-------
∂25-Sep-80 1155 FFL
To: JMC, FFL
Betty Scott would like to speak with you when you come in today.
∂25-Sep-80 1317 CSD.BSCOTT at SU-SCORE NSF Supplement for Doyle
Date: 25 Sep 1980 1303-PDT
From: CSD.BSCOTT at SU-SCORE
Subject: NSF Supplement for Doyle
To: JMC at SU-AI
cc: CSD.BScott at SU-SCORE
Had a call today from NSF, John. They want to know whether you want to
withdraw the supplement (it was never formally submitted as a new proposal
as they requested) for Jon Doyle's salary. Hersche put Jon on the payroll
against ARPA funds. NSF knows this--probably found out through auditors
or from the accounting office here.
How do you want me to respond to NSF?
Betty
-------
Who called from NSF? I didn't realize that it was in any way alive.
I have to submit a new proposal for my own support starting next June,
and I'm working on it. I want to discuss whether to include Doyle
activities in it. I intend to phone them tomorrow.
∂25-Sep-80 1443 CT
∂25-Sep-80 1440 JMC dover
Until program tells you that it has done the listing, the writeup
should tell you that the program doesn't.
CT -- right. sorry about that; I was suppressing all display from DOVER
FOO.BAR mode.
∂25-Sep-80 1441 JK Nozaki's paper
Thanks. I am a little confused - he seems to claim that he already has it
for all reals?
Hmm. I must have forgotten that.
∂25-Sep-80 1859 Denny Brown <CSD.DBROWN at SU-SCORE> CS200 - request for speakers
Date: 25 Sep 1980 1856-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Subject: CS200 - request for speakers
To: CSD-Faculty: ;,
CSD-Research-Associates: ;
cc: csd.tajnai at SU-SCORE, sis at SU-AI
Informal description of CS200 for potential lecturers
CS200 meets Thursday, 2:45 to 4:00, in 420-041 (Psychology) during fall
quarter. I give the first lecture (on 10/2). There are nine more class
meetings. (Skip 11/27, finish 12/11). Contact me or Carolyn to request a
date for you or your group to do a presentation.
CS200 is intended to be an introduction to the research being done in the
Stanford CS Department. It is intended for entering students, in both the
PhD and Masters programs. The course is organized as a series of lectures
by CS faculty members. One unit of pass is available for attendance. The
course is intended to be (primarily) a forum for new students to meet
faculty members and to learn about current faculty interests and
activities. It is also a good way for a faculty member to attract
students to a pet project.
Each week a faculty member will be invited to present a lecture on what
his/her field is all about, what s/he is doing, and what sort of ongoing
projects s/he is involved in. The style of each lecture will certainly
vary depending on the speaker and his/her field of interest. I will
suggest the following questions as typical of the sort a speaker might
want to address.
Definition/description of the field of interest. What are the classical
questions of the field? What are the classical tools applied to discover
answers to the questions? In what courses, in what books, can one find a
good introduction to the field? What is a representative open problem?
(I.e. what is the speaker working on right now?) What contribution, if
any, can a new grad student make to a research project? How does a
student get involved in project activities? (E.g. SIGLUNCH for those
interested in HPP activities.) Can someone whose primary interests are
not in the field make a contribution to the research? (E.g. Graphics
programming in the NA world.) Where is the interesting research going on?
Who is doing it? How does the Stanford CS dept. stack up compared to
other places?
-------
∂25-Sep-80 2003 TOB
John
Please comment on whether this will serve for the note that
you asked me to write for the facully meeting.
I made it brief, assuming that they are busy people and want a concise
outline. I could flesh it out with details of our program, industrial
contacts, anecdotes making concrete what are summarized below.
Tom
This note presents the case that the computer science department
should participate in a strong way in an institute for
automation and productivity research.
Most of us think that computer science is important to a lot of fields,
but we don't necessarily believe that computer science should be
involved with them.
The essential considerations for consideration for an academic
program are basic research problems in the field, support for
research; education programs; training and hiring of graduates;
and relations with industry.
I think that computer science is important to automation,
and I think that automation is important to computer science.
We propose a leading role in an institute here at Stanford, together with
Mechanical Engineering, Aeronautics and Astronautics, Business, and
Electrical Engineering. SRI is also interested in collaborating.
We plan to do for mechanical production what computer science
and computer systems research plan to do for VLSI.
VLSI generates interesting research problems. Technical problems of
VLSI production do not seem soluble by "more of the same", and despite
substantial industrial research in the area, VLSI research has
an important place at Stanford in the long payoff / high risk / high payoff
area of basic research and experimental system
building. The same is true of production of mechanical systems.
As in VLSI, fundamental computer science problems stand in the way
of increasing industrial productivity. Automation research generates
good academic problems and provides extensive support for research to
solve them.
In productivity research, computer science
lies at the heart of major research issues. CAD/CAM is a major
theme of industrial automation research, i.e. computer-aided design/
computer-aided manufacturing. These are not just
using computer science technology, but bring major research
issues in productivity.
Within computer-aided design, key research issues are:
geometric modeling;
numerical analysis for segmentation and surface representation;
representation of function for creative design;
formalizing geometry;
computational geometry (analysis of algorithms for intersection,
hidden surface);
computer graphics,
expert systems for utilizing knowledge of design experts and experts in process
planning,
planning systems and problem-solving.
Computer-aided manufacturing introduces research areas in:
perception and cognition in robotics;
analysis of scheduling and resource allocation algorithms;
(Rivest and students have worked on this at MIT)
planning and error-recovery;
language and programming environments for robotics;
concurrency and synchronization;
real-time, real-world non-deterministic systems;
distributed systems;
I find that many people concerned with automation are concerned about
computer system problems; discussions are remarkably similar to those
centering around ADA.
Computer science has a central role in a teaching program
for modern productivity research. McCarthy has suggested that an MS
program in manufacturing center about robotics and numerical control,
disciplines which would provide students with the basis to apply to
other problems they encounter.
Graphics is a significant part of a computer science curriculum; while
the robotics group has work in this area, we should expand our effort and
provide a graphics course.
Computer science course series for PhD students in automation would
overlap with the MS program in computer science: systems courses,
AI, numerical analysis.
The automation program would bring stronger contact with a large segment
of industry which now has little contact with computer science.
Stanford has long had close contact with local industry. There is
now national interest in university-industry cooperation.
Manufacturing of mechanical things is a different segment of industry
than computer forum.
∂26-Sep-80 0354 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #86
Date: 26 SEP 1980 0643-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #86
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Friday, 26 Sept 1980 Volume 2 : Issue 86
Today's Topics:
Electronic Banking - Security/Reliabity & Cashless Society,
Home Information Retrieval - Radio Shack Videotex
Computers and the Deaf - TDD Design & CA D.E.A.F. Surcharge
----------------------------------------------------------------------
REM@MIT-MC 09/25/80 09:22:53
I share concerns about EFT having programming bugs, thus I don't want
to jump into it 100 percent any more than I want to jump 100% into
nuclear power just because it seems safe (see ENERGY at MIT-MC for
those discussions) or into anything else. But I'd like to see EFT
moving faster than it is so I can put maybe 10% of my funds into it
and see how it works. If it crashes, I lose access to 10% for as
long as it takes to resolve it. As it proves its safety, I'd shift
more and more into it.
Comparing to Arpanet mail has a flaw. Whereas Arpanet mail gets
created when it is typed, duplicated at mailing-lists or manual
forwarding, and consumed at the destination; money is supposed
to be conserved. If ten dollars is debited from my account, it
better damn well appear in some other account via a legitimate
transaction or as physical money in my hand. If ten dollars is
credited somewhere, it had better been debited somewhere else or
else physically deposited. Assuming there are physical receipts
at deposit and withdrawl spots, all the rest of the EFT should
exactly balance, and the slightest misbalance would trigger an
audit to find out the bug in the program. If some records are
lost, but it is known that Robert Maas's account was debited by
$35.72 on a given day, which he (I) claim(s) was payment of a
dentist bill, and if the dentist's account shows an unexplained
$35.72 credit on approximately that same day, it can pretty well
be assumed that my bill was paid and the computer should be
corrected to show it paid. If the dentist's account wasn't
credited, but his bank showed an unexplained $35.72 credit, it
can reasonably be concluded that the bank owes the dentist that
amount on behalf of my dentist bill. I think with the usual
reports of misbalances, and a computer program to take all the
misbalances and compute a likely scenerio and a proposed fix,
that EFT can be made reliable. I'd like to see if I'm right
someday...
------------------------------
Date: 25 Sep 1980 1000-PDT
From: Chesley at SRI-KL
Subject: EFT reliability
I can't resist putting in a few cents worth on the question of EFT
reliability, since I worked for Bank of America's telecommunications
group for a while.
What they have at the moment might be described as closer to a
fancy telegraph system than to EFT, but it's getting closer every
day, and is used for funds transfer messages (huge amounts of money
are transfered via the system every day).
B of A is very paranoid about reliability, and VERY VERY PARANOID
about accountability. A message may go thru several machines along
the way (including two 11/70's, two General Automation machines, and
a soon-to-be-in-service Tandem). Reliable transmission is done on a
link-by-link basis (i.e., not end-to-end), with each machine "safe
storing" the message on disk before acknowledging receipt to the
sending machine; some of the machines keep two copies on two separate
disks in case of hardware trouble with one of the disks. Each machine
journals every message on a separate disk (again, some keep two copies
on separate disks). Once a day all the journaled messages are moved to
tape, and sent to the giant IBM machine, where they get cross-indexed
six ways to hell and put on microfiche. Then the microfiche is stored
in a safe place (as far as I could tell) forever.
The intent of all this is three-fold: (1) to keep the auditors
happy; (2) to be able to track down customer complaints (e.g., "what'd
you guys do with my $20,000??!!"); and (3) to be able to determine
specifically which machine is buggy in the event of a problem.
Then there's security, but B of A probably wouldn't like me going
into detail there.
So I wouldn't worry about security and accountability (at least
with B of A). If anything, their concern in that area is keeping
them from seeing better solutions to problems (for instance, I feel
that they'd be better off using end-to-end techniques).
--Harry...
------------------------------
Date: 25 Sep 1980 1200-PDT
From: ROODE at SRI-KL (David Roode)
Subject: Cashless society
Since I got an automatic teller card 3 years ago, I find myself
dealing MORE in cash. I no longer bother writing a check in a
retail store, because I can more easily get cash out of the teller.
However, I consider this a step toward a cashless society, because
it is a step toward the mode of operation modeled by those automatic
gas stations where you buy tokens and use the tokens to operate the
pumps. The cash is merely an ephemeral expedient in completing my
transaction. The time-weighted average of my cash content can be
said to have decreased. It is also noteworthy that more and more
stores take plastic money. There are two different grocery stores
I know of that do.
What I think will be the next step is the intercommunication of
each bank's ATM's with those of some, several or all other banks.
I understand this has long been the practice in Japan, and I have
heard rumblings from UCB about doing this in the U.S., at least
with banks in other states where they are forbidden to operate.
Compare to Massachusetts where I understand Banks are limited to
single-county operation, but there tend to be different banks in
different counties owned by the same holding company which all
cooperate (within the holding company) on ATM's and other matters.
Since these will be independently operated segments of a global
network, I think reliability will be less of a problem than some
people have mentioned lately.
The following logical step is for, say, Grocery stores to install
ATM's for customer convenience on a widespread basis, and then it
will be a simple matter for various of these installations to be
modified so that customers can dispense with dealing with the
tokens in favor of cash register to ATM communication. The
tokens I refer to are paper money.
------------------------------
REM@MIT-MC 09/25/80 09:40:38
That Radio-Shack Videotex for $400 total sounds nice. It doesn't
have 24 by 80 text nor answer-mode apparently, but for $400 it sounds
decent. Considering it to be equivalent to a $200 TV-typewriter with
a $200 modem, plus the 90-day warantee thrown in for free and a store
within 5 miles of you, sounds quite tolerable for anyone who can't
afford $600-$800 for a "real terminal" plus $200 for modem or $400 for
auto-everything modem. ($800-1200 total) Maybe CPUC should designate
Videotex as the default terminal to give the deaf, if it lives up to
its promotion.
------------------------------
Date: Thu, 25 Sep 80 17:00-EDT
From: bmstrazz at DNGC
Subject: more comments
To Roberts at USC-ISI
Regarding the Videotex of Radio Shack; I wrote them asking them
to see how if any can it be applicable to the deaf. That was two
months ago. I am still waiting for their reply.
To Roode at SRI-KL
To be on the conservative side, there are about 50,000 Weitbrecht
Baudot modems, but if one wants to be liberal with figures, then the
count is 75,000. No one knows exact figures. As for number of Bell
103 modems, leave me out; its not my ball game.
The deaf people of Washington DC, San Francisco, and Boston have our
own electronic mail systems. Four possibilites are in store for our
ASCII future:
1. value added networks
2. source or some kind of electronic mail systems
3. phone company - their lines
4. micro-system network
Barry Strassler
------------------------------
Date: Thursday, 25 September 1980 15:10-EDT
From: John A. Pershing Jr. <JPERSHING at BBNA>
Subject: ASCII vs. Baudot; Bell103 vs. Weitbrecht
To Barry Strassler: The purpose of my "leg weights" analogy wasn't
to suggest that "hearies"[sic] be given earplugs; I was suggesting
that "equality" is an ethereal notion which tends towards silliness
as we try harder and harder to take the notion literally. "Equitable
access" is a much better slogan.
