perm filename S81.IN[LET,JMC] blob
sn#602257 filedate 1981-07-29 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00204 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00021 00002 ∂01-Apr-81 1419 Navarro at SRI-AI consultant agreement
C00023 00003 ∂01-Apr-81 1550 Masinter at PARC-MAXC
C00028 00004 ∂01-Apr-81 1831 HPM
C00030 00005 ∂01-Apr-81 1943 RPG Ballpark
C00031 00006 ∂01-Apr-81 2150 RWG at MIT-MC (Bill Gosper)
C00033 00007 ∂02-Apr-81 2255 JK
C00034 00008 ∂03-Apr-81 1002 LEWIS at SRI-AI (Bil Lewis) WCCF gathering
C00036 00009 ∂03-Apr-81 1309 HPRINTZ at BBNE Newsletter
C00037 00010 ∂03-Apr-81 1321 FFL
C00038 00011 ∂03-Apr-81 1337 FFL
C00039 00012 ∂04-Apr-81 1631 LLW Inverting Signs of the Times
C00041 00013 ∂04-Apr-81 2130 MINSKY at MIT-ML (Marvin Minsky)
C00043 00014 ∂06-Apr-81 0807 DOLESE at RUTGERS arpanet access to Rutgers Technical Reports
C00047 00015 ∂06-Apr-81 0902 FFL
C00048 00016 ∂06-Apr-81 0907 FFL
C00049 00017 ∂06-Apr-81 1207 LES
C00050 00018 ∂06-Apr-81 1352 FFL
C00052 00019 ∂06-Apr-81 1451 NEUMANN at SRI-KL [DUCKETT at USC-ISIE: Verkshop II/IPTO Contribution]
C00058 00020 ∂06-Apr-81 1735 CSD.ULLMAN at SU-SCORE Vaughan Pratt
C00059 00021 ∂06-Apr-81 1740 CSD.ULLMAN at SU-SCORE binford case
C00060 00022 ∂06-Apr-81 1745 TOB thanks
C00061 00023 ∂06-Apr-81 1736 100 : REM via SU-TIP Bookstore and women
C00063 00024 ∂07-Apr-81 0030 LGC Today's A-T Discussion
C00065 00025 ∂07-Apr-81 0505 Darden@SUMEX-AIM Washington
C00067 00026 ∂07-Apr-81 0811 Betty Scott <CSD.BSCOTT at SU-SCORE> Next Week
C00068 00027 ∂07-Apr-81 0843 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) comp programming problem
C00069 00028 ∂07-Apr-81 0903 TOB ARPA proposal
C00070 00029 ∂07-Apr-81 0916 DEK VRP
C00071 00030 ∂07-Apr-81 0924 RPG Rules
C00075 00031 ∂07-Apr-81 1321 CLT
C00076 00032 ∂07-Apr-81 1323 JONL at MIT-MC (Jon L White) Proposed ''mini'' benchmark, with interpretation.
C00089 00033 ∂08-Apr-81 0200 ZM Pratt
C00090 00034 ∂08-Apr-81 1049 FFL Comprehensives Committee meeting
C00091 00035 ∂08-Apr-81 1245 Buchanan@SUMEX-AIM Jonathan King's application
C00092 00036 ∂08-Apr-81 1430 CSD.ULLMAN at SU-SCORE
C00093 00037 ∂08-Apr-81 1552 TOB arpa proposal
C00094 00038 ∂08-Apr-81 1729 Vaughan Pratt <CSD.PRATT at SU-SCORE> Apollo presentation
C00095 00039 ∂08-Apr-81 1951 DCL
C00096 00040 ∂09-Apr-81 1042 CSD.JEANIE at SU-SCORE April 21 Faculty Meeting
C00097 00041 ∂09-Apr-81 1108 FFL
C00098 00042 ∂09-Apr-81 1150 Vaughan Pratt <CSD.PRATT at SU-SCORE> Apollo presentation
C00099 00043 ∂09-Apr-81 1324 FFL
C00100 00044 ∂09-Apr-81 1417 FFL
C00101 00045 ∂09-Apr-81 1451 FFL
C00102 00046 ∂09-Apr-81 1806 SQU Thanks
C00104 00047 ∂10-Apr-81 0933 Buchanan@SUMEX-AIM Jonathan King's application resolve
C00106 00048 ∂10-Apr-81 1053 COHEN at PARC-MAXC Re: Benjamin Cohen thesis
C00107 00049 ∂10-Apr-81 1205 George J. Carrette <GJC at MIT-MC> Rules
C00109 00050 ∂10-Apr-81 1051 HEDRICK at RUTGERS Re: Rules
C00113 00051 ∂10-Apr-81 1818 JD arpa proposal
C00114 00052 ∂10-Apr-81 1958 Janofsky.Tipi at MIT-Multics (Bill Janofsky) Re: Student's credulity (JMC@SU-AI msg in HNT V3#77)
C00116 00053 ∂10-Apr-81 1958 Janofsky.Tipi at MIT-Multics (Bill Janofsky) Re: Domestic Computers
C00118 00054 ∂11-Apr-81 0935 LES ARPA proposal
C00119 00055 ∂12-Apr-81 0242 Stephen C. Hill <STEVEH at MIT-MC> Oops
C00120 00056 ∂12-Apr-81 1008 CLT today
C00121 00057 ∂12-Apr-81 1155 CLT
C00122 00058 ∂12-Apr-81 1357 CLT
C00123 00059 ∂12-Apr-81 1916 Robert Elton Maas <REM at MIT-MC>
C00124 00060 ∂13-Apr-81 0944 JD proposal file
C00125 00061 ∂13-Apr-81 1119 LGC Discussion Today?
C00126 00062 ∂13-Apr-81 1239 RPG Groundrules (reprise)
C00131 00063 ∂13-Apr-81 1347 FFL
C00132 00064 ∂13-Apr-81 1357 Brian K. Reid <CSL.BKR at SU-SCORE> address space
C00133 00065 ∂13-Apr-81 1404 COHEN at PARC-MAXC Reply to Comments
C00134 00066 ∂13-Apr-81 1457 FFL Thesis Committee meeting for Vic Scheinman
C00135 00067 ∂13-Apr-81 1736 Mark Crispin <Admin.MRC at SU-SCORE>
C00136 00068 ∂13-Apr-81 2327 the tty of Geoffrey S. Goodfellow [HEDRICK at RUTGERS: report on Lisp conference]
C00160 00069 ∂14-Apr-81 0118 V. Ellen Golden <ELLEN at MIT-MC> REM
C00164 00070 ∂14-Apr-81 0122 V. Ellen Golden <ELLEN at MIT-MC>
C00165 00071 ∂14-Apr-81 0123 V. Ellen Golden <ELLEN at MIT-MC>
C00167 00072 ∂14-Apr-81 0151 V. Ellen Golden <ELLEN at MIT-MC> Betsy
C00171 00073 ∂14-Apr-81 0205 V. Ellen Golden <ELLEN at MIT-MC>
C00172 00074 ∂14-Apr-81 0859 LLW A NIL-Potent Advice Taker?
C00174 00075 ∂14-Apr-81 1057 CSD.CRANGLE at SU-SCORE newell article
C00175 00076 ∂14-Apr-81 1215 CLT
C00178 00077 ∂14-Apr-81 1704 Richard M. Stallman <RMS at MIT-AI>
C00181 00078 ∂14-Apr-81 2219 Richard M. Stallman <RMS at MIT-AI>
C00182 00079 ∂14-Apr-81 2304 ZM Pratt
C00183 00080 ∂15-Apr-81 1645 CSD.GOLUB at SU-SCORE Thanks
C00184 00081 ∂15-Apr-81 1914 CSD.DOYLE at SU-SCORE Proposals concerning AAAI publications
C00191 00082 ∂15-Apr-81 2038 Woody Bledsoe <ATP.Bledsoe at UTEXAS-20> THANKS
C00192 00083 ∂15-Apr-81 2138 CLT arpa
C00193 00084 ∂15-Apr-81 2225 CLT
C00194 00085 ∂15-Apr-81 2253 CLT
C00195 00086 ∂16-Apr-81 0017 CLT Frege
C00196 00087 ∂16-Apr-81 0106 JK
C00198 00088 ∂16-Apr-81 0815 JK
C00200 00089 ∂16-Apr-81 1051 FFL
C00201 00090 ∂16-Apr-81 1134 Walker at SRI-AI Re: Proposals concerning AAAI publications
C00205 00091 ∂16-Apr-81 1248 TOB
C00206 00092 ∂16-Apr-81 1443 JJW EKL proof
C00207 00093 ∂16-Apr-81 2325 Robert Elton Maas <REM at MIT-MC> Non-loneliness
C00211 00094 ∂17-Apr-81 0510 Darden@SUMEX-AIM trip to Washington
C00212 00095 ∂17-Apr-81 1536 Untulis at SRI-AI Badge application
C00213 00096 ∂18-Apr-81 1123 CLT
C00214 00097 ∂18-Apr-81 1628 RPG
C00215 00098 ∂18-Apr-81 1743 LGC Tomorrow
C00216 00099 ∂18-Apr-81 2210 LLW Talking About RPG
C00217 00100 ∂18-Apr-81 2250 CLT
C00218 00101 ∂19-Apr-81 0051 RPG Meeting
C00219 00102 ∂19-Apr-81 1054 JD analyst
C00221 00103 ∂19-Apr-81 1555 Darden@SUMEX-AIM Re: Monday
C00223 00104 ∂20-Apr-81 0235 CLT ARPA
C00224 00105 ∂20-Apr-81 1231 CLT phon
C00225 00106 ∂20-Apr-81 1824 BYY situations
C00226 00107 ∂20-Apr-81 2118 BYY acl paper
C00253 00108 ∂21-Apr-81 1040 FFL
C00254 00109 ∂21-Apr-81 1100 UNTULIS at SRI-AI [VIVIAN: Re: [Untulis: McCarthy account]]
C00255 00110 ∂21-Apr-81 1104 FFL supplemental pay in the summer
C00256 00111 ∂21-Apr-81 1809 REM via SU-TIP My crunch-and-spindle program
C00257 00112 ∂22-Apr-81 1148 FFL
C00258 00113 ∂22-Apr-81 1310 Buchanan@SUMEX-AIM agenda item for AAAI from Jon Doyle
C00288 00114 ∂22-Apr-81 1328 CLT
C00289 00115 ∂23-Apr-81 0752 CERF at USC-ISI Re: verification
C00291 00116 ∂23-Apr-81 0927 Konolige at SRI-AI (Kurt Konolige) Re: tomorrow
C00292 00117 ∂23-Apr-81 1401 RPG
C00294 00118 ∂23-Apr-81 2218 CLT LTE
C00295 00119 ∂23-Apr-81 2237 Daul at OFFICE (Response to message)
C00297 00120 ∂24-Apr-81 0001 LLW Summer Appointment Offer
C00299 00121 ∂24-Apr-81 0041 LLW Summer Arrangements
C00302 00122 ∂24-Apr-81 0133 LLW Divine Intervention
C00306 00123 ∂24-Apr-81 0136 LLW Outrageous Terms
C00307 00124 ∂24-Apr-81 1138 FFL
C00308 00125 ∂24-Apr-81 1211 ROD my dissertation
C00310 00126 ∂24-Apr-81 1307 LGC NSF Discussion
C00311 00127 ∂25-Apr-81 0858 Darden@SUMEX-AIM (Response to message)
C00313 00128 ∂25-Apr-81 1112 Susan L. Gerhart <GERHART at USC-ISIF> M. Davis paper
C00315 00129 ∂26-Apr-81 1719 ROD signature pages
C00316 00130 ∂27-Apr-81 0606 Darden@SUMEX-AIM Re: 8th day
C00317 00131 ∂27-Apr-81 0908 FFL
C00318 00132 ∂27-Apr-81 0900 JMC*
C00319 00133 ∂27-Apr-81 0900 JMC*
C00320 00134 ∂27-Apr-81 0931 FFL
C00321 00135 ∂27-Apr-81 1303 CLT
C00322 00136 ∂27-Apr-81 1302 CLT
C00323 00137 ∂27-Apr-81 1501 FFL
C00324 00138 ∂27-Apr-81 1502 FFL
C00325 00139 ∂27-Apr-81 1503 FFL
C00326 00140 ∂27-Apr-81 1549 JAK Departure
C00327 00141 ∂27-Apr-81 2353 DCL Amalgamated ARPA
C00328 00142 ∂28-Apr-81 0426 Darden@SUMEX-AIM (Response to message)
C00329 00143 ∂28-Apr-81 1117 FFL
C00330 00144 ∂28-Apr-81 1202 DGCOM at USC-ISIC FOL
C00332 00145 ∂28-Apr-81 1300 RPG Summer Student
C00333 00146 ∂28-Apr-81 1255 RPG
C00335 00147 ∂28-Apr-81 1523 KEETON at USC-ISI translation algorithms
C00336 00148 ∂28-Apr-81 1629 TW AI question
C00337 00149 ∂28-Apr-81 2259 RWW nsf grant
C00338 00150 ∂28-Apr-81 2346 JK ramsey's theorem
C00339 00151 ∂29-Apr-81 1100 CSD.ULLMAN at SU-SCORE
C00341 00152 ∂29-Apr-81 1134 FFL
C00342 00153 ∂29-Apr-81 1136 FFL on TTY62 (at TV-141) 1136
C00343 00154 ∂30-Apr-81 1608 CLT
C00344 00155 ∂01-May-81 1509 CERF at USC-ISI Re: verification
C00345 00156 ∂01-May-81 1635 Bill.Scherlis at CMU-10A Dana Scott
C00346 00157 ∂01-May-81 1905 Bill Gosper <rwg at MIT-MC>
C00348 00158 ∂02-May-81 2227 Steve Kudlak <FFM at MIT-MC> REM
C00351 00159 ∂04-May-81 1051 FFL Speaking at U. of N.M. when you are at Sandia Lab
C00352 00160 ∂06-May-81 1403 CSD.ULLMAN at SU-SCORE IBM grant
C00353 00161 ∂07-May-81 1308 Vaughan Pratt <CSD.PRATT at SU-SCORE> Etude documentation
C00355 00162 ∂08-May-81 0647 Navarro at SRI-AI your badge
C00356 00163 ∂12-May-81 1319 Moore at SRI-CSL (J Moore) University of Texas
C00357 00164 ∂14-May-81 1226 CSD.ULLMAN at SU-SCORE Leo Guibas
C00360 00165 ∂14-May-81 1318 Randall Davis <KRD at MIT-AI>
C00364 00166 ∂16-May-81 1745 CSD.GENESERETH at SU-SCORE AI Qual
C00367 00167 ∂16-May-81 2149 CSD.LENAT at SU-SCORE Re: AI Qual
C00370 00168 ∂17-May-81 1918 ENGELMORE at USC-ISI Visit
C00371 00169 ∂18-May-81 1019 TW
C00373 00170 ∂18-May-81 1641 CSD.GENESERETH at SU-SCORE
C00374 00171 ∂19-May-81 1155 Bobrow at PARC-MAXC A new feature for the AI journal
C00379 00172 ∂19-May-81 1359 Feldman@SUMEX-AIM Job Application
C00380 00173 ∂19-May-81 1403 Feldman@SUMEX-AIM Job Application
C00382 00174 ∂27-May-81 1917 Richard M. Stallman <RMS at MIT-AI>
C00384 00175 ∂28-May-81 0801 FFL
C00385 00176 ∂28-May-81 1511 Denny Brown <CSD.DBROWN at SU-SCORE> Office space plans
C00387 00177 ∂29-May-81 1607 JK hello John:
C00388 00178 ∂01-Jun-81 0138 Richard M. Stallman <RMS at MIT-AI>
C00389 00179 ∂02-Jun-81 2105 Richard M. Stallman <RMS at MIT-AI>
C00390 00180 ∂02-Jun-81 2126 Richard M. Stallman <RMS at MIT-AI>
C00392 00181 ∂04-Jun-81 0038 POURNE@MIT-MC
C00394 00182 ∂05-Jun-81 1118 FFL
C00395 00183 ∂08-Jun-81 0949 FFL
C00396 00184 ∂08-Jun-81 1504 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) grade for John Kelley
C00398 00185 ∂10-Jun-81 1057 Walker at SRI-AI AAAI office space
C00400 00186 ∂12-Jun-81 1450 LOUNGO at RUTGERS Rutgers Technical Reports
C00403 00187 ∂14-Jun-81 1357 LWE ns files
C00404 00188 ∂17-Jun-81 1628 CSD.ULLMAN at SU-SCORE Adopt an Orphan this Summer
C00405 00189 ∂18-Jun-81 0047 Bill Gosper <RWG at MIT-MC> crinkle
C00406 00190 ∂18-Jun-81 2107 BYY terminal
C00407 00191 ∂19-Jun-81 1349 RFN
C00408 00192 ∂22-Jun-81 2302 OTA Renewing the space program
C00410 00193 ∂23-Jun-81 0012 Jerry E. Pournelle <POURNE at MIT-MC> nets
C00412 00194 ∂24-Jun-81 0317 Jerry E. Pournelle <POURNE at MIT-MC> space and communications
C00413 00195 ∂25-Jun-81 1013 Wiederhold at SUMEX-AIM ARPA consolidated contract
C00415 00196 ∂26-Jun-81 0639 JRA lisp course
C00416 00197 ∂27-Jun-81 0510 Robert Elton Maas <REM at MIT-MC> Go endgame algorithm, new breakthrough
C00422 00198 ∂29-Jun-81 0043 Kenneth Kahn <KEN at MIT-AI> an answer to your questions about Omega
C00425 00199 ∂29-Jun-81 1420 Betty Scott <CSD.BSCOTT at SU-SCORE> ARPA Supplement for LISP Project
C00427 00200 ∂29-Jun-81 1543 Wiederhold at SRI-AI schedule for visit by Machado nvalex-arpa
C00430 00201 ∂30-Jun-81 0427 JRA lisp course dates
C00431 00202 ∂30-Jun-81 1004 FFL
C00432 00203 ∂30-Jun-81 1005 FFL
C00433 00204 ∂30-Jun-81 1126 WIEDERHOLD at SRI-AI updated schedule
C00437 ENDMK
C⊗;
∂01-Apr-81 1419 Navarro at SRI-AI consultant agreement
Date: 1 Apr 1981 1419-PST
From: Navarro at SRI-AI
Subject: consultant agreement
To: jmc at SU-AI
cc: navarro
I am filling out the consultant agreement and there are a few things I need
to know.
1. Your address and telephone number.
2. Other clients for which you have worked at this rate of of pay.
3. Field of expertise (we need a biography).
Please get this information to me ASAP.
You can reach me at NAVARRO@SRI-AI.
Thanks. Georgia
-------
846 Lathrop Dr.
Stanford CA 94305
857-0672 home, 497-4430 Stanford
I worked for Information International at this rate several years
ago, but I haven't been doing consulting work recently.
You can FTP BIOJMC[PAT,JMC] from SAIL, but I'll send you a paper copy.
∂01-Apr-81 1550 Masinter at PARC-MAXC
Date: 1 Apr 1981 15:49 PST
From: Masinter at PARC-MAXC
To: LispTiming at SU-AI
These are numbers that I generated in late 1977, measuring instruction
counts for various Interlisp-10 operations. One thing to be careful of in
measuring Interlisp-10 is to watch whether the functions are swapped
or not.... it makes a big difference. I suggest Interlisp-10 timings should
be made (at least once) with NOSWAPFLG set to T before the timed
program is loaded in.
------ begin forwarded message -------
I have just made some measurments of how many instructions it takes
to do various things in LISP, and I thought they might be of general
interest.
All measurements are in number of PDP-10 instructions, taking
no account of the relative speed of those instructions.
Measurements for Maxc (which has some special PDP-10 mode
instructions to improve function call performance) are given in
parentheses.
To call a non-swapped compiled function (not in a block) which has
all of its args LOCALVARS takes 50 instructions. (28 on Maxc)
{note that in this and subsequent figures, the time "to call" something
also includes the time to return from it}
To call a SUBR of 1 argument takes 56 instructions. (30 on maxc)
To call a function in the same block where the called function
has all of its args LOCALVARS takes 4 instructions + 1 for each
formal argument.
If the called function has any of its arguments SPECVARS then
it takes 57 instructions plus 12 for each SPECVAR arg and 2 for
each non-specvar arg. To bind variables with a PROG is roughly
the same. (this is 25+9/specvar on Maxc)
Block entry takes 69 instructions, i.e. (BLOCK (FOO FOO)) then
to call FOO will take 19 more instructions than (BLOCKS (NIL FOO
(LOCALVARS . T)))
(this is 45 on Maxc, i.e. about 17 more for block entry).
{you want to do the former if FOO calls itself recursively, though}.
To do a BLKAPPLY* takes 80 instructions + 3 per entry on blkapplyfns
which must be skipped (i.e. if you BLKAPPLY 'FOO and FOO is the
third entry on BLKAPLYFNS then this is 6 extra instructions).
(same on Maxc)
To call a SWAPPED function takes at least 86 additional instructions
per call. This is independent of whether the called function is
a block or a simple function, etc.
A LINKED function call takes 10 more instructions than a non-linked
function call. You should therefore always put (NOLINKFNS . T)
in your blocks declaration unless you have a specific
reason for wanting the calls linked.
∂01-Apr-81 1831 HPM
To: JMC
CC: MLB
∂31-Mar-81 2000 JMS
∂31-Mar-81 1646 MLB
∂23-Mar-81 0920 JMS
was it you who wrote the KAFFEE program? If so, you might
want to try running it again. It logs you out after 1 minute (!)...
Originally, yes. But HPM moifie it for his own use uring his thesis
daze. You might ask him what's become of it...
HPM - I made a version that lasted an hour (and counted down in
its jobname). Some people found it objectionable and modified it to
log out almost immediately (this change was perpetrated anonomously
about once a day). To counter, I protected the DMP file, and
that kept it safe for a while (but obviously not indefinitely).
I'd say there is social pressure against the running of KAFFEE.
Safest thing is to write your own.
∂01-Apr-81 1943 RPG Ballpark
One thing I'd like to get a grasp on pretty quickly is whether
my working for you next fall on the Advice Taker is still
within the realm of possibility. Otherwise I'd like to keep
exploring other possibilities. Staying at Stanford and working
on this project seems like it would be a top choice, so the
rest depends on mutual ability to accomodate the other's point
of view a bit.
-rpg-
∂01-Apr-81 2150 RWG at MIT-MC (Bill Gosper)
Date: 2 APR 1981 0050-EST
From: RWG at MIT-MC (Bill Gosper)
To: DTB at MIT-MC, BROWN at MIT-MC, RP at MIT-MC, RZ at MIT-MC
To: REM at MIT-MC, MACDON at MIT-MC, JHD at MIT-MC, ES at MIT-MC
To: jmc at SU-AI, mlb at SU-AI
CC: RWG at MIT-MC
Tnx to mlb, we have on p56 of melzak, v1:
Polynomial(f(x), f(y), f(x+y)) = 0 implies f is elliptic (or degenerate
case thereof). But, uncharacteristically, he gives no references except
to call this "the theorem of Weierstrass".
Tnx to es, we have a footnote on p519 of Whittaker and Watson (with no
clues in the index to its existence):
"A thm due to Weierstrass states that an analytic function, f(z),
possessing an addition thm in the strict sense must be either..."
<clumsy statement equiv to elliptic fcn or degenerate case>
"See Forsyth, Theory of Fcns, (1918), Ch. XIII." I shall.
∂02-Apr-81 2255 JK
∂02-Apr-81 2243 JMC arpa2.doc[doc,jk]
Well, I think it needs a bit more. I put the "Ketonen plans ..." in a separate
final paragraph, but the sentence isn't liftable by Engelmore, because it
refers to "these techniques". If you're tired, perhaps you can give it
some more thought tomorrow.
-------
I think you were looking at the wrong page - anyhow, I killed the page mark
so it's all on the same page now.
∂03-Apr-81 1002 LEWIS at SRI-AI (Bil Lewis) WCCF gathering
Date: 3 Apr 1981 0958-PST
From: LEWIS at SRI-AI (Bil Lewis)
Subject: WCCF gathering
To: jmc at SU-AI
John,
Jim McG suggested that I include you in this msg.
