perm filename W82.OUT[LET,JMC] blob
sn#864699 filedate 1982-04-01 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00251 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00018 00002
C00019 00003 ∂04-Jan-82 1411 JMC John romano
C00020 00004 ∂04-Jan-82 1602 JMC
C00021 00005 ∂04-Jan-82 2111 JMC
C00022 00006 ∂05-Jan-82 0003 JMC Note to Herb Caen
C00023 00007 ∂05-Jan-82 0012 JMC
C00024 00008 ∂05-Jan-82 0024 JMC Using your system at Stanford
C00026 00009 ∂05-Jan-82 1544 JMC
C00027 00010 ∂05-Jan-82 2211 JMC
C00028 00011 ∂05-Jan-82 2235 JMC
C00029 00012 ∂06-Jan-82 0057 JMC
C00030 00013 ∂06-Jan-82 0059 JMC
C00031 00014 ∂06-Jan-82 1101 JMC Competitive life
C00032 00015 ∂06-Jan-82 1114 JMC
C00033 00016 ∂06-Jan-82 1811 JMC
C00034 00017 ∂07-Jan-82 0039 JMC
C00035 00018 ∂07-Jan-82 0059 JMC computer use for CS258
C00036 00019 ∂07-Jan-82 1155 JMC life
C00037 00020 ∂07-Jan-82 1545 JMC admissions
C00038 00021 ∂07-Jan-82 1616 JMC
C00039 00022 ∂07-Jan-82 2112 JMC
C00040 00023 ∂08-Jan-82 0811 JMC
C00041 00024 ∂08-Jan-82 0848 JMC ekl problem
C00042 00025 ∂08-Jan-82 0914 JMC
C00043 00026 ∂08-Jan-82 1001 JMC
C00044 00027 ∂08-Jan-82 1421 JMC funds for computing
C00045 00028 ∂08-Jan-82 1546 JMC Computer accts
C00046 00029 ∂09-Jan-82 0019 JMC lispx inelegances
C00048 00030 ∂09-Jan-82 0020 JMC more on lispx
C00049 00031 ∂09-Jan-82 0126 JMC
C00050 00032 ∂09-Jan-82 0150 JMC more lispx
C00051 00033 ∂10-Jan-82 0044 JMC
C00052 00034 ∂10-Jan-82 0046 JMC
C00054 00035 ∂10-Jan-82 1735 JMC more flat
C00056 00036 ∂10-Jan-82 2333 JMC
C00057 00037 ∂11-Jan-82 1834 JMC
C00058 00038 ∂11-Jan-82 2328 JMC bug?
C00059 00039 ∂12-Jan-82 0008 JMC
C00060 00040 ∂12-Jan-82 0135 JMC
C00062 00041 ∂12-Jan-82 1602 JMC proposal
C00063 00042 ∂12-Jan-82 2134 JMC
C00064 00043 ∂12-Jan-82 2150 JMC
C00065 00044 ∂13-Jan-82 0259 JMC
C00066 00045 ∂13-Jan-82 0323 JMC lisp axioms
C00067 00046 ∂13-Jan-82 1051 JMC
C00068 00047 ∂13-Jan-82 1101 JMC
C00069 00048 ∂13-Jan-82 1142 JMC
C00070 00049 ∂13-Jan-82 1155 JMC
C00071 00050 ∂13-Jan-82 1327 JMC system
C00072 00051 ∂13-Jan-82 1417 JMC
C00073 00052 ∂13-Jan-82 1440 JMC
C00074 00053 ∂13-Jan-82 1853 JMC proof done but could be improved
C00075 00054 ∂13-Jan-82 2136 JMC new bug
C00077 00055 ∂13-Jan-82 2156 JMC more on bug
C00078 00056 ∂14-Jan-82 1950 JMC improved lname
C00079 00057 ∂16-Jan-82 1705 JMC
C00080 00058 ∂16-Jan-82 2026 JMC
C00081 00059 ∂17-Jan-82 2313 JMC your axioms
C00082 00060 ∂18-Jan-82 1056 JMC
C00083 00061 ∂18-Jan-82 1110 JMC
C00085 00062 ∂18-Jan-82 1341 JMC
C00086 00063 ∂18-Jan-82 1344 JMC
C00087 00064 ∂18-Jan-82 1647 JMC
C00088 00065 ∂18-Jan-82 2205 JMC
C00089 00066 ∂18-Jan-82 2332 JMC m and b
C00090 00067 ∂19-Jan-82 1252 JMC
C00091 00068 ∂19-Jan-82 1252 JMC
C00092 00069 A partial message was sent by mistake. Here's the whole message.