Also, from your own comments (along with those of others), it sounds
like the TDD manufacturers do, indeed, have a captive market. If all
other deaf people have Baudot TDDs, then someone wanting to purchase
such a device is tied to the dozen-or-so TDD manufacturers (DEC and
IBM surely cannot provide such a device). Just because some of these
manufacturing firms are run by deaf people doesn't seem to be a valid
reason to saddle the deaf community with Baudot/Weitbrecht for the
next couple of decades.
Which brings me to my flame for today... What should (can??) be done
in a situation where there is (a) substantial investment in technology
X, but (b) X is being quickly outdated by techology Y? Granted, most
(all?) TDDs currently in the field are Baudot/Weitbrecht; however,
ASCII has essentially taken over the communications field (and, at
300 baud, Bell103 seems to be the de-facto standard).
One possibility is to continue with Baudot. There's lots of 'em out
there, and there ARE a few vendors. (By the way, what is the story
on TDDs in other countries? Do they exist? Are they Baudot?) If
the idea catches on, then the market may actually be able to support
the entry of a couple of more vendors (unless the PUCs make that
impractical). Unfortunately, if I want to communicate with a deaf
person, I either have to buy a second terminal, or build a widget
that interconverts ASCII and Baudot. Analogously for a deaf
programmer who wants to work at home.
Or... We can adopt a new (possibly de-facto) standard, such as
ASCII, which seems to have a chance of surviving for a few years.
Similarly for the modem protocol. In this case, owners of (current)
TDDs lose, unless someone produces an ASCII/Baudot converter widget
fairly cheaply (anybody got a good guess on the cost of one, using
current-technology chips?). This is a pretty "heavy" tradeoff --
adopt Baudot and keep the deaf segregated from the (ever-growing)
community of terminal owners; or, adopt ASCII and ostrasize owners
of old TDDs.
How "standard" is ASCII (and Bell103) anyway? Is it in use in Europe,
or do they have their own standard code for information interchange?
Perhaps the communications field is moving so fast right now that this
whole issue should be postponed (e.g. held at the status-quo) for a
few years, at which point we adopt X.25 (or whatever). It very well
could be the case that Bell103 is about to become as outmoded as
Weitbrecht...
-jp
------------------------------
Date: 25 Sep 1980 2010-PDT
From: Stevan Milunovic <Milunovic at SRI-KL>
Subject: Talking Chip
In the first issue of DISCOVERY on pg. 82 is a brief article
describing a "talking chip" developed by Votrax Co. of Troy
Michigan. The chip is capable of synthesizing speech from typed
input by transformation to phonemes. The chip sells for $12. An
envisioned product is a hand-held device with a keyboard and
built-in loudspeaker. The device would accept typed input then
produce the synthesized speech (which sounds slightly clipped)
when the 'speak' button is depressed. The user should be able
to "speak" as fast as he can type. Now all that's missing to
permit the deaf population to easily communicate with people
via voice is the reverse transformation i.e. voice into text.
Before locking the deaf into data-only devices for communication,
perhaps devices such as these should be added to permit voice
communication as well. The value of any mass communication
technology is proportional to the interconnections provided,
and since more people still predominantly use voice such devices
would have more intrinsic value than any ASCII terminal. After
all, I'm sure the deaf would like to communicate with society at
large - not just those with data terminals at their disposal.
------------------------------
Date: 24 September 1980 1154-EDT (Wednesday)
From: Robert.Frederking at CMU-10A (C410RF60)
Subject: Minimum possible TDD; McCarthy's new organization
I would just like to chime in with those who say that what
the commission wants for deaf people is EXACTLY what is necessary
to bring them supposedly equal access and no more. This is a general
trait of our society. "I'll be damned if I pay for someone with my
TAX DOLLARS to have something really good, instead of just adequate."
This has occurred in other areas, most notably with public housing
for poor people, where the government REALLY does prefer to pay more
to get something equal than a little less for something superior.
I think it is probably all tangled in with the work ethic, and not
letting someone get something for free that someone else would work
to get. I brought this up in an earlier message in an argument that
automation won't by itself help poor people, unless a very strong,
deeply rooted set of values is expunged from the average American.
I present this as a piece of evidence.
On a vaguely related topic, John McCarthy's organization to
promote progress is completely unnecessary, whether you like the
idea or not. Progress has very powerful driving forces behind it
(including lots of profit for lots of people), and you generally
need a massive show of force to stop even a small project if it
represents "progress".
------------------------------
KLH@MIT-AI 09/26/80 01:27:06 Re: Meanwhile, back ATT the stockade...
I should have mentioned this before; thanks to Earl Craighill
for the memory jog. In late June we were at the "opening ceremony"
PR bash that AT&T put on for its new TDD operator service (mentioned
by Marotta in a previous HN issue), and some AT&T brass showed up.
One of the questions put to them was "what do you think about the
California SB597 bill?".
The answer was quite definite, and is rather interesting. In
essence, they uniformly expect nothing to happen... "don't worry
about that, it won't go anywhere". Their expectation is apparently
that the public, or certain consumer groups fighting utility rate
increases, will scream in protest and join battle in the courts to
either delay implementation for years or rescind the bill altogether.
Any bets? If that's true, why aren't these groups fighting the
ZUM plans, which according to Lauren are being well documented as a
way to increase profits? Or did they really mean that if matters went
beyond a certain point, AT&T would maternally step into the courtroom
and sew things up?
I dunno. I have a hard time believing everything certain people
say, but there you are. I suppose there's no knowing which datums are
useful to Kremlin watchers...
--Ken
------------------------------
End of HUMAN-NETS Digest
************************
∂26-Sep-80 0803 CSD.BSCOTT at SU-SCORE
Date: 26 Sep 1980 0802-PDT
From: CSD.BSCOTT at SU-SCORE
To: JMC at SU-AI
In-Reply-To: Your message of 25-Sep-80 1417-PDT
A Mrs. Bolding called from NSF about the supplement. She said it should
have been officially withdrawn if you didn't want it considered. Shall I
do this? Mrs. Bolding's number is (202) 357-7345. Sorry I don't have
her full name.
Betty
-------
∂26-Sep-80 1030 JRA
price 22.50 us, can, and mexico; o.w. $25
This is for the LISP conference proceedings.
∂26-Sep-80 1330 FFL
To: JMC, FFL
William Stockton of the NY Times, who spoke with you early in September,
called to say he would be in Palo Alto October l6, all day, and October l7
in the morning. I made an apppointment with you for him at 3 p.m. on
October l6. If that is not alright, I can phone him to change it.
(212 556 714l\
∂26-Sep-80 1412 CSL.CRC.LYD at SU-SCORE Visitors
Date: 26 Sep 1980 1407-PDT
From: CSL.CRC.LYD at SU-SCORE
Subject: Visitors
To: CSD-Faculty: ;
September 26, 1980
The following persons are visitors to the Center for Reliable
Computing, Computer Systems Laboratory, during Academic Year
1980/81. They are located in ERL226D, ext. 71448, if you would
like to meet with them and discuss areas of mutual interest.
Mr. Hao Dong
Mr. Hao Dong is a visiting scholar from Harbin Shipbuilding
Engineering Institute, Harbin, People's Republic of China and he will
be here for approximately one year. He taught courses on digital
circuits and the design of combinational and sequential circuits. He
has also designed a control computer and other electronic systems for
automatic control. He is interested studying fault-tolerant design
techniques such as concurrent testing and self-checking.
Mr. Fumiyasu Hirose
Mr. Fumiyasu Hirose is a visiting scholar from Fujitsu Laboratories,
Ltd., Kawasaki, Japan. He is visiting the Computer Systems
Laboratory for the period, September 1, 1980 through August 31, 1981.
He has been involved with work in the area of computer-aided design
at Fujitsu Laboratories and is interested in studying testing graph
isomorphism, minimum feedback arc set problems, and partitioning
graphs and interval graphs while he is here.
Prof. Okihiko Ishizuka
Prof. Okihiko Ishizuka is a visiting professor from Miyazaki
University, Miyazaki, Japan. He is visiting the Computer Systems
Laboratory for acacdemic year 1980/81. He has been involved in
reseach concerning network synthesis of multithreshold logic and has
developed circuit implementation of a multivalued, multi-threshold
network. While he is here he is interested in pursuing studies in
the areas of multivalued logic and threshold logic, development of
multivalued, multi-threshold networks using IIL circuits, information
processing fields, and the reseach and education systems used in
universities in United States.
Mr. Akira Kanuma
Mr. Akira Kanuma is from the Semiconductor Device Engineering
Laboratory of Toshiba Corporation of Kanagawa, Japan. He is visiting
the Center for Reliable Computing for the period, September, 1980
through August 1981. His work at Toshiba has been involved in the
area of design and evaluation of microprocessors and he is
interested in pursuing studies on high-speed arithmetic circuits,
fault-tolerant computer systems, and utilization of data flow
concepts in multimicroprocessor systems.
!
-------
∂26-Sep-80 1646 FFL
To: JMC, FFL
I wanted to print a copy of NSF[e80,jmc] so that it would be easier to
find the changes in the a,b,c,d, references to the years in the bibliography.
When I pubbed it, it told me that line l6 on page 4 was too long. I think there
is more there than the completion of the word "duration", so I am afraid to
fool with it. Would you look at it, please.
∂27-Sep-80 0330 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #87
Date: 27 SEP 1980 0617-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #87
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Saturday, 27 Sept 1980 Volume 2 : Issue 87
Today's Topics:
Home Information Retrieval - Source and RD,
Electronic Banking - Security/Reliability & Possible Services,
Computers and the Deaf - TDD query & Doing Something
& CA PUC Recommendations & TELEX/TWX Query
----------------------------------------------------------------------
KRAUSS@MIT-MC 09/26/80 08:50:30 Re: Readers Digest buys into Source
According to the 9/25 Washington Post, Readers Digest has bought a
51% interest in The Source. Further information in the 9/26 Post
says that the price was $3 million, a lot of money for a business
with no assets and only a marketing concept (Source is entirely
a resale operation: communications from Telenet, computing from
a time-sharing service called Dialcomm, and data bases from all
over). The article also discloses a messy court battle between
Bill Von Meister, who developed the idea for the Source (and
some years ago developed the idea for Mailgram) and Jack Taub
who ousted Von Meister in a financial/power struggle last year.
Anyway, Readers Digest is clearly moving into electronic
journalism/communications/information transfer.
---Jeff Krauss---
------------------------------
REM@MIT-MC 09/26/80 08:37:18
Chesley's description of Bank of America's multiple-backups and
logs of messages sounds like just what I'd design if I were doing
it. Makes debugging easy since you have all the info available,
while you don't have to actually use that info if nothing goes
wrong so you can use methods that are write-mostly like spooling
to magtape or dumping to a disk without bothering to compile an
index as you go along (indexes compiled later by big IBM
coordinator/crossreferencer). Although I might be tempted to use
end to end instead of by-hop accounting, by-hop has the advantage
of identifying any lossage to a particular machine thus allowing
it to be taken offline for debugging, whereas end to end allows
the particular mistake to be fixed but doesn't discover the cause
so that it can be fixed (doesn't even identify which machine has
the bug, thus requiring the whole net to be brought down just to
locate one bug). I think I'll call B of A and see what they have
to offer here in San Francisco bay area, maybe give them a try....
(By the way, that industry-blurb about 24-hour ATM service that
appeared a couple nights ago may be true for some ATM services,
but as you've probably heard before Crocker Bank (Calif.) runs
their ATMs only from 6am to after midnight (about 1 am) and
probably other banks do too, so that 24-hour part isn't univer-
sally true, I wish it was. Why do IBM computers used by bank
ATMs need 5 hours of PM every AM? Does it really make them run
that much more reliably than DEC machines with their weekly or
sporadic PM? I haven't observed a great reduction in frequency
of system crashes. <flame>)
------------------------------
Date: Friday, 26 September 1980 10:28-EDT
From: John A. Pershing Jr. <JPERSHING at BBNA>
Subject: Cashless Society [literally]
The only real complaint I have about automated tellers is that they
are TOO CONVENIENT. Like David Roode <ROODE at SRI-KL>, I find that
I write relatively fewer checks; instead, I tend to just drop by the
ol' ATM every couple of days or so and whip out a few more bucks (I
haven't transacted with a human teller since I bought some traveller's
checks two years ago). I have become very used to this mode of
operation over the course of the last 7 years; however, on occasion
the ATM has given new meaning to the phrase "Cashless Society" by
going down just after a Friday bank closing, leaving me with $1.37
for the weekend.
-jp
P.S.: When the machines are UP (which they are most of the time), they
seem to perform flawlessly. Considering the age of the system,
I find this performance pleasantly surprising. (Has anybody out
there had their account butchered by Coolidge's ATMs?)
------------------------------
Date: Friday, 26 September 1980 10:35-EDT
From: John A. Pershing Jr. <JPERSHING at BBNA>
Subject: TDDs
Could you give a brief description (capabilities, gross limitations)
of major TDDs in use today? Are they essentially Teletypes, or tiny
CRTs, or what? What is the character set (aside from A-Z)?