I am going to show up at the WCCF Sat late morning, and give
my talk on natural language interfaces to databases in the 4:00-5:30
session in Rm 2 (where ever that might be).
Talking to Jim McGrath, the most logical place to meet seems
to be at the only booth that I know of, which is the one run by the
GRAPHICS GATHERING. I'm going to be there at 2:00pm shooting the S←←←
with whomever. So what do you say?
Bil
-------
∂03-Apr-81 1309 HPRINTZ at BBNE Newsletter
Date: 3 Apr 1981 1608-EST
From: HPRINTZ at BBNE
Subject: Newsletter
To: JMC at SU-AI
Dear John McCarthy,
In her book "Machines Who Think," Pamela McCorduck mentions a
newsletter that you write, available to anyone on the ARPANET. I am a
new employee of BBN, where I work on the ARPANET. Can you put me on
the mailing list? My computer-mail address is HPrintz@BBNE.
Thank-you,
Harry Printz
-------
I didn't continue with the public files, partly from lack of feedback.
∂03-Apr-81 1321 FFL
To: JMC, FFL
Elliott Bloom would like to have lunch with you at 12:30, Faculty Club,
Thursday, April 9. I am supposed to report your answer to his secretary.
∂03-Apr-81 1337 FFL
To: JMC, FFL
Bob Wagoner from Physics called to remind you that you are giving a
seminar at his home at 2:30 on Sunday.
∂04-Apr-81 1631 LLW Inverting Signs of the Times
To: minshy at MIT-AI
CC: LLW at SU-AI, RAH at SU-AI, JMC at SU-AI
∂04-Apr-81 1544 MINSKY at MIT-ML (Marvin Minsky)
Date: 4 APR 1981 1845-EST
From: MINSKY at MIT-ML (Marvin Minsky)
To: MINSKY at MIT-ML, llw at SU-AI, rah at SU-AI, jmc at SU-AI
Hey folks, dynamioc compression is getting popular
already. Why don't we get that paper out
[Marvin, I pioneered (unclassified aspects of) dynamic compression a dozen
years ago--the field is known these days as inertial confinement fusion,
and it can certainly use all the popularity it can get (especially in
OMB). However, what we're working on now is dynamic *tension*! We'll
finish it up just as soon as we can--at the moment, however, Rod is busy
single-handedly oiling what Teller has recently labeled as `a major pivot
point of the current geopolitical scene,' and hasn't been able to put the
finishing touches on the Starbridge quite as rapidly as either of us had
hoped. Meanwhile, what happened to your visit out here of `the end of the
month?']
∂04-Apr-81 2130 MINSKY at MIT-ML (Marvin Minsky)
Date: 5 APR 1981 0031-EST
From: MINSKY at MIT-ML (Marvin Minsky)
To: jmc at SU-AI
Noam's number is 862-6160
Couqueaux in Marseille is involved in what Chomsky considers major
new things in linguistics.
The Chomsky world:
U.S. and Russia are not afraid of each other, but are allies.
U.S. gov't hopes for Soviet invasion of Poland.
Capitalists want invasion of Poland to suppress Solidarity so that
austerity measures leading to repayment of foreign debts will occur.
Reagan administration wants to increase "state sector" of the economy.
A possible inversion: A fears that B will do X and tries to get allies
to prevent B from doing X. Difficulties in getting these allies lead A
to hope that B will start doing X so that the allies will feel more
threatened.
∂06-Apr-81 0807 DOLESE at RUTGERS arpanet access to Rutgers Technical Reports
Date: 6 Apr 1981 1103-EST
From: DOLESE at RUTGERS
Subject: arpanet access to Rutgers Technical Reports
To: Erman at USC-ISIB
cc: Shortliffe at SUMEX-AIM, Driefus at WHARTON-10, csd.bennett at SU-SCORE,
Mittal at RUTGERS, Chandrasekaran at RUTGERS, JSmith at RUTGERS,
Deolankar at RUTGERS, Wilkins at SRI-KL, Bruce at BBNA, Webber at BBND,
Friedland at SUMEX-AIM, Plondon at USC-ISIB, ERM at MIT-AI, RDG at SU-AI,
Pressburger at SCI-ICS, csd.gardner at SU-SCORE, Fagan at SUMEX-AIM,
Fikes at PARC-MAXC, JMC at SU-AI, Clancey at SUMEX-AIM, KRD at MIT-AI,
Hamilton.es at PARC-MAXC, cs.Amsler at UTEXAS-20, ChiNguyen.es at PARC-MAXC,
lisa at UTEXAS-11, kwh at MIT-AI
Recently I have received several requests for arpanet access to our
technical reports.
In general, we don't have entire reports available online (even those that
individual authors do have online aren't necessarily useful, since they don't
always contain the figures).
However, what we do have available are the following three files:
<LIBRARY>tech-reports.all (This file contains the entire listing of Rutgers
technical reports from year one to the present).
<LIBRARY>publication-order.form (This file contains the most recent listing
of technical reports that we are presently publicizing).
<LIBRARY>online-tecrpts.doc (This file contains an abstract of each of the
reports listed in publication-order.form)
This message will be a duplicate for some of you. Our present policy is to
release a new listing of technical reports every other month or when we have
a sufficient amount to warrant release.
The files mentioned above are available using FTP user account anonymous
with any password.
If after you have looked at these files and would like to order you can either
send me a message via arpanet or a letter requesting the reports you want.
The mailing address is:
Technical Reports Librarian
Rutgers University
Department of Computer Science
Hill Center - Busch Campus
New Brunswick, New Jersey 08903
I hope that this clarifies our policy on accessing information on our
technical reports. If you have any questions or comments please send
me a message via arpanet mail. The next scheduled release of Rutgers
Technical Reports should be toward the end of May.
Thank you.
Cathy Dolese <DOLESE@RUTGERS>
-------
∂06-Apr-81 0902 FFL
To: JMC, FFL
Dr. Ralph Foster of Northeast U. in Oklahoma called to ask your recommendation
for a LISP processor tthat will run with a H-P 3000, Series 3.
918 456 55ll, Ext. 521.
∂06-Apr-81 0907 FFL
To: JMC, FFL
∂03-Apr-81 0122 JMC
If you haven't sent Pournelle report to Levinthal yet, please hold it.
I sent the latest - the "final" report - to E. Levinthal.
∂06-Apr-81 1207 LES
∂06-Apr-81 0135 JMC arpa proposal
To: LES
CC: FFL
Do you know what file previous ARPA proposal might be?
-----------
The main file is ARPA.PUB[P,LES]. It calls in files on the various
projects. The latter have been deleted for reason of space saving, but they
may be PUMPKINed back into existence if you wish.
∂06-Apr-81 1352 FFL
To: "@COMPRE.[1,FFL]" at SU-AI
COMPREHENSIVES COMMITTEE
At the meeting on April 3, the following area assignments were made.
Please bring questions for review at the next meeting.
MTC Floyd and Weening
Systems-Hardware Owicki and Ossher
AA Spencer
NA Golub
AI Winograd and Clarkson
Please also think of changes to the Winter 1981-82 Reading List and send
them to Joe Weening (JWW at Sail) before the next meeting.
The next meeting will be a a brown bag lunch on Monday, April 13,
in Room 402 MJHall at 12:20 p.m. Please reply by April 7 if you are
unable to be there.
∂06-Apr-81 1451 NEUMANN at SRI-KL [DUCKETT at USC-ISIE: Verkshop II/IPTO Contribution]
Date: 6 Apr 1981 1429-PST
From: NEUMANN at SRI-KL
Subject: [DUCKETT at USC-ISIE: Verkshop II/IPTO Contribution]
To: VERKshop: ;
OK, gang, despite Vint's presumably still being worried about his
income tax return, he was this year again able to come in early in
filing his VERkshop return. I am redistributing it at this time to
help jog you all into action, if you haven't yet thought about what
you might want to send in for redistribution. The deadline for
getting your thoughts into the main handouts to be distributed is
probably Wednesday AM, 15 April. You may subsequently revise your
contributions if you wish them to appear in SEN -- the deadline for
that is 21 June. Peter
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
A View of Verification Technology
Vinton G. Cerf
DARPA/IPTO
2 April 1981
INTRODUCTION:
Verification technology has been the subject of research for a number of
years, now, supported by a variety of agencies for a variety of reasons.
Research problems have ranged from specification language design to the
implementation of automatic theorem provers. The simple premise of this
brief note is that further progress is possible, but that it should be
guided by deliberate application of the technology to a series of
incresingly demanding and real problems. Artificial problems should be
used only if they are needed to test some aspect of a new system or
techniques.
If a parallel with compiler technology may be drawn, serious progress
was only made in efficient compilation when users became dependent on
compilers for serious, production work.
PRACTICAL APPLICATIONS:
IPTO is supporting the use of AFFIRM for specifying and proving theorems
about communication protocol behavior. These theorems geneally have to
do with functionality, consistency, progress and the like. Debugging of
protocol specifications has been aided by this effort. For example,
USC-ISI identified a bug in the TCP specification document by this
means.
The State Delta system at ISI is being applied to microcode protocol
and, eventually distributed algorithm verification. Each new feature
planned during the state delta development is driven by the needs of
each new, increasingly demanding application. An important area of
interest is the ability to provide proofs of the correctness of
microcodes used to enforce certain aspects of computer or system
security.
RECOMMENDATIONS:
The verification community should review the applications now driving
their development and research - some common applications should be
attempted to permit comparisons among the various verification systems.
The AB protocol has been a popular candidate, but more elaborate
problems are needed.
A fundamental issue has to do with the ability of verification systems
to scale - that is, to accommodate larger programs/protocols without
exhibiting uncontrolled growth in processing time and/or storage
requirements. We need to understand which of the various existing or
evolving systems are organized for production work.
Furthermore, the problem of integration of PV systems into normal
programming support environments must be addressed. How does a PV
system map into the traditional edit/compile/debug/document paradigm of
software development?
-------
∂06-Apr-81 1735 CSD.ULLMAN at SU-SCORE Vaughan Pratt
Date: 6 Apr 1981 1734-PST
From: CSD.ULLMAN at SU-SCORE
Subject: Vaughan Pratt
To: rwf at SU-AI, dek at SU-AI, zm at SU-AI, jmc at SU-AI
There is a possibility that Vaughan would be available to accept an offer
for Sept. 1981 and that we would like to make such an offer. I would like the
four of you to serve as an ad-hoc committee to evaluate the case, with Zohar
as chairman. Please let me know if you are unable to serve in this capacity.
-------
∂06-Apr-81 1740 CSD.ULLMAN at SU-SCORE binford case
Date: 6 Apr 1981 1737-PST
From: CSD.ULLMAN at SU-SCORE
Subject: binford case
To: jmc at SU-AI
It went through the A&P committee to the provost!
-------
Tom: The A&P committee is the largest single hazard in appointments.
It isn't sure, but it's very probable now barring some effect of
prospective changes in high-level personnel.
∂06-Apr-81 1745 TOB thanks
John
Thanks for the news. I am happy to have the good news.
Tom
∂06-Apr-81 1743 JMC
∂06-Apr-81 1740 CSD.ULLMAN at SU-SCORE binford case
Date: 6 Apr 1981 1737-PST
From: CSD.ULLMAN at SU-SCORE
Subject: binford case
To: jmc at SU-AI
It went through the A&P committee to the provost!
-------
Tom: The A&P committee is the largest single hazard in appointments.
It isn't sure, but it's very probable now barring some effect of
prospective changes in high-level personnel.
∂06-Apr-81 1736 100 : REM via SU-TIP Bookstore and women
I dropped by Printers Inc. bookstore today and looked around.
Several attractive women there, mostly in places where it would be
rather obvious and unacceptable for me to get their attentin and say "hi".
One really pretty (beautiful) YL was there, but sitting down at a table
with a drink (chocolate or whatever) in front of her and with
her head bowed down to read a book in her lap, I wanted to meet her
but I would have had to get her attention first just to see her
face directly and say "hi" to her. She didn't look up the whole several
minutes I was admiring her, so I gave up and left. I wish I knew how
to approach women without annoying them, but under the circumstances today
I feared anything I tried would be annoying so I didn't even try.
On the way there I said "hi" to a pretty YL on bicicle waiting
for signal, and she said "hi" back, but although I looke back a lot
she didn't follow me with her eyes at all and I felt uncomfortable
stopping my rapid walking to try to meet her, so nothing happened
except exchange of "hi"s. SIgh....
∂07-Apr-81 0030 LGC Today's A-T Discussion
RPG may need to attend a meeting on LISP at SRI today at 3:30pm, which would
preclude his participation in an advice-taker discussion at 4:30. He and I
would both prefer that he be present for the advice-taker discussion. Several
alternatives are possible: a) he and I will be meeting at 11:30am; you might
wish to join us then, or in the early afternoon. b) you might want to hold
open the 4:30 slot in case RPG decides on the basis of new information that he
doesn't need to attend the LISP meeting. c) you might prefer simply to
postpone today's advice-taker discussion 'till Friday. What do you think?
-- Lew
I think I can be in at 11:30, and if so, I'll join you.
∂07-Apr-81 0505 Darden@SUMEX-AIM Washington
Date: 7 Apr 1981 0506-PST
From: Darden@SUMEX-AIM
Subject: Washington
To: JMC@SAIL
Hi, John. If you send any messages about your April 20 Washington
trip between April 10 and April 16 I may not answer them, since I
am going to a conference that weekend at Princeton and then on to
Boston to give a paper in the Boston Colloquium for Philosophy of
Science. Since I plan to visit Patrick Winston at MIT to talk about
analogy, I may be able to check messages that day, but I don't know.
I wonder how long it will be before terminals are easily available to
one while traveling. The Post just had an article about GTE's
successes with Telenet for business communications.
See you soon. Bye, Lindley.
-------
∂07-Apr-81 0811 Betty Scott <CSD.BSCOTT at SU-SCORE> Next Week
Date: 7 Apr 1981 0810-PST
From: Betty Scott <CSD.BSCOTT at SU-SCORE>
Subject: Next Week
To: JMC at SU-AI
cc: CSD.BScott at SU-SCORE
John, I decided yesterday to take next week off to help wind up some family
affairs. Thought I would let you know this now in case you will need
budgets prepared this week. I'll do my best to get them done for you.
Betty
-------
∂07-Apr-81 0843 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) comp programming problem
Date: 7 Apr 1981 0838-PST
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: comp programming problem
To: jmc at SU-AI
One of the phd minor students turned in his programming problem. Do you want
to take it now, or wait until I collect more?
Carolyn
-------
∂07-Apr-81 0903 TOB ARPA proposal
To: DCL, JMC
I have a second draft of the ARPA proposal which is in
good shape. I have been in contact
with the ARPA program manager and we agree in
general. He has not responded yet in detail to
the first draft, but I should hear soon. In
short, I am essentially done.
∂07-Apr-81 0916 DEK VRP
To: RWF at SU-AI, ZM at SU-AI, JMC at SU-AI, JEF at SU-AI
I'll gladly serve on committee re a Pratt appointment, especially since
its easier to be a committee member than a committee chairman
∂07-Apr-81 0924 RPG Rules
To: lisptiming at SU-AI
I have sent out the first benchmark, and already there are a number
of issues that need to be faced. First is that some systems (SAIL,
for instance) only reliably report EBOX time (excluding memory
time). Fortunately SAIL does make available some form of EBOX + MBOX
time, though it is unreproducible. When possible I want to see the most
information you can give me with as many distinctions as possible. If your
system can give a breakdown of memory references, cache write-through time,
page fault time, EBOX time,... please give me that. When I send out the
benchmarks I include the SAIL times so that you can get some idea of how
long it all takes. From now on I will provide EBOX time, EBOX + MBOX time,
and GCTIME. Because I only provide that does not mean that is all I want to
see.
Slightly more importantly is the issue of `cheating'. If the sources of
benchmarks wish to allow specializations of their programs to the test data,
they should make remarks to that effect. If someone cares to make such
specializations they must be cleared by the author and me. This isn't because
I like to be in control so much as I want to understand what is being gained
by the specialization and what features of the target LISP make such
specializations necessary and/or desirable.
For example, in the first benchmark several of the functions are implicit
LEXPRs, which in MacLisp means that there are more than 5 arguments. This
means that the arguments are passed on the stack rather than through registers.
Since this takes longer than the register convention (in this case) I want that
feature timed. In the test data I sent out, some of the arguments are provably
constantly (). Chuck Hedrick at Rutgers (cleverly) noticed this and specialized
the functions. I want to specifically disallow that specialization (since the
LISP he had allows LEXPRs). [So do it again, Chuck.]
-rpg-
∂07-Apr-81 1321 CLT
∂07-Apr-81 1224 JMC travel agent
Call Dina Bolla Travel 329-0950 and ask for Franklin Hersch.
Actually, I think it would be more efficient to just go together
and see him at some mutually convenient time. Otherwise it will
just result in chaos.
∂07-Apr-81 1323 JONL at MIT-MC (Jon L White) Proposed ''mini'' benchmark, with interpretation.
Date: 7 APR 1981 1611-EST
From: JONL at MIT-MC (Jon L White)
Subject: Proposed "mini" benchmark, with interpretation.
To: lisptiming at SU-AI
It hardly seems appropriate to run timing tests without including a word
or two about the discussion last fall which generated (in Masinter's words)
so much more "heat" rather than "light", namely the TAK function sent my way
in September 1980 by Mr. Shigeki Goto of the Electical Communication
Laboratories, Nippon Telegraph and Telephone Co., in Tokyo.
(DEFUN TAK (X Y Z)
(COND ((GREATERP X Y)
(TAK (TAK (SUB1 X) Y Z)
(TAK (SUB1 Y) Z X)
(TAK (SUB1 Z) X Y) ))
(T Y) ))
The test case, named TARAI-4, is to measure the timing of (TAK 4 2 0)
The value of this trivial function can be seen, not in a competition
between lisps for "speed", nor in a condemnation of one dialect for
the "kludges" which must be performed in order to get such a trivial
thing to run reasonably fast, but rather in the analysis of the basic
issues which trying to time it brought out. After receiving many responses
from around the community, I mailed out a note in which was discussed
what I thought were some fundamental issues, and I'd like to send that
note to this group for consideration. The original note from Mr. Goto and
a lengthy series of communications about the timings for his test case,
especially from people in the Interlisp community, is in the file
JONL;GOTO NOTE
on the MIT-MC machine.
Date: 22 October 1980 11:08-EDT
From: Jon L White <JONL at MIT-MC>
Subject: Response to Goto's lisp timings
To: . . .
As Larry Masinter mentioned in his comment on the Goto
timings, comparisons between LISPs are likely to generate
more heat than light; but the replies did throw a little
more light on some things, especially the runtime variabilities
of an Interlisp program, and I thought I'd summarize them
and pass them along to the original recipients of the note.
However, I'd like to say that the general concern with speed
which I've encounterd in the past has been between MacLISP and
FORTRAN, rather than with some other lisp; and several Japanese
research labs have done AI research still in FORTRAN.
Just in case you're put off by looking at even more meaningless
statistics, I'd also like to aprise you that following the little
summary is a brief technical discussion of three relevant points
disclosed by the TAK function (out of the many possible points at
which to look). These technical points may be new to some of you,
and even beyond the LISP question you may find them useful; the key
words are (1) UUOLINKS, (2) Uniform FIXNUM representation, and
(3) Automatic induction of helpful numeric declarations by a compiler.
Almost everyone familiar with Interlisp recognized that
the ECL people had not requested "block" compilation in the TARAI-4
example, and several persons supplied results from various
20/60's around:
default compilation rewritten code, with
correspondent timings block-compiled timing
Date: 19 OCT 1980 2127-PDT 9.8ms 1.8ms
From: MASINTER at PARC-MAXC2
Date: 20 Oct 1980 1039-PDT 16.ms 2.ms
From: CSD.DEA at SU-SCORE (Doug Appelt)
Date: 20 Oct 1980 at 2257-CDT 0.83ms (for UCILISP only)
From: tyson at UTEXAS-11
<Goto's original timings on ICILISP> 2.9ms
<Goto's original timings on Interlisp> 15.0ms
<myself, for MacLISP on 20/50) 0.556ms
There seems to be some unexplained discrepancy between Tyson's timing
and that of Goto, as well as between Masinter's and Appelt's default-
compilation timings; but the "best-possible" Interlisp timings for
a re-written function (replacing GREATERP by IGREATERP) and using
the "right" block declarations seem consistent at around 2ms. Indeed,
as Shostack suggest in his note of "20 Oct 1980 1036-PDT" there is
quite a bit of variablity in timing Interlisp functions depending on
just the right set of declarations etc (even for such a simple function).
A point which, however, seems to be missed is that the notion of
"block" compilation requires a decision at compile-time as to what
kind of function-linkage would be desirable (I presume that spaghetti-
stack maintainence is the worst offender in runtime slowdown here);
by comparison, the decision between fast and slow function linkage
in MacLISP is made dynamically at runtime, so that only one kind of
compilation be needed. Indeed, by not burdening the novice with the
understanding of yet one more inscrutable compilation question
("block" versus what?), the novice needn't be unduly penalized for
not becoming a "hacker"; the above timings show a penalty of a factor
between 5 and 10 for ignoring, or under-utilizing, the "block" question.
(1) UUOLINKS:
The following strategy, which we call the UUOLINKS hack, may have
first been introduced into the old LISP 1.6:
Arguments are set up for passing to a function and an instruction
in a specially-managed hash table is XCT'd.
In case a fast link is acceptable, the first usage of this linking
will convert the hash entry to a PUSHJ P,... -- if not
acceptable, it will remain a slow interpretive route.
Two copies of the hash-table are kept -- one is never altered by
the linker, so that at any given point in time, all the "fast"
links may be restored to the unconverted slow interpretive route
(which may yet again be "snapped" to fast).
Typically, a hash table size of 512. is satisfactory, but some
applications require 1024. or more (in particular, MACSYMA).
Indeed as Boyer (@SRI-KL) mentioned in his note of "21 Oct 1980 2055-PDT",
the fast route -- whether by Interlisp's block compiler, or by MacLISP's
runtime "snapper" -- does not leave much debugging help lying around
on the stack; at least with this UUOLINKS approach, one can make the
decision while in the middle of a debugging session, without penalty.
The time cost of using the slow linkage seems to be a factor of between
2 and 5.
(2) Uniform FIXNUM representation
Many years ago we changed MacLISP's representation of FIXNUM so
that it would be uniform; unlike the other PDP10 lisps with which I
am familiar, we do not code some fixnums (small ones) as "immediate"
pointers and others (larger ones) as addresses. Also, there is a
read-only page or two which "caches" fixnum values of about -300. to
+600., so that number consing of small numbers won't actually be
allocating new cells; e.g. interpreting a form like
(DO I 0 (ADD1 I) (GREATERP I 100.) ...)
Although I took a lot of flak for flushing the INUM scheme in favor
of the uniform scheme, consider the advantage for compilation strategy,
as seen in these representative code sequences for (IGREATERP X Y):
INUM scheme: MOVE A,-3(P)
JSP T,UNBOX
SAVE TT,somewhere
MOVE A,-4(P)
JSP T,UNBOX
CAME TT,somewhere
...
Uniform scheme: MOVE TT,@-3(P)
CAME TT,@-4(P)
...
(3) Automatic induction of helpful numeric declarations by a compiler.
As Masinter and Boyer pointed out, most Interlisp programmers
would probably be using "IGREATERP" rather than "GREATERP" (the MacLISP
correspondent is "<" ). But a compiler can accumulate static information
and do this change automatically; at the very least, it could give
out warning checks such as "Floating-point argument used with FIXNUM-only
operation". Providing the capability for compile-time typing of variables
is probably the only way to meet the FORTRAN challenge -- which must be
met since much useful AI research needs both symbolic and numeric
capabilities. Larry's MASTERSCOPE is a very similar sort of automated
induction scheme.
∂08-Apr-81 0200 ZM Pratt
I need your list. Thanks Zohar
∂08-Apr-81 1049 FFL Comprehensives Committee meeting
To: HLO at SU-AI, JMC at SU-AI, FFL at SU-AI
Gene Golub tells me that he has an NA luncheon on Mondays and cnanot
attend the Committee meeting.