C00096 00070 ∂19-Jan-82 1551 JMC paper
C00097 00071 ∂19-Jan-82 1650 JMC
C00101 00072 ∂19-Jan-82 1654 JMC
C00102 00073 ∂20-Jan-82 1254 JMC organization meeting for Porto course
C00103 00074 ∂20-Jan-82 1357 JMC
C00104 00075 ∂20-Jan-82 1835 JMC
C00106 00076 ∂20-Jan-82 2036 JMC
C00107 00077 ∂21-Jan-82 0051 JMC Hurd title
C00108 00078 ∂21-Jan-82 2102 JMC
C00109 00079 ∂21-Jan-82 2318 JMC
C00110 00080 ∂22-Jan-82 2034 JMC
C00111 00081 ∂22-Jan-82 2322 JMC
C00112 00082 ∂23-Jan-82 0055 JMC Umnov and Roth
C00113 00083 ∂23-Jan-82 0125 JMC
C00114 00084 ∂23-Jan-82 1104 JMC
C00115 00085 ∂23-Jan-82 1801 JMC Gosper
C00116 00086 ∂23-Jan-82 1814 JMC running out of screeen while running out of E
C00118 00087 ∂23-Jan-82 2316 JMC improved string search
C00121 00088 ∂24-Jan-82 0006 JMC
C00122 00089 ∂24-Jan-82 0121 JMC flap about Soviet robotics visitor
C00123 00090 ∂24-Jan-82 1752 JMC
C00124 00091 ∂24-Jan-82 2333 JMC
C00144 00092 ∂25-Jan-82 0201 JMC
C00145 00093 ∂25-Jan-82 0548 JMC
C00146 00094 ∂25-Jan-82 2120 JMC
C00147 00095 ∂26-Jan-82 1034 JMC
C00148 00096 ∂26-Jan-82 1309 JMC
C00149 00097 ∂27-Jan-82 1521 JMC
C00150 00098 ∂27-Jan-82 1526 JMC
C00153 00099 ∂27-Jan-82 1816 JMC
C00154 00100 ∂29-Jan-82 1532 JMC
C00155 00101 ∂29-Jan-82 1533 JMC
C00156 00102 ∂29-Jan-82 1535 JMC
C00157 00103 ∂29-Jan-82 1536 JMC
C00158 00104 ∂29-Jan-82 1546 JMC
C00159 00105 ∂29-Jan-82 2144 JMC
C00160 00106 ∂29-Jan-82 2150 JMC report and next friday
C00161 00107 ∂29-Jan-82 2207 JMC
C00162 00108 ∂30-Jan-82 0032 JMC report to LLL
C00163 00109 ∂30-Jan-82 1711 JMC Antonio Porto
C00164 00110 ∂30-Jan-82 2040 JMC
C00165 00111 ∂30-Jan-82 2041 JMC
C00166 00112 ∂31-Jan-82 1146 JMC
C00167 00113 ∂31-Jan-82 1722 JMC
C00171 00114 ∂31-Jan-82 1724 JMC
C00175 00115 ∂31-Jan-82 1925 JMC
C00176 00116 ∂31-Jan-82 1952 JMC
C00177 00117 ∂01-Feb-82 2320 JMC
C00178 00118 ∂01-Feb-82 2322 JMC
C00184 00119 ∂01-Feb-82 2325 JMC
C00185 00120 ∂01-Feb-82 2329 JMC Tops-20 Prolog Tape
C00188 00121 ∂02-Feb-82 0036 JMC
C00189 00122 ∂02-Feb-82 0144 JMC
C00190 00123 ∂02-Feb-82 0149 JMC
C00191 00124 ∂02-Feb-82 1020 JMC
C00192 00125 ∂02-Feb-82 1048 JMC
C00193 00126 ∂02-Feb-82 1620 JMC
C00194 00127 ∂02-Feb-82 1717 JMC
C00195 00128 ∂02-Feb-82 1753 JMC
C00196 00129 ∂02-Feb-82 1755 JMC
C00197 00130 ∂02-Feb-82 1841 JMC