After the ARPANET guys finish removing the EBCDIC support from the
TIPs, maybe they can be convinced to replace it with Baudot? Just
a thought... (After all, what are computers for anyway?) Or does
this capability already exist for benefit of the DeafNet?
-jp
------------------------------
KLH@MIT-AI 09/27/80 02:20:00 Re: PUC: on "doing something"...
In response to people who have asked whether there is anything
that they can do personally to influence the PUC proceedings:
I believe there will be two formal opportunities to comment
(refer to the "Administrative Law Judge's Ruling" of the 2nd PUC
background message). As I understand it, Judge Daly will be the
deciding authority, and he need not merely rubber-stamp the PUC's
suggestions; that is why they are called "recommendations". So
it's quite possible for letters to have an effect, especially if
they address specific points of the PUC statement.
Roger is absolutely right, however, about the dangers of
exposing the ARPAnet. I suggest that interested people should
simply send a message to "KLH at SRI-KL" describing how they
may be contacted, and I will try to follow up.
--Ken
------------------------------
KLH@MIT-AI 09/27/80 02:20:00 Re: PUC Recommendation Overview
Okay, folks -- the PUC recommendation arrived today. I haven't
had time to assimilate it completely, but want to provide a brief
summary and some extracts. There were three "exhibits":
(1) Testimony of PUC engineer responsible, in Q & A format.
(2) Paper describing D.E.A.F. and rate recovery mechanism.
(3) Illustrative tariff for TDD's and surcharge.
The author, who will most likely represent the PUC on the
witness stand, is Paul Popenoe, Jr. -- a "Principal Utilities
Engineer" with the CPUC who has worked there since 1952 and is Chief
of the Communication Division's Rate Branch. Since I'll be quoting
him, I should explain that exhibit #1 is in a "Question and Answer"
format where the testifying person furnishes BOTH the questions and
the answers! This appears to be the usual roundabout legal method of
making presentations, written or otherwise.
In general the recommendations are not too surprising.
First, the number of TDD's required is now estimated at
90,000 which produces an initial surcharge of 15 cents instead of
25. This will be applied to all subscribers uniformly, with a couple
of unimportant exceptions. There is an explanation for using this
lower number which I think is irrelevant since all along they harp
on the necessity to adjust the rate up or down according to actual
experience; I figure they just wanted to lower the initial surcharge
so it would look more acceptable at first glance.
The fund would be administered by a "Committee" of three
people, one from each of Pac Tel, General Tel, and the Calif.
Independent Telephone Association (representing everyone else).
No deaf organizations will be involved in any way or step that
I can see.
The speech impaired would receive equipment on the same basis as
the hearing impaired, since their requirements are almost identical.
This is sensible, and all the evidence does seem to show that "this
group is actually very small".
Now for the juicy part on "basic equipment". I'm going to
extract here the relevant written testimony exactly, to avoid
distortion. I apologize for the resulting length and obfuscation,
but the wording can be crucial.
------------------------------
Exhibit 2, D.E.A.F. Description
Section H. Basic Equipment Chargable to Fund
Each certified deaf user shall be furnished with a telecommuni-
cations device meeting the following minimum specifications:
1. Keyboard send.
2. Visual electroluminescent display.
3. Hard-copy printout on 2.25 inch paper tape in lieu of visual
display for users who cannot communicate adequately with
visual display.
4. Baudot 5-unit code.
5. Acoustic coupler for connection with telephone network.
6. Portable instrument not to exceed 12 pounds weight with
carrying case.
7. Alternate A.C. power line or self-contained battery pack
operation.
8. A direct-coupled, non-portable instrument in lieu of
portable, battery-operated, acoustically-coupled instrument
where required for oral transmission and visual reception.
Each telephone company shall make available to certified deaf
users equipment meeting the foregoing minimum specifications at no
charge in addition to the basic telephone rate. There shall also
be provided to each user a non-portable signaling device containing
a relay which will operate a lamp or other device furnished by the
user to indicate when the telephone is ringing. The cost of
providing such minimum equipment shall be charged to the fund. Other
optional features or devices shall be charged to the user at rates
which reflect the incremental fully-allocated cost of providing
such features. Revenues and expenses associated with such optional
equipment shall be treated as normal utility revenues and expenses.
An initial set of batteries and roll of paper tape, where required,
shall be provided with each device furnished to the certified deaf.
Rechargeable batteries may be furnished where appropriate. Provision
of additional batteries and paper shall be the responsibility of the
user.
Where there is available to telephone companies equipment
features or combinations of features beyond the minimum features
above, and where those features are available at substantially no
greater cost to the utility (within %5) of the cost for a minimum
package, equipment incorporating such features may be furnished as
a basic device at no additional charge to the certified deaf. In
any event, charges to be made both to users of equipment and to the
Fund shall be specified in tariffs filed with the Public Utilities
Commission.
The California Public Utilities Commission does not endorse any
particular product. Telephone companies may purchase any equipment
which meets the foregoing minimum standards and which is acceptable
to the utility in terms of quality and price.
------------------------------
Exhibit 1, Testimony of Paul Popenoe Jr. (Q & A format)
Q.6: In Section H, Basic Equipment Chargable to Fund, on page 5
you list specifications for equipment which should be furnished to
deaf persons at no additional charge and which would be charged to
the fund. Please explain the basis of your recommendation.
A.6: I have carefully reviewed the testimony of the deaf groups
and other parties in this proceeding. Based on that testimony it
is my conclusion that both a visual display and a hard copy device
should be offered to the deaf public to permit selection of the
device which best meets their needs. At the same time, I have
attempted to establish specifications which would result in minimum
cost to the rate paying public. For instance, I recommend that the
basic hard copy device to be furnished utilize a 2.25" paper tape.
The testimony has indicated that both the printing mechanism and
the paper are much lower in price for that size than for wider paper
because of the extensive use in the calculator industry. Therefore,
I recommend this as the minimum equipment which will take care of the
basic needs of the deaf. In other areas I recommend that the basic
device to be furnished be portable as the testimony has indicated
to me that deaf persons require a device with which to communicate
not only at home but also from public telephones and their place of
business. However, for the oral deaf an alternate provision should
be made which will allow direct connection to the telephone line to
permit use of the telephone handset for speaking while receiving on
the printer or display. In addition to the basic communication
devices a signaling device is also necessary as set forth in my
recommendation.
Q.7: What is your basis for recommending the Baudot code?
A.7: I believe that practically everyone in this proceeding
would agree that Baudot capability is a requirement of any device
to be immediately issued in this program. There is now an extensive
network of the deaf using machines with the Baudot code. Any new
machines must work with that network. Furthermore, Baudot machines
are simple and well understood by the deaf who use them. Baudot is
widely used in commerce and is the standard code for the domestic
and international Telex networks. As International Telegraph Code
No. 2 Baudot is the recognized CCITT code for keyboard send and
receive devices. The issue then is whether the inclusion of ASCII
should be mandated for inclusion in a basic device. My answer is
no. ASCII is basically designed for use in communicating with
machines. The ASCII code, or CCITT code No. 5, has provision for
128 characters or functions of which 34 are control functions, 32
are graphic symbols, and the remainder are alphanumerics. ASCII
introduces many complexities that are not needed for deaf communi-
cation and that would substantially add to the cost of providing
the service. Furthermore, there is no way I can see ASCII as
providing the equivalent of basic telephone service to the deaf.
------------------------------
Exhibit 3, Illustrative Tariff
RATES, 2. Equipment furnished to Certified Deaf or Speech Impaired
A. Telecommunications Terminal Devices
Suitable for transmission of signals at speed up to 60 words
per minute. Arranged for Keyboard send and receive operation,
includes carrying case, integrated terminal data set, etc.
NRC MR
(1) Display Type Terminal (with a 32 character
display field)
- For certified subscriber, one only - -
- Charge to DEAF Trust $600. $ 7.00
(2) Printer Type Terminal (with 2.25" paper)
- For certified subscriber, one only - -
- Charge to DEAF Trust $800. $12.00
(3) Printer Type Terminal (with 6" paper)
- For certified subscriber, one only $ 40. $ 8.00
- Charge to DEAF Trust $800. $12.00
B. Supplemental Equipment
(1) Relay switch, baseboard mounted, to permit control of a
customer-provided 5 ampere noninductive commercial power
device during ringing signal periods.
- For certified subscriber, one only - -
- Charge to DEAF Trust $ 50. $ 1.40
[NOTE: The NRC column stands for "Non-Recurring Charge", meaning
capital investment. The MR column is "Monthly Rate", the recurring
operating costs associated with the service. Note that the user has
to pay something extra for the 6" paper printer. The charges include
taxes, shipping, handling, accounting, installation, maintenance,
losses and other costs associated with the program. This is why
they seem high. The .15 surcharge for 90,000 TDD's was computed
assuming $800-type devices. --KLH]
------------------------------
So there we are. I will forbear from lengthening this message
further with comments on the extremely poor eyesight and double
standard of P.P., reserving that for another edition. However, I
have an URGENT INFO REQUEST to make of Human-Nets readers: if any
of you are familiar with the current/projected situations regarding
domestic/international Telex/TWX networks, or even know where to
find good references, please send something! (To me, with CC to
Human-Nets if of general interest.)
Thanks,
--Ken
------------------------------
End of HUMAN-NETS Digest
************************
∂28-Sep-80 0123 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #88
Date: 28 SEP 1980 0410-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #88
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Sunday, 28 Sept 1980 Volume 2 : Issue 88
Today's Topics: Radio Growth, Electronic Banking - Possible Services,
Home Information Retrieval - Yellow Pages,
Computers and the Deaf - TDD Design & Baudot Phase Out
& CA D.E.A.F. Surcharge & Speech Synthesis
----------------------------------------------------------------------
REM@MIT-ML 09/27/80 16:48:10 Re: radio vs. human-net
The Radio Advertising Board claims radio increased 200 percent in the
past ten years and that it will be the fastest growing communications
medium in the next year or whatever. I guess they don't consider data
communication to exist (I'm sure it's grown more than 200 percent in
ten years and will continue to grow faster than radio, eventually
maybe even surpassing radio in total volume). Oh well..
------------------------------
Date: 27 September 1980 10:49 edt
From: Frankston at MIT-Multics (BOB at SAI-Prime)
Subject: EFT and 35.72
One problem in trying to track down errors and reconcile checkbooks
is the lack of UID's (much the same as I've flamed about mail). If
you've ever had to track down an error in automatic deposit paychecks
you'd appreciate the difficults in matching transactions simply by
amounts.
But that is the standard means to reconciliation -- the assumption
is that check amounts are unique. This is totally unnecessary. We
already have a start with check numbers appearing on statements and
transaction UID's for charge cards.
------------------------------
Date: 27 September 1980 13:41-EDT
From: Daniel L. Weinreb <dlw at MIT-AI>
While I have not yet had my account butchered by Coolidge Bank's "Cool
Cash" machines, I have often run into the $1.37-left-for-the-weekend
phenomenon, which I guard against by keeping some cash stashed away
in my home for just such a purpose. The machines are down so often
that I and many of my friends refer to them as "Cool Crash" machines.
I had much better luck with the Wells Fargo machines, when I was
living in California (until just recently). Ours (the Livermore
branch) had excellent uptime and never gave me any trouble.
------------------------------
Date: 19 SEP 1980 1023-PDT
From: HAAS at USC-ECL
Subject: INFORMATION ASSISTANCE TERMINALS
This is my first time responding to messages in the HUMAN-NETS
mail list. I would like to point out, in regards to information
assistance terminals (re: Kerns) that that is already being done
(at least the R&D is started) in a company here in California.
This is the same company that is computerizing the Information
Assistance Services (Information Operator). The system is in the
process of being implemented in area code (213) (Los Angeles).
The next area will be 714, then 415 (San Francisco). The system
works great, searching for any given entry with a response time
of <= 1 sec. Pretty good when you consider that there are over
6 million entries for Los Angeles alone. They are investigating
making the information service available to the individual con-
sumer who happens to have a computer terminal at home (instead
of calling and talking to an operator, you dialup with your
terminal and ask the computer yourself!
Anyways, look for an increase of quality from the Los Angelese
Directory Assistance Operators (now stop that chuckling back
there.....).
See you all later.....
ray
------------------------------
Date: 19 September 1980 06:10-EDT
From: Frank J. Wancho <FJW at MIT-MC>
Subject: Reverse White Pages
There has long been a service to businesses in various cities which
provides the white pages by name, by street address, and by phone
number, along with a variety of other information such as credit
rating (at one time), zoning, etc.
I suspect that most of you were not even aware that such information
is available to businesses in the first place. Nevertheless, it would
be interesting to see a discussion on the merits (or lack thereof) of
such a database available to the general dialup public, given that
businesses already have access to this information.
--Frank
------------------------------
Date: 27 SEP 1980 0506-EDT
From: FREND at MIT-AI (Michael A. Bloom)
Subject: ASCII-Baudot Conversion
Your query re development of a chip to do ASCII/Baudot conversion
toggled a vague memory of having seen something along those lines.
Doing some checking uncovered the following in my 1974 copy of the
National Semiconductor MOS circuits catalog:
There is a ROM code converter called the MM5220BL Baudot to ASCII
converter described on pages 8-8 and 8-9. This chip does however
convert in one direction only...