∂08-Apr-81 1245 Buchanan@SUMEX-AIM Jonathan King's application
Date: 8 Apr 1981 1236-PST
From: Buchanan@SUMEX-AIM
Subject: Jonathan King's application
To: jmc@SU-AI
cc: csd.dbrown@SCORE
John,
after you and I talked about admitting King from Mathematics, I made a
note that he was going to talk with you and you would send me a note
to admit him (or not). Have you decided whether to take him on?
thanks,
bgb
-------
I cannot take King as a student as our interests don't sufficiently overlap.
∂08-Apr-81 1430 CSD.ULLMAN at SU-SCORE
Date: 8 Apr 1981 1429-PST
From: CSD.ULLMAN at SU-SCORE
To: CSD-Faculty: ;
We are in the process of collecting applications for systems faculty.
Some responses to our ad have materialized; I have looked at them and
sent for references on those that look plausible. I would appreciate
people looking at the files (in Jeanie's office) and flagging any that
I overlooked and are worth pursuing, and also assessing the letters of
reccommendation already received.
-------
∂08-Apr-81 1552 TOB arpa proposal
To: DCL, JMC
I am finished, waiting for a response from Druffel.
We had agreed on budget, this relates to making the
proposal read well to get a good reception.
What do I do now?
∂08-Apr-81 1729 Vaughan Pratt <CSD.PRATT at SU-SCORE> Apollo presentation
Date: 8 Apr 1981 1724-PST
From: Vaughan Pratt <CSD.PRATT at SU-SCORE>
Subject: Apollo presentation
To: @sun at SU-AI
There will be a presentation of Apollo's 68000-based system of distributed
personal computers tomorrow (Thursday) at 1.00 pm. Place to be announced.
-------
∂08-Apr-81 1951 DCL
To: TOB, JMC
∂08-Apr-81 1552 TOB arpa proposal
To: DCL, JMC
I am finished, waiting for a response from Druffel.
We had agreed on budget, this relates to making the
proposal read well to get a good reception.
What do I do now?
REPLY:
What we need to do is to get all four proposals together, probably in
a reasonably consistent formatt (fonts, one column, etc.), and make up
a cover page etc etc.
-David
∂09-Apr-81 1042 CSD.JEANIE at SU-SCORE April 21 Faculty Meeting
Date: 9 Apr 1981 1038-PST
From: CSD.JEANIE at SU-SCORE
Subject: April 21 Faculty Meeting
To: CSD-Faculty: ;
cc: csd.bscott at SU-SCORE, csd.jeanie at SU-SCORE
A faculty meeting has been scheduled for Tues., April 21 at 1:30. The
meeting will be in MJH 252.
Agenda:
1. CSD participation in CIS
2. Dicussion of Instructional Computing
3. Discussion of MS enrollment limits.
-------
∂09-Apr-81 1108 FFL
To: JMC, FFL
Zohar came in requesting a list of names relating to Pratt. He says Ullman
is pushing him and he would like to have it as soon as possible.
∂09-Apr-81 1150 Vaughan Pratt <CSD.PRATT at SU-SCORE> Apollo presentation
Date: 9 Apr 1981 1145-PST
From: Vaughan Pratt <CSD.PRATT at SU-SCORE>
Subject: Apollo presentation
To: @sun at SU-AI
The Apollo presentation will be in MJH 402 at 1 pm. Very sorry about the
short notice for the place.
-------
∂09-Apr-81 1324 FFL
To: JMC, FFL
Dr. Jain of NSF called. Wishes to talk to you about revising your
budget proposa for Basic Research in AI. 202 357 7345.
∂09-Apr-81 1417 FFL
To: JMC, FFL
Call from Douglas LeBlanc of Walt Disney Productions. Interested in
having the lab do some consulting work with robotics application.
213 956 6526.
∂09-Apr-81 1451 FFL
To: JMC, FFL
Please call Martha Coyote at Stone Candles. 57 465 428
∂09-Apr-81 1806 SQU Thanks
I wanted to say this in person but I haven't been around when you have, apparently,
so mail will have to suffice.
I've decided to stay in the mathematics department because of several
reasons. One) Andy is going to Berkley. Two) I spoke with Jim Boyce and other
graduate students and with Bob Floyd and it seems that the only person working
in the area I'm interested is in the electrical engineering department.
I'm writing this because I wanted to thank you for writing a recom-
mendation for me and particularly for offering to give an oral rec. That was
awfully kind of you and it picked my spirits up when they were kinda low.
If you don't mind, I may still drop by now and again to discuss/exchange
problems.
tHE oTHER
jonathan king
Good luck in the Math Dept., and I look forward to further discussions.
∂10-Apr-81 0933 Buchanan@SUMEX-AIM Jonathan King's application resolve
Date: 10 Apr 1981 0911-PST
From: Buchanan@SUMEX-AIM
Subject: Jonathan King's application resolve
To: jmc@SU-AI
cc: csd.dbrown@SCORE
Mail-from: ARPANET host SU-SCORE rcvd at 9-Apr-81 2048-PST
Mail-from: ARPANET site SU-AI rcvd at 9-Apr-81 1812-PST
Date: 09 Apr 1981 1808-PST
From: Jon King <SQU at SU-AI>
Subject: Application
To: csd.buchanan at SU-SCORE
Bruce: I wanted to thank you for your help,time, and advice. I've decided
to stay in the math dept. (assuming I had a choice!) because with Andy leaving
it seems the only person working in the area I'm interested in is Prof. Gill
in the electrical engineering dept..
Thanks again for your time.
tHE oTHER
jonathan king
-------
∂10-Apr-81 1053 COHEN at PARC-MAXC Re: Benjamin Cohen thesis
Date: 10 Apr 1981 10:18 PST
From: COHEN at PARC-MAXC
Subject: Re: Benjamin Cohen thesis
In-reply-to: Your message of 10 Apr 1981 0042-PST
To: John McCarthy <JMC at SU-AI>
cc: cohen
Professor McCarthy,
I'm Benjamin Cohen and I'd very much appreciate your comments positive or
negative. I enjoyed yesterday's discussions at SRI very much and would be
pleased to give you the most recent version which expands a bit more on
definability. I'll leave it in your box at the Computer Science department.
Ben
∂10-Apr-81 1205 George J. Carrette <GJC at MIT-MC> Rules
Date: 10 April 1981 14:19-EST
From: George J. Carrette <GJC at MIT-MC>
Subject: Rules
To: HEDRICK at RUTGERS
cc: lisptiming at SU-AI, RPG at SU-AI
I too was suprised at the chastising about breaking rules that
went on here considering its relevance that classic law of programming,
"if a function has more than three arguments then they are
probably in the wrong order."
In point of fact, all the common so-called "lexprs," PROGN,
TIMES, PLUS, LIST, are specially handled by the compilers
of the lisp system's I'm familiar with.
Furthermore, the maclisp and lispm programs that have
used user-defined multi-argument constructions have invariably
developed in the direction of passing arguments by "keywords"
where these keyword are pre-processed in some form or another
at compile-time.
Ah, there are some interesting possible ways of optimizing
keyword argument calling in VAX NIL. Maybe we can talk about
things like this at some time.
-gjc
∂10-Apr-81 1051 HEDRICK at RUTGERS Re: Rules
Date: 10 Apr 1981 1301-EST
From: HEDRICK at RUTGERS
Subject: Re: Rules
To: RPG at SU-AI
cc: lisptiming at SU-AI
In-Reply-To: Your message of 7-Apr-81 1224-EST
OK, but you will notice that I was also the only person who got the
thing done by the deadline you specified. If we are going to end up
doing major conversions for each test, and furthermore if conversions
are going to have to be approved by you, you may find fewer volunteers
than originally planned. The reason for eliminating the extra args was
of course that turning the things into LEXPR's would be a pain in the
neck. This is because in UCI Lisp LEXPR's do not refer to arguments by
name, but as (ARG x). I can obviously write a program to do this, but
that was not feasible in the amount of time I had before the meeting.
Furthermore, it looked to me like the functions that had more than 5
arguments were used for preparing the data, but that only PAIRS1
and PAIRS2 were actually called large numbers of times. Now that the
issue has been brought up, I will of course test LEXPR's, but would
be surprised if there is any change in performance.
Even if LEXPR's change the performance, I am not sure that is the right
way to do the conversion to UCI Lisp. It would be very unusual for a UCI
Lisp programmer to use an LEXPR in order to handle more than 5
arguments. They are normally used for indefinite arguments, when it
makes sense to number them. If named variables are replaced with (ARG 1),
(ARG 2), ..., this obviously makes the program somewhat opaque. Another
possible method is to lambda-bind variables to the required values and
refer to them globally. While this is less than ideal, I claim that it
is better than (ARG n). Making a list of the extra arguments is also
possible, but (CADR ARGS) is little better than (ARG 2).
At this point we start having to ask what the purpose of this project
is. If it is to see how the dialects vary, then I claim that nothing is
accomplished by forcing us to convert into code that we would not in
fact write that way in the dialect. It seems to me that it is perfectly
legitimate for me to say that UCI Lisp simply does not support EXPR's
with more than 5 arguments, and that I would find some other way to do
my task.
-------
∂10-Apr-81 1818 JD arpa proposal
John,
If you can tell me what file has the arpa proposal in it, I
will try to modify it over the next few days to include me.
I looked through your directory as you suggested, but couldn't
find it.
Jon
The file with the old proposal hasn't been restored yet. Perhaps tonight.
∂10-Apr-81 1958 Janofsky.Tipi at MIT-Multics (Bill Janofsky) Re: Student's credulity (JMC@SU-AI msg in HNT V3#77)
Date: 10 April 1981 17:11 est
From: Janofsky.Tipi at MIT-Multics (Bill Janofsky)
Subject: Re: Student's credulity (JMC@SU-AI msg in HNT V3#77)
To: HUMAN-NETS at MIT-AI
cc: JMC at SU-AI
I plead quilty as charged to being an optimist. However, I must take
umbrage at the assertion that I'm overly optimistic. It presumes that I
was overcome by some powerful emotion. Actually, as a curmudgeon
working at being an iconoclast, I simply wished to counter the
impressions left by the discussion of Dick Gregory's lecture. (Besides,
it's more fun than agreeing with everyone.) Furthermore, at no time did
I ever assert that there was a "return to rationality" in progress, on
campus or elsewhere. Only that there might be cause for hope . . .
which there always is!
Regards, Bill J.
∂10-Apr-81 1958 Janofsky.Tipi at MIT-Multics (Bill Janofsky) Re: Domestic Computers
Date: 10 April 1981 17:26 est
From: Janofsky.Tipi at MIT-Multics (Bill Janofsky)
Subject: Re: Domestic Computers
To: JMC at SU-AI
In-Reply-To: Your msg of 30 Mar in HNT V3#77
I agree! There is nothing I've yet identified that could be used as a
*GOOD* reason for a domestic computer. All of the (few) other responses
I received seem to confirm the absolute necessity of external
communications interfaces and inply a need for useful public information
sources. Unfortuneately, there are very few currently available, public
information sources, even fewer applicable to "domestic use."
∂11-Apr-81 0935 LES ARPA proposal
The text of your last proposal has been retrieved:
∂09-Apr-81 2256 Great Pumpkin Restored Files
Here is a status report on your PUMPKIN requests:
Tape File Status
P1730 FR[P,LES] Restored
∂12-Apr-81 0242 Stephen C. Hill <STEVEH at MIT-MC> Oops
Date: 12 April 1981 05:42-EST
From: Stephen C. Hill <STEVEH at MIT-MC>
Subject: Oops
To: JMC at SU-AI
cc: STEVEH at MIT-MC
The error that you point out has been corrected in a message to
HUMAN-NETS. I never was very good at decimal places on my
slide-rule or identifying the correct units. That's why I left
the field!
Steve
∂12-Apr-81 1008 CLT today
its a nice day, perhaps we could take a little walk or something
this afternoon
∂12-Apr-81 1155 CLT
Good, will you come by the office?
∂12-Apr-81 1357 CLT
I gather there is going to be more shuttle on TV this afternoon,
if you would rather watch that its ok.
∂12-Apr-81 1916 Robert Elton Maas <REM at MIT-MC>
Date: 12 April 1981 21:32-EST
From: Robert Elton Maas <REM at MIT-MC>
Sender: ←←←050 at MIT-MC
To: JMC at SU-AI
That was nice of you to invite to hang out at Printers Inc, but you
caught me at the wrong time, alas. Feel free to call me again the
next time you feel like hanging out there and want me to join you.
Moulin Rouge is one of my top 20 favorite movies of all time, and this
was the first time I've seen it all the way from the start, also I
was rather tired that evening. Next time I'll probably not be busy.
∂13-Apr-81 0944 JD proposal file
Are you sure you restored the correct file? The one you referred
me to fr[s81,jmc] bears almost no resemblance to the proposal you
gave me on paper. In particular, I could find no trace of the sections
I had planned to modify in the old proposal. Could you perhaps restore
them as well so I can modify them?
∂13-Apr-81 1119 LGC Discussion Today?
Will you have some time after 4:30pm today for a discussion of the advice taker
with RPG and me?
∂13-Apr-81 1239 RPG Groundrules (reprise)
To: lisptiming at SU-AI
I believe that at the outset of this project I stated what my goals
were, and the order of importance. But I will reiterate them anyway:
1. I want to provide for each major type operation (variable lookup,
function call,...) a relative, hopefully total, order on the various LISPs
and LISP systems around. I also hope that there is enough standardized
timing facilities around to at least get some idea of the approximate
absolute speeds as well.
2. I want to determine, myself, what the qualitative differences between
the various LISP systems around are. I hope to do this by watching how various
programs and constructs are handled by the translators. At first I hoped
that the ``translator volunteers'' would be just that, but now it seems I
will need to do most of the translations myself, which is ok as long as
I can merely provide the framework and not exact working code. If you
want a NIL/MacLisp person to propose the exact program whose timing is
universally reported as the performance of your favorite LISP system, then
I might be more willing to do everything myself.
3. Having ``rules'' is absolutely fair. First, one certainly cannot look
at specializations of a program to the data. Moreover, innate laziness (no
`major conversions' to quote Hedrick) dictates that the programs should be
examined as little as possible. Arguing style is totally irrelevant.
Suppose I wanted to test function call time and proposed factorial in its
recursive form, it is totally opposed to the spirit of what that tests to
translate it into an iterative program no matter what absolute style
considerations you bring to bear. One of the things I wanted to test
with this program is how functions with more than 5 arguments behave
and are handled with different systems. This is totally fair.
As pointed out, the standard > 5 argument LEXPR conversion is
(defun foo (a1 ... an) ...) =>
(defun foo n ((lambda (a1 ... an) ...) (arg 0) ... (arg n)))
I was flip with Chuck because we spent 2 delightful years at Illinois
together a few years back.
4. Perhaps I should have sent out more help with this program, and
in the future I will, but another point of this benchmark was to test
the testing system.
From now on I will provide with each benchmark a description of each
`primitive' in the program(s) that is not in McCarthy along with translation
tips for those facilities that are not universal. As time goes on and I
become truly familiar with all the systems, I will provide specific help
for each LISP.
-rpg-
∂13-Apr-81 1347 FFL
To: JMC, FFL
Man called who has BS in physics and has been working in computer science. He
wants to go back for his PhD at UCSC and is interested in working in AI. He
would like to speak with you about course work and direction. He would like
an appointment or to speak with you on the phone. Will Hogan, 408 425 0659
or 0763.
∂13-Apr-81 1357 Brian K. Reid <CSL.BKR at SU-SCORE> address space
Date: 13 Apr 1981 1352-PST
From: Brian K. Reid <CSL.BKR at SU-SCORE>
Subject: address space
To: jmc at SU-AI
The machine that we were discussing at dinner the other night
has a 22-bit address space. No more than 20 bits of physical
memory can be connected at once, though.
-------
∂13-Apr-81 1404 COHEN at PARC-MAXC Reply to Comments
<contents moved to COHEN[S81,JMC].>
∂13-Apr-81 1457 FFL Thesis Committee meeting for Vic Scheinman
To: JMC, FFL
Bernard Roth asks if you can meet on Thursday, Apr. 23, at 2 p.m. with
the thesis committee. If not, can you meet at any other time on Thursday.
If not, can you meet on Friday. He would appreciate a reply as soon as
possible.
Friday is possible. Thursday is not.
∂13-Apr-81 1736 Mark Crispin <Admin.MRC at SU-SCORE>
Date: 13 Apr 1981 1724-PST
From: Mark Crispin <Admin.MRC at SU-SCORE>
Postal-Address: 12155 Edgecliff Place; Los Altos Hills, CA 94022
Stanford-Phone: (415) 497-1407
To: JMC at SU-AI
In-Reply-To: Your message of 9-Apr-81 1733-PST
If the file is protected with public access (such as 775252 or
777752), then SCORE FTP will autologin as ANONYMOUS to get it.
Most TOPS-20 sites support user ANONYMOUS with any password to
get publicly accessible files; the autologin is a local hack.
-- Mark --
-------
∂13-Apr-81 2327 the tty of Geoffrey S. Goodfellow [HEDRICK at RUTGERS: report on Lisp conference]
Date: 13 Apr 1981 2317-PST
Sender: GEOFF at SRI-CSL
Subject: [HEDRICK at RUTGERS: report on Lisp conference]
From: the tty of Geoffrey S. Goodfellow
Reply-To: Geoff at SRI-CSL
To: Lynch at ISIB, Tom at SUMEX-AIM, EAF at SAIL, JMC at SAIL
Cc: Masinter at XEROX-PARC
Message-ID: <[SRI-CSL]13-Apr-81 23:17:57.GEOFF>
Gents, FYI - Thought you'd like to see this interesting report on
the LISP Conference,
Begin forwarded message
===========================
Mail-from: ARPANET host RUTGERS rcvd at 13-Apr-81 2257-PST
Date: 10 Apr 1981 1503-EST
From: HEDRICK at RUTGERS
Subject: report on Lisp conference
To: bboard@Rutgers
I just came back from a conference at SRI. The conference was called
by ARPA to discuss the future of Lisp in the AI community, as well as
the proposal for ARPA to buy Dolphin computers for use within this
community. It lasted from 8:30 am to about 10pm on 8 April.
One of the major concerns that motivated the meeting was a feeling that
there are suddenly many projects to reimplement Lisp, and that we may
end up with many half-baked, incompatible implementations. There had
been some hope for getting some coherence among them. As far as I can
see, these fears are somewhat justified, and nothing was accomplished
towards creating coherence. There were in fact 13 Lisp implementation
projects listed. (Some of these have been finished, however, so that
number is a bit too big.) Fortunately, none of them are creating new
languages. Rather they are reimplementations of existing dialects
for new hardware. Thus there is somewhat less chaos than the number
13 would imply. Here they are. All results are publically available
unless stated otherwise.
Interlisp:
SRI portable Interlisp [SRI]. not yet funded
Interlisp-D [Xerox]. For Dophin, a Lisp machine with bit-mapped
display. Finished.
Interlisp-Jericho [BBN]. For BBN's Jericho, a high-performance
personal computer. in progress. Results will be proprietary.
VAX Interlisp [USC-ISI]. They hope to have something running by
June, but it will be up to a year before the whole environment
is working. This is the most critical of all the projects
as far as most users are concerned.
MACLisp:
Lisp machine [MIT]. A version of Lisp extended to make heavy use
of bit-mapped display, and having a number of new language
features. Finished some time ago, but development continues.
Results are proprietary, with MIT licensing two commerical
companies to distribute them.
NIL [MIT]. Intended to have everything that the Lisp machine has
that will be useful on a conventional machine. Also
high-performance compiler. Will have something by end of the
summer, but development will be ongoing. Mainly intended
for VAX, but probably will be done for extended-addressing 20.
S1-NIL. NIL for the S1. This is a multi-CPU supermachine sponsored
by the military (Navy?).
Spice Lisp [CMU]. Dialect of MACLisp for personal computers. Will
use microcode assist. First implementation will be on VAX and
extended addressing DEC-20 by simulating the proposed microcode.
It is unclear whether this will be a usable Lisp or not, and
is intended mainly for debugging until the personal machine
hardware is available. [see below for comments on this]
Franz Lisp [Berkley]. MACLisp dialect for VAX. finished. Many
people seem to be unenthusiastic about this, but it seems to
be a solid implementation. Maybe a trifle slower than it might
be and somehow not as "sexy" as a typical MIT product.
Other dialects:
Standard Lisp (Utah) - This is really a research project in portability.
They are trying to write as much of Lisp as possible in Lisp.
The compiler has extensions to allow it to be used for system
programming. Currently a very small part is still written in
assembly language. They should have an implementation for
extended-address DEC-20 within 6 months.
Elisp (Rutgers) - This is a recoding of R/UCI Lisp for extended
addressing DEC-20. This should be finished by the end of the
summer.
MDL [MIT] - This is not really a Lisp dialect. It is intended as
a successor to Lisp. It has more data types and has been
generally cleaned up. they are working on a portable
implementation. There has been a DEC-20 implementation for
years. They now have an implementation that makes some use
of extended addressing. when the portable implementation is
finished, it will be used to bring up a fully extended version
for the DEC-20.
Of all these, the project that generated the most interest was
clearly the VAX Interlisp. Many people are committed to both the VAX
and Interlisp, and are in bad shape until this happens.
Now some comments as to why there are 13 projects. I will not comment
much on the "other dialects". MDL is its own thing, and a portable
implementation for it makes perfect sense. Similarly, Utah's research
is a very interesting project. In my opinion it is a more promising
approach than those being used by the Interlisp or MACLisp people, and
these folks would have been well advised to be cooperating more closely
with Utah than they are. Our project is strictly a short-term fix to a
critical problem, and requires minimal effort. That is its only
possible justification. In the long run we will probably want to choose
a dialect that is going to be available on personal machines.
Now for the Interlisp and MACLisp projects. They are all attempts to
implement their language for new hardware. In the case of MACLisp they
also are trying to clean up and develop the language further.
The Interlisp projects are coordinated with each other, at least in the
sense that they are all implementing the same language. Also, much of
the Lisp code is shared. However apparently the Interlisp "virtual
machine" (i.e. the part done in assembly language) is relatively large,
and the user interface (e.g. debugger) depends upon details of the stack
that are different for different machines. Thus transporting Interlisp
is a fairly major project. As far as I can see, the projects other than
SRI's were reimplementations of the virtual machine for particular
hardware, with no particular thought given to portability. I think SRI
hopes to improve this. Whether their results will help anyone else
depends upon who sponsors them, since certain funding arrangements could
result in a proprietary product. There is a single manual that applies
to all Interlisp versions.
The MACLisp projects are not particularly coordinated. Each of them is
implementing different dialects, with separate (or non-existent) manuals.
In general the Lisp Machine seems to have the most features, and the
other dialects are probably more or less subsets of it. Of the current
projects, there are considerable differences in style:
Franz Lisp is the most conservative. They wanted something up
quickly, using existing technology. It was done in C, which
should help its transportability.
Lisp Machine is the most radical. First, it is standalone, with
microcode support. Second, it has every language feature one
can imagine except for spaghetti stacks. Finally, it supports
the bit-mapped display. They believe that many of its features
could only be done on special-purpose hardware. This might
possibly be transportable to another microcodable machine with
similar capabilities, though no particular thought was given
to portability.
Spice Lisp and NIL are in some sense compromises. They are attempts
at cleaning up the old language design, and taking many of the
good ideas of Lisp Machine Lisp, but for somewhat more
conventional machines. At the moment these projects talk to
each other, but are planning to implement somewhat different
dialects. They were strongly encouraged to talk to each other.
They are both giving thought to portability, though SPICE is
only intended to be portable among personal machines with
certain capabilities.
The big question is, why so many projects? As far as I can see,
here are the justifications:
MDL - this is a new language design, with generally good ideas.
It has been very influential on newer developments in Lisp.
Utah - this is a research project in portability, and seems to
be doing very good work. In retrospect, the AI community
would be much better off if they had decided to do their
research using either Interlisp or MACLisp. As it is,
they have attempted to create a new dialect, and no one
is interested. Probably their work will be ignored, much
to everyone's hurt, unless they decide to change to
another dialect, which I do not expect.
Franz Lisp and Elisp - these are short-term projects to transport
dialects to machines with lots of users that desperately
needed the results. They should die away eventually if
other projects succeed.