C00198 00131 ∂02-Feb-82 2054 JMC
C00199 00132 ∂02-Feb-82 2056 JMC
C00200 00133 ∂03-Feb-82 1845 JMC sign
C00201 00134 ∂03-Feb-82 2335 JMC
C00206 00135 ∂03-Feb-82 2355 JMC
C00207 00136 ∂04-Feb-82 0157 JMC
C00208 00137 ∂04-Feb-82 1025 JMC
C00209 00138 ∂04-Feb-82 1032 JMC
C00210 00139 ∂04-Feb-82 1511 JMC
C00212 00140 ∂04-Feb-82 1624 JMC
C00213 00141 ∂04-Feb-82 2325 JMC
C00214 00142 ∂05-Feb-82 0036 JMC
C00215 00143 ∂05-Feb-82 0959 JMC
C00216 00144 ∂05-Feb-82 1719 JMC
C00217 00145 ∂06-Feb-82 2340 JMC
C00218 00146 ∂07-Feb-82 0223 JMC
C00219 00147 ∂07-Feb-82 1047 JMC
C00220 00148 ∂07-Feb-82 2302 JMC
C00221 00149 ∂08-Feb-82 0111 JMC
C00222 00150 ∂09-Feb-82 0404 JMC
C00223 00151 ∂09-Feb-82 0404 JMC account for Herbert Stoyan
C00224 00152 ∂09-Feb-82 0407 JMC
C00225 00153 ∂09-Feb-82 0408 JMC Stoyan account
C00226 00154 ∂09-Feb-82 0410 JMC
C00227 00155 ∂09-Feb-82 0944 JMC
C00228 00156 ∂09-Feb-82 1149 JMC
C00229 00157 ∂09-Feb-82 1150 JMC consulting bill
C00230 00158 ∂10-Feb-82 2345 JMC
C00231 00159 ∂11-Feb-82 1512 JMC
C00232 00160 ∂11-Feb-82 1704 JMC
C00233 00161 ∂12-Feb-82 1327 JMC
C00234 00162 ∂12-Feb-82 1455 JMC guest account for Martin Davis
C00235 00163 ∂12-Feb-82 1457 JMC account
C00236 00164 ∂12-Feb-82 2001 JMC message files
C00237 00165 ∂14-Feb-82 0119 JMC
C00238 00166 ∂14-Feb-82 1326 JMC
C00239 00167 ∂14-Feb-82 1328 JMC
C00240 00168 ∂14-Feb-82 1344 JMC
C00241 00169 ∂14-Feb-82 1659 JMC
C00242 00170 ∂14-Feb-82 2039 JMC
C00243 00171 ∂14-Feb-82 2055 JMC
C00244 00172 ∂14-Feb-82 2150 JMC
C00245 00173 ∂15-Feb-82 0128 JMC
C00246 00174 ∂15-Feb-82 1351 JMC
C00247 00175 ∂16-Feb-82 1128 JMC Davis address
C00248 00176 ∂16-Feb-82 2118 JMC
C00249 00177 ∂17-Feb-82 0047 JMC
C00250 00178 ∂17-Feb-82 0103 JMC
C00251 00179 ∂17-Feb-82 0138 JMC
C00252 00180 ∂17-Feb-82 0145 JMC
C00253 00181 ∂17-Feb-82 0149 JMC
C00254 00182 ∂17-Feb-82 0155 JMC
C00255 00183 ∂17-Feb-82 0241 JMC
C00256 00184 ∂17-Feb-82 0307 JMC
C00257 00185 ∂17-Feb-82 0312 JMC
C00258 00186 ∂17-Feb-82 0348 JMC film in space
C00259 00187 ∂17-Feb-82 0350 JMC
C00260 00188 ∂17-Feb-82 1553 JMC
C00261 00189 ∂18-Feb-82 1841 JMC
C00262 00190 ∂18-Feb-82 2241 JMC
C00264 00191 ∂18-Feb-82 2309 JMC more on simulation
C00265 00192 ∂20-Feb-82 1432 JMC
C00266 00193 ∂20-Feb-82 1432 JMC disjunctive knowledge