National also states in its index:
"for information on the following code converters
contact National, Santa Clara:
MM3501TL ASCII to BAUDOT, BAUDOT to ASCII converter
,
,
MM4221TM/MM5221TM ASCII to BAUDOT,BAUDOT to ASCII converter
,
<chips that convert other codes>
. "
Again, this catalog is over 6 years old. There is no telling what
new items they may have added to their product lines.
Michael Bloom <FREND@AI>
------------------------------
Date: Fri, 26 Sep 80 09:42-EDT
From: bmstrazz at DNGC
Subject: pot pourri of comments
The SUPERPHONE is a new TDD, already demonstrated but not yet in
production. They have a voice to text feature. Exactly how is
this going to revolutionize the deaf technology, I'd like to see.
Speaking of European deaf TDDS, luckily a deaf leader from Switzerland
was in to see me last week. He said that Europe is divided into two
standards; a bloc of nations follow the 110 baud, 8 bit mode while
another bloc uses 300 baud, 8 bit mode. He does not recall if there
are Baudot terminals used by the deaf there.
The deaf should go ASCII in the future, that's for sure. But the
question is which kind of ASCII.
If it is attested that Baudots are still plentiful, then where
are these machines???? My agents have been trying to beg, steal
or borrow Baudot machines to feed their starved Baudot customers.
Interesting to see what Ken Harrenstien has to say about ATT's
thinking towards the California law. Their response to me was
simple "who is going to pay for it????"
Barry Strassler
------------------------------
REM@MIT-MC 09/26/80 09:04:49
Here's a simple way to phase out Baudot terminals without making the
transition unbearable - we geographically segregate and phase in the
ASCII terminals in one geogrphical area at a time, providing gateways
at the boundary between Baudot and ASCII regions. Thus users within
a single region can talk to each other without trouble, users within
a transition region have a brief time when they must use gateways
to talk among themselves, perhaps a month or two, and users making
expensive long distance calls can actually save money by using a
gateway that they will need if their call is to an opposite-equipment
region (assuming gateways provide efficient trunking of calls or
store&forward or other service; even if not, the overhead of the
gateway isn't much worse than the long distance call in the first
place). Abandoned Baudot terminals in regions that are converting
to ASCII can be shipped en masse to other regions that haven't yet
converted, to be used as replacements for broken units or to fill
the needs of new users in those backward regions. Because all units
are being shipped from the same point at the same time, and to a
limited number of destinations, this massive relocation of terminals
would be less expensive than if each were shipped individually. If
the upgrading were globally smooth, with one or two centers of ASCII
starting up and spreading to adjoining regions instead of hopscotching
around, the relocation of Baudot terminals might be even
cheaper/simpler.
Somebody want to compare the economics of this with
(1) saddling deaf people with new Baudot terminals for next 20 years,
requiring gateways EVERYWHERE if the deaf are to obtain access
to ASCII services and the non-deaf are to obtain acces to Baudot
services,
(2) not redistributing, but merely making all new units be ASCII,
thus again require gateways EVERYWHERE to allow deaf with Baudot
to talk to deaf with ASCII next door.
(Note, I'm saying Baudot vs. ASCII -- Same question arises for
Weitbrecht vs. Bell-103 -- recent feedback from the deaf indicates
Weitbrecht modems are easier to use in two-party talking than
Bell-103, thus conversion to ASCII might happen right away while
conversion to Bell-103 might be delayed.)
------------------------------
REM@MIT-MC 09/26/80 09:22:18
Addenda... the geographical phase-in I suggested has an advantage
I didn't think of: If halfway through the conversion we find that
ASCII has become obsolete and everyone is converting to x.25, we
can stop replacing Baudot by ASCII and start replacing Baudot by
x.25. Then when the whole country is non-Baudot we can (if x.25
is still standard) start replacing ASCII by x.25. At no point
will geography be fragmented. Local areas will be homogeneous
except for this month's transition region, and gateways will be
available for inter-region communication just as they exist now
for inter-national communication (except I think you have to go
through some commercial network like TYMNET or DATAPAC currently;
I'm not sure whether explicit gateways or merely going through a
commercial network is the best way for deaf to talk to somebody in
another geographical region where a different modem or character
code is used).
Of course this phase-in of new equipment is merely for the bottom
of the line user, the deaf and us lowly 300-baud programmers and
personal computer network people. People who can afford Bell-212
or other fast modems will of course use them instead of whatever
the bottom-of-line is in their geographical area, and the burden
will be upon them to provide gateways if they want to talk to deaf
or 300-baud people. (Some of these gateways already exist, such as
PDP-10 computers and commercial networks that have dialup ports for
both 300-baud and 1200-baud, so that persons with differing equipment
can log into the same system and leave mail for each other or even
talk live with each other.)
------------------------------
Date: 27 Sep 1980 0537-PDT (Saturday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: TDD security deposits?
Just out of curiosity, will users of the "free" TDD equipment be
required to put down some sort of security deposit before they get
them? Will they have to return them to the phone company if they
leave the service area or are they "gifts"?
--Lauren--
------------------------------
Date: 26 Sep 1980 0438-PDT (Friday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: TDD, ZUM, and Votrax
1) TDD Discussion
I think the bottom line will most certainly NOT be technical. We can
talk all day about which modem and which protocol is better, and it
will not make any real difference. The controlling element is the
PERCEPTION of the public, not the realities by any means. The public
(in this case defined to be the mass of telephone subscribers) is
probably not going to be happy about having ANY charge added to their
bill that will not benefit them directly. The 911 tax in California
at least theoretically will someday benefit the majority of users.
There will probably be even more uproar when people discover that
they have to pay federal, state, and city taxes on the TDD surcharge.
The amount will not be the issue -- it will be argued on principle.
I personally feel that the best way to avoid this problem is to supply
terminals out of general revenues, and not to try pull the full amount
solely from telephone subscribers. General revenues are frequently
used to help minority groups of various sorts. On the other hand, a
surcharge to customers of the type envisioned by the California bill
is rather unusual and thusly subject to problems.
One more point on this topic -- I think that the terminals provided
should really be the minimum required -- be it ASCII, Baudot, or
whatever. I suspect this will mean a non-hardcopy device. We are
dealing with public monies here -- people are always free to go out
and buy more elaborate equipment with their own money. We cannot
achieve equity -- if an attempt is made to provide too much, the
whole thing will backfire and the result will be: nothing! That
$600 figure for units, by the way, seems ridiculous to me: Sears
is selling their closed captioning boxes at COST for $250, and
they are considerably more complex than a simple TDD. I wonder
how many of these TDD manufacturers plan to sell units at cost?
2) ZUM
The California ZUM rates are not being effectively attacked by
consumer groups because they do not perceive them as a threat.
On the face, they look like LOWER phone rates. It is only if
you understand the long range telco ideas that the real "dangers"
appear. The public at large does not have such access. By the
time they find out what is REALLY going on, it may well be too
late.
The "danger" in this case is the demise of flat rate residential
phone service -- as discussed at length in previous HN digests
(including a HN subscriber poll on the topic).
3) VOTRAX
In the INFO-MICRO mailing list a short time ago, I mentioned the
Votrax chip, and several discouraging points I discovered when I
learned about the details from Federal Screw Works, the manufacturer:
a) The chip will theoretically be $12 in OEM quantities -- but Federal
Screw refuses to confirm this -- or say anything firm about pricing
at all.
b) If you want to buy a sample chip, you have to buy their $15,000
development system (a simple micro with a text to phoneme trans-
lator), which comes with two chips. Around November, they will
offer a protyping board for "under one thousand dollars".
c) They are completely uninterested in dealing with other than very
large firms who will make initial minimum orders of 5000 units.
They will not release samples.
I asked interested persons in INFO-MICRO to call Federal Screw (an
appropriate name at that) and see what they could find out -- but
EVERYONE is getting the same brushoff so far. If anyone else out
there wants to give it a try, the person to talk to is Alfred
Lubienski (general manager) at (313) 588-2050. You will probably
end up talking to a marketing guy by the name of Max Bishop, who
seems to be a nice guy who is unable to get the higher ups (like
Lubienski) to change the policy on these chips. It seems they are
very concerned about destroying the rest of their product line,
among other things. If anyone does call, I strongly suggest not
saying anything to suggest you know the points above -- just act
like a person who wants to buy a sample. Do not mention Human-Nets
or anything like that, or they will just brush you off as part of
a coordinated campaign to change their policy. They are very
discouraging people when it comes to this chip.
For those of you familiar with Votraxes, the chip is functionally
equivalent to a VS-6 Votrax unit: 64 phonemes, 2 bits of inflection.
If anyone gets any new info out of Federal Screw, I'd sure appreciate
it if they'd drop me a note detailing what they've learned...
Enough for now. Back to the limbo.
--Lauren--
------------------------------
End of HUMAN-NETS Digest
************************
∂28-Sep-80 1000 JMC*
massy, 493-9277, 7-2232
1. precise staff plan. Costs will be mainly staff.
Who will be chiefs, and who will be Indians?
2. How many users will it support?
3. Reservations about Stanford leadership.
No monopoly
4. The fallback position is to offer the services of the
machine for general sale at Stanford.
Additional remark: the amount of disk file is likely to prove
insufficient. Rates for disk file should be part of the plan.
It is very important that they be low enough to encourage permanent
storage of documents.
The plan should include what documentation is provided. Arthur Samuel's
manual for E is a model of what is wanted. Maybe Samuel could be
induced to do some of it.
∂28-Sep-80 2336 RMS at MIT-AI (Richard M. Stallman)
Date: 29 SEP 1980 0234-EDT
From: RMS at MIT-AI (Richard M. Stallman)
To: JMC at MIT-AI
Hello. I have been reading the magazine you gave me.
I have done some thinking but I don't see
any especially interesting avenue to persue yet.
I want to try to work out the precise differences
between Reiter's defaults and Doyle's non-monotonic
logic. I am curious about it, but Reiter's paper says
he is going to write about it soon.
For the most part I don't think that the precise details
of circumscription are as good a way of thinking about
what is goin on than Reiter's defaults.
The circumscription method may be useful for
interfacing to a particular form of reasoning program,
one wedded inextricably to first order logic, by
performing the non-monotonic step and coming out with
more statements of first order logic. But I don't think
that inflexible reasoning programs like that are the right way
to implement something. The logic is good mainly for figiiuring
out what a program will or won't be able to do, or figuring
out what knowledge is to be represented, and for that use,
the detailed method of circumscription has some slightly
confusing/paradoxical characteristics.
I hope I'll be able to talk about this with you
tomorrow morning; if not then, then Tuesday.
1. I'll be in late in the morning, since I have to take my car in.
2. I saw Reiter last week, and he didn't mention having any new paper.
3. You are missing the fact that both Reiter and McDermott-Doyle
treat defaults as axioms, i.e. originating form the programmer,
whereas circumscription is a rule the program can decide to apply
at any time, i.e. it can generate defaults.
4. Moreover, circumscription is semi-decideable. One can generate
the set of possible conclusions that can be reached by circumscription,
while non-monotonic logic is not r.e., since the definition depends
on fixed points among models.
5. Your conclusion about logic being suitable mainly for meta-reasoning
about what programs can do would follow if something else were guaranteed
to be uniformly better for expressing the data in the program.
6. I will be interested to hear about paradoxical aspects of circumscription,
because not all is clear to me.
∂29-Sep-80 0841 FFL
To: JMC, FFL
Room 301 in Jacks Hall is reserved for your seminar on October 6,
4:l5 to 5:30. I assume that you will decide on your permanent time and
place at that first meeting. Carolyn tells me that the rooms are in
high demand, so do you wish me to reserve Mondays at that hour,pending a
final decision?
∂29-Sep-80 0928 CSD.ULLMAN at SU-SCORE arpa contract
Date: 29 Sep 1980 0926-PDT
From: CSD.ULLMAN at SU-SCORE
Subject: arpa contract
To: csl.fb at SU-SCORE, csl.lantz at SU-SCORE, jmc at SU-AI
I have just gotten word that my proposal for silicon compilation will
not be funded until january. Do you three have any objection to my
continuing to use some of the "McCarthy money" to cover my project during the
fall quarter? The amounts I would like to spend are 25% salary for
me and also for Bob Floyd, supplements to fellowship holders Karplus,
Strong, and Dolev, and the attendant computer services for the latter 3.
With overhead, this probably comes to $25K. Please let me hear any objections.
-------
I certainly can have no objections. It seems to me you should check
with ARPA.
∂29-Sep-80 0930 FFL SEMINAR - John MccCarthy
To: JK at SU-AI, CG at SU-AI, PLS at SU-AI, DRH at SU-AI, DBL at SU-AI,
ZM at SU-AI, RWW at SU-AI, CLT at SU-AI, JD at SU-AI, GAC at SU-AI,
LGC at SU-AI, TCS at SU-AI, TW at SU-AI, JMC at SU-AI,
FFL at SU-AI
The organization meeting of an open weekly seminar entitled "Formalization
of Knowledge and Action," will be held in Room 301, Jacks Hall ( Computer
Science Department), at 4:l5 p.m., on Mondays, starting, October 6.
∂29-Sep-80 0940 FFL Seminar
To: JMC, FFL
Already have a repl from J. Ketonen that Mondays and Wednesdays are
out for him for a seminar. He would like to know if the time can be
changed.