Interlisp - these projects are simply transporting Interlisp to
other machines, which is always reasonable. The only
real criticism here would be that it is a shame that
they can't develop technology to produce a new implementation
more quickly. But this is what SRI proposes to do. In my
opinion it is critical for that project to be done in such
a way that the results are public.
MAClisp - it is unfortunate that so much work is being done on MAClisp
in an uncoordinated way. There is some evidence that each
project is made up of "true believers" who believe that there is
little merit to the others. We heard a somewhat amusing comment
by one. He said it wasn't true that there was chaos among the
MAClisp people. It was just that there were 4 different
projects going in 4 different directions....
Everyone seems to believe that it is a good idea for there to be ongoing
work on both Interlisp and MACLisp. Interlisp tends to be more
conservative in its development. It has good user facilities and
well-coordinated support. But there was a surprising concensus (that
included most of the Interlisp users) that
- Interlisp is quite expensive compared to Maclisp
- Interlisp as a dialect lacked a number of important features compared
to Maclisp, and had no important features missing in Maclisp.
Note that this comment refers only to facilities commonly used
in programs, not to the user support environment. Many people
believe that in user support Interlisp is better than MACLisp
(except on the Lisp Machine).
Thus what was keeping people with Interlisp is
- good user facilities
- the fact that all implementations are compatible
- good, complete documentation
- good support
To the outside observer it appeared that in the long run Maclisp might
in fact take over if it could do the following:
- supply the same user facilities as Interlisp, or document the
fact that it already has them (if it does - I take no position
on this question)
- agree on a common language definition, with extensions for the
Lisp Machine and other who need it
- produce a complete manual, showing all the user facilities and
being common among implementations. A good form for this
might be a loose-leaf binder, so that they could provide
additional pages and chapters for the Lisp machine, and
let you select which debugger and editor you wanted.
- somehow assure users outside MIT that there was a central support
organization that would respond to their concerns. (This seems
to be true, but there may be a PR problem.)
- possibly do something to isolate non-MIT users from the hack-of-the
week club, while preserving the fact that Maclisp will continue
to develop more aggressively than Interlisp
I am not convinced that they can do this if left to themselves, but
there was a proposal that ARPA might somehow be able to cause it to
happen.
Finally there is the question of what will happen next as far as hardware
acquisition. We polled representatives of various user groups. Most
of them have DEC-10's and -20's. They all say the DEC-10 is dead, but
are still doing most of their work on it. However there are also a
significant number of people who use VAX as their primary system. Some
of these are using Franz Lisp. It seems fine. But many are desperate
for VAX Interlisp. Few people are currently using personal computers
for much, or even have definite plans to use them. The main places that
are highly committed are MIT, where most Lisp work is using Lisp
machines, BBN, which is committed to Jericho, CMU, which is committed
to Spice (in a somewhat indefinite future), and HPP(Sumex), which is
committed to Dolphins. Others are hedging their bets with one or two
experimental machines, or are just plain waiting.
On the other hand, there are a number of complaints of a serious
shortage of cycles. The most vocal is HPP/Sumex. Thus there is some
pressure on ARPA to do something immediately. HPP is pressuring them to
buy a large group of Dolphins. A number of other people have said that
if ARPA does go this way, they will follow. But few people other than
HPP feel strongly enough to commit to Dolphins with their own money,
unless ARPA does. If ARPA does, then it will become a de facto standard
and no one wants to be left out. People are concerned that the Dophin
is a bit too slow (apparently about 1/2 the power of a Lisp Machine).
The feeling is that a year ago it would have been a great idea, but its
time may have passed. The problem is that there is no obvious other
candidate. People advised ARPA to look around carefully to see if there
wasn't something better. One problem is that whatever it is must
support Interlisp. Thus the Lisp machine will only work if they will
put Interlisp-D on it. This might be technically feasible, but the Lisp
Machine people aren't interested in doing it. (Whether the smell of that
much money will arouse their appetite remains to be seen.) The Jericho
isn't quite ready, nor are any other personal computers, though many are
on the horizon. The only other thing that could be done quickly would
be a single user VAX, which DEC might be able to build if they would, or
a small Foonly (a PDP-10 equivalent). DEC has a single user PDP-10,
which would be an ideal solution, but decided not to build it. The
guess is that even ARPA isn't a large enough customer to cause DEC to do
something it otherwise wouldn't do. If there is nothing better than the
Dolphin, I don't have much of a guess whether ARPA will get it or not.
I think they may flip a coin.
The problem is that the proposed buy of Dolphins would be a large
commitment, and might make it hard for ARPA to take advantage of a
better system that may show up in a year. It is not just that the
hardware will be expensive. It is that by the time you add in the
software and hardware support, the total is 3 times the actual hardware
purchase price.
At that point the meeting ended, with no conclusion.
========
===========================
End forwarded message
∂14-Apr-81 0118 V. Ellen Golden <ELLEN at MIT-MC> REM
Date: 14 April 1981 04:17-EST
From: V. Ellen Golden <ELLEN at MIT-MC>
Subject: REM
To: JMC at SU-AI
I agree with you that REM is not as desperate as he makes out, but he keeps
making it out on MC, and people bug me about his plans. I have renamed
one of them so it won't print out when someone "fingers" him, and send him
mail about it (I don't like to modify someone else's files without warning
them), at SAIL, which has somewhat better security than MC. He forwarded
that message to FFM (Steve Kudlak) at MC. So my attempt to "speak privately"
with him got put on MC's disks anyway. He has done the same with two or
three other messages I have tried to send him in private. I am (of course)
3000 miles away from him, and thus am in a very poor position. I know that
he is really a lot more noise than action, but he is putting messages here
on our disks which are disturbing (a) me, (b) other ITS users who in turn
disturb me by asking me to "do something" (what can I REALLY do at this
remove, except call someone on the West Coast, like RWG, which I have done
in the past, for instance), (c) some of his friends on the west coast, who
(explitive deleted) seem unable to figure out that they should seek assistance
there instead of from me... I guess I am supposed to be "House Mother" to the
net. I have tried calling him long distance, but I have not reached him
(at the time I found out he was out riding buses... I gather to scare people
who tried to call him, since that was when he threatened suicide the last
time). I admit I am in a poor position, since I probably should be refusing
access to MC to both REM and this Betsy person, but I cannot find reasons
for doing it to either of them except for the general nuisance value of the
"non-relationship" between them, which at this point is on REM's part. But
he is the only one of the pair who has any history of usefulness to MC.
See my dilemma? And this is REM's birthday. I have sent him a "Happy
Birthday" message, and also told him to clean up his act on MC. He has
replied in what I call "Typical REM fashion" meaning that he corrected my
spelling and grammar...
SIGH.
∂14-Apr-81 0122 V. Ellen Golden <ELLEN at MIT-MC>
Date: 14 April 1981 04:21-EST
From: V. Ellen Golden <ELLEN at MIT-MC>
To: jmc at SU-AI
Well, I just replaced REM's latest plan with a "birthday cake"
(something JPG and I send to people on their birthdays).
I guess I won't tell him, since the renaming is the same as I did
before, and he has already seen the cake, since I sent it to him
before... you can see the effect by doing :f rem@mc)
∂14-Apr-81 0123 V. Ellen Golden <ELLEN at MIT-MC>
Date: 14 April 1981 04:22-EST
From: V. Ellen Golden <ELLEN at MIT-MC>
To: jmc at SU-AI
Whoops... I'm not much of a Lisp programmer, I sent one too many
right parentheses in that last message... remove one ) at the end and
replace it with a period.
That's a fine birthday cake, and I'll not worry my head trying to figure
out how REM will take it. I don't see that people being disturbed by
fingering REM is quite reason enough to take action even to replace the
plan by something else. I too find REM's appeals disturbing - comparable
to the continual guilt-mongering begging of the listener supported
radio and TV stations. The remedy is the same in either case - switch
to another channel.
I'm curious about Betsy, however. I don't know if I've met her although
she's a user of SCORE, but REM's last flap was about her. I supposed that
she was an innocent user on whom REM had become fixated, but I gather
from your last message that she is also somewhat of a nuisance. In
what way?
∂14-Apr-81 0151 V. Ellen Golden <ELLEN at MIT-MC> Betsy
Date: 14 April 1981 04:50-EST
From: V. Ellen Golden <ELLEN at MIT-MC>
Subject: Betsy
To: jmc at SU-AI
Well, she is a nuisance only in that she (while claiming to want to
learn EMACS (which she can do on SCORE) and MACSYMA (ok...)) spends
most of her time on MC talking to people. Since spying on this system
is possible and easy, REM tends to watch her and send her messages. I
will not try to play psychiatrist for her any more than for REM, but
apparently he originally send her a message of a sexual nature (as he
has for most female types on ITS... I went through it myself, but did
not cause such a flap, being a bit older and more experienced, when he
started in on me, I just quietly ignored him until he stopped...).
Betsy started yelling to all her friends (mostly male, given the
population of computers) about how he was bugging her. They in turn
started yelling to me (the "House Mother"). She apparently did meet
him, talk to him, etc, and then decided he was too much (she does live
in Palo Alto). I get only the accounts from REM and Betsy (COOKIE@MC,
G.BETS@SCORE) of course, which makes it hard to evaluate. I started
by trying to tell Betsy to ignore REM. This did not work. (She
"loves everybody and wants to help...") She finally figured out that
"helping" REM was beyond her (she is also in poor health, which she
tells me whenever I suggest she take stronger action, like "If REM is
outside your house and won't go away, call the police"... she claims
she "can't stand it" with allusions to her health). In short, she is
a problem in that it seems to me that she should make herself scarce
(actually lately she finally is to some extent) so as to provide the
LEAST stimulation to REM as possible. REM at least has a history of
some useful work on MC, "COOKIE" has yet to do much beyond talk to
some of our systems programmers who would be better employed
programming.
I am sorry for the length of this message, and for involving you
in this. Rest assured that your mail to me is being deleted so as not
to remain in my mail file (which has the unfortunate property of being
like a BBOARD... EVERYONE reads ELLEN's mail). If you should wish to
say anything to me in less public fashion, you may use ELLEN@MIT-XX
which does NOT forward to MC, and which I read at least once a day.
∂14-Apr-81 0205 V. Ellen Golden <ELLEN at MIT-MC>
Date: 14 April 1981 05:05-EST
From: V. Ellen Golden <ELLEN at MIT-MC>
To: jmc at SU-AI
Thank you for your interest and your curiosity, and no, I do not
send messages which I think need to be deleted usually... my mail
to REM was special, as he has two mail boxes, and I chose the "secure"
one for a reason, and he chose to disregard my discretion.
Cheerio!
∂14-Apr-81 0859 LLW A NIL-Potent Advice Taker?
To: JMC
CC: LLW
∂14-Apr-81 0054 JMC sharing Dick Gabriel
I would like to support him to work on an Advice Taker, but I expect to have
money enough only for half of his time, and anyway he wants to put much of
his time into LISP. Is there any possibility that S-1 could support him
half time, say through the contract with Stanford to work on LISP?
[John: Yes, but only at academic rates (which are far below those of his
Schlumberger offer, and even much below ours)--I doubt that he could be
seriously interested. Also, the arrangement presently being finalized
with Dick and Guy Steele is for the two of them to work together at CMU
from now 'til Summer's end on NIL development, Guy as a sub-contractor PI
and Dick as a 'remote work assignment' Lab employee. I'm open to any
reasonable proposal that gets NIL going in the foreseeable future, but
Dick's working half-time at Stanford before Summer's end seems marginal.
How about post-Summer possibilities, perhaps using NIL on a Mark IIA?
Lowell]
∂14-Apr-81 1057 CSD.CRANGLE at SU-SCORE newell article
Date: 14 Apr 1981 1056-PST
From: CSD.CRANGLE at SU-SCORE
Subject: newell article
To: jmc at SU-AI
Your secretary cannot find the article on representation. If you are
able to find it, could you let me borrow it please.
Colleen Crangle
-------
∂14-Apr-81 1215 CLT
To: "@A.DIS[1,CLT]"
I have agreed to put together the FR section of the ARPA proposal.
The basic form will be that of previous proposals. (The 1979 proposal
should appear soon as FR79[1,CLT] ) In the meantime, there is a hard
copy in my office.
A tentative outline for the text is:
0. Intro
1. ANALYST
2. Breif summary of accomplishments, applications, work completed
[references to relevant publications]
3. Individual sections
JMC
JK
JD
RPG-LGC
CLT
4. The FR Group (mini bios)
5. References
What I need from each of you is a pointer to a file containing
(1) Accomplishments
brief summary of work done during last contract
with references to any reports or papers (completed or underway)
(2) Description of proposed work
more details of previous accomplishments if you wish
A summary of proposed work in ARPAs favorite format
Goals & Milestones
(3) A sentence or two describing your qualifications (for section 4.)
See old proposal for examples (2.5)
(4) List of references to be included in section 5.
Of course all this is needed yesterday, if not before. Actually
Friday at the latest, earlier would be great. Sorry, but I am leaving
in a couple weeks, so it must be put together in all due haste.
I will to put together a draft this weekend.
Any constructive suggestions a welcome.
--- Carolyn ---
∂14-Apr-81 1704 Richard M. Stallman <RMS at MIT-AI>
Date: 14 April 1981 20:04-EST
From: Richard M. Stallman <RMS at MIT-AI>
To: JMC at MIT-AI
I was paid by MIT for the rented car and for the actual transportation.
You would know better than I what a plane ticket costs.
I'm pretty sure I could get travel approval if they didn't have
to pay anything additional for the trip.
I gave a talk here about a month ago, and I plan to go over the tape
and write down some of what I said. I've had a few ideas
since I came back. For one thing, I can put my defaulting
scheme into modal logic as well as metalogic. If I use
an operator triangle which means should-be-provable,
then use an inference rule triangle p goes to p for any p,
then this can be used in non-monotonic Doyle-McDermott style
logic so that defaults can be represented as possible p implies
triangle p (rather than the usual possible p implies p).
Also I realized that in meta logic when a theory refers to itself,
the result is not a disaster, just equivalent to modal logic.
Also, I figured out how to formulate a restricted reasoning domain
in which certain defaults apply, then talk about the results outside.
It's the non-monotonic equivalent of an implication: the statement
that if p is added to theory T then q would be a non-monotonic
consequence.
∂14-Apr-81 2219 Richard M. Stallman <RMS at MIT-AI>
Date: 15 April 1981 01:06-EST
From: Richard M. Stallman <RMS at MIT-AI>
To: jmc at SU-AI
I think I would like to come for only one month.
∂14-Apr-81 2304 ZM Pratt
To: RWF at SU-AI, DEK at SU-AI, JMC at SU-AI, JEF at SU-AI
CC: ZM at SU-AI
Letters will be sent to:
Rabin, Paterson, Milner, Burstall, Aho, Scott, Karp, Meyer,
Lipton, Reynolds, Constable, Hartmanis
Comparison list:
Lipton, Chandra, Paterson, Milner, Sethi, Valiant, Burstall
M.Fischer, Rivest, Meyer, Rabin, Constable, Kowalski, Parikh
Any comments?
∂15-Apr-81 1645 CSD.GOLUB at SU-SCORE Thanks
Date: 15 Apr 1981 1641-PST
From: CSD.GOLUB at SU-SCORE
Subject: Thanks
To: comp: ;
Thanks for comming to the meeting today.
We'll see you next Wednesday at 3:30 in Margaret Jacks.
GENE
-------
∂15-Apr-81 1914 CSD.DOYLE at SU-SCORE Proposals concerning AAAI publications
Date: 15 Apr 1981 1907-PST
From: CSD.DOYLE at SU-SCORE
Subject: Proposals concerning AAAI publications
To: bobrow at PARC-MAXC, bledsoe at UTEXAS-11, csd.buchanan at SU-SCORE,
csd.feigenbaum at SU-SCORE, feldman at SUMEX-AIM, jmc at SU-AI,
minsky at MIT-AI, newell at CMU-10A, nilsson at SRI-AI, reddy at CMU-10B,
simon at CMU-10A, walker at SRI-AI, tw at SU-AI, phw at MIT-AI
To the Executive Council,
American Association for Artificial Intelligence
April 15, 1981
Ladies and Gentlemen:
In January, 1981 I proposed to you (through Bruce Buchanan) that the
AAAI, as the scientific society of the field of artificial
intelligence, take up its proper responsibility for facilitating free
scientific communication in AI by publishing a monthly announcement of
titles and abstracts of papers. That proposal also mentioned other
possible reorganizations of scientific publications in our field, but
did not attach any urgency to their consideration. In this letter, I
extend my earlier proposal to urge publication of a scholarly journal
by the AAAI.
AI needs a scholarly journal dedicated to the communication of
detailed discoveries. For a decade now, the journal Artificial
Intelligence (henceforth AIJ) has served much of this need. However,
recent experiences of mine have awakened me to the unpleasant fact
that the AIJ is operated principally as an economic venture under the
direction of a private corporation, rather than as a communications
effort under the control of the scientific community it serves. To
illuminate the problem, I will first describe my experience with the
AIJ, point out the problem underlying it, and then show how my current
proposal would go far towards remedying the situation.
My experience was this. Recently, the AIJ granted, without
ever consulting me, exclusive rights to a third party to reprint an
article of mine in an anthology. However much I support the
publication of the anthology, this seems to be professionally
unjustifiable behavior on the part of the AIJ. I bear no grudge
against the editor of the AIJ, one of yourselves (Dr. Bobrow), for I
feel the matter to be more one of the institutional structure involved
than of personal responsibility. The AIJ receives copyright powers
over all papers submitted to it, and this seems to be neither unusual
nor at fault, as this power may often be necessary to protect the life
of the journal. However, as far as I know, there are no public
policies concerning the operation of the AIJ, and any unannounced
policies there may be would seem to be dictated more by the economic
interests of the North-Holland publishing company than by the
community of scholars whom the AIJ ostensibly serves. I would find it
extremely surprising were an organization devoted primarily to
encouraging scientific communication to grant exclusive reprinting
rights to individual articles, unless in exceptional circumstances,
and only after consultation with the authors involved. In most
fields, some fraction of papers worthy of being anthologized are
worthy of being reprinted in several different contexts, as they
sometimes make several different contributions to the field. But
unannounced exclusive delegations of authority to third parties can
only be justified on unusual economic gains to the journal or
anthology publisher.
Unfortunately, I cannot believe that my experience will be the
last such case of private economic interests overcoming public
communication responsibilities in the AIJ, no matter what the
intentions of its editors. The best way I see to avoid this problem
is that of the AAAI publishing its own scholarly journal, a journal
solely under the control of those it serves. Such a journal could of
course demand copyrights, but its policies about reprinting and other
matters could be matters for public discussion and review.
That, ladies and gentlemen, is the substance of my proposal.
I feel very stronly about this matter, and would be willing to help in
the effort of establishing such a journal, should you agree to it.
Thank you very much for your consideration.
Sincerely,
Jon Doyle
Computer Science Department
Stanford University
Stanford, CA 94305
415-857-1892
JD@SAIL
-------
∂15-Apr-81 2038 Woody Bledsoe <ATP.Bledsoe at UTEXAS-20> THANKS
Date: 15 Apr 1981 2236-CST
From: Woody Bledsoe <ATP.Bledsoe at UTEXAS-20>
Subject: THANKS
To: JMC at SU-AI
cc: ATP.Bledsoe at UTEXAS-20
JOHN, THANKS FOR GIVING A RECOMMENDATION FOR ME IN REGARDS TO
THE ASHBEL SMITH PROFESSORSHIP HERE AT UP (WHICH IS NOW HAVE).
I AM VERY PLEASED. BEST REGARDS, WOODY
-------
∂15-Apr-81 2138 CLT arpa
I've arranged for us to meet with Jon Doyle tomorrow around 2pm.
If that's not suitable let me know so I can tell Jon.
Unfortunately, I forgot I'm at SRI. Can you make it 5:15 or after?
∂15-Apr-81 2225 CLT
∂15-Apr-81 2203 JMC
Unfortunately, I forgot I'm at SRI. Can you make it 5:15 or after?
Thanks alot!
If JD isn't willing to hang around til then what about Friday, say at 11,
or 1?
Please check before committing yourself.
I've checked. I'll cancel my lunch with Elliott Bloom if necessary to
make this meeting. I prefer tomorrow if JD can do it, because it's sooner.
∂15-Apr-81 2253 CLT
∂15-Apr-81 2228 JMC
I've checked. I'll cancel my lunch with Elliott Bloom if necessary to
make this meeting. I prefer tomorrow if JD can do it, because it's sooner.
Just talked to JD. Can't make it late tomorrow. We tentatively set
1:30 Friday (modulo JDs car getting oiled by then). I'll have to
leave around 3. But that should be enough time. Presumably
you can get your lunch by 130?
∂16-Apr-81 0017 CLT Frege
I got a threatening letter from the math library regarding
`Frege and the philosophy of mathematics'. Could you please
put it on my table, so I can return it tomorrow? Thanks.
∂16-Apr-81 0106 JK
∂15-Apr-81 1720 JMC parser
For reasons I will be glad to enumerate, it would be very helpful if
your parser would allow operators comprising more than one operator
symbol, e.g. ∧' or ∧∧. All it needs to do is to consider a string
of operator symbols as designating a compound operator, there being
no necessary semantic relation between ∧ and ∧', etc. This would
also permit expressing ≤ as =< by people with impoverished character
sets.
-----
The current parser does allow operators with more than one letter in them
provided that they do not contain any special characters. The special
characters can then be scanned without any delimiters; i.e.
p∨q instead of p ∨ q and so on. I'll see what I can do to make the
parser more flexible in this regard. There are potential ambiguities:
For example, the string ∧¬ : what is meant by p∧¬q?.
I admit I hadn't thought of this last, and I suppose that any of the
reasonable solutions is acceptable.
(1) it is interpreted as a single symbol so that ∧ ¬p must be so written.
(2) it is interpreted the same as ∧ ¬.
(3) it is interpreted as ∧¬ if this has been defined otherwise as ∧ ¬.
∂16-Apr-81 0815 JK
∂16-Apr-81 0110 JMC
I admit I hadn't thought of this last, and I suppose that any of the
reasonable solutions is acceptable.
(1) it is interpreted as a single symbol so that ∧ ¬p must be so written.
(2) it is interpreted the same as ∧ ¬.
(3) it is interpreted as ∧¬ if this has been defined otherwise as ∧ ¬.
--------------------
o.k. - I think I will take (3); at the same time I will extend the type
checking to allow user to introduce as many ambiguities they like at their
own risk - the "first fit" will be recognized in some reasonable way.
This will take some time to do.
By the way, likewise with combinations like ∃'.
∂16-Apr-81 1051 FFL
To: JMC, FFL
Leslie Dugan said he would call again this afternoon. If you have time, he
would appreciate a call at 55 495 5669.
∂16-Apr-81 1134 Walker at SRI-AI Re: Proposals concerning AAAI publications
Date: 16 Apr 1981 1112-PST
From: Walker at SRI-AI
Subject: Re: Proposals concerning AAAI publications
To: CSD.DOYLE at SU-SCORE, bobrow at PARC-MAXC,
To: bledsoe at UTEXAS-11, csd.buchanan at SU-SCORE,
To: csd.feigenbaum at SU-SCORE, feldman at SUMEX-AIM,
To: jmc at SU-AI, minsky at MIT-AI, newell at CMU-10A, nilsson,
To: reddy at CMU-10B, simon at CMU-10A, tw at SU-AI,
To: phw at MIT-AI
cc: Erman at ISIB
Jon's proposal certainly constitutes a reasonable agenda item for the
Council to consider. Which reminds me that we ought to be arranging
both for a meeting of the Council and for a meeting of the AAAI, its
"Annual Meeting", which the by-laws require to occur and which "in years
when there is no Annual Conference shall be held at a time and place
selected by the Executive Council", such notice to be sent out not less
than 60 days before the date of such meeting. It would seem reasonable
for both of these meetings to take place at the IJCAI, the only
convocation likely to include a reasonable percentage of our members.
As I understand it, Ed's Presidential Address will be presented there.
I have spoken to Lou Robinson about the conference schedule, and there
seems to be some flexibility. We might in fact sequence our meeting
before or after the IJCAI business meeting. If there is no objection
in principle to such an action, rather than having the Council as a
whole make these plans, I suggest that Ed and I handle matters. In
any case, other agenda items for both the Council and the Annual meetings
should be forwarded to Ed for inclusion.