C00267 00194 ∂20-Feb-82 1521 JMC commute mileage for consultant
C00268 00195 ∂20-Feb-82 1603 JMC
C00269 00196 ∂20-Feb-82 1629 JMC
C00270 00197 ∂20-Feb-82 1634 JMC
C00271 00198 ∂20-Feb-82 1657 JMC letter for Teller
C00272 00199 ∂20-Feb-82 1718 JMC
C00273 00200 ∂20-Feb-82 1835 JMC
C00274 00201 ∂20-Feb-82 2142 JMC
C00275 00202 ∂20-Feb-82 2148 JMC iii trips Monday and Wednesday
C00276 00203 ∂20-Feb-82 2154 JMC L.A. trip
C00277 00204 ∂20-Feb-82 2214 JMC
C00278 00205 ∂21-Feb-82 0007 JMC
C00279 00206 ∂21-Feb-82 0105 JMC
C00280 00207 ∂21-Feb-82 0157 JMC
C00281 00208 ∂21-Feb-82 1409 JMC
C00282 00209 ∂22-Feb-82 0041 JMC
C00283 00210 ∂22-Feb-82 0403 JMC
C00284 00211 ∂23-Feb-82 1644 JMC
C00285 00212 ∂23-Feb-82 2246 JMC
C00286 00213 ∂24-Feb-82 0043 JMC
C00288 00214 ∂24-Feb-82 0044 JMC
C00289 00215 ∂25-Feb-82 1110 JMC
C00291 00216 ∂25-Feb-82 1421 JMC
C00292 00217 ∂25-Feb-82 1430 JMC Kuck paper
C00293 00218 ∂25-Feb-82 1743 JMC
C00294 00219 ∂25-Feb-82 1748 JMC
C00295 00220 ∂25-Feb-82 1751 JMC
C00296 00221 ∂25-Feb-82 1753 JMC
C00297 00222 ∂25-Feb-82 2122 JMC
C00298 00223 ∂25-Feb-82 2123 JMC
C00299 00224 ∂26-Feb-82 0016 JMC
C00300 00225 ∂26-Feb-82 1202 JMC
C00301 00226 ∂26-Feb-82 1337 JMC
C00302 00227 ∂26-Feb-82 1637 JMC
C00303 00228 ∂27-Feb-82 1536 JMC
C00304 00229 ∂27-Feb-82 1759 JMC
C00305 00230 ∂28-Feb-82 1433 JMC
C00306 00231 ∂28-Feb-82 1820 JMC Hayes quote
C00307 00232 ∂28-Feb-82 1832 JMC Wolfram
C00308 00233 ∂28-Feb-82 1848 JMC
C00309 00234 ∂01-Mar-82 0300 JMC meta-cognition: reasoning about knowledge
C00310 00235 ∂01-Mar-82 1156 JMC
C00311 00236 ∂01-Mar-82 1228 JMC
C00312 00237 ∂03-Mar-82 1546 JMC
C00313 00238 ∂03-Mar-82 1551 JMC
C00314 00239 ∂03-Mar-82 1700 JMC Common LISP discussion
C00315 00240 ∂03-Mar-82 1812 JMC
C00316 00241 ∂05-Mar-82 0112 JMC
C00319 00242 ∂05-Mar-82 0114 JMC
C00320 00243 ∂05-Mar-82 0318 JMC dinner with Ehud Shapiro
C00321 00244 ∂05-Mar-82 1416 JMC
C00324 00245 ∂05-Mar-82 1418 JMC
C00325 00246 ∂05-Mar-82 1419 JMC
C00328 00247 ∂05-Mar-82 1424 JMC
C00329 00248 ∂05-Mar-82 1449 JMC
C00330 00249 ∂05-Mar-82 1449 JMC reprint
C00334 00250 ∂07-Mar-82 1517 JMC Golux
C00335 00251 ∂07-Mar-82 1520 JMC
C00338 ENDMK
C⊗;
∂04-Jan-82 1411 JMC John romano
To: JMM
this hp student from Colorado 303 598-1900x2131 wants to know how
he's doing.