∂29-Sep-80 0941 FFL
To: JMC, FFL
I have mailed the Seminar announcements to the CSD and to SRI. I will
not pub the announcements on the bulletin boards until I have verification
of the date from you.
∂29-Sep-80 1018 FFL SEMINAR -JOHN MCCARTHY
To: JK at SU-AI, CG at SU-AI, PLS at SU-AI, DRH at SU-AI, DBL at SU-AI,
ZM at SU-AI, RWW at SU-AI, CLT at SU-AI, JD at SU-AI, GAC at SU-AI,
LGC at SU-AI, TCS at SU-AI, TW at SU-AI, JMC at SU-AI,
FFL at SU-AI
Messages have come saying Monday, Tuesday and Wednesday are bad for the
seminar on "Formalization of Knowledge and Action." Is there serious
objection to Thursdays at 4:l5. Please register your answer by Wedesday
morning so we may make a decision.
∂29-Sep-80 1105 Keith A. Lantz <CSL.LANTZ at SU-SCORE> Re: arpa contract
Date: 29 Sep 1980 1059-PDT
From: Keith A. Lantz <CSL.LANTZ at SU-SCORE>
Subject: Re: arpa contract
To: CSD.ULLMAN at SU-SCORE
cc: CSL.LANTZ at SU-SCORE, fb at SU-AI, jmc at SU-AI
In-Reply-To: Your message of 29-Sep-80 0926-PDT
Since I'm now PI (I think), I have to sit down with Betty Scott and
Forest at least and figure all this stuff out. I need some Research
Assistants (maybe even Associates if they're around) and should start
learning about budgets. Given the amount of the contract it seems that
$25K is available, however. If any of you want to help me out in this
regard, let me know. I will be setting up an appointment with Betty
sometime this week.
Keith
-------
∂29-Sep-80 1139 CSL.FB at SU-SCORE Re: arpa contract
Date: 29 Sep 1980 1135-PDT
From: CSL.FB at SU-SCORE
Subject: Re: arpa contract
To: CSD.ULLMAN at SU-SCORE, csl.lantz at SU-SCORE, jmc at SU-AI
In-Reply-To: Your message of 29-Sep-80 0926-PDT
Sounds fine to me. -Forest
-------
∂29-Sep-80 1343 CSD.ULLMAN at SU-SCORE
Date: 29 Sep 1980 1338-PDT
From: CSD.ULLMAN at SU-SCORE
To: JMC at SU-AI
In-Reply-To: Your message of 29-Sep-80 1016-PDT
I already checked with arpa. Thanks.
-------
∂29-Sep-80 1628 CG
I'm off tomorrow morning. If for any reason you should need to get in
touch with me, my number in New Mexico will be 505-455-2464.
The address is:
Rt. 1, Box 411
Espanola, N.M. 87532
I gave the proposal to Betty in final form. However, just in case, you
might want to know that the source files for the proposal are:
NSF.PUB[THE,CG] and CG.BIO[THE,CG]. I'll be back on October 30th.
See you then.
∂29-Sep-80 1952 AAM Fenaughty & Griffith
Thank you for calling him. I'll get in touch with him again soon.
Allan
∂30-Sep-80 0107 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #90
Date: 30 SEP 1980 0355-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #90
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Tuesday, 30 Sept 1980 Volume 2 : Issue 90
Today's Topics:
Electronic Banking - Auto-teller security,
Computers and the Deaf - Modems & History & Speech Synthesis
& Telex/TWX & Baudot/ASCII/X.25,
Electronic Mail - NBS Standards, Telephone Services - Phone Book
----------------------------------------------------------------------
Date: 29 Sep 1980 0358-PDT (Monday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: Big Brother -- VADIC -- Captioning
Three unconnected comments:
1) The mirror at the B of A tellers is almost certainly to let
personnel inside look out at teller activity. There is no
reasonable way that "B of A central" could manage a video
hookup to the teller locations. Just not practical or
economical at this time. There MIGHT be a still camera in
there to snap pictures of people attacking the machine, but
I doubt it.
2) VADIC modems (1200 baud full-duplex) are indeed very expensive, as
are the less desirable (for various technical reasons) Bell 212's
which have much the same functionality. There is little hope of
the prices on these units falling drastically in the near term,
mainly because people have been perfectly willing to pay the high
prices, even for fairly large quantity purchases. And those things
are selling like the proverbial hotcakes! UDS and VADIC recently
fought a lawsuit over UDS's use of VADIC modem patents. UDS won
this round, but there may be appeals yet to come. VADIC does not
really want to license its technique.
3) While it is true that Teletext is a more general technique than
the PBS designed closed captioning system, the captioning boxes
have a number of benefits that really do seem to make them better
for their selected purpose. I have both a captioning box and a
Teletext (CEEFAX) demonstration system setup on which I perform
experiments, so I have been in a fairly unique position to judge
them side by side.
Once again, I remind interested persons that my TELETEXT
mailing list (now running over 100 persons) exists specifically
to discuss issues of TELETEXT, Closed Captioning, Digicasting,
and related technologies. To join up, please send a message to
TELETEXT-REQUEST@AI, which will be forwarded automatically to me.
--Lauren--
------------------------------
Date: 28 Sep 1980 1825-PDT
From: ROBERTS at USC-ISI
Subject: Votrax chip
I had a demo of the votrax unit a couple of weeks ago. The startling
thing about it is that it transmits speech at a bit rate of only
70 bps, and all you need to get from the 70 bps stream back to speech
is the proverbial $12 chip. Of course, getting from the typed input
to the 70 bps is a little more complicated. That's where the $10-15K
computer and software come in. The votrax sales guy had a cute little
unit, somewhat smaller than a TI700, on which you could type in sen-
tences and hear the result come out the speaker. Adjustable controls
included talking speed and pitch. Interesting note: the only voice
available is a male speaker. Not sex discrimination, they claim a
female voice would take much more development on their part. Since
they've never lost a sale due to not having a female voice, they're
not too interested in investing the development funds, I gather.
(maybe that IS discrimination?)
How does it sound? Well, not too bad. Actually, when you consider
that the stuff is flowing at a 70 bps rate, it sounds damned good.
However, they still have some work to go. One interesting point
is that if you have a complex or unusual word, you're better off
spelling it phonetically. One example that we played around with
was 'pizza', which sounded much better when typed in as 'piTzza'.
I gather that their main market right now is toy manufacturers,
so you may see this product on the toy shelves by Christmas. They
definitely are interested in high volumes, however. I typed in some
spanish just to hear what would come out, and it was a total disaster.
However, by typing it in phonetically I got some pretty intelligible
espanol. All in all, an interesting product that I expect we'll be
hearing much more from. It'll be interesting to see how the battle
between the LPC speech synthesizers and the phoneme-based ones works
out.
ollie o.
------------------------------
Date: 29 Sep 1980 1134-PDT
From: ROUNDS at OFFICE-1
Subject: TWX Interfaces
Many thanks to TK for the info on TWX terminals. I had always
wondered how they differed from the "regular" portable ASCII
terminals we started out using, and still use a great deal.
The question now becomes: "How do we call up an ASCII TWX machine
(in the 410-910 area code range) with a normal ASCII computer-type
terminal at an ordinary area code phone?" Can we do something when
making the call to turn off the "Baudot-to-ASCII" converter unit
which would otherwise be "spliced into" the connection?
Conversely, how can we force the "ASCII-to-Baudot" converter to
be "spliced in" when using a terminal to call a TWX number in an
ordinary area code?
Having these two capabilities would make every ordinary ASCII terminal
a TWX machine too, a great thing from my point of view, but maybe not
so wonderful to Ma Bell, who could lose revenue by people not having
to buy/lease TWX machines.
Or am I dreaming? Are there other differences (carrier tones, etc.)
which make the two incompatible?
I'd really like TWX access from this terminal. Don't know quite what
I'd use it for, but I'd like to have it...
Regards, Will Martin (ROUNDS at Office-1)
------------------------------
Date: Monday, 29 September 1980 10:44-EDT
From: John A. Pershing Jr. <JPERSHING at BBNA>
Subject: X.25 vs deaf vs ASCII vs Baudot
Hold It!!! I didn't intend to turn the TDD discussion into a
discussion on the technical feasibility of X.25. I merely used
X.25 as a paradigm of technical change. (By the way, X.25 is a
transmission-ish protocol, analogous to Bell-103 modem control.
It makes no assumption about the underlying character code except
that it utilizes 8-bit octets.)
It appears that there is no "right" decision. Whatever we "know"
as being the absolutely, perfectly correct technology today will
undoubtedly be outmoded in another decade. There IS a lot of
merit in the testimony (transcribed the other day by KLH) that
Baudot is "better" because existing TDDs talk Baudot. (This
assumes, of course, that the statement about existing TDDs is
correct -- I have no way of verifying it one way or the other.)
Sigh. I guess silicon is still a bit too expensive for us to have
our cake and be able to eat it too. We'll have to wait a few years
for ASCII/Baudot/EBCDIC terminals with Weitbrecht/103/212/Vadic
modems connected to Phone/X.25/Ethernet couplers.
-jp
------------------------------
Date: 27 September 1980 15:14 edt
From: Sibert at MIT-Multics (W. Olin Sibert)
Subject: NBS Computer Standards
To: header-people at MIT-AI, msggroup at MIT-ML
The following cheerful news is brought to us by the September 29 issue
of Computerworld. Copyright (c) 1980 by CW Communications, Inc.
Following ANSI guidelines:
NBS to Give Format Standards Over Four Years
Chicago - The National Bureau of Standards (NBS) will release over
the next four years a series of format standards aimed at improving
computer communications and integration and at aiding electronic
mail transmissions between automated offices.
The NBS-sanctioned standards - five in all - will be voluntarily
developer by the user and vendor community, following American
National Standrads Institute (ANSI) guidelines and processing
requirements, according to James H. Burrows, director of the
NBS' Institute for Computer Science and Technology (ICST). The
ICST manages the governmentwide computer standards program and
provides technical assistance to federal agencies.
At the recent National Symposium on Office Automation, sponsored by
the Data Processing Management Association's Education Foundation
(DMPA) here, Burrows gave a brief outline of each of the proposed
protocol standards. The NBS standards will include:
1) A message interchange format standard, which will allow users of
one system to send and receive messages from a foreign system.
2) A flexible disk format standard to establish common file
formatting and labelling for flexible disks.
3) A text editing directives standard that will establish a common
set of user directives for text editing systems.
4) A text formatting directives standard to provide directives for
text formatting systems.
5) A message processing directives standard establishing a common
set of directives for computer-based message systems.
The latter three standards will provide a minimal level of
functionality and help users switch from one text editing,
formatting, or message system to another, Burrows explained.
As a flagship for these five standards, the NBS is planning to release
several guidelines geared to help users plan for, select, and evaluate
computer based office machinery. The first guideline, "requirements
analysis for office automation systems", which Burrows said will be
available in late October or early November, will recommend a process
to measure the benefits of office automation.
Drafts of the requirements analysis guideline have already been
circulated to a number of federal agencies and vendors for comments.
------------------------------
It all sounds kinda familiar, doesn't it? Remember NETED? Standard
mail systems, described in an RFC too old for me to remember? Since
the ARPAnet technical community has already been through exactly
this same process, perhaps we should provide some input; presumably
Mr. Burrows is the person to talk to.
------------------------------
Date: 27 Sep 1980 1623-EDT
From: DDEUTSCH at BBNA
Subject: NBS Computer Standards
Cc: header-people at MIT-AI, msggroup at MIT-AI
I believe that Marvin Sirbu mentioned (at least to MsgGroup) last
spring that NBS was having BBN develop a message format standard.
This is the very same standard as was mentioned in the CW article.
We have finally made it to the "draft standard" stage. NBS will
soon be publishing a document for public review. The lessons
learned on the ARPAnet have not been ignored by either NBS or
the draft standard's authors.
After the public has made its comments and suggestions the draft
standard will be revised into a final version.
Debbie
------------------------------
Date: 29 Sep 1980 0841-EDT
From: POGRAN at BBND
Subject: Re: NBS Computer Standards
To: header-people at MIT-AI, msggroup at MIT-AI
NBS/ICST has been an ARPANET participant just about as long as anyone
else; I think we can safely assume that they are fully aware of the
ARPANET experience in these areas. In particular, I think the fact
that the goals, and the phrasing, sounded so familiar to us is an
indication that NBS is aware. What NBS must now do is LISTEN to all
the input they're soliciting from industry -- and still come up with
technically valid standards, a la ARPANET.
Ken Pogran
------------------------------
Date: Monday, 29 September 1980 11:12-EDT
From: John A. Pershing Jr. <JPERSHING at BBNA>
Subject: NBS Computer Standards
To: header-people at MIT-AI, msggroup at MIT-AI
It seems the one lesson to be learned from the ARPA Net (which NBS
apparently hasn't learned) is that there IS such a thing as too many
protocols. With DOD issuing their "standard" protocol, ANSI issuing
theirs, yet another coming from Europe (CCITT), plus the extant ARPA
Net, ChaosNet, PRNet, PUP, etc., etc., etc., this country needs
another "standard" protocol about as much as I need another hole
in my head.