Since nominations should be sent out at least 60 days in advance, too,
I suggest that once we have decided on timing, the two items be mailed
out together. And I would suggest that that be done sooner rather than
later.
Don
-------
∂16-Apr-81 1248 TOB
John
Dr. William Gevarter, NASA now visiting at NBS, said that NASA is looking
to place Dr. Roger Cliff in a major AI center. Dr. Cliff is a computer
scientist at NASA, Goddard who has worked in design of spacecraft
computers. NASA would pay his salary plus money for expenses such as
computer resources. He would like to know about the possibility of Dr.
Cliff visiting at Stanford for one year. This is in the talking stage now.
Tom
∂16-Apr-81 1443 JJW EKL proof
See APPEND.TXT[EKL,JJW] for an EKL proof of the associativity of append.
Actually there's more stuff on this file; it covers roughly sections
3.1-3.5 of LISP: Programming and Proving.
∂16-Apr-81 2325 Robert Elton Maas <REM at MIT-MC> Non-loneliness
Date: 17 April 1981 02:23-EST
From: Robert Elton Maas <REM at MIT-MC>
Subject: Non-loneliness
To: REM-DIARY-READERS at MIT-MC
cc: JMC at SU-AI, GOSPER at PARC-MAXC
The past 22 hours have been quite curious. Believe it or not, I actually
haven't been lonely/depressed this whole time even though nothing really
nice has happened the whole time. I've felt at times sort of like I
did in 1968, like any minute now I was going to become depressed again,
but although I got very slightly depressed I haven't really been hurting
the whole time. It started when I went to bed Wednesday evening about
23:10 with abdominal soreness and very-tiredness, had pain in rectum&anus
as I tried to sleep and finally about 01:00 woke up with insomnia. But
this insomnia wasn't acompanied with loneliness or depression like my
insomnia usually has been for the past 13 years, and I haven't felt
really hurting any time since that insomnia started. -- I have some
conjectures about what might have caused this state of non-deprssion:
(1) Wednesday morning about 1-2 am Betsy finally explained to me what
I did that upset her and made her not want to be my friend, and it took
24 hours for the relief (over finally ending 7 weeks of wondering what
went wrong between us) to set in my mind, (2) Wednesday evening FFM and
I went out to hang around bookstores to pick up women, and there was this
employee at Plowshare who had really pretty blue eyes and was sort of
friendly to me, (3) whatever combination of food that made me
sick in abdomen also had an opposite effect on my mind, (4) the book I
bought at Plowshare (Goedel/Escher/Bach) happens to be the achilles-heel
for my depression record that has been playing inside me for 19 years
and almost continuously for the past 13 years, (5) all of the above,
(6) this is totally random I happened to not be depressed for totally
random fluctuations in my mind for a longer time than normally I get
such fluctuations and there isn't any identifiable cause. -- At folk
dance tonite I started to get depressed a couple times when women sort
of rejected, but never got really hurt/depressed/lonely, just slightly.
Most of the time tonight, when women asked me how I was I could honestly
say "ok" for the first time in years!
∂17-Apr-81 0510 Darden@SUMEX-AIM trip to Washington
Date: 17 Apr 1981 0508-PST
From: Darden@SUMEX-AIM
Subject: trip to Washington
To: JMC@SAIL
Hi, John. I am back from Boston. What are the plans for Monday,
April 20?
-------
∂17-Apr-81 1536 Untulis at SRI-AI Badge application
Date: 17 Apr 1981 1536-PST
From: Untulis at SRI-AI
Subject: Badge application
To: jmc at SU-AI
John,
Please stop by my office the next time you are over at SRI in order
to fill out some of the information for a badge request.
Thanks
Chuck
-------
∂18-Apr-81 1123 CLT
I think the material preceeding the analyist example in FR should be
rewritten. Perhaps even from scratch. Will you do it?
Also, I find the fragments of possible solutions and proposed work
that appear in the analyst example non-optimal. Wouldn't it be better
to just use the example to motivate interest in and explain technicals problems
and use the other material in the proposed work sections, to relate back.
Or it this other style something that pleases ARPA types.
∂18-Apr-81 1628 RPG
∂18-Apr-81 1616 JMC
To: LGC, RPG, JD, CLT
How about 3pm Sunday for getting together and finishing it off.
I'll check with my boss, who is planning a large Easter food consortium
sometime tomorrow. I certainly can authorize LGC to make judgments for me
on the proposal if he can make it and I cannot at that time.
-rpg-
∂18-Apr-81 1743 LGC Tomorrow
It would be better for my family life if we met at or after 4pm on Easter
Sunday, but if everyone else is prepared to meet earlier than that, I will
do my best to cooperate. Would it cause any problems if I arrived at 3:30?
I'll have a new, ANALYST-oriented version of FRAT.TXT[EP,LGC] later on today,
and will ask RPG to edit it if he's available. -- Lew
P.S. A control typo garbled the previous msg; please flush it.
∂18-Apr-81 2210 LLW Talking About RPG
To: JMC
CC: LLW
∂18-Apr-81 2207 JMC
Do we need to talk about Dick Gabriel?
[I don't think we need to, at the moment. If it seems to you that we
should, please give me a ring at 422-0758 at your convenience.]
∂18-Apr-81 2250 CLT
To: JMC, LGC, JD, JK
In FR[1,CLT] there is a section `.SS Summary of work completed'
I have made entries for myself and people no longer around.
If you have accomplishments, work completed, etc., that would
fit better here that in your individual section of the proposal
send me the appropriate text.
∂19-Apr-81 0051 RPG Meeting
To: JMC
CC: LGC
Lew and I can meet at your office from 4:30 until about 6:00 tomorrow afternoon.
We've thought about the Analyst and agree that it's more ambitious and,
at least for me, more fun sounding, but we are both nervous about
it being a little too big a little too fast.
-rpg-
∂19-Apr-81 1054 JD analyst
I'm sorry, but I am busy all day today. I am still considering your
proposals, and will discuss them with you later.
By the way, I'm not sure just what technique you were thinking about
for this property list business, but people have been using property
lists of arbitrary objects for over a decade now in many, many programs.
It certainly is not a "not yet implemented" technique. It even appears
as a standardized feature of languages like Conniver and QA4. I suspect
you should check out the facts, lest your proposal seem to hinge on a
well-known "innovation". I'd also recommend not discussing trivial
details of datastructures in a theoretical proposal. It's at the wrong
level of description.
I guess we will meet later.
Jon
∂19-Apr-81 1555 Darden@SUMEX-AIM Re: Monday
Date: 19 Apr 1981 1553-PST
From: Darden@SUMEX-AIM
Subject: Re: Monday
To: JMC@SU-AI
In response to your message sent 18 Apr 1981 2320-PST
Hi, John, I will be happy to meet you at BWI Monday night and have
you stay with me. I do have a small guest bedroom and since I
don't teach on Tuesday I can take you out to the National Bureau
of Standards; I assume you have to go to their place near Rockville.
On Wednesday I will be quite pressed for time, but I assume something
can be worked out to get you back to the Bureau or to the airport.
The limosome for BWI leaves from Greenbelt, so that will be
convenient for you.
What is your airline and flight number? Is there anything in
particular that you like to have in the refrigerator or for breakfast?
Looking forward to seeing you.
Bye, Lindley.
-------
That will all be lovely. My flight is UA108 which arrives at BWI
at 1035pm. My breakfast tastes are entirely omnivorous. I too
look forward to it.
∂20-Apr-81 0235 CLT ARPA
To: "@A.DIS[1,CLT]"
The `semi-final' draft of FR is complete. XSP A.XGP[1,CLT] or
ET FR[1,CLT]. It will be sent to press later today. Let me
know if you have any corrections or modifications.
(There is a copy on my desk if XSP or ET is broken.
Fran can let you in.)
∂20-Apr-81 1231 CLT phon
howcum FIND can't find phon[1,jmc]?
∂20-Apr-81 1824 BYY situations
could you send me the references for some of your papers
that use situations? I need them for my acl paper qutie soon. Thanks, Jon
∂20-Apr-81 2118 BYY acl paper
SOME COMPUTATIONAL ASPECTS OF SITUATION SEMANTICS
Jon Barwise
Can a realist model theory of natural language be computationally
plausible? Or, to put it another way, is the view of lingusitic meaning
as a relation between expressions of a natural language and things (objects,
properties, etc.) in the world, as opposed to a relation between expressions
and proceedures in the head, consistent with a computational approach to
understanding natural language? The model theorist must either claim
that the answer is yes, or being willing to admit that humans transcend
the computationally feasible in their use of language?
Until recently the only model theory of natural language that
was all well developed was Montague Grammar. Unfortuanetly, it was based
on the primitve notion of "possible world" and so was not a realist
theory, unless you are prepared to grant that all possible worlds are
real. Montague Grammar is also computationall intractible, for reasons to
be discussed below.
John Perry and I have developed a somewhat different approach to
the model theory of natural language, a theory we call "Situation Semantics".
Since one of my own motivations in the early days of this project was to
use the insights of generalized recursion theory to find a computationally
plausible alternative to Montague Grammar. it seems fitting to give a progress
report here.
1. Model-theoretic Semantics "versus" Procedural Semantics
First, however, I can't resist putting my two cents worth into this
continuing discussion. Procedural semantics starts from the observation
that there is something computational about our understanding of natural
language. This is obviously correct. Where some go astray, though, is in
trying to identify the meaning of an expression with some sort of program
run in the head. But programs are the sorts of things to HAVE meanings,
not to BE meanings. A meaningful program sets up some sort of relationship
between things - perhaps a function from numbers to numbers, perhaps something
much more sophisticated. But it is that relation which is its meaning, not
some other program.
The situation is analogous in the case of natural language. It is
the relationships between things in the world that a language allows us to
express that make a language meaningful. It is these relationships that are
identified with the meanings of the expressions in model theory. The
meaningful expressions are procedures that define these relations that are
their meanings. At least this is the view that Perry and I take in situation
semantics.
With its emphasis on situations and events, situation semantics shares
some perspectives with work in A.I. on representing knowledge and action
(McCarthy (1968) e.g.) but its differs in some crucial respects.
It is a mathematical theory of linguistic meaning, one that replaces the
view of the connection between language and the world at the heart of Tarski-
style model theory with one much more like that found in J.L. Austin's "Truth".
For another, it takes seriously the syntactic structures of natural language,
directly interpreting them, without assuming an intermediary level of "logical
form".
!2. A Computation Obstruction at the Core of First-Order Logic.
The standard model-theory for first-order logic, and with it the
derivative model-theory of indices ("possible worlds") used in Montague
Grammar in based on Frege's supposition that
that the reference of a sentence could only be taken
as a truth value; that all else specific to the sentence is lost at the
level of reference. As Quine has seen most clearly, the resulting view
of semantics is one where to speak of a part of the world, as in (1),
is to speak of the whole world and of all things in the world.
(1) The dog with the red collar belongs to my son.
There is a philosophical position that grows out of this view of logic,
but it is not a practical one for those who would implement the resulting
model-theory as a theory of natural language. Any treatment of (1)
that involves a universal quantification over all objects in the domain
of discourse is doomed by facts of ordinary discourse, e.g., the fact that
I can make a statement like (1) in a situation to describe another
situation without making any statment at all about other dogs that come
up later in a conversation, let alone about the dogs of Tibet.
Logicians have been all too ready to dismiss such philosophical
scruples as irrelevant to our task - especially shortsighted since the same
problem is well known to have been an obstacle in developing recursion
theory, both ordinary recursion theory and the generalizations to other
domains like the functions of finite type.
We forget that only in 1938, several years after his initial work
in recursion theory, did Kleene introduce the class of PARTIAL recursive
functions in order to prove the famous Recursion Theorem. We tend to
overlook the significance of this move, from total to partial functions,
until its importance is brought into focus in other contexts. This is
just what happened when Kleene developed his recursion theory for functions
of finite type. His initial formulation restricted attention to total
functions, total functions of total functions, etc. Two very important
principles fail in the resulting theory - the Substitution Theorem and
the First Recursion Theorem.
This theory has been reworked by Platek (1963), Moschovakis (1975),
and by Kleene (l978, 80) using partial functions, partial functions of
partial functions, etc., as the objects over which computations take place,
imposing (in one way or another) the following constraint on all objects
F of the theory:
Persistence of Computations: If s is a partial function and F(s) is defined
then F(s') = F(s) for every extension s' of s.
In other words, it should not be possible to invalidate a computation
that F(s) = a by simply adding further information to s. To put it yet
another way, computations involving partial functions s should only be
able to use positive information about s, not information of the form that
s is undefined at this or that argument. To put it yet another way, F
should be continuous in the topology of partial information.
Computationally, we are always dealing with partial information
and must insure persitence (contituity) of computations from it.
But this is just what blocks a straightforward implementation of the
standard model-theory - the wholistic view of the world which it
is committed to, based on Frege's initial supposition.
When one shifts from first-order model-theory to the index or
"possible world" semantics used in Montague's semantics for natural
language, the wholistic view must be carried to heroic lengths. For
index semantics must embrace (as David Lewis does) the claim that
talk about a particular actual situation talks indirectly not just
about everything which actually exists, but about all possible objects
and all possible worlds. And it is just this point that raises serious
difficulties for Joyce Friedman and her co-workers
in their attempt to implement Montague Grammar in a working system.
The problem is that the basic formalization of possible world semantics
is incompatible with the limitations imposed on us by partial information.
Let me illustrate the problem that arises in a very simple instance.
In possible world semantics, the meaning of a word like `talk' is
a total function from the set I of ALL possible worlds to the set of
ALL TOTAL functions from the set A of ALL possible individuals to
the truth values 0,1. The intuition is that b talks in `world' i if
meaning(`talk')(i)(d) = 1.
It is built into the formalisim that each world contains TOTAL information
about the extensions of all words and expressions of the language.
The meaning of an adverb like `rapidly' is a total function from
such functions (from I into Fun(A,2)) to other such functions. Simple
arithmetic shows that even if there are only 10 individuals and 5
possible worlds, there are (2exp50)exp(2exp50) such functions and
the specification of even one is completely out of the question.
The same sorts of problems come up when one wants to study the
actual model-theory that goes with Montague Semantics, as in Gallin's
book. When one specifies the notion of a Henkin model of intensional
logic, it must be done in a totally `impredicative' way,
since what constitutes an object at any one type depends on what the
objects are of other types.
For some time I toyed with the idea of giving a semantics for
Montague's logic via partial functions but attempts convinced me
that the basic intuition behind possible worlds is really inconsistent
with the constraints placed on us by partial information. At the same time
work on the semantics of perception statements led me away from possible
worlds, while reinforcing my conviction that it was crucial to
represent partial information about the world around us, information
present in the perception of the scenes before us and of the situations
in which we find ourselves all the time.
!3. Acutal Situations and Situation-types.
The world we perceive and talk about consists not just of objects,
nor even of just objects, properties and relations, but of objects
having properties and standing in various relations to one another;
that is, we percieve and talk about various types of situations from
the perspective of other situations.
In situation semantics the meaning of a sentence is a relation
between various types of situations, types of discourse situations
one the one hand and types of `subject matter' situations on the other.
We represent various types of situations abstractly as PARTIAL functions
from relations and objects to 0 and 1. For example, the type
s(belong, Jackie, Jonny) = 1
s(dog, Jackie) = 1
s(smart, Jackie) = 0
represents a number of true facts about my son, Jonny, and his dog.
(It is important to realize that s is taken to be a function from
objects, properties and relations to 0,1, not from words to 0,1.)
A typical situation-type representing a discourse situation
might be given by
d(speak, Bill) = 1
d(father, Bill, Alfred) = 1
d(dog, Jackie) = 1
representing the type of discourse situaton where Bill, the father of
Alfred, is speaking and where there is a single dog, Jackie, present.
The meaning of
(2) The dog belongs to my son
is a relation (or multi-valued function) R between various
types of discourse situations and other types of situations. Applied
to the d above R will have various values R(d) including s' given below, but
not including the s from above:
s'(belong, Jackie, Alfred) = 1
s'(tall, Alfred) = 1.
Thus if Bill were to use this sentence in a situation of type d, and if
s, not s', represents the true state of affairs, then what Bill said would
be false. If s' represent the true state of affairs, then what he said
would be true.
Expressions of a language have a fixed linguistic meaning, independent
of the discourse situation. The same sentence (2) can be used in different
types of discourse sitations to express different propositions. Thus, we
can treat the linguistic meaning of an expression as a function from discourse
situation types to other complexes of objects and properties. Application
of this function to a particular discouse sitation type we call the inter-
pretation of the expression. In particular, the interpretation of a
sentence like (2) in a discourse situation type like d is a set of various
situation types, including s' above, but not including s. This set of types
is called the proposition expressed by (2).
Various syntactic categories of natural language will have various
sorts of interpretations. Verb phrases, e.g., will be interpreted by
relations between objects and situation types. Definite descriptions will
interpreted as functions from situation types to individuals. The difference
between referential and attributive uses of definite descriptions will
correspond to different ways of using such a function, evaluation at a
particular accessible situation, or to constrain other types with its domain.
4. A fragment of English involving definite and indefinite descriptions
At my talk I will illustrate the ideas discussed above by
presenting a grammar and formal semantics for a fragment of English that
embodies definite and indefinite descriptions, restrictive and non-
restricitve relative clauses, and indexicals like "I", "you", "this"
and "that". The aim is to have a semantic account that does not go thru
any sort of first-order "logical form", but operates off of the
syntactic rules of English. The fragment incorporates both
referential and attributive uses of descriptions.
The basic idea is that descriptions are interpreted as functions
from situation types to individuals, restrictive relative clauses
are interpreted as functions from situation types to sub-types, and
the interpretaton of the whole is to be the compositon of the functions
interpreting the parts. Thus, the interpretations of "the", "dog", and
"that talks" are given by the following three functions, respectively:
f(X) = the unique element of X if there is one,
= undefined, otherwise.
g(s) = the set of a such that s(dog, a)=1
h(s) = the `restriction' of s to the set of a such that
s(talk,a)=1.
The interpretation of "the dog that talks" is just the compositon of these
three functions.
From a logical point of view, this is quite interesting. In first-order
logic, the meaning of `the dog that talks' has to be built up from the
meanings of `the' and `dog that talks', not from the meanings of `the dog'
and `that talks'. However, in situation semantics, since compositon of
functions is associative, we can combine the meanings of these expressions
either way: f.(g.h) = (f.g).h. Thus, our semantic analysis is compatible with
both of the syntactic structures argued for in the linguistic literature,
the Det-Nom analysis and the NP-R analysis.
One point that comes up in Situation Semantics that might interest
people at this meeting is the reinterpretaton of compositionality that it
forces on one, more of a top-down than a bottom up compositionality. This
makes it much more computationally tractible, since it allows us to work
with much smaller amount of information. Unfortunately, a full discussion
of this point is beyond the scope of such a small paper.
Another important point not discussed is the constraint placed by
the requirement of persistence discussed in section 2. It forces
us to introduce space-time locations for the anakysis of attributive
uses of definite descriptions, locations that are also needed for the
semantics of tense, aspect and noun phrases like `every man', `neither
dog', and the like.
5. Conclusion
The main point of this paper has been to alert the readers to
a perspective in the model theory of natural language which they
might well find interesting and useful. Indeed, they may well find
that it is one that have in many ways already adopted for other reasons.
REFERENCES
1. J.L. Austin, "Truth", Philosophical Papers, Oxford, 1961, 117-134
2. J. Barwise, "Scenes and other situations", J. of Philosophy, to appear, 1981
3. J. Barwise and J. Perry, "Semantic innocence and uncompromising situations",
Midwest Studies in Philosophy VI, to appear 1981
4. J. Barwise and J. Perry, Situation Semantics: A Mathematical Theory of
Linguistic Meaning, book in preparation
5. S.C. Kleene, "Recursive functionals and quantifiers of finite type
revisited I", Generalized Recursion Theory II, North Holland,
1978, 185-222; and part II in The Kleene Symposium, North Holland,
1980, 1-31
6. J. McCarthy, "Programs with common sense", Semantic Information Processing,
(Minsky, ed.), M.I.T., 1968, 403-418
7. R. Montague, "Universal Grammar", Theoria, 36 (1970), 373-398
8. Y.N. Moschovakis, "On the basic notions in the theory of induction",
Logic, Foundations of Mathematics and Computability Theory, (Butts
and Hintikka, ed), Reidel, 1976, 207-236
9. J. Perry, "Perception, action and the structure of believing", to appear.
10. R. Platek, "Foundations of Recursion Theory", Ph.D. Thesis, Stanford
Universty, 1963
___________________________________________________________________________
Stanford University (Department of Philosophy)
University of Wisconsin, Madison (Departments of
Mathematics and Computer Science)
∂21-Apr-81 1040 FFL
To: JMC, FFL
Arnold Sanns of Behavioral and Brain Sciences would like to talk with you
about refereeing an article. Collect 609 921 7771.
∂21-Apr-81 1100 UNTULIS at SRI-AI [VIVIAN: Re: [Untulis: McCarthy account]]
Date: 21 Apr 1981 1054-PST
From: UNTULIS at SRI-AI
Subject: [VIVIAN: Re: [Untulis: McCarthy account]]
To: jmc at SU-AI
Date: 21 Apr 1981 0953-PST
From: VIVIAN
Subject: Re: [Untulis: McCarthy account]
To: UNTULIS
cc: heeter, sad
In-Reply-To: Your message of 16-Apr-81 1306-PST
JMC for John McCarthy on the 2060 has been set up. Password is OR7-1511.
Vivian
-------
---------------
-------
∂21-Apr-81 1104 FFL supplemental pay in the summer
To: JMC, FFL
Betty Scott asked me to tell you that we are planning to pay your two
summer months in July and August. Your leave month would occur in
September.
∂21-Apr-81 1809 REM via SU-TIP My crunch-and-spindle program
To: JMS
CC: JMC
There appears to be some source for it on [CSP,SYS] but somebody in
their infinite wisdom and paranoia has protected it so I, its author,
can't even get directory information about it to confirm whether it
is the correct program. CRU3.FAI[CSP,SYS]
∂22-Apr-81 1148 FFL
To: JMC, FFL
Martha Coyote called to ask if you had mailed the check.
4l5 465 4528.
∂22-Apr-81 1310 Buchanan@SUMEX-AIM agenda item for AAAI from Jon Doyle
Date: 22 Apr 1981 1227-PST
From: Buchanan@SUMEX-AIM
Subject: agenda item for AAAI from Jon Doyle
To: AAAI committee:
cc: csd.barb@SCORE
Mail-from: ARPANET host SU-SCORE rcvd at 21-Apr-81 1138-PST
Date: 21 Apr 1981 1136-PST
From: CSD.DOYLE at SU-SCORE
Subject: copy of the original proposal
To: buchanan at SUMEX-AIM
January 8, 1981
Executive Council
American Association for Artificial Intelligence
Ladies and Gentlemen:
In forming and nurturing the AAAI, you have taken timely
actions deserving of the field's applause and support. For the first
time, the field has a forum for discussion and presentation of its own
interests. Through your leadership, the field can now take direct
measures to improve itself. In this letter, I submit to you for your
consideration two proposals aimed at improving scientific
communication in artificial intelligence. The first proposal is for
the creation of a monthly newsletter of abstracts and titles. The
second proposal concerns reorganizations of responsibilities for
publications in the field.
For some time, I and others I have talked with have been
troubled by the typical isolation of research groups in artificial
intelligence. Workers in AI seem woefully ill-informed of each
other's work. This problem leads to poor scholarship, excessive
duplication of effort, and worst of all, an inability to build on the
work of others. While we cannot abolish scientific ignorance and
multiple discoveries, we should at least feel bound to investigate
simple steps to reduce ignorance and duplication when possible.
A major contributing factor to this problem of isolation is
that much, even most, of the work in AI is never published except as
local technical reports or as conference papers. When there were only
a few centers of research and only one or two conferences, these media
were tolerable. But currently, there are not only several large
centers of research, but also many smaller centers and numerous
isolated but interested individuals. There are ever-multiplying
conferences and specialty workshops. This growth in the number of
informal and semi-formal repositories for research results spells the
doom of effective scientific communications in AI unless serious
measures are taken. It is now possible to keep track of what is
published only by means of near super-human effort on the part of
researchers, namely hearing of and writing off for all conference
proceedings, and regularly enquiring about technical reports from all
known centers of research.
Fortunately, there is a simple measure taken by many other
scientific societies which makes more feasible keeping track of what
is published. This measure is the publication of a monthly
announcement of abstracts and titles. The purpose of the A/T monthly
is to provide a timely list of all current informal papers, work in
progress, research reports, talks, conference papers, journal
articles, and books, much like the monthly accessions list of an ideal
technical library. The tables of contents would be printed for all
relevant technical journals and conferences of any nationality
appearing in the preceding month, or possibly even from the page
proofs or advance announcements of the journals and conferences. Some
editing may be necessary since the list of journals in which
AI-relevant material occasionally appears is quite long, owing to the
interdisciplinary nature of AI and the multi-disciplinary origins of
its researchers. (E.g. ACM journals and conferences, IEEE journals
and conferences, IFIP conferences, Cognitive Science, IJMMS, Science,
Technology Review, Omni, Biological Cybernetics, ACL, and assorted
other journals in philosophy, logic, education, psychology, etc.) The
abstracts would be limited length, standard format notices of results
or papers prepared and submitted by the authors or their agents. The
content of these abstracts would be completely unrestricted, with only
format and possibly frequency controlled.
For example, the American Mathematical Society publishes three
monthlies towards these ends: Current Mathematical Publications, which
publishes tables of contents from page proofs of journals; Abstracts,
which publishes the standard form abstracts or short notices from
conferences or independent of conferences; and Notices, which, among
other things, prints in advance the titles of papers presented at
conferences.
The AMS Abstracts are made uniform by means of standard
camera-ready preparation forms available free to authors from any
mathematics department or from the AMS. One possible improvement on
their scheme might be to publish a submission form in each issue. In
the abstracts, typographical conventions of capitalized names and
underlined titles makes skimming of the abstracts easy. In addition,
each abstract indicates whether or not the author is willing to supply
preprints or reprints, so that notices can be made of work in
progress. To aid authors, the AMS offers the further service of
typing or re-typing abstracts for a standard fee.
I think it crucial that the ease of submission of abstracts
not be dependent on access to the Arpanet. For the next few years, at
the very least, the population of AI researchers without Arpanet
access will grow substantially, and nothing could sabotage the value
of this publication more than effectively excluding participation by
these people. Those with Arpanet access normally have convenient
runoff facilities, so requiring camera-ready copy for short,
fixed-form abstracts is little burden for them.
AMS Abstracts are also grouped by subject categories under
major headings of the Mathematics Subject Classification. The authors
choose which heading best classifies their abstract. Fortunately, AI
is probably still small enough so that one month's worth of abstracts
can be reduced to managable size with only a rudimentary
classification scheme. As the constant evolution of the Mathematics
Subject Classification indicates, the details of the classification
are rarely crucial. It is worthwhile developing and periodically
revising a standard classification scheme for AI, but I will not
pursue that in this proposal.
In summary, the proposed A/T monthly is easy to use for
several reasons. (A) It is monthly, so authors need never worry about
great delays caused by missed submission deadlines. (B) It is easy to
skim, because simple typographical conventions illuminate the subject
category, author, and title in different ways. (C) Finally, it is
easy to make submissions because of the fixed format and standard
submission forms.
While these reasons show the monthly is easy to use, the two
most difficult tasks lie in making it useful. The lesser task is that
of ensuring regular publication. As the Sigart Newsletter disaster
shows, an irregular publication receives no submissions from authors.
I suggest that regular publication requires a professional (i.e. paid)
editor or producer, rather than an overworked volunteer whose true
duties lie elsewhere. This is, of course, a matter for some study.
But the more important task in ensuring the utility of the monthly is
that of developing habits of near universal submission on the part of
authors. The very parochialism which the abstracts are aimed at
ameliorating makes developing such habits difficult. Once developed,
however, the system is self-regulating, in that authors can see it is
to their advantage to give notice of their work. When everyone finds
out what is going on by reading it in the abstracts, work left out is
work ignored.
This concludes the substance of the first proposal.
Currently, the Sigart Newsletter attempts to fulfil this
purpose, among others. How well does it do? I submit that it is
inadequate for several reasons. (A) Sigart Newsletter is a quarterly
publication, with a month-long ACM lead time for printing and
distribution. This means that an author submitting an abstract can
wait up to 4 or more months before the abstract appears, rather than
up to a month or month and a half in the case of a monthly. (B)
Sigart Newsletter occasionally pre-empts all regular features in favor
of special issues, which can lengthen the delay of abstract appearance
indefinitely. (C) Sigart Newsletter has no standard form for abstract
submissions, thus putting the burden of typing and formatting the
abstracts on a basically elsewhere-committed editorial staff. The
amount of work involved in this makes finding willing and competent
editors very difficult. (D) Sigart Newsletter lacks the respect
needed to ensure universal submissions. This seems partly due to
several reasons: (D1) The admitted ``breezy'' nature of the rest of
Sigart Newsletter encourages serious researchers at the larger centers
to ignore it as worthless. (D2) The fiascos of production in
1977-1979 encouraged ignoring it, since it did not exist for such a
long time. (D3) Finally, some part of its poor reputation may be due
to a hostility to ACM, which will be discussed in the second proposal,
below. In summary, the availability of frequent abstract notices is
too important to depend on the fragility of Sigart Newsletter's
sometimes irregular, sometimes preempted, and always infrequent
publication.
My second proposal in this letter is to suggest that the
principal responsibility for scientific communications in AI lies with
the field's scientific society, namely AAAI, and that accepting this
responsibility requires reorganizing the publication systems of the
field. The most serious defect of the current organization is the
abdication of responsibility properly that of AAAI in favor of a
subsidiary of the professional/scientific society of another
discipline, namely ACM's SIGART.
As an analogy, there is the field of mathematics, and the
field of biology, and a subspecialty of mathematical biology. It does
little harm to have a small mathematical biology field expedite its
communications as subsidiaries of either the mathematical or
biological scientific societies. But it makes no sense to have, for
example, all mathematical abstracts appear in a publication of a
subsidiary of biology. That responsibility belongs to the
mathematical field itself.
It is true that AI developed largely as a subfield of computer
science, but that is no excuse for continuing to restrict the conduct
and communications of our science by obesiance to the ACM bureacracy.
I have seen some of the difficulties in getting ACM's permission to do
things, and those are extra difficulties AI does not need. I have
also known several researchers, and I am sure there are more, who do
not belong to ACM, and have little interest in joining one of its
subsidiaries. Should we require AAAI members to join Sigart or ACM to
get scientific communications properly due from AAAI?
It may be that Sigart has served its purpose. It served as an
organ of the field as best it could until the field recognized itself
with its own scientific society. It may thus be best to dissolve
Sigart. However, that is a drastic step, and there may still be a
role for Sigart, for communicating ideas between the fields of AI and
CS. But this should imply a duplication or paraphrasing of
information disseminated by the two scientific societies, rather than
a supplantation of the main responsibilities of the scientific
societies to their members. Sigart might be for the computer
scientist who wishes to keep up with the main events in AI, or for the
AI worker who wants to stay abreast of developments in CS, but it
should not be the main vehicle for the computer scientist finding out
about CS or for the AI worker keeping track of AI.
If AAAI accepts principal responsibility for scientific
communications in AI, it is prudent to reflect on the sorts of
information communicated by other scientific societies as a way of
sorting out requirements on publications. By an examination of
Science, American Scientist, Artificial Intelligence, CACM, JACM,
Sigart Newsletter, AMS Notices, Abstracts, Bulletin, Current
Mathematical Publications, and Mathematical Reviews, I produced in
Table 1 a list (not necessarily complete) of categories of
communications. Along with each of these categories, I have listed
three publication frequencies: ideal frequencies, tolerable
frequencies, and current frequencies. I have entered M (monthly) and
Q (quarterly) in a conservative fashion for the ideal and tolerable
columns (e.g. monthly scholarly papers wouldn't be bad, and letters
and queries probably should be more frequent than quarterly). On the
other hand, I have been very generous in the current column (e.g.
calls for papers are published in a quarterly, but only once, rather
than with new and previous listings). None of these frequencies are
hard and fast, but I would think it a great hindrance to a field as
rapidly growing as is AI not to have a monthly announcement of at
least current publications and abstracts.
Ideally, I would divide up these categories into publications
as shown in Table 2, subject to revision as the field grows. In
principle, these divisions might be combined. For example, the
Notices and Titles and Abstracts might be one publication, and might
even incorporate the Reviews as well, if reviews of scholarly articles
are shifted to the scholarly Journal.
Currently, as mentioned above, AAAI leaves many of these
categories to the quarterly Sigart Newsletter, and devotes its AI
Magazine (as best I can tell) to the research news/surveys/expository
articles category. While I strongly recommend that AAAI take
responsibility for the monthly publication of those categories
currently left to Sigart, I do not recommend as strongly taking the
scholarly journal responsibility away from Artificial Intelligence.
AAAI should stand ready to begin its own scholarly journal whenever
necessary. The only reason for which I might suggest starting one now
is that too few papers in the field are submitted for scholarly
publication, and that another journal might encourage more
publications and more careful expositions of work that currently ends
up as informal reports.
Together with reflection on the organization of published
scientific communications, AAAI might well reflect on the
organizations of its conferences. At the international level there is
no federation of AI societies (i.e. AAAI, AISB, GI, CSCSI/CECIO,
etc.), but there is an international biennial conference, IJCAI.
There are also numerous specialty conferences, often of international
character. (Computer science analogues of these might be IFIP and its
working group workshops.) At the national level, we have the AAAI and
its annual conference. I think that soon AI will grow large enough
for AAAI to reconsider its abandonment of its annual meeting every
fourth year to IJCAI. At the regional level, there are no
organizations, and no conferences. The AMS has frequent regional
meetings, whose titles and abstracts are communicated to members in
other regions by means of the titles and abstracts monthlies. This
facilitates communications in areas which do not demand large travel
budgets. Steps hinting of this in AI have been taken in the TINLUNCH
meetings at SRI, the SIGLUNCH meetings at Stanford, and the AISNE
meetings in New England: informal meetings for graduate students or
others to present their work briefly. Perhaps AAAI ought to sponsor
such regional meetings along lines similar to the AMS, in which anyone
who wants to can submit an abstract and talk for 15 minutes (total -
presentation and questions), with no published proceedings outside of
the titles and abstracts published in the monthlies. I suspect AI has
grown up and out of MIT, CMU, and Stanford enough to make a number of
regional meetings worthwhile by stimulating the growing numbers of
researchers outside those three centers.
In conclusion, I propose both the development of a monthly
announcement of titles and abstracts, and the acceptance of
responsibility for scientific communication in AI by AAAI. I see
little hope of improving the quality of communications, and hence the
level and pace of research, without creating channels encouraging it.
I submit that the current communications channels are inadequate for
this purpose.
Respectfully yours,
Jon Doyle
Department of Computer Science
Stanford University
Stanford, California 94305
----------------------------------------------------------------------
Table 1.
Categories of Communications
Category Ideal Toler. Current
----------------------------------------------------------------------
Editorials/messages from president Q Q Q
Letters M Q Q
Queries M Q -
Society news/reports on meetings M Q Q
Calendar of events (new and previous) M Q Q
Calls for papers/deadlines (new and previous) M Q Q
Announcements/professional activities M Q Q
Personal notices/deaths M Q Q
News and comments M Q Q
Research news/surveys/expository articles M Q Q
Scholarly articles - long and short Q Q Q
Reviews of scholarly articles Q Q -
Book reviews - telegraphic and detailed M Q Q
Books received Q Q -
Current publications M M Q
Abstracts of papers M M Q
Advertisements - jobs, equipment, and books M Q Q
----------------------------------------------------------------------
Table 2
Proposed Publications of AAAI
Notices - monthly
Editorials/messages from president
Letters
Queries
Society news/reports on meetings
Calendar of events (new and previous)
Calls for papers/deadlines (new and previous)
Announcements/professional activities
Personal notices/deaths
News and comments
Advertisements - jobs, equipment, and books
Titles and Abstracts - monthly
Current journal contents
Current conference contents
Abstracts of papers and short notices
Reviews - quarterly
Research news/surveys/expository articles
Reviews of scholarly articles
Book reviews - telegraphic and detailed
Books received
Book advertisements
Journal - quarterly
Scholarly articles - long and short
(Reviews of scholarly articles)
-------
-------
∂22-Apr-81 1328 CLT
I moved all files relating to FR/ARPA to S81,JMC. They are
FR79 -- the 79 proposal text
FR81 -- the 81 proposal text
FR81.PUB -- .pub fr81.pub ;produces the xgp file
ARPA.PUB -- LESs old file, + instructions for updating some data bases
∂23-Apr-81 0752 CERF at USC-ISI Re: verification
Date: 23 Apr 1981 0750-PST
Sender: CERF at USC-ISI
Subject: Re: verification
From: CERF at USC-ISI
To: JMC at SU-AI
Message-ID: <[USC-ISI]23-Apr-81 07:50:02.CERF>
In-Reply-To: Your message of 22 Apr 1981 2249-PST
John,
happy to continue our discussion - I am tied up the 23rd,
but suggest that a page or two of what you have in mind to do would
be a helpful start. I could not tell from our short discussion
Tuesday whether the work you had in mind would be applicable
in ay direct way to problems current facing me, but I don't mind
exploring with you some more.
Vint
∂23-Apr-81 0927 Konolige at SRI-AI (Kurt Konolige) Re: tomorrow
Date: 23 Apr 1981 0922-PST
From: Konolige at SRI-AI (Kurt Konolige)
Subject: Re: tomorrow
To: JMC at SU-AI
In-Reply-To: Your message of 22-Apr-81 2233-PST
I had assumed that you would be over early next week, and haven't
done much with the catalogs. But I will be around until 3pm. --kk
-------
∂23-Apr-81 1401 RPG
John I mentioned to Engelmore that we could use some money for
Lisp timing, mentioning computer time and a grad student, and
starting this summer. Here's what he said. The inclusion in
the JMC proposal is why I mention it at this point. How do I
figure out summer funding for people and computer time?
-rpg-
∂23-Apr-81 1345 ENGELMORE at USC-ISI Re: Lisp Timing
Date: 23 Apr 1981 1340-PST
Sender: ENGELMORE at USC-ISI
Subject: Re: Lisp Timing
From: ENGELMORE at USC-ISI
To: RPG at SU-AI
Message-ID: <[USC-ISI]23-Apr-81 13:40:26.ENGELMORE>
In-Reply-To: Your message of 23 Apr 1981 1300-PST
Dick, I may be able to help, but I need to know what you have in mind by
"a very small amount of ARPA funding", over what period of time, and what
would get accomplished. If you're thinking of a few thousand dollars for
the student during the summer, I can probably handle that (but I need to
fast, because we're rapidly spending all our FY81 money, and it will be
essentially all out the door by May 1). If you're thinking of longer term
and higher level support, then that should be part of the proposal that
JMC is currently preparing to submit to ARPA.
Bob
∂23-Apr-81 2218 CLT LTE
To: RPG, JMC
The LISP Timing Evaluation Project supplement has been
shipped. See TIM81.*[s81,jmc] or .xsp TIM81.XGP
∂23-Apr-81 2237 Daul at OFFICE (Response to message)
Date: 23 Apr 1981 2227-PST
From: Daul at OFFICE
Subject: (Response to message)
To: JMC at SU-AI
cc: DAUL
In response to your message sent 23 Apr 1981 2136-PST
Re: Women And ARPAnet
Good point about looking at the directory. That is the HARD way. I just
thought I might take the shortcut and see if anyone had statistics already.
Thanks for your response, --Bill
-------
Inspecting a few random pages gave 19 out of 99. The formula
sigma = sqrt(n*p*q) suggests a standard deviation of 4. Several
of the women were known to me as secretaries. Of course, the
directory is heavily loaded towards officials and their administrative
helpers.
∂24-Apr-81 0001 LLW Summer Appointment Offer
To: JMC
CC: LLW, CDW
Dear John:
The Lab Express Mail-ed an offer of Summer Appointment to you yesterday.
I presume that you received it (at your home address) this past day. If
you have any questions or problems regarding it, please give me a ring
before you leave the country--it would be good to get the processing going
on it as soon as possible.
Thanks,,
Lowell
I received the offer today. My main problem is with the dates.
I won't be back from France until early in July, and I'll need to spend
some time at Stanford for a while - probably I should be there the month
of July. My best time would be September. One alternative is to put
this on the form. The other is to formally accept for the latest date,
but then put it off, which I would do only to keep the process going.
I have a question concerning pay, which has been proposed at precisely
my academic rate. Is how LL generally pays, or is it characteristic of
this particular summer program? It isn't an important concern for this
summer.
∂24-Apr-81 0041 LLW Summer Arrangements
To: JMC
CC: LLW
∂24-Apr-81 0028 JMC
I received the offer today. My main problem is with the dates.
I won't be back from France until early in July, and I'll need to spend
some time at Stanford for a while - probably I should be there the month
of July. My best time would be September. One alternative is to put
this on the form. The other is to formally accept for the latest date,
but then put it off, which I would do only to keep the process going.
I have a question concerning pay, which has been proposed at precisely
my academic rate. Is how LL generally pays, or is it characteristic of
this particular summer program? It isn't an important concern for this
summer.
[You could always accept for the latest date, and then have presently
unforeseen problems come up later which will delay your start date. I would
informally recommend this approach.
On Summer pay for faculty, the Lab always pays at exactly the academic rate
(which is a political necessity, since it is required to do this by
University policy for UC faculty, and it is politically infeasible to pay
faculty from other universities in a more marketplace-y fashion, as doing so
would discriminate sharply against UC faculty). To compensate for this,
there are reasonably generous relocation and/or per diem allowances for
Summer faculty (or at least there were the last time I looked).]
OK, I've signed it, but we'll see how far $12 per diem goes.
∂24-Apr-81 0133 LLW Divine Intervention
To: JMC
CC: LLW
Dear John:
Before you get away until well into the Summer, I would appreciate your
briefly looking into the case of a CSD doctoral student, Rod Brooks, who
has written a rather impressive, souped-up scene analysis dissertation for
Tom Binford. Brooks was one of the three graduate students (Steele and
Gabriel being the other two) who did the first part of the NIL/LISP for
the S-1 two Summers ago, and Steele and Gabriel are very keenly hoping
(along with myself) that he will be able to join them this Summer to
finish the job at CMU.
The problem which apparently needs a bit of divine intervention (of at
least a showing of High Level Interest) is that, while Brooks' thesis has
been completed for a while now, Binford has apparently made it
unmistakeably clear that he will keep asking for changes-and-extensions
until Brooks completes other, peripherally related work for him, which
Binford expects will keep Brooks `productively occupied' through the end
of this Summer. Having such additional work effective extorted from him
in return for Binford's signature on his thesis will of course keep Brooks
from working with Gabriel and Steele, which will impair the likelihood of
getting a good LISP system on the Mark IIA into the foreseeable future,
which I will keenly regret and which I expect would not please you
greatly, either. (Brooks has been given the `utilities, garbage
collector, I/O, etc.' portion of the Summer's effort, which will be
difficult to get someone competent and knowledgable to undertake
subsequently.)
Brooks, being young and naive, is slaving furiously to complete Binford's
(present) list of work-which-must-be-done-prior-to-thesis-signing, in the
hope that he will be released from bondage upon its completion and
can then scamper off to CMU for the Summer's work. I'm not nearly as
optimistic as he is that the list won't simply be extended very
substantially, as many times as necessary to keep him grinding away thru
Summer's end.
Could you possibly inquire into this matter before you leave the country?
I'm not necessarily asking that you pick up Brooks' thesis, read it and
sign it, but the hint that you might be thinking of doing so, before or
soon after your departure, might induce a good bit more reasonable
(and proper) behavior on Binford's part.
Thanks,
Lowell
∂24-Apr-81 0136 LLW Outrageous Terms
To: JMC
CC: LLW
∂24-Apr-81 0108 JMC
OK, I've signed it, but we'll see how far $12 per diem goes.
[***$12***???? I'll inquire immediately as to what they think they're
doing. They give Lab employees $20/day, in addition to lodging, when
on travel!]
∂24-Apr-81 1138 FFL
To: JMC, FFL
Doris Hyde of Information International asks that you call.
213 390 8611.
∂24-Apr-81 1211 ROD my dissertation
Somebody mentioned to me yesterday that you are about to leave for
France for a couple of months. I've been hoping to get my dissertation
into signature shape by the June 12 deadline. I haven't yet revised it
from the version I gave you before the oral exam as I been doing some
other work for Tom. However I plan on getting back to it next week
(writing chapters 2 and 10, and rewriting 7 and 8). Do you have any
criticisms of the draft you have, or changes that you'd like to see
included? I can mail a later version to you in about a month. The
signature page may cause some logistic problems --- I believe a common
solution is to leave signed pages with Carolyn Tajnai to be released
only when she gets word from the remote reader. If this is agreeable
to you I will make up the signature pages.
It's all true, so do as you propose.
∂24-Apr-81 1307 LGC NSF Discussion
Would 5pm today be a good time for us to discuss some basic facts, goals,
advice, etc. concerning my NSF grant? (The computer is going down for
a purge at that time; also, our discussion will require of you little
if any knowledge of my program or previous knowledge.) If 5pm is not a good
time, perhaps you can suggest a better one. -- Lew
∂25-Apr-81 0858 Darden@SUMEX-AIM (Response to message)
Date: 24 Apr 1981 0451-PST
From: Darden@SUMEX-AIM
Subject: (Response to message)
To: JMC@SU-AI
cc: darden@SUMEX-AIM
In response to your message sent 23 Apr 1981 2158-PST
I guess we all do what we are able to do, but sometimes it isn't so
easy to know whether something is possible or not.
I had lunch with Herb Simon yesterday (this has been my AI people
week) and he is thinking of applying Bacon to genetics. He and I
both agree that a way of finding out how to do scientific reasoning
is to look at the historyof science and try to figure out the
reasoning patterns. But he hasn't yet learned the importance of
looking at the incorrect ideas as well as the correct ones, nor
gotten to your insight that considering what was possible, but
not done at the time, would be very useful (plus fun). In other
words, he is just considering the successful path through the tree,
not all the other actual and possible branches.
Enjoyed talking to you. Keep in touch. Bye, Lindley
-------
∂25-Apr-81 1112 Susan L. Gerhart <GERHART at USC-ISIF> M. Davis paper
Date: 25 Apr 1981 1105-PST
From: Susan L. Gerhart <GERHART at USC-ISIF>
Subject: M. Davis paper
To: mccarthy at SU-SCORE
cc: gerhart at USC-ISIF
I'd like to follow up on the paper you mentioned at the Verkshop.
If you just have the title our library can chase it down but a copy
of the paper would be appreciated, if it's available.
I have a student down here from Berkeley who's interested in the structure
of proofs for a possible thesis area.
Thanks.
Susan Gerhart
-------
She wants the paper about obvious deductions. I think you have a copy.
∂26-Apr-81 1719 ROD signature pages
I have left copies of the signature page in you MJH mailbox.
I've told Carolyn Tajnai to expect them.
I will mail you a revised dissertation in four weeks or less.
∂27-Apr-81 0606 Darden@SUMEX-AIM Re: 8th day
Date: 27 Apr 1981 0606-PDT
From: Darden@SUMEX-AIM
Subject: Re: 8th day
To: JMC@SU-AI
In response to your message sent 26 Apr 1981 1926-PDT
I doubt that an in vivo attempt at introducing systhesized RNA would
have worked; probably they couldn't have gotten it in and had the
cell actually use it. But Lederberg would be the person to ask, since
bacterial mating is his area.
"See you after we return"???
-------
leaving for Marseille Thursday and returning around July 1
∂27-Apr-81 0908 FFL
To: JMC, FFL
Prof. Donald Morrison of the U. of New Mexico and former visiting scholar
here called. He has learned that you will be speaking at Sandia on
Sept. 9, and wants to know if you can go to thhe U. of NM and give a talk
which will be carried to three Universities by TV. Would like you to call
to speak with him about it. Any time on Mondays, best Tues and Thurs after
2 p.m.. 507 277 3244.