-jp
------------------------------
Date: 28 Sep 1980 0339-PDT (Sunday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: upside-down phone books
Businesses have indeed for years been able to RENT what are popularly
known as "upside-down" phone books, which are listings indexed by
phone number and/or address. They do not include "unlisted" numbers
of course, which in a city like L.A. leaves out almost half the
residential subscribers! Usually the businesses renting these books
are required to physically return the old set before they are given
a new one. I have worked with these books in the past, and they are
rather amusing.
------------------------------
Date: 16 Sep 1980 0725-PDT
From: ADPSC at USC-ISI
Subject: Yellow Pages (and their white counterparts)
I am glad (as Lauren pointed out on 9/15), that the central LA yellow
pages comes out in two volumes. It's about time that TPC got smart
about the whole thing and arranged books the way that (most) people
tend to look for things (alphabetically).
I live in the DC suburbs and the problem of trying to find things in
the phone book can be frustrating. The problem is not so great when
looking for someone or something in VA, as it is separated from MD
and DC by a river (a convenient breaking point if ever I've seen one.)
However, when trying to find something in the MD suburbs, you can
often be in for an adventure. Some of the localities very near the
state/district line often are full of businesses and people who refer
to themselves as living "in DC". This is all well and good for mail
purposes, since zip codes negate a lot of ancillary silliness, but
the DC phone book ends at the DC line and doesn't help you too much
when you want a Chevy Chase number.
I used to live outside of Boston, which had an even worse problem.
The phone books were arranged geographically. This was all well and
good if you knew which town you were looking for and where it was
located in relation to the center of the city. South Boston was easy
enough to find (the southern suburbs book), and the North End was no
big deal (the northern book). But just where do you find East Boston?
My knowledge of geography had always told me that east of Boston is
a lot of water (aka the Atlantic.) Luckily for my sanity, I moved.
As for the Yellow pages, I have never been able to figure out their
indexing system to my satisfaction. I am willing to believe that
part of this may be due to my unusual outlook on life in general,
but one can never be too sure.
This is admittedly a minor set of problems and they only bother me
now and again. Mostly I get around it by writing down the numbers
I call frequently and filing them in my own way. There is something
cathartic about being able to grumble about it to the world.
don
------------------------------
End of HUMAN-NETS Digest
************************
∂30-Sep-80 0851 CSD.FROG at SU-SCORE Welcoming Lunch for New CSD Graduate Students
Date: 30 Sep 1980 0847-PDT
From: CSD.FROG at SU-SCORE
Subject: Welcoming Lunch for New CSD Graduate Students
To: CSD-Faculty: ;
Informal Pre-Game Patio Lunch
Professor and Mrs. Herriot's Home
11:00-1:00 p.m.
Saturday, October 4, 1980
1302 Channing Avenue, Palo Alto
We hope as many as possible of the CS faculty and their spouses can join us
in welcoming the new CS graudate students. Please let me know by October 1st
whether or not you can come. Thank you.
Jack Herriot
-------
∂30-Sep-80 1411 Nilsson at SRI-KL TEX work
Date: 30 Sep 1980 1410-PDT
From: Nilsson at SRI-KL
Subject: TEX work
To: LTP at SU-AI
cc: Nilsson, JMC at SU-AI
Luis, Is there any more news about finding someone to TEX John
McCarthy's files for his book of collected papers? I don't know
how fast John wants to proceed on this, but as far as I am concerned,
we may as well get cracking. As I mentioned, I have an alternative
way to process John's files into typeset camera-ready form if
TEXing it is impractical. --Nils
-------
∂30-Sep-80 1421 Nilsson at SRI-KL Book
Date: 30 Sep 1980 1419-PDT
From: Nilsson at SRI-KL
Subject: Book
To: JMC at SU-AI
cc: Nilsson
John, I presume you are still interested in having Tioga publish
your book of philosophical/AI papers. One of these days I'll send
over a contract form for you to look at and comment on and eventually
sign. Do you have any strong preferences for TEXing your files if
there are alternatives that appear just as good or better to me?
--Nils
-------
∂30-Sep-80 1426 DRH seminar schedule
Thursdays at 4:15 are fine with me!
∂30-Sep-80 1704 SEK phone number
If you don't have it already, it's 329-9081, and it is in my finger plan.
∂30-Sep-80 2356 VRP via Dialnet dialnet
To: JMC, LES
I now have enough C code to use dialnet for telnetting to sail via
the dialnet dialup (324-2149). (I'm using it right now.) The whole thing
is a page and a half of code. I don;t check parity on received
packets, but it is only a few extra lines to do so, so I'll add that
soon (which might also help with the occasional hanging I get).
Will you do mail or file transfer?
∂01-Oct-80 0023 VRP via Dialnet dialnet
To: JMC, LES
Depends on how hairy it is. I've just about run out of time, with classes
starting.
∂01-Oct-80 0327 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #91
Date: 1 OCT 1980 0615-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #91
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Wednesday, 1 Oct 1980 Volume 2 : Issue 91
Today's Topics:
Communicating via Network - Teleconferencing,
Electronic Mail - NBS Standards, Telephone Services - Phone Book,
Computers and the Deaf - Speech Synthesis & Telex/TWX & TDD Design
----------------------------------------------------------------------
Date: 30 September 1980 05:06-EDT
From: Frank J. Wancho <FJW at MIT-MC>
Subject: Teleconferencing
Since we have an extensive cable system here (El Paso), I wondered
why one of our Holiday Inns had their own dish set up in their
parking lot. A call to them produced the following information:
1. They prefer to get HBO and our cable company distributes
ShowTime.
2. Apparently the cable company is not franchised to distribute
to motels (though most apartments have it, so I'm not certain
that this restriction is fact).
3. During the day various (large?) companies can hold video
teleconferences, sales meetings, seminars, classes, and
whatnot by gathering at their local HI and forego the
traditional business trip. (Presumably, this is a one-way
video deal with conference calls supplying the reverse
channel.)
--Frank
------------------------------
Date: 30 Sep 1980 (Tuesday) 0933-EST
From: WATKINS at NBS-10
Subject: The NBS Standards Program in Computer Based Office Systems
cc: header-people at MIT-AI, msggroup at MIT-AI
As manager of the NBS program in Computer Based Office Systems,
I would be happy to receive any comments either on the general
program of standards planned or any specific product. It would
greatly expedite things, if comments were directed to me (either
via netmail, USPS mail, or phone) rather than Jim Burrows or any
of our contractors.
The success of this standards program, as any, is dependent
upon sound technical inputs from vendors, Government, users,
and implementors.
I look forward to hearing from the ARPANET community and
certainly their voice of experience.
Shirley Ward Watkins
(301) 921-3516
National Bureau of Standards
Bldg. 225 B226
Washington, D. C. 20234
------------------------------
Date: 30 September 1980 08:35-EDT (Tuesday)
From: Andrew G. Malis <MALIS at BBNE>
Subject: ADPSC@USC-ISI's comments about the Boston phone book
The Boston area white pages are split into four books: A Boston book
that includes the entire city of Boston (including the North End,
South Boston, East Boston, etc., along with Cambridge, Brookline and
Somerville), and a book each for the north, west, and south SUBURBS.
Each book has both a map and a list on the back cover showing which
localities are in each of the books. The entire area is served by
one Yellow Pages.
I have never had any problem finding any number that I am looking
for. In fact, I prefer it this way; I have seen the Houston phone
book, which is two huge books split alphabetically for the whole
metropolitan area. Try looking for an Al Smith whose address you
are not quite sure of - there will probably be a hundred of them,
spread throughout a several-hundred-square-mile area!
Andy
------------------------------
Date: 30 Sep 1980 0211-PDT (Tuesday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: phone books -- votrax -- twx
Sigh. I really do hate sending messages consisting largely of a
number of seemingly unrelated topics -- but I am finding that various
discussion topics seem to be forking out in parallel. If this process
continues at the current rate, it will no doubt surpass the ability
of my so-called brain to assimilate. If you hear about a mysterious
explosion sometime soon in the West Los Angeles area, you'll know
that overload has finally taken its toll.
Be that as it may -- once more into the breach:
1) Phone books. The split apart phone books may have some good
points, but they have a number of bad points as well. For
example, to get a reasonable coverage of yellow pages for my
area, I now have to request four extra books. The "Neighborhood"
book they give me by default is so limited that it is totally
useless to me. If I were a business who did alot of yellow page
advertising, I'd be VERY upset about the whole situation. Now
I'd have to advertise in a pile of books (paying for separate
ads in each) where one would have sufficed before. Separate
ads for each little neighborhood, and for the "business" and
"consumer" books for each (in many cases) would be necessary.
2) The votrax. Just to prevent any confusion, the chip in question
is really no different (speech quality-wise) from the model VS-6
unit which has been around for a number of years and has seen
considerable use in commercial systems. The only difference is
that four circuit boards have been replaced with one chip.
The $15,000 price for their development system is rather
outrageous no matter how I look it. Software that handles
text to phoneme translation has existed for quite some time.
I have worked extensively with the Bell Labs "speak" program
for the votrax, which performs exactly that task under Unix
systems. The Naval Postgraduate School has published a set of
translation rules which several people have translated into
very good text to phoneme systems. There is even a program
for the 8080 to perform this task, called "Anglophone" or
some such, which is available commercially. It costs some-
thing around $100 as I recall. The simple micro/software and
terminal being offered by Federal Screw for $15,000 can hardly
REALLY be worth that -- even giving them a good margin for
profit. Oh well.
By the way, Federal Screw Works manufactures another (more
expensive) model synthesizer (the ML-1 [multi-lingual-1])
which does better than the VS-6 with non-English speech.
Not great -- but definitely better.
3) TWX. Sorry, but you cannot reach TWX area codes from your home
phone. The area codes involved (910, 810, etc.) are segregated
inside the toll network from the rest of the system.
Interestingly, while TWX numbers CAN call regular DDD numbers
(I've seen it done), a regular DDD number CANNOT call a TWX
number. There are complications in charging and control that
have made this necessary. The result is that the issue of turning
on/off ASCII<=>BAUDOT converters is a non-issue since you cannot
reach them in any case. I believe, however, that Western Union
does offer a service which allows you to dialup a computer with
your ASCII terminal and input messages into their computer. The
computer then sends the messages off into the TWX/Telex network
(TWX uses ASCII, Telex uses BAUDOT). They charge quite a premium
for this service as I recall.
Tah tah for now...
--Lauren--
------------------------------
Date: 30 SEP 1980 0737-EDT
From: CBF at MIT-MC (Charles Frankston)
Subject: TWX/Telex and ASCII
My sense has been shattered in recent days reading references from
people who believe ASCII is more "complicated" than Baudot code to
those who want to know how to control the "Baudot-to-ASCII" converter
unit in a TWX terminal. So for the benefit of my sense and in partial
answer to KLH's question about TWX/Telex I will threfore pontificate
on these matters to the best of my ability; I was hoping there would
be persons on this list with far more knowledge of these subjects
than I, since my knowledge of TWX/Telex matters is based on rather
casual interest.
Firstly, there is no ASCII-to-Baudot unit in a TWX terminal.
TWX terminals are ASCII/Bell 103 machines. It is indeed possible
(though not publicly documented to call a TWX machine directly
from the direct dial network (the phone system); Western Union
does it for testing purposes. You can however for an additional
charge get a legitimate connection from your TWX terminal to the
direct dial network, controlled by a switch. (eg. you flip the
switch and then use your TWX terminal to call your local TIP or
whatever).
For the uninformed: The TWX network was set up in the 1960's by
Western Union. It is primarily used for business to exchange
small text messages, such as orders, inquiries etc. The TWX
network is in fact just another telephone network where instead
of most installations consisting of phone sets, they consist of
terminals. The terminals are usually equipped with dials for
placing the call. The system uses an entirely different
address space from the phone system (though as I said there are
unpublicised patches to and from the phone system). There are I
believe, on the order of 100,000 TWX subscribers in the country.
As I remember the rates per call were comparable or worse than
long distance voice rates (even though they could obviously use
far less bandwidth). On the other hand, most business's probably
realize that TWX calls tend to be shorter and to the point. The
network is domestic US (probably also Canada?) only.
The most common terminal is the venerable ASR-33, which is still about
the cheapest setup one can get. You typically want an ASR unit (ASR
means you have some means, such as paper tape, for preparing a message
beforehand) so one spends the least time possible dialed up (at $3 a
minute or whatever). However, DECwriter's with internal memory, TI's
with bubbles etc. can all be obtained. One can in fact purchase just a
line and hook up your own equipment. (You can even hook up a computer
and send out automatically generated mailgrams). The network is
defined to use 110 ASCII communication. If you and another party
mutually agree to start talking 300 baud, I think there are reasons
it might not work, but I have never tried. (Ie. they are free in the
contract of the network to digitally (or mechanically!) regenerate the
signal, I don't know if they do or not, it may work on local loops but
not on a longer call). There are numbers you can call to type directly
to a human being for directory assistance, ordering supplies for your
TTY33 etc. Nothing stops people from typing directly at each other,
but most often a business just leaves the machine collecting messages
and calls back at their own convenience if at all. If you really
expect to get into a conversation, remember, a voice phone call has
higher bandwidth, and is probably cheaper.
The International Telex network has been in operation longer than
TWX and it uses 5 level baudot code at about 5 characters per second.