∂27-Apr-81 0900 JMC*
hersch about tickets and Concorde
seraphima letter
susie about zella, mail, tuesday, borrowing stuff, bills
feigenbaum or ullman about shapiro
cerf about support for verification
bank, money
bills, payem
zella about cleanup, possibility of working for Susie, getting supplies, mail
police about empty house
hifi, tv rug, military register, xerox→ jerry. What else to hide or lend?
Jerry about xerox
cerf
∂27-Apr-81 0900 JMC*
fredkin about iii and briabrin
∂27-Apr-81 0931 FFL
To: JMC, FFL
Please call Richard Schroeppel at III. 213 390 86ll
∂27-Apr-81 1303 CLT
∂27-Apr-81 1208 JMC
I found it and called it in. It's B1225574.
good, thanks
∂27-Apr-81 1302 CLT
∂27-Apr-81 1204 JMC
Franklin Hersch 329-0950 needs your passport number by 4:30pm.
my passport is on the table in my room, I won't be home before
4:30
∂27-Apr-81 1501 FFL
To: JMC, FFL
Please call Prof. Suppes at 7-3lll.
∂27-Apr-81 1502 FFL
To: JMC, FFL
please call Ed Fredkin at 412 682 4444, if within the hour.
∂27-Apr-81 1503 FFL
To: JMC, FFL
please call Wes Clark at 617 864 5229 at your convenience.
∂27-Apr-81 1549 JAK Departure
Your secretary told me that you age going to be leaving shortly. I need to speak
with you about a couple of incompletes. Will you be around some time tomorrow?
thanks-
John
∂27-Apr-81 2353 DCL Amalgamated ARPA
To: MAS, JMC
John,
HHow goes the proposal?
Today, Tuesday, we are going to start putting the
hardcopies together and adding the necessary glue
(cover sheet, budget summary etc.)
Does Marilyn have your hardcopy now?
-David
I think she has had it for some days. At least Fran isn't asking me
for anything now.
∂28-Apr-81 0426 Darden@SUMEX-AIM (Response to message)
Date: 28 Apr 1981 0410-PDT
From: Darden@SUMEX-AIM
Subject: (Response to message)
To: JMC@SU-AI
In response to your message sent 27 Apr 1981 0902-PDT
Bon voyage.
-------
∂28-Apr-81 1117 FFL
To: "@COMPRE.[1,FFL]" at SU-AI
Comprehensives Committee:
Gene's secretary has copies of the revised algorithms and
data structures section.
Tom Spencer
∂28-Apr-81 1202 DGCOM at USC-ISIC FOL
Date: 28 Apr 1981 1202-PDT
From: DGCOM at USC-ISIC
Subject: FOL
To: jmc at SU-AI
cc: dgcom
Prof. McCarthy,
We discussed in Washington the possibility of my examining FOL.
If you could send me a manual, open an account, etc. I would
appreciate it. More generally, I am interested in the logical
and epistemological issues in AI. I've spent the last ten years
studying philosophy. I am now preparing to reenter the world of
research and am looking for fruitful things to work on. I've spent
the last eight months working in program verification in order to get
aquainted with computer culture. I'd be interested in any reprints,
advice, etc. that you could pass on.
Richard Platek
-------
I will have information about both Weyhrauch's FOL and Ketonen's EKL
sent, but please send a U.S. Mail address to my secretary Fran Larson,
FFL@SU-AI.
Please send Platek when he supplies an address copies of my AI and
philosophy papers.
∂28-Apr-81 1300 RPG Summer Student
Should I talk to Feigenbaum about this summer or should I just bag it?
It doesn't help us out financially much to get his money, but it
might be helpful to have Stanford not look so fragmented. A student
would sure help me out with this thing, too. Feigenbaum and I aren't
exactly pals: do you forsee problems with me negotiating with him?
-rpg-
∂28-Apr-81 1255 RPG
∂28-Apr-81 1253 ENGELMORE at USC-ISI Re: Lisp Timing
Date: 28 Apr 1981 1249-PDT
Sender: ENGELMORE at USC-ISI
Subject: Re: Lisp Timing
From: ENGELMORE at USC-ISI
To: RPG at SU-AI
Message-ID: <[USC-ISI]28-Apr-81 12:49:10.ENGELMORE>
In-Reply-To: Your message of 28 Apr 1981 1133-PDT
The stark reality is that there is just no way I can push through a funding
action for FY81, no matter how small it is. All our FY81 funds will be gone
as of the end of this week. Can you find another source of support within
the Stanford CSD? Maybe Ed Feigenbaum can help.
I don't foresee any problems with FY82 support for the half-time student.
Bob
I'll talk to Ed this week; I already had some student interest for this
summer based on rumor vs advertisement. Thanks for looking into it, though.
-rpg-
∂28-Apr-81 1523 KEETON at USC-ISI translation algorithms
Date: 28 Apr 1981 1523-PDT
Sender: KEETON at USC-ISI
Subject: translation algorithms
From: KEETON at USC-ISI
To: mccarthy at SU-AI
Cc: keeton, woodward
Message-ID: <[USC-ISI]28-Apr-81 15:23:47.KEETON>
Dr. McCarthy,
As I mentioned at the Verkshop, I am interested in algorithms for
translating between high level notations and simple logical
languages. Consequently, I would very much appreciate receiving
a description of your translation algorithm for getting from
programming notation to First Order Logic.
Thanks very much, Jim Keeton-Williams
∂28-Apr-81 1629 TW AI question
To: "@COMPRE.[1,FFL]" at SU-AI
Gene's secretary now has new versions of the long AI question
and a solution for it.
∂28-Apr-81 2259 RWW nsf grant
Please don't forget or neglect to sign the papers in frans office necessary
the contract renewal before you go. Thanks. Richard
∂28-Apr-81 2346 JK ramsey's theorem
To: JMC, JJW
There is a slightly cleaned up proof of ramsey's theorem in ramsey.txt.
∂29-Apr-81 1100 CSD.ULLMAN at SU-SCORE
Date: 29 Apr 1981 1100-PDT
From: CSD.ULLMAN at SU-SCORE
To: CSD-Faculty: ;
I have a letter from Jean Fetter offering to contribute some funds for
affirmative action activities to any faculty member willing to visit one
of a number of schools she has targeted as having reasonable numbers of
minority students that might be eligible for our programs. The list is
not unreasonable, e.g., it includes MIT, and she will pay up to $500
expenses for such a visit. Please see me if you are interested. If you
are taking a trip, you might want to check the list for a nearby school, even
if you are not ready to make a trip especially for this purpose.
-------
∂29-Apr-81 1134 FFL
To: JMC, FFL
Prof. Kreisel is very anxious to have you join him and Prof. Engeler of
zurich for lunch at the Faculty Club at l p.m. today. Pls. call him
at 327 5925.
∂29-Apr-81 1136 FFL on TTY62 (at TV-141) 1136
Please read your message from Prof. Kreisel. Fran
∂30-Apr-81 1608 CLT
Colmerauer, Prof.A.,
Groupe d'Intelligence Artificielle
Unite d'Enseignement et de Recherche
70, Route Leon-Lachamp
13 - Marseille (9 e)
FRANCE
(office) 01133 + 91 41 32 48,
(home) 01133 + 42 70 04 41,
(dept #?) 01133 + 91 41 01 40
∂01-May-81 1509 CERF at USC-ISI Re: verification
Date: 1 May 1981 1508-PDT
Sender: CERF at USC-ISI
Subject: Re: verification
From: CERF at USC-ISI
To: JMC at SU-AI
Message-ID: <[USC-ISI] 1-May-81 15:08:17.CERF>
In-Reply-To: Your message of 30 Apr 1981 1634-PDT
John,
glad to talk further with you on the verification stuff -
I am still unclear on how the work you described on Elephant
would be applied in systems DoD might have to work with -
demonstrations of capability which are of interest to DoD
in this area are often hard to come by - got any ideas?
Vint
∂01-May-81 1635 Bill.Scherlis at CMU-10A Dana Scott
Date: 1 May 1981 1925-EDT (Friday)
From: Bill.Scherlis at CMU-10A
To: :INCLUDE: GOSSIP.DST[C397BS71] at CMU-10A
Subject: Dana Scott
Message-Id: <01May81 192546 BS71@CMU-10A>
Dana Scott has accepted a University chair at Carnegie-Mellon. He will
be here starting in Fall, 1981. Regards,
Bill
∂01-May-81 1905 Bill Gosper <rwg at MIT-MC>
Date: 1 May 1981 22:03-EDT
From: Bill Gosper <rwg at MIT-MC>
To: rpg at SU-AI
cc: jmc at SU-AI
Date: 1 May 1981 19:58-EDT
From: Kent M. Pitman <KMP at MIT-MC>
Subject: [REM: forwarded follow-up]
To: ELLEN at MIT-MC, RWK at MIT-MC, RWG at MIT-MC
Date: 1 May 1981 19:38-EDT
From: Robert Elton Maas <REM>
To: KMP
You rotten fink, you called JMC and had him send the pigs after me.
Well, I got them off my back, and now can complete my plans. I'll give
Betsy the weekend to reconsider.
From here, I can derive amusement. I hope you are similarly compensated. Should
we send REM an ultimatum?
∂02-May-81 2227 Steve Kudlak <FFM at MIT-MC> REM
Date: 3 May 1981 01:26-EDT
From: Steve Kudlak <FFM at MIT-MC>
Subject: REM
To: KMP at MIT-MC, CSTACY at MIT-MC, RWK at MIT-MC, VP at MIT-MC,
MJA at MIT-MC, RWG at MIT-MC, GZ at MIT-MC, SJK at MIT-MC,
rww at SU-AI, jmc at SU-AI, cs.webb at UTEXAS-20
cc: FFM at MIT-MC
I have spoken with REM and he is distressed but ok...
It would be useful if officers of the law did not appear
at his door. These people apparently are not very adept
in human relationships and only make the matter worse by
causing more fear and ridiculous behavior than they solve.
Telling someone "Why don't you just kill yourself quietly,
instead of making a big fuss about it" is counterproductive
and even when there is some provocation for such a remark.
I think Robert has finally seen that some relationships due
to no fault of the people involved are just not good for him
or the other people either. I am trying to arrange for Robert
to form other relationships, even if they are thru therapeutic
channels.
If necessary I can arrange professional help, however, such
relationships if entered into are best entered into voluntarily
for maximum benefit to occur.
We are all human and sometimes do some strange and moderately ugly
things, however hopefully we see thru them and become better
people after such traumas.
If need be...I can be called at (415)948-5753
Have fun
Sends Steve
P.S. Note this is not phrased well and I wish I could have
said it better....
Have fun
Sends Steve
∂04-May-81 1051 FFL Speaking at U. of N.M. when you are at Sandia Lab
To: JMC, FFL
Two offices have called you about this. Dr. Carney and Prof. Morrison.
The latter says you may recall that he was a visiting scholar here at one
time. I gave Dr. Carney's secretary the message that you did not wish to
make a decision about this until you get back on July lst and suggested
that Dr. Carney and Prof. Morrison get together on that matter.
Dr. Carney 505 277 2600.
∂06-May-81 1403 CSD.ULLMAN at SU-SCORE IBM grant
Date: 6 May 1981 1401-PDT
From: CSD.ULLMAN at SU-SCORE
Subject: IBM grant
To: csl.lantz at SU-SCORE, csd.golub at SU-SCORE, jmc at SU-AI
I just got my notice that I have to report on what we did with the money.
Can you guys give me a paragraph each on what you have done (or will do)
with your shares? I need these by a week from Friday.
-------
∂07-May-81 1308 Vaughan Pratt <CSD.PRATT at SU-SCORE> Etude documentation
Date: 7 May 1981 1312-PDT
From: Vaughan Pratt <CSD.PRATT at SU-SCORE>
Subject: Etude documentation
To: @sun at SU-AI
Five Etude reports are available for borrowing from room 428 (usual place
on bookshelf on left, waist high). Includes current annual report, two
papers for next month's Text Processing conference, an overview, and
a copy of the Oct. 80 Seybold Report describing XDS, Etude, etc.
Etude is Michael Hammer's word processor. The CLU implementation completed
at the end of 1979 is now retired, as is CLU, the latter being ultimately
to be replaced by Extended CLU. The next implementation of Etude will be
in MDL (an extension of Lisp running on MIT-DM), and is intended for
"abstract workstations" in the sense that it should run on such
workstations as the Apollo Domain, the Nu terminal, the Sun terminal, etc.
-------
∂08-May-81 0647 Navarro at SRI-AI your badge
Date: 8 May 1981 0648-PDT
From: Navarro at SRI-AI
Subject: your badge
To: jmc at SU-AI
cc: navarro
I have your security badge to get into SRI. When you come back, please
see me. Thanks.
... Georgia
-------
∂12-May-81 1319 Moore at SRI-CSL (J Moore) University of Texas
Date: 12 May 1981 1317-PDT
From: Moore at SRI-CSL (J Moore)
Subject: University of Texas
To: TAYLOR at PARC, JMC at SU-AI
Bob and I have reconsidered our earlier rejection of
the Texas offer and have decided to take them up on it.
They have managed to regenerate the offer (at our request)
and we have accepted. We will disappear from the Bay Area
sometime this summer and officially start at UT in
September.
J
-------
∂14-May-81 1226 CSD.ULLMAN at SU-SCORE Leo Guibas
Date: 14 May 1981 1223-PDT
From: CSD.ULLMAN at SU-SCORE
Subject: Leo Guibas
To: or.dantzig at SU-SCORE, csd.feigenbaum at SU-SCORE, rwf at SU-AI,
csd.golub at SU-SCORE, csd.herriot at SU-SCORE, dek at SU-AI, zm at SU-AI,
jmc at SU-AI, csl.crc.ejm at SU-SCORE, miller at SRI-KL,
baskett at PARC-MAXC, csd.oliger at SU-SCORE, csd.winograd at SU-SCORE,
csd.buchanan at SU-SCORE
I propose to appoint Leo Guibas of PARC a consulting associateprofessor.
The appointment is without salary unless he teaches courses for us, as
he has been doing, in which case we would offer Xerox 25% salary for his
services. As Leo is known to us, and has a distinguished record
(Betty has his resume), I do not expect much objection to the appointment
in principle, although you are welcome to raise them, of course. What
concerns me is the rank "associate." Leo has never held an academic
position, so we must decide the correct rank. I is my contention that
his record is sufficiently good that he would be a serious candidate
for an associate professorship here if he were available.
While the consulting professorship does not necessarily imply that, I don't
want to bestow it capriciously. Therefore, I shall wait a week and
proceed only if I receive no "no" votes and a reasonable number of "yes"
votes.
-------
∂14-May-81 1318 Randall Davis <KRD at MIT-AI>
Date: 14 May 1981 15:43-EDT
From: Randall Davis <KRD at MIT-AI>
To: konolige at SRI-AI
cc: KRD at MIT-AI, nilsson at SRI-AI, jmc at SU-AI
Kurt:
Read over your TN232 recently and liked it very much (tho I don't
claim to be much of a logic hacker at all). It appears to put down
a very nice foundation for the reasoning about beliefs and actions stuff.
A couple of minor typos that you've probably heard about:
-- on page 27 you say PO1 several times and I suspect you mean LT1.
-- page 35, middle paragraph, after the ref., you seem to have been
trying to make up your mind about how many examples had been done.
A comment about the wise men problem: it seems to involve not only
reasoning about belief, but about capabilities. The difference in
speed of reasoning of the agents is both necessary and closely bounded.
To see this, imagine that the differences in intelligence are in fact
very large. Then consider #1's standard reasoning:
"if mine were black, then #2 would see a black and a
white, and he could say
'if mine were black, then #3 would see two
blacks and he would have responded by now'."
BUT: if #1 is VERY dumb, then #2 can't make that inference. And (much more
plausibly), if #3 is significantly faster than #2, then he can't say "since #2
hasn't responded, mine can't be black, or #2 would have solved his (simpler)
problem already."
The problem itself is made plausible by the fact that each level of
hypothesization reduces the problem complexity (eg, the nth guy is left with
a trivial observation and inference), but the time issue is a tricky one.
In effect, there are hidden assumptions that each agent believes that each
other agent is only a little bit dumber, and hence given the hypothesized
simpler problem to solve, would have solved that simpler problem already.
You get around this in your formalization because you poll the men. But in
principle at least, the problem is insoluble without assumptions about speed of
reasoning. The first guy could be SO dumb that even when seeing 2 black spots
it takes him arbitrarily long to make the obvious inference [maybe he's running
a resolution theorem prover]; my claim gets more plausible as we move up each
level of embedding [the second guy's problem is a little harder, the third more
difficult still, and so forth if there were more people involved.]
Agree, or have I overlooked something?
cheers
Randy
∂16-May-81 1745 CSD.GENESERETH at SU-SCORE AI Qual
Date: 16 May 1981 1743-PDT
From: CSD.GENESERETH at SU-SCORE
Subject: AI Qual
To: csd.buchanan at SU-SCORE, csd.feigenbaum at SU-SCORE,
csd.lenat at SU-SCORE, JMC at SU-AI, TW at SU-AI
As de facto chairman of the qual committee this year, I am considering
slightly revamping the procedure, and I would like your comments.
Heretofore, students were asked write critiques of papers before the
exam. Presumably, the discussion of these critiques helped to break
the ice in the exams, gave the students a sense of mastery over some
small area of research, and gave the examiners some material by which
to judge the critical faculty and understanding of the student. In
practice last year, the examiners hardly looked at the critiques, and
many of the students disliked the requirement. I would like to
propose as a substitute that the students be required to submit a 1-2
page summary of their special are of interest, outlining any
significant work they have already completed and any concrete plans
for thesis research. In practice last year, this is the subject that
all of my committees started with. If there are no objections, I will
publicize this policy next week.
It appears that the best dates for the exam are Thursday and
Friday in the second week of June. Can you all make it then?
mrg
-------
∂16-May-81 2149 CSD.LENAT at SU-SCORE Re: AI Qual
Date: 16 May 1981 2145-PDT
From: CSD.LENAT at SU-SCORE
Subject: Re: AI Qual
To: CSD.GENESERETH at SU-SCORE, csd.buchanan at SU-SCORE,
csd.feigenbaum at SU-SCORE, JMC at SU-AI, TW at SU-AI
In-Reply-To: Your message of 16-May-81 1743-PDT
I can certainly make it any of those days.
The assignment of articles was intentionally done OUTSIDE the
area of specialization of the student, and in general the students
are more than qualified in their subarea. What you are suggesting,
Mike, is a good supplementary procedure: in addition to an
"ice-breaking" article outside their field, each student is
asked to prepare a brief (1-2 paragraph) summary of their
past, current, and proposed future (if known) research in AI,
including the names of faculty advisers past and present.
One of the main things we want to know, in a borderline case,
is: Is there evidence that this student can carry out
productive research, and is anyone on our faculty willing
to serve as a thesis adviser to them?
Perhaps the trouble last year was that the articles and the
students' reviews were not available to he committee members
until a couple days ahead of exam time; let's try to give
more lead time this year.
Doug
-------
∂17-May-81 1918 ENGELMORE at USC-ISI Visit
Date: 17 May 1981 1915-PDT
Sender: ENGELMORE at USC-ISI
Subject: Visit
From: ENGELMORE at USC-ISI
To: JMC at SU-AI
Message-ID: <[USC-ISI]17-May-81 19:15:41.ENGELMORE>
John,
I'm bringing Ron Ohlander to Stanford on June 2nd for an introduction
to the ARPA-sponsored AI work (Ron will be taking over as AI program manager
when I leave at the end of June). Will you be around to say hello?
Bob
I'm in France till July 1.
∂18-May-81 1019 TW
To: JMC at SU-AI, TW at SU-AI, buchanan at SU-SCORE,
feigenbaum at SU-SCORE, lenat at SU-SCORE,
genesereth at SU-SCORE
My feeling is that in the abstract, there is good reason not to make
the change you suggest. The hurdles are designed to show (in order)
breadth in all of CS (the comp), breadth in all of AI (the Quql) and
depth in one topic (the dissertation). By giving an article outside
the area of specialization, we could probe for breadth. In practice,
this hasn't been followed very well, as you observed. I don't object
to making the change this year, but we might do some thinking about the
breadth issue in general. --t
∂18-May-81 1641 CSD.GENESERETH at SU-SCORE
Date: 18 May 1981 1639-PDT
From: CSD.GENESERETH at SU-SCORE
To: TW at SU-AI, JMC at SU-AI, buchanan at SU-SCORE, feigenbaum at SU-SCORE,
lenat at SU-SCORE, genesereth at SU-SCORE
In-Reply-To: Your message of 18-May-81 1019-PDT
The majority of you seem to feel that the critique should be
retained; so I won't change the policy. However, I'll make
an effort to get the papers and critiques distributed early enough
to be of use.
mrg
-------
∂19-May-81 1155 Bobrow at PARC-MAXC A new feature for the AI journal
Date: 19 May 1981 11:41 PDT
From: Bobrow at PARC-MAXC
Subject: A new feature for the AI journal
To: "@[ivy]<BOBROW>ai-board.dl"
cc: Dake
Reply-To: Bobrow
Dear AIJ board member:
The following is a proposal for a new feature for the AI journal. I solicit your
comments and suggestions, especially on people to fill the roles described.
The AIJ has been the journal of record for the field of Artificial Intelligence.
However, the field is growing, and there are a number of areas which overlap
with artificial intelligence that have journals containing papers of interest to the
readers of the AIJ. In order to better inform our readers of what is going on, I
believe it would be useful to have area "Correspondents". A Correspondent reads
the journals in an area, and selects articles which would be of interest to our
readers. Every six months or so, we would publish a column, listing references,
with perhaps a sentence or two describing the content of a paper.
I have identified a number of areas I believe would be of interest, and in three
cases found high quality people who have agreed to take on the job of
correspondent. The areas are listed below. Please comment on whether these
areas are appropriate in your opinion, and if so, whom you would like to see fill
the role of correspondent. They should be someone normally reads the right set
of journals, whose taste you trust, who knows what would be of interest to the
AI reader, and might be willing to take on the job. Other areas which have
large overlap with AI and have journals which publish articles of interest to the
AI reader should also be suggested.
AREAS CORRESPONDENT
Vision Mike Brady
Robotics Tomas Lozano Perez
Natural Language Barbara Grosz
Cognitive Science
Logic and Philosophy
Computer Science
(should this be divided up?
e.g. programming languages, hardware, data bases?)
Here are some related questions to consider.
Should papers other than those in published journals be included in our
selected bibliographies? For example, conference proceedings and/or technical
reports.
Should people be able to send papers to the correspondent for their
attention? I think the rule must be that this is a selected bibliography, selected
by the correspondent on the basis of his/her judgement on the relevance and
interest of this article for an AI audience. The correspondent has complete
discretion to be discriminatory (use discrimination). I want to make this job not
too burdensome for the correspondents, and at the same time get good coverage.
Should full abstracts be included, just the sentence that I suggested or
only the bibliographic reference?
In addition to comments on the above proposal, I solicit names for new members
to be added to the AI board, and suggestions for agenda items at our next
meeting. That meeting will take place in Vancouver in conjunction with IJCAI,
though a time and place has not yet been set. Please let me know if you will be
at IJCAI.
Thank you,
danny bobrow
∂19-May-81 1359 Feldman@SUMEX-AIM Job Application
Date: 19 May 1981 1354-PDT
From: Feldman@SUMEX-AIM
Subject: Job Application
To: jmc@SU-AI
cc: feldman@SUMEX-AIM
Dear Professor McCarthy:
I am sending a copy of my just-completedPh.D.
thesis. Please consider this part of my application material
for a faculty position in your Department.
-------
Please make sure your application is primarily considered by Jeff Ullman
and Ed Feigenbaum. In any case, I'll be gone till July.
∂19-May-81 1403 Feldman@SUMEX-AIM Job Application
Date: 19 May 1981 1400-PDT
From: Feldman@SUMEX-AIM
Subject: Job Application
To: jmc@SU-AI
cc: feldman@SUMEX-AIM
Dear Professor McCarthy:
I am sending you a copy of my just-completed Ph.D.
thesis. Please consider this part of my application
material for a faculty position in your Department.
Sincerely yours,
Donald Perlis
-------
There is a letter on my desk at the left from Donald Perlis asking about a job
in the department. Please write him that I welcome a look at his thesis, but
he should make sure that his application is considered by Ullman and Feigenbaum.
∂27-May-81 1917 Richard M. Stallman <RMS at MIT-AI>
Date: 27 May 1981 22:18-EDT
From: Richard M. Stallman <RMS at MIT-AI>
To: jmc at SU-AI
Date: 27 May 1981 0833-PDT
From: Susan Hill <CSD.HILL at SU-SCORE>
Subject: SAIL account
To: rms at SU-AI
We have reason to believe that you are allowing a number of people to use your
account. Please change your password and use the account only for yourself.
Further non-personal use will be cause for cancelling the account.
It looks like I will not be able to continue to have a SAIL account.
Perhaps you can persuade them to stop caring (since they have not
cared for the past six years). If not, then coming to visit for a
month would mean being out of touch with the net, and I would
therefore decide to cancel the visit.
I'm not willing to go along with their terms because it would violate
the golden rule so greatly that I couldn't feel comfortable with
myself. If I can't avoid them, I'll just have to stay away.
∂28-May-81 0801 FFL
To: JMC, FFL
Your three research funds are almost completely used. Please do not
make any further commitments with them without speaking to me about
them.
∂28-May-81 1511 Denny Brown <CSD.DBROWN at SU-SCORE> Office space plans
Date: 28 May 1981 1512-PDT
From: Denny Brown <CSD.DBROWN at SU-SCORE>
Stanford-Phone: (415) 497-2274
Subject: Office space plans
To: CSD-Faculty: ;
cc: tob at SU-AI, reg at SU-AI, csd.bscott at SU-SCORE, csd.tajnai at SU-SCORE
It's the time of year to beginning planning office assignments for
the summer and for next year. In order to do a reasonable job,
I need to have LOTS of data about what your plans are.
Please send hard or electronic messages with your plans for summer
and next year, for yourself and/or for your empire.
Faculty members who expect new employees or visitors or ...
should be sure to contact me.
I'd also like to get a PhD student volunteer to coordinate student
requests and suggestions.
_
-------
∂29-May-81 1607 JK hello John:
I m on campus for the afternoon and thought i,d stick my head in
and see if you'd remember me. sorry to miss you. i see your name
in the papers and all that from time to time. Bob de Zafra,Princeton
50-54.
∂01-Jun-81 0138 Richard M. Stallman <RMS at MIT-AI>
Date: 1 June 1981 04:39-EDT
From: Richard M. Stallman <RMS at MIT-AI>
To: GJS at MIT-AI, jmc at SU-AI, jd at SU-AI
RMS;META > is the beginning of a paper on the metalogic stuff I was
working on. I hope to add more to it soon.
∂02-Jun-81 2105 Richard M. Stallman <RMS at MIT-AI>
Date: 2 June 1981 23:36-EDT
From: Richard M. Stallman <RMS at MIT-AI>
To: jmc at SU-AI
I'm going to be around Stanford for a little while
in two or three weeks. Are you around?
∂02-Jun-81 2126 Richard M. Stallman <RMS at MIT-AI>
Date: 3 June 1981 00:26-EDT
From: Richard M. Stallman <RMS at MIT-AI>
To: RWG at MIT-AI, GEOFF at MIT-AI, MCLURE at MIT-AI, LPD at MIT-AI,
KLH at MIT-AI, MRG at MIT-AI, Feinler at SRI-KL, jmc at SU-AI,
tvr at SU-AI, jd at SU-AI, reg at SU-AI, don at SU-AI,
g.eldre at SU-SCORE, csd.spencer at SU-SCORE,
g.mulligan at SU-SCORE, Goldstein at PARC-MAXC,
dekleer at PARC-MAXC
I'm planning to be around Stanford for a week to two weeks
in a week or two. Can any of you offer me or find me a place
I can stay? Places to stay just for a couple of days would
be very welcome, since I don't have to spend the whole time in
one place. All I need is a place to put a sleeping bag.
∂04-Jun-81 0038 POURNE@MIT-MC
From: POURNE@MIT-MC
Date: 06/04/81 03:38:06
POURNE@MIT-MC 06/04/81 03:38:06
To: JMC at MIT-MC, MINSKY at MIT-MC, TAW at MIT-MC, JEP at MIT-MC
We are printing up the space council report, which I am
pleased to say has had considerable impact in washington.
To flesh out the report I am including some brief signed
papers; criswell on moon mines, for example. Signed papers will
be appendices, with notation that they are not Council documents
and are included for intrinsic interest at the diescretion of
the chairman.
If you have short pieces applicable to the deep space situation
which you can include in the council report printed version,
please send to taw at SAIL with copy to me. The idea is to
apear respectable and sound but visionary. That TOWER IS NOT
WHAT I WANT> Easier I can sell a case of plague than that tower...
∂05-Jun-81 1118 FFL
To: JMC, FFL
Carolyn Tajnai said to tell you that she gave Bossack an "L", so he has
graduated but you will have to decide on a grade when you return.
here's where i got to june 23
∂08-Jun-81 0949 FFL
To: JMC, FFL
Chris Goad is scheduled for a half-time appointment to be paid out of his
NSF grant, for which we have not yet received approval. Betty Scott would
like to know if you intend to carry him an additional half-time starting
in October. If so, I do not know where there is space on your grants.
Perhaps you have plans of which I am not aware.
Chris plans to support himself from other sources for the other half time.
∂08-Jun-81 1504 CSD.TAJNAI at SU-SCORE (Carolyn Tajnai) grade for John Kelley
Date: 8 Jun 1981 1500-PDT
From: CSD.TAJNAI at SU-SCORE (Carolyn Tajnai)
Subject: grade for John Kelley
To: JMC at SU-AI
cc: FFL at SU-AI
John,
Autumn quarter you initially gave John Kelley a B for CS206, but
changed it to an INC so he could improve his grade. He needed a
grade to graduate from Law School, so I reinstated the B. If he
turned in more work and you wish at a later time to improve it,
let me know. Otherwise, the B will stand. Since we are runnning
down to the wire on degree candidates (the Law School has an earlier
deadline than the regular University), I took the responsibility to
do this. I hope it meets with your approval.
Carolyn
-------
∂10-Jun-81 1057 Walker at SRI-AI AAAI office space
Date: 10 Jun 1981 1053-PDT
From: Walker at SRI-AI
Subject: AAAI office space
To: balzer at ISIB, bobrow at PARC, buchanan at SUMEX-AIM,
To: erman at ISIB, feigenbaum at SUMEX-AIM, feldman at SUMEX-AIM,
To: rick at RAND-AI, jmc at SU-AI, minsky at MIT-AI,
To: newell at CMUA, nilsson at SRI-AI, reddy at CMUB,
To: sacerdoti at SRI-AI, schank at ISI, simon at CMUA,
To: tenenbaum at SRI, athompson at USC-ECL, walker at SRI-AI,
To: tw at SU-AI, phw at MIT-AI, robinson at SRI-AI
Ed Feigenbaum and I, with the advice of Marty Tenenbaum and after discussions
with Lou Robinson, have leased an office for the AAAI. Essentially what we
are doing is holding some space in a particularly desirable location in
anticipation of a more extensive discussion of the larger issues at the
Council meeting in August. The facilities, which will be co-located with
those of Lou Robinson's company, Garcia-Robinson, will allow us to handle
the demands of our participation in the IJCAI operation more easily.
Don
-------
∂12-Jun-81 1450 LOUNGO at RUTGERS Rutgers Technical Reports
Date: 11 Jun 1981 1025-EDT
From: LOUNGO at RUTGERS
Subject: Rutgers Technical Reports
To: erman at USC-ISIB
cc: shortliffe at SUMEX-AIM, driefus at WHARTON-10, bewnnett at SU-SCORE,
mittal at RUTGERS, chandrasekaran at RUTGERS, jsmith at RUTGERS,
deolankar at RUTGERS, wilkins at SRI-KL, bruce at BBNA, webber at BBND,
friedland at SUMEX-AIM, plondon at USC-ISIB, erm at MIT-AI, rdg at SU-AI,
pressburger at SCI-ICS, csd.gardner at SU-SCORE, fagan at SUMEX-AIM,
fikes at PARC-MAXC, jmc at SU-AI, clancey at SUMEX-AIM, krd at MIT-AI,
hamilton.es at PARC-MAXC, cs.amsler at UTEXAS-20, chinguyen.es at PARC-MAXC,
lisa at UTEXAS-11, kwh at MIT-AI, utgoff at RUTGERS, turock at RUTGERS,
ecg.rich at DEC-MARLBORO, loungo at RUTGERS
The newest technical reports are available for access via FTP with user
account <anonymous> with any password. The file names are:
<LIBRARY> publ-order-form.doc - This contains
the list of our newest technical reports.
<LIBRARY>online-tecrpts.doc - This contains
an abstract of each of the reports.
If you wish to order copies of any of these reports please send mail
via the ARPANET to DOLESE@RUTGERS OR LOUNGO@RUTGERS. Thank you!!
CL
-------
∂14-Jun-81 1357 LWE ns files
To: JMC
CC: REG
help!!!
my ns files seem to have been purged on may 30?
today is the first time i've logged on in almost
two months, so i didn't notice earlier (had to go
to Germany at end of April due to family illness
and have been ill since returning), also had
not been on in April due to apparent bad modem at
my end (what a sad story--sorry).
Anyhow: is there a way to recover, or are the files
of May 30 lost for good??? Please do tell...
Thanks as ever--and I apologize for all the trouble.
∂17-Jun-81 1628 CSD.ULLMAN at SU-SCORE Adopt an Orphan this Summer
Date: 17 Jun 1981 1625-PDT
From: CSD.ULLMAN at SU-SCORE
Subject: Adopt an Orphan this Summer
To: CSD-Faculty: ;
The CSL is undergoing renovation, and a number of our colleagues
have been kicked out of their offices. If you will be away between
now and about Aug. 1, for any period of time, would you offer the
use of your space to a needy CSL orphan? They'll thank you
for it, and you'll be glad you did.
-------
∂18-Jun-81 0047 Bill Gosper <RWG at MIT-MC> crinkle
Date: 18 June 1981 03:39-EDT
From: Bill Gosper <RWG at MIT-MC>
Subject: crinkle
To: geoffm at RAND-AI
cc: ES at MIT-MC, jmc at SU-AI
i think jmc's fcn was sum(g(4↑n x)/2↑n,n>=0), where g(x) is the continuous
triangle wave which is 0 on even integers and 1 on odds, i.e.
g(x) = abs(x - 2 floor((x+1)/2)). Then, for all k and x, there is a pair
of dyadic rationals within 2↑(1-2k) of x on which jmc(x) differs by
2↑(-k) + O(4↑-k).
∂18-Jun-81 2107 BYY terminal
John, John Perry is interested in using this for a while to learn what it is
all about. I thought I would pass this terminal on to him, and leave it set
up on my logic grant, so that I could communicate with him on our work. Is that
alright? You could change the login to something more approprriate for him.
Say login JP with password Frnch. Jon
∂19-Jun-81 1349 RFN
To: "@FULLPR.JU[1,RFN]" at SU-AI
TO: CSD Full Professors
FROM: Jeff Ullman
Full professors met June 16 and voted to propose an Adjunct Professorship
jointly with the dept. of medicine for Gio Wiederhold. Comments to
Jeff Ullman are solicited.
∂22-Jun-81 2302 OTA Renewing the space program
To: JMC at SU-AI
CC: OTA at SU-AI, TAW at SU-AI, pourne at MIT-MC
I am editing your paper "Renewing the Space Program" for Jerry Pournelle
for inclusion as a signed paper in the Citizens Advisory Council on Space
Policy. I presume he has talked to you about this. Anyway the file I
started from is SPACE.TEX[W81,JMC]. The paragraph that talks about
"Private Enterprise" has most of a sentence missing: "...opportunity. We
should The orbital hotel...". Could you try and fix this up?
Also there are several scribled in comments on the hard copy that Jerry sent
me. I have know idea whose comments they are, what do you suggest I do about
them?
Thanks,
Ted Anderson
If you are still editing my paper, I'll be glad to look at what you have.
∂23-Jun-81 0012 Jerry E. Pournelle <POURNE at MIT-MC> nets
Date: 23 June 1981 03:12-EDT
From: Jerry E. Pournelle <POURNE at MIT-MC>
Subject: nets
To: JMC at SU-AI
Some time ago you sent me some general information on your
information utility project. I'd like to write about it for
BYTE magazine; also, I'm getting some actual interest from
publishers who'd like to receive books mss. (well not really
mss) over the phone for their computer typesetters.
Is your system operational and can I learn much about it?
I'll be in Palo Alto Wednesday, over at Possony's place; we've
planned a dinner to include Colin Mick whom I think you know;
possibly you could join Stefan and me?
Sorry to have missed you in Palo Alto. Just got back from France.
∂24-Jun-81 0317 Jerry E. Pournelle <POURNE at MIT-MC> space and communications
Date: 24 June 1981 06:18-EDT
From: Jerry E. Pournelle <POURNE at MIT-MC>
Subject: space and communications
To: JMC at SU-AI
ROGER. By then the REPORT will be out, but we'll also be having
another meeting of the Council.
Hans Mark working out fine. Senate confirmations soon. All's
well. Have fun. JEP
∂25-Jun-81 1013 Wiederhold at SUMEX-AIM ARPA consolidated contract
Date: 25 Jun 1981 1012-PDT
From: Wiederhold at SUMEX-AIM
Subject: ARPA consolidated contract
To: jmc at SU-AI, tob at SU-AI, dcl at SU-AI, zm at SU-AI,
wiederhold at SUMEX-AIM
cc: mas at SU-AI, mcc at SU-AI, kaplan at SRI-KL, sagalowicz at SRI-KL,
csd.betty at SU-SCORE
John Machado from Navelex will be the project officer for our consolidated
ARPA contract. He wants to get up-to-date on the projects and will be in the
Bay Area ( also in Berkeley and at LLL ) July 27th to Aug. 1.
He wants to set up appointments and asked my to coordinate a schedule.
Could you let me know your unavailability and perhaps preferred times
during that week and the amount of time you need to cover past progress and
future work.
I will try to make up a schedule and get it to John by Monday if possible.
I'll warn Jayne ( 497-0685) in case telephone comm. is needed.
Thanks Gio
-------
∂26-Jun-81 0639 JRA lisp course
john,
i have tentatively scheduled you to talk on the morning of july 10
at santa clara. topic is of your choice. is this ok with you?
July 10 is fine. Is "Proving correctness of LISP programs" a suitable
topic? If so, that's it.
∂27-Jun-81 0510 Robert Elton Maas <REM at MIT-MC> Go endgame algorithm, new breakthrough
Date: 27 June 1981 08:10-EDT
From: Robert Elton Maas <REM at MIT-MC>
Subject: Go endgame algorithm, new breakthrough
To: RWG at MIT-MC, JMC at SU-AI, ALS at SU-AI, WESTCOTT at BBNC
cc: ELLEN at MIT-MC, REM at SU-AI
Tonight I realized a crucial fact that allows automatic pruning of
the search tree so that only 1↑N instead of 2↑N paths have to be analyzed
for behaviour for various values of the threshold T when only partial
values (from T=0 up to some small value of T) are needed at some superior
node relative to the search-tree being expanded. Heuristically it means
that large threats are distinguished from small threats in a rigorous
way so that large threats may be dismissed as "need immediate reply"
while small threats must be fully expanded as functions of T (but such
threats are just the ones that are easy to completely analyze as
functions of T because they become gote-constant above their threshold
which is small in these cases). Now that I've solved this problem
mathematically, all we need is a team of well-paid LISP programmers to
actually work out all the programming details to implement my algorithm.
Then we should have a program that plays better than any human in the
micro-endgame (many small skirmishes, nothing large except threats that
turn up partway through a small skirmish and must be answered because they
are so large compared to everything else on the board).
A few days ago I started handling large ko's which turn up in the midst
of small-skirmish analysis, and found that it's not too bad to just keep
separate functions (of T) for all the different possibilities (end of
game where there are no ko threats or number of threats are equal, black
can force-win the ko by having more threats, white can force-win the ko
by having more threats). This assumes all threats are either gigantic (thus
useful in winning the ko) or tiny (thus useless). If some ko threats are
about the same size as the ko itself, things get messy and I choose not to
handle that case at present. In many cases, large ko's that turn up in
the midst of small-skirmishes have differing values depending on who wins
them but backing up to the superior node the differing values merely mean
differing values for a threat, but all these differing values are big enough
to force an immediate reply, thus their differing values collapse and the
value of the superior node doesn't depend on who has ko threats.
Are all of you acquainted with the original description/writeup of my
threshold method for analyzing endgame skirmishes in Go?
P.s. the crucial fact is trivial, once you see it. Except for very
small values of the threshold in the case of a ko, where the slope of
the graph might be plus or minus 2, the slope ranges between -1 and +1
always. Thus for widely separated starting points (T=0) the graphs
can't possibly cross each other unto T exceeds half the separation
in starting points. Thus the constant-gote value can't take effect
until then, thus to obtain one-sided value you don't have to consider
the constant-gote value, thus you have to explore only the branch you're
interested in, not the other which crosses it where the constant-gote value
starts. Thus if this happens at each node, you search 1 long branch
instead of 2↑N long branches.
∂29-Jun-81 0043 Kenneth Kahn <KEN at MIT-AI> an answer to your questions about Omega
Date: 29 June 1981 03:39-EDT
From: Kenneth Kahn <KEN at MIT-AI>
Subject: an answer to your questions about Omega
To: JMC at MIT-AI
I tried out the questions you asked me about Omega at dinner last week on the
Hi.
Omega-discussion on MIT-AI and got the response:
BEPPE@MIT-AI 06/26/81 12:22:36 Re: adjectives
You can say that:
1. John is (a Criminal)
2. John is (a BadCriminal)
(a BadCriminal) is (a Criminal)
The next example is more interesting. A viewpoint (which I call Indictment) can be used
to avoid the mistake of asserting that any alleged criminal is a criminal, while still
being able to have that John inherits the properties of criminals:
3. (John is (a Criminal)) in (an Indictment)
Ciao, Beppe
Your June 29 note is a bit sketch, because I have forgotten about Omega, but
I get the general idea. I'll send you a note soon about map coloring strategies
as control for the logic of a simple PROLOG program. Also Pereira and Porto
tinkered with my eval, and there are now "official versions" with and without
cut although the version without (by Porto) has some constructs that may
be somewhat impure. More on this later. Regards to Sten-Ake.
∂29-Jun-81 1420 Betty Scott <CSD.BSCOTT at SU-SCORE> ARPA Supplement for LISP Project
Date: 29 Jun 1981 1412-PDT
From: Betty Scott <CSD.BSCOTT at SU-SCORE>
Subject: ARPA Supplement for LISP Project
To: JMC at SU-AI
cc: CSD.BScott at SU-SCORE
Bob Engelmore called me a week or so ago to say that ARPA has no more money
for this year, so it will not be possible to supplement your contract for
work this summer for the LISP Timing Evaluation Project. He did say,
however, that he would have no objection to funding the project after 10/1,
so that if you could (or want to ) do the work this summer, and be funded
"after the fact" for "evaluation of the results," he thought this would be
possible. It's just that there is no more money available to be awarded
in this fiscal year.
Betty
-------
I'm back. See if the attached note from Betty Scott presents any opportunities.
We can certainly lend the money out of my unrestricted funds. If Engelmore
has already been in touch with you, let me know what was decided or propoed.
∂29-Jun-81 1543 Wiederhold at SRI-AI schedule for visit by Machado nvalex-arpa
Date: 29 Jun 1981 1544-PDT
From: Wiederhold at SRI-AI
Subject: schedule for visit by Machado nvalex-arpa
To: jmc at SAIL, tob at SAIL, dcl at SAIL, zm at SAIL,
To: wiederhold at SUMEX-AIM
cc: csd.betty at SCORE, pickering at SUMEX-AIM, kaplan,
cc: sagalowicz, rowe, ark at SAIL, des at SAIL, navathe,
cc: jed at SAIL, csd.whang at SCORE, polak at SAIL,
cc: henke at SAIL
To: jmc@SAIL, tob@SAIL, dcl@SAIL, zm@SAIL, wiederhold
cc: mas@SAIL, kaplan@SRI-KL, sagalowicz@SRI-KL,
cc: csd.betty@SCORE
Provisional Schedule
for John Machado, Navelex project officer for KBMS
and for other consolidated ARPA contracts Oct.81-Sept.83.
Stanford July 27th - 28
KBMS Monday July 27th 1981 8:30 to noon
Principals Wiederhold, Kaplan, Sagalowicz
Participants Davidson, Whang, Navathe, Shaw,
(if again operational : Beetem )
(absent to Aug. 3: Rowe, Keller)
Lunch noon:1pm
Verification Monday July 27th 1:30 to 5:pm
Principal David Luckham ( subject to being at Stanford,
expects to return Aug. 1 )
Participants Polak and von Henke.
McCarthy's project Tuesday July 28th 8:30 to noon
Principal John McCarthy
Participants ...
(Zohar Manna : no response yet)
(Tom Binford : no response yet)
-------
∂30-Jun-81 0427 JRA lisp course dates
john,
i have to make up the course packets thursday morning. if i don't
hear otherwise, i'll leave your slot in the 10:45-12:15 friday july 10
session. you can use all this session if you like. any idea of title?
∂30-Jun-81 1004 FFL
To: JMC, FFL
John Allen tells me that his LISP class will meet on July l0, tentatively
from 10:45 - 12:15 at SC University.
∂30-Jun-81 1005 FFL
To: JMC, FFL
G. Wiederhold is setting up meeting with Machado from ARPA with PIs and their
researchers. You and yours are scheduled for Tuesday, July 28, 8:30-noon.
We'll do our show-and-tell as scheduled July 28, 8:30-noon.
∂30-Jun-81 1126 WIEDERHOLD at SRI-AI updated schedule
Date: 30 Jun 1981 1122-PDT
From: WIEDERHOLD at SRI-AI
Subject: updated schedule
To: jmc at SU-AI, tob at SU-AI, zm at SU-AI, dcl at SU-AI, wiederhold at SUMEX-AIM
cc: kaplan at SRI-AI
From: Wiederhold at SRI-AI
Subject: schedule for visit by Machado nvalex-arpa
To: jmc at SAIL, tob at SAIL, dcl at SAIL, zm at SAIL,
To: wiederhold at SUMEX-AIM
cc: csd.betty at SCORE, pickering at SUMEX-AIM, kaplan,
cc: sagalowicz, rowe, ark at SAIL, des at SAIL, navathe,
cc: jed at SAIL, csd.whang at SCORE, polak at SAIL,
cc: henke at SAIL
To: jmc@SAIL, tob@SAIL, dcl@SAIL, zm@SAIL, wiederhold
cc: mas@SAIL, kaplan@SRI-KL, sagalowicz@SRI-KL,
cc: csd.betty@SCORE
!
Provisional Schedule
for John Machado, Navelex project officer for KBMS
and for other consolidated ARPA contracts Oct.81-Sept.83.
Stanford July 27th - 28
Monday July 27th 1981 8:30 to noon
"KBMS, Management of Distributed Knowledge "
Principals Wiederhold, Kaplan, Sagalowicz
Participants Davidson, Whang, Navathe, Shaw,
(if again operational : Beetem )
(absent to Aug. 3: Rowe, Keller)
Lunch noon:1pm
Monday July 27th 1:30pm to 5:00
"Analysis and Verification of High Order Language Programs"
Principal David Luckham ( subject to being at Stanford,
expects to return Aug. 1 )
Participants Polak, von Henke, Germani
Larsen, Stevenson, Treitel, Osher, and Finlayson
Tuesday July 28th 8:30 to noon
"Basic Research in Artificial Intelligence and Formal Reasoning"
Principal John McCarthy
Participants Creary, Doyle, Goad, Ketonen, Gabriel, Tolcott
Lunch open
Tuesday July 28th 1:30 pm to 4:00
"Image Understanding"
Principals Binford, McCarthy
Participants Liebes
(Zohar Manna : will be out of town until Aug. 9, no response yet)
"A System for Deductive Programming"
Principal Manna
Participants Wolper, Malachi
-------