(I believe at 50 baud (remember 1 start & 1.5 stop bits etc.)). Since
5 levels or 32 different combinations cannot represent even all the
letters and all the numbers, codes are assigned to mean "shift to
numbers (and/or figures)" and "shift back to letters". Anyone who
thinks Baudot is simpler than ASCII obviously hasn't tried to decipher
a message with a dropped or forgotten shift character. Telex calls
are placed in an analagous fashion to TWX calls. There are "country"
codes one would look up to lets say dial Thailand from Camaroon,
except my guess would be you'd have to route that one through London.
I suspect internationally there are a lot more than the 100,000 I
guessed for TWX, but I wouldn't put it above 500,000. Remember for
much international commerce, reliable phone connections are not
available.
Most Telex terminals use a 3 row keyboard with the numbers above
"qwertyuip". If you are in figures mode and you try to type a letter
that doesn't have a figure above it, some don't allow it mechanically,
others type the wrong thing. There are newer fancier non-mechanically
(and actually even mechancial) units with 4 row keyboards tha automa-
tically generate the appropriate shifts. I would be interested to
know which variety of keyboard the likely modern manufacture TDD's
(as opposed to recycled TTY 28's) tend to have.
TWX can talk to Telex and visa-versa. The international Telex lines
from the US are all owned by either Western Union, ITT, RCA Global,
or Tran Telecommunications. All four of these carriers have numbers
on both TWX and Telex that act as gateways to the other network, doing
necessary code translation, buffering and flow control. They each all
also offer store and forward services where they will gobble down a
whole message and then periodically try to get through to a machine
you couldn't get through to (busy, or probably more frequently, the
lines to Qwaziskwaziland were down). Naturally, there is a charge
for this. It will all appear on your monthly bill. Don't ask me
how or via what convenents the carriers manage to exchange billing
information, but this (plus probably FCC regulations and Justice
Dept consent decrees from eons ago) is obviously why there is no
way to originate calls from the AT&T direct dial phone network,
even though technically feasible with TWX.
As for the future of TWX/Telex, most people seem to feel the
possibilities for TWX growth in the U.S. are poor do to all the
fascinating alternatives that seem to be popping up. (Such as
facsimile transmission, communicating word processors that dial
each other up etc., not to mention mail services modelled after
the Arpanet such as Telemail). My opinion is that no one really
has much idea was is happening and that most people (corporations
etc.) are sitting on the fence trying to see which technologies
are going to win the battle. In the meantime those who can afford
the investments slowly offer one little service based on this
technology and another based on that making the end-user's choice
ever more complicated. A good example: ITT recently announced a
store and forward facsimile service capable of translating between
most of the major (normally incompatible) facsimile machines. It
can also take ASCII input from a TWX terminal and generate up a
fax with the text on it!
Anyway, except as background information, I don't think much of this
is applicable to the CA PUC deaf communcations issue. My own opinion
on that issue is that it is a mistake to specify baudot devices. It
is clear (and not denied by the PUC) that a lot is going to happen in
data communications over the next several years, and that most of it
will be happening in ASCII. Somone asked how many ASCII terminals
were out there. I don't think anyone has good numbers, and I don't
believe what I read in most of the magazine surveys anymore. It is
well known DEC alone manufactured over 100,000 of one type of ASCII
terminal (the VT100) last year and could not meet the demand. My
guess would be over a million ASCII terminals. Much more significant
than these numbers to the deaf community is the fact that around
500,000 personal computers (Radio Shack, Apple, Commodore and
others) were sold in this country in the last 2 years, and sales
are increasing. These machines all have optional ASCII/Bell 103
interfaces. It is clear that these machines are increasingly going
to be used to talk to either each other, or services marketed over
phone lines (ie. the Radio Shack/Compuserve deal). In my mind the
fundamental issue is whether deaf people are being subsidized merely
to talk to each other and a few essential services that will probably
have TDD numbers, or also be permitted to talk to the world at large.
If the issue were just Baudot/ASCII conversion that could probably
be finessed, but I doubt all sorts of places are going to want to
install Weitbrecht modems. On the other hand, there is the problem
of the existing deaf community. I think that makes it obvious that
if Novation can demonstrate their technology, the the regulations
should then in fact require dual mode terminals.
- Charles Frankston
------------------------------
End of HUMAN-NETS Digest
************************
Telex and TWX
If I remember correctly TWX formerly belonged to AT&T and Telex to
Western Union. We used TWX, which used ASCII instead of Baudot for remote
time-sharing terminals before dataphones and voice couplers were
available. At some time in the 60's, Western Union was about to go out of
business (partly because their technology was obsolete compared to AT&T),
which would have made AT&T even more subject to antitrust, so the FCC
induced AT&T to give TWX to Western Union which now operates both networks
but using lines rented from AT&T. Thus monopoly is created in the name of
competition.
A few years ago I looked into the advertised Mailgrams from
computers. As might be expected, the local Western Union sales office had
never heard of what their company was nationally advertising. When we
finally made contact, they explained that they would come and program our
computer to send Mailgrams into their network. The idea that we might
program our own computer if they would tell us the protocols was too
exotic for them. Anyway, they explained that the service was intended to
send identical Mailgrams to large numbers of people and the minimum charge
was $25. Since our idea was to allow net messages by Mailgram to people
not on the net, there was no meeting of minds. Maybe things are better
now.
∂01-Oct-80 0909 VRP via Dialnet Dialnet performance measurement
To: MRC at SU-AI, LES at SU-AI, JMC at SU-AI, FB at SU-AI
I did a simple performance measurement of Dialnet at 9 am this morning
(Wed. Oct. 1). Typing out the file dln.dl[1,vrp] took 170 seconds.
By comparison, typing out the same file three minutes later using a
conventionally attached 1200/1200 Vadic (SRI's, 326-9129)
took only 50 seconds. The difference was that typeout was
continuous in the latter case, i.e. as fast as the modem could go, whereas
there were long pauses with the former.
If dialnet were to use a 300 baud modem typing out the same file would
presumably require 200 seconds. Thus the use of a 1200 baud modem adds
only 15% to the performance of dialnet, which would seem to defeat the
purpose of running dialnet at 1200 baud.
Although I haven't performed the measurement, I noticed subjectively that
typeout at 2 am was about the same as at 9 am, so the performance
problem can't be attributed entirely to a heavy load on waits.
Is there any way to speed up dialnet? If not then I will not be able to use it.
∂01-Oct-80 1014 JMC*
db
∂01-Oct-80 1012 VRP via Dialnet Dialnet performance
To: MRC at SU-AI, JMC at SU-AI, LES at SU-AI, FB at SU-AI
I tried a different acknowledgment strategy, in which my acknowledgments
were always of the most recently received packet but one. This
dramatically improved dialnet's performance. I suspect the problem is
that I am acknowledging the current packet too soon,
which presumably gets dialnet worried to the point of
losing performance over it (I don't have a more concrete theory).
∂01-Oct-80 1108 FFL
To: JMC, FFL
Ben Cosme would like to see you for a few minutes this afternoon preferably, or
Friday afternoon. Is this possible? (7-4132 or 7-4049 Wong)\
∂01-Oct-80 1227 Boyer at SRI-KL (Bob) Recommendation Request
Date: 1 Oct 1980 1207-PDT
From: Boyer at SRI-KL (Bob)
Subject: Recommendation Request
To: JMC at SU-AI
J Moore and I are applying for posts in the Computer Science
Department at the University of Texas.
Would you be willing to recommend either or both of us?
-------
Yes, I would be glad to recommend both of you.
∂01-Oct-80 1238 VRP via Dialnet Dialnet
To: MRC at SU-AI, LES at SU-AI, JMC at SU-AI, FB at SU-AI
Well, I think I've finally managed to converge on a fairly bug-free
telnet user. I had a residual bug that was causing occasional hanging,
which I finally traced to my not doing DLE doubling properly (the
manual wasn't too clear at what level DLE doubling happened).
Anyway, it works just dandy now, and I'm very pleased with it.
When I get some time I'll think about mail and ftp servers. I looked at
MRC's implementations of these (MAISER.MID[D,MRC] - Mark, where does the
FTP server live?) and (the mail server) didn't seem too hairy. I like
the idea of the messages for the mailer etc. being structured
data rather than unstructured strings - should make the implementation
pretty clean.
∂01-Oct-80 1339 FFL
To: JMC, FFL
Paul Armour sends the message that title you wants is "Turing and Von Neumann's
Dispute Over the Purpose of an Electronic Calculator - as Artificial
Intelligence or as Calculating Device."
Wm. Aspray's phones; Office - 413 597 2158, Home - 413 458 3971
Williams College, Williamstown, MA.
∂01-Oct-80 1755 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
There will be a PROGRAM VERIFICATION AND ANALYSIS SEMINAR this
quarter at the usual time and place (Tuesdays, 2.30 pm, ERL 237).
The first session is planned for Tuesday, Oct. 7; an announcement
with abstract will be mailed out separately.
At present, you are on the net mailing list for the seminar. If you
prefer to be taken off the list, or know of someone who would like
to be added, please contact me.
Also let me know if you are interested in giving a talk in the seminar.
Friedrich von Henke (FWH at SAIL)
I will be interested in giving a talk on new developments in Elephant
late in the quarter.
∂01-Oct-80 2258 MINSKY at MIT-AI (Marvin Minsky) hole through world
Date: 2 OCT 1980 0001-EDT
From: MINSKY at MIT-AI (Marvin Minsky)
Subject: hole through world
To: MINSKY at MIT-AI, jmc at SU-AI, llw at SU-AI
About tube through Earth.
1. Good thinking. That idea has not been taken seriousy before.
2. It seems to me that if the heat can be pumped out fast enough,
then maybe you don't need the dynamic pressure-resistor. For, if you
can keep a few meters solid around the hole, perhaps it will have
enough compressive strength.
This isn't clear to me. Presumably there is a critical pressure above
which there is no "solid" rock anyway, so it will flow. In that case,
the dynamic centrifugal pipe is necessary.
zHowever, in that case, we must assume that the dynamic pipe is itself
under enough compression that its material would flow if it were free
to. So we have to think of the rotating member as a fluid whirlpool.
Question: is it possible that Helium III -- or whatever it is -- the
one that's superfluid, would be frictionless enough at that speed? It
would have to be very cool, of course...
3. There is no reason to suppose that the tube would remain straight,
and whatever large-scale flows there are inside the earth would drag
it around. Nevertheless, we could conceive that whatever process
drills it in the first place might be maintained in the steady state
to keep up with drift.
marvin
∂01-Oct-80 2309 RWG at MIT-MC (Bill Gosper)
Date: 2 OCT 1980 0208-EDT
From: RWG at MIT-MC (Bill Gosper)
To: jmc at SU-AI
I just learned of a possibly useful function DESOLVE which converts your
rabbits and wolves into
(D24) [LAPLACE(R1, T, S) = - (E1 H (B2 LAPLACE(R2 W2, T, S) - R2(0))
+ (B1 S - A2 B1) LAPLACE(R1 W1, T, S) - R1(0) S + R1(0) A2)
2 2
/(S + (- A2 - A1) S - E1 E2 H + A1 A2),
LAPLACE(R2, T, S) = - (S (B2 LAPLACE(R2 W2, T, S) - R2(0))
+ A1 (R2(0) - B2 LAPLACE(R2 W2, T, S)) + B1 E2 H LAPLACE(R1 W1, T, S)
2 2
- R1(0) E2 H)/(S + (- A2 - A1) S - E1 E2 H + A1 A2),
LAPLACE(W1, T, S) = (H (%D2 F1 LAPLACE(R2 W2, T, S) + W2(0) F1)
+ (%D1 S + %C2 %D1) LAPLACE(R1 W1, T, S) + W1(0) S + W1(0) %C2)
2 2
/(S + (%C2 + %C1) S - F1 F2 H + %C1 %C2),
LAPLACE(W2, T, S) = (S (%D2 LAPLACE(R2 W2, T, S) + W2(0))
+ %C1 %D2 LAPLACE(R2 W2, T, S) + %D1 F2 H LAPLACE(R1 W1, T, S) + W1(0) F2 H
2 2
+ W2(0) %C1)/(S + (%C2 + %C1) S - F1 F2 H + %C1 %C2)]
where R1 is R1(t) etc. can this shed any light on the stability?
I'm afraid I'll need some help in decrypting this.
∂02-Oct-80 0024 POURNE at MIT-MC (Jerry E. Pournelle) article
Date: 2 OCT 1980 0324-EDT
From: POURNE at MIT-MC (Jerry E. Pournelle)
Subject: article
To: JMC at SU-AI
I love the title and concept. Pleased to hear that you'll have
it. Best,
JEP
∂02-Oct-80 0234 HUMAN-NETS@MIT-AI HUMAN-NETS Digest V2 #92
Date: 2 OCT 1980 0527-EDT
From: HUMAN-NETS@MIT-AI
Sent-by: DUFFEY at MIT-ML
Subject: HUMAN-NETS Digest V2 #92
To: (@FILE [DSK:AIRAN;DUFFEY HNTXX6]) at MIT-ML
HUMAN-NETS AM Digest Thursday, 2 Oct 1980 Volume 2 : Issue 92
Today's Topics:
WORLDnet - 2-way Cable, Electronic Banking - Auto-teller security,
Computers and the Deaf - Speech Synthesis & Telex/TWX & Talking
& Modems & CA PUC Recommendations,
Telephone Services - Bill Fraud & Advanced Call Services & Phone Book
----------------------------------------------------------------------
Date: 1 October 1980 12:17 edt
From: Janofsky.Tipi at RADC-Multics
Subject: Two Way TV Cables
Would you believe using your Cable TV link for TV games? Well
Mattel and General Instruments have a joint venture trying to exploit
TV cable systems for 'renting' TV games rather than buying them. I
just saw an article in Electronic Engineering Times (29 Sept., pg. 83)
announcing the testing of this idea, called PlayCable, in 1000 rural
homes. The next step is a test marketing activity in Toronto and 'a
major West Coast metro area'. It seems that you will be able to
'rent' these games (i.e. request them over the cable) for $10 to $15
per month rather than purchase them directly. The initial selection
is 15 games but they're planning on regular changes to the list of
games available. Anybody for a game of PONG?
If this is the direction our cable systems are going, there won't
be anything left for WorldNet by the time someone gets around to
building it - all the cable channels will have been taken. (Home
information retrieval - SURE! - use your encyclopedia!)
Bill J.
------------------------------
Date: 1 October 1980 12:24 edt
From: Sibert at MIT-Multics (W. Olin Sibert)
Subject: cameras at automated tellers
Actually, the mirrors are probably concealing ordinary cameras of the
one shot every fifteen seconds variety, of the same sort as have been
in banks watching human tellers and customers since time immemorial.
------------------------------
Date: 1 Oct 1980 at 2002-CDT
From: wilcox at UTEXAS
Subject: Votrax SC-01
I have received word from Votrax that they intend to release a board
including their new SC-01 speech IC sometime in November, at a cost
of apx. $1500. Release of the chip in single quantities is scheduled
for Feb. 81 at a presently undetermined price. They added that they
had not originally intended to release the chip as a end-user item at
all, but changed their minds due to 'MASSIVE CONSUMER INTEREST AND
INQUIRIES'.
--JIM
------------------------------
Date: 01 Oct 1980 0354-PDT
From: John McCarthy <JMC at SU-AI>
Telex and TWX
If I remember correctly TWX formerly belonged to AT&T and
Telex to Western Union. We used TWX, which used ASCII instead
of Baudot for remote time-sharing terminals before dataphones
and voice couplers were available. At some time in the 60's,
Western Union was about to go out of business (partly because
their technology was obsolete compared to AT&T), which would
have made AT&T even more subject to antitrust, so the FCC
induced AT&T to give TWX to Western Union which now operates
both networks but using lines rented from AT&T. Thus monopoly
is created in the name of competition.
A few years ago I looked into the advertised Mailgrams from
computers. As might be expected, the local Western Union sales
office had never heard of what their company was nationally
advertising. When we finally made contact, they explained that
they would come and program our computer to send Mailgrams into
their network. The idea that we might program our own computer
if they would tell us the protocols was too exotic for them.
Anyway, they explained that the service was intended to send
identical Mailgrams to large numbers of people and the minimum
charge was $25. Since our idea was to allow net messages by
Mailgram to people not on the net, there was no meeting of
minds. Maybe things are better now.
------------------------------
Date: 1 October 1980 01:58 edt
From: Frankston at MIT-Multics (BOB at SAI-Prime)
Subject: Reading speed
A written conversation is not limited by the reading speed, but by
the scanning speed. While 120 cps might be good for reading speed,
it is greatly inadequate for scanning speed -- the more normal mode
for reading stuff that is not worth reading.
------------------------------
Date: 30 September 1980 1246-EDT (Tuesday)
From: The Twilight Zone <Doug.Philips at CMU-10A>
Subject: Many people talking at once on one windowed screen
In regards to REM's message which was in regards to Joe Newcomer's
message:
I have seen and used a 'talk' program here at CMU that allowed at most
5 people to talk (type) at once. Each person had three or four lines
(I don't remember which) that was their virtual window. As it turned
out what usually happened was that either you would type along and
ignore (most) of the other windows, hoping that someone was listening
to you [occasionally several people would do this all at once, and the
cursor would flash and thrash, then they would stop to read everone
elses, <goto [>] on the other hand you would get something like a real
conversation where only one person would type at a time, and when they
were done (or more likely had stopped to check their typo frequency)
some one else would start typing. It was quite amusing for a while,
but really doesn't get you anything over a more conventional type
setup like the link mechanism on ITS.
Doug
P.S. please compensate for tense changes as appropriate.
------------------------------
Date: 30 September 1980 03:46 edt
From: Frankston at MIT-Multics (BOB at SAI-Prime)
Subject: Weitbrecht vs 103
How feasible is to to have a modem that can receive both codes and
automatically select (like the Vadic triples do)? Can this be done
for sending. Another aspect is automatic ASCII/Baudot detection.
(For sending, this might involve testing the answerback drums --
unless the modern CRT versions have omitted this useful feature)
(If an anachronism is something that is supposed to be in the past
that can't be because of a modernism, what is the term for the
opposite, something from the past that persists?)
------------------------------
REM@MIT-MC 09/30/80 20:13:23
The PUC reccommendation on D.E.A.F. sounds reasonable. Despite
our fears that they would be required NOT to provide any more than
equivalence even if it turned out to be cheaper, they specifically
stated that extra capabilities would be provided at no charge if
they cost the same as minimal capabilities. The one reservation I
have, of course, is their specification of Baudot. Too bad they
didn't think of converting California to ASCII, sending old Baudot
terminals to other states. Oh well...
------------------------------
Date: 21 Sep 1980 2127-PDT (Sunday)
From: Lauren at UCLA-SECURITY (Lauren Weinstein)
Subject: 3'rd party billings
Some time ago, in the course of a discussion of the Bell System
Voice Storage System in HUMAN-NETS, some mention was made of the
issues of "uncontrolled billings" -- that is, charges which could
be caused to appear on a person's phone bill by some OTHER person.
In addition to new charges possible from the Voice Storage System,
this category also includes collect, credit card, and third party
calls. The following news item concerning the latter two types of
calls seems relevant:
Phone Fraud: Long-Distance Cheaters Cost Millions Yearly
By TOM CRANE Associated Press Writer
NEW YORK (AP) - Telephone customers and phone companies throughout
the nation are being ripped off for more than $30 million annually by
people who charge calls to randomly selected numbers or give false
credit-card numbers.
The latest figures from American Telephone & Telegraph Co. show a
known loss nationally of $18.6 million per year on interstate phone
calls fraudulently billed to unrelated third-party numbers and $13.5
million in credit-card falsifications.
Not included are additional millions paid by phone subscribers
who do not check the long-distance calls on their bills and pay for
other people's calls without realizing they have been victimized.
New York Telephone Co. alone reported fraudulent third-party
billings running at a rate of $2.6 million annually and an additional
$2.1 million in credit-card frauds involving in-state long-distance
calls.
Further, John Quinn, a spokesman for New York Telephone, said
phone companies have no way of telling how extensive the ripoff
may be in fraudulently billed local calls.
AT&T spokesman Jay Grossman said the identified fraudulent
billings, brought to the attention of the company by complaints
from customers who discover they have been charged for calls made
by someone else, amounted to 2.3 percent of the $72.5 million per
month worth of third-party interstate calls.
Both the national and local figures were projections based on
the records of March 1979, the latest month for which accurate
figures exist, the phone companies said.
The question of fraud was explored with the utilities after a
reporter started checking his monthly home phone bill and found
that three of four bills contained improper third-party billings for
long-distance calls that inflated the bill by more than 20 percent.
Billed to his New York number were mystery calls from Los Angeles
to San Francisco, Hollywood to Arizona, and Dallas to Newark, N.J.
In each case, when he informed New York Telephone Co., a company
representative called the locations involved and then deducted the
charges and taxes from his bill, with an apology.
The cost of the fraud is paid by phone subscribers, either
through misbilling or eventual rate increases to cover losses.
Phone companies instruct their operators not to execute long-
distance calls billed to third-party numbers without telephoning the
third party to verify authorization. And third-party billing calls
are not supposed to be accepted from coin telephones between 10 p.m.
and 8 a.m.
Phone companies are not at a complete loss, however. They offer
statistics that show toll fraud is not as safe as some cheaters may
think.
Checks of call destinations and studies of caller patterns - which
phone officials will not discuss publicly - resulted last year in 631
arrests for illegal third-party billings. Of that number, 554 persons
- 87.8 percent - were convicted and subjected to federal penalties of
up to five years in prison and fines ranging to $5,000.
For illegal credit-card billings, AT&T had 225 persons arrested
in 1979 and won 198 convictions.
All 50 states have similar laws, with varying penalties, against
in-state telephone fraud.
------------------------------
This is hardly an encouraging situation. It seems reasonable that
some of the automated systems envisioned for the CCIS telephone data
network will help stem part of these false billings. It is hard to
imagine how to stamp them out entirely without refusing to accept
such calls completely, which is not considered reasonable at this
time. One can only hope that whole new avenues of problems do not
result from widespread implementation of the Voice Storage System
services.
--Lauren--
------------------------------
Date: 30 September 1980 0845-EDT (Tuesday)
From: Joe.Newcomer at CMU-10A
Subject: Split Yellow Pages
Pittsburgh recently split its yellow pages into "For People" and
"For Business". I found out very quickly that all the interesting
places I like to deal with are classified as people who only are of
interest to "Business", not "People". This is the first I realized
that I are not a people. It seems several other folks were upset;
this year's edition is "Business/Industrial" and "Consumer" (although
again I didn't realize that business and industry did not consume).
Fortunately, once having insisted on a REAL yellow pages, I now
continue to get it without asking. However, it is a real pain to
try to second-guess which one the item I want will be in; I usually
have to check both.
joe
------------------------------
Date: 30 Sep 1980 11:00:01 EDT
From: Bernie Cosell <cosell at BBN-UNIX>
Subject: boston phone books
It is fortunate that Don moved, because the North End, South Boston
and East Boston are all in the main directory. I don't know how it
is now, but as of about a few years ago (when I gave up trying to
track all this), the Boston area was "served" by about ten yellow
pages directories (not to mention the four white pages). In their
wisdom, NET decided that the three suburban white pages regions were
too large and so they subdivided each into several yellow pages zones.
A few times I actually hassled someone at the phone company to order a
complete set, but it was an awful lot of trouble (for me, I don't much
care how they felt about it). Things apparently stay bad as you work
west of Boston: the state seems to be covered by a virtually endless
series of disconnected tiny regional directories, making it virutally
impossible to locate a phone number if you have either only a vague
notion of where the person lives or only a vague notion of the state's
topography in the area. I have to admit that on the instances when
I've needed it, the information operators have been very helpful, but
when I've been looking for a town that they didn't recognize it was
pretty much hopeless. I originally came from New York City and I was
always impressed by the efficiency of covering all of Long Island
with two (large) directories - I don't know if the endless tiny towns
publish their own small directories, but outsiders can safely ignore
the problem.
Is Massachusetts the common case, or is New York? That is, is it the
norm around the country to have an almost endless collection of tiny
directories, or to have the directories based on larger (and, to my
view, more useful) chunks of territory?
/Bernie
------------------------------
End of HUMAN-NETS Digest
************************
∂02-Oct-80 0809 FFL Expense report for David Waltz
To: JMC, FFL
Am I to go ahead putting the expense three quarters to ARPA and
one quarter to Tom Binford's NSF?
∂02-Oct-80 0811 FFL Meeting day and time for Seminar
To: JMC, FFL
I have had no objections to Thursday at 4:15 from anyone except
Louis Creary. He says he may have to leave early on that day if the
seminar is long, which he does not want to do. I need to let the others
know definitely. Lew suggested moving the hour to 3:30. Perhaps even
4 p.m. instead of 4:15 would help?
∂02-Oct-80 0813 FFL
To: JMC, FFL
I pubbed CBCL and you said you wanted to give me a note to go with it
to Massy in the Business School.
∂02-Oct-80 0948 FWH PV+A Seminar
To: "@SEM.DIS[SEM,VER]" at SU-AI
PROGRAM VERIFICATION AND ANALYSIS SEMINAR
PLACE: ERL 237
TIME: 2:30 p.m., Tuesday, October 7
SPEAKER: Olaf Owe, Stanford University and University of Oslo
TITLE: PROVER - the Oslo PROgram VERifier
ABSTRACT:
An overview of the verification system developed at the
University of Oslo will be given.
The talk will concentrate on the verification condition
generator and the simplifier, explaining some features that are not
found in the Stanford Verifier - such as how to generate so called
"effect function theorems" without requiring any pre- or post-conditions
(only loop invariants). Also some "framing" techniques will be shown,
as well as an ad-hoc simplification method, useful for proving
quantified loop invariants.
∂02-Oct-80 1001 SOL via MIT-TIP we be inBay Area oct 16-17
To: JMC, ME
I'll be in your neck of the woods again, on a CATv project,
and would like to drop by. Either Oct 16 or17 is my current
schedule. Richard
∂02-Oct-80 1006 JMC*
rochester
∂02-Oct-80 1011 TOB ARPA
John
Marianne found out from Betty Scott that new ARPA
money has not come in yet, (approval?). And also
that we can't spend the money which remains from
the previous year. Do you know anything about
it? I am going to follow up with Betty.
Tom
$
∂02-Oct-80 1100 JMC*
Rochester and Teller