∂04-Jan-82 1602 JMC
To: FFL
Use Diffie's business address.
∂04-Jan-82 2111 JMC
To: RWG
New version of life[f81,jmc].
∂05-Jan-82 0003 JMC Note to Herb Caen
To: FFL
It tells me that It will keep raining until the last environmentalist
takes the brick out of his toilet tank.
∂05-Jan-82 0012 JMC
To: FFL
todoro.7
∂05-Jan-82 0024 JMC Using your system at Stanford
To: boyer at UTEXAS-20, moore at UTEXAS-20
I am thinking of having my course in theory of computation use your
prover this quarter for some problems. Are there any problems (technical
or proprietary) in importing the system to SCORE?
Alternatively, do you think SRI might be willing to let people use
it over the net on the Foonly they bought for you? Do you see any technical
problems and whom would I ask about it?
I have another problem about sublis(pattern, alist) and match(pattern,
expression, alist) being partial inverses that I have tried once and
failed. However, I'll try a couple more ideas before I bother you with
it.
∂05-Jan-82 1544 JMC
To: FFL
WADA.1
∂05-Jan-82 2211 JMC
To: FFL
lehman.1
∂05-Jan-82 2235 JMC
To: FFL
lindze.3
∂06-Jan-82 0057 JMC
To: RPG
Thanks for mail file info.
∂06-Jan-82 0059 JMC
To: rwg at MIT-MC
What number can I phone you at PARC?
∂06-Jan-82 1101 JMC Competitive life
To: RWG at SU-AI, rwg at MIT-MC
All gliders but one become homogeneous in color but
w
w
wbb
oscillates between 3 white and two black and the reverse. The geometric
phase and the color phase are synchronous. Of course, it isn't clear
that mixed gliders can be emitted by any kind of mixed glider gun.
Where do you prefer to receive mail?
∂06-Jan-82 1114 JMC
To: DON at SU-AI
LIFE[F81,JMC] proposes a competitive version.
∂06-Jan-82 1811 JMC
To: JJW
Please arrange with boyer%utexas-20 to ftp their prover to SCORE.
∂07-Jan-82 0039 JMC
To: RWW
What would be involve in reviving LCF for class experimentation?
∂07-Jan-82 0059 JMC computer use for CS258
To: csd.golub at SU-SCORE
As you recall, I arranged for SAIL use for CS206 in the Fall. However,
I recoiled in horror when I found out how many students there were in the
course and had them use LOTS. I want to retroactively consider that I
took a rain check and have the students use SAIL and also SCORE. Different
proof systems that I want them to use require the different operating systems.
There are only nine students in the class and they all turned out to have
SAIL accounts already; I didn't ask about SCORE.
∂07-Jan-82 1155 JMC life
To: rwg at MIT-MC
I had gone to bed when your message arrived only noticed it as I logged
in which made it disappear, so I didn't get it all. I assume a block
is a square of four points. In my version of competitive life, a square
doesn't oscillate, because a cell flips only if a majority of neighbors
including itself are of the opposite color. I think this is more likely
to lead to a decisive victory for one color than making it more oscillatory.
The oscillatory glider configuration was obtained on paper looking at
all ten gliders with 3 whites and two blacks.
∂07-Jan-82 1545 JMC admissions
To: csd.golub at SU-SCORE
Why do we allow any GRE but engineering?
∂07-Jan-82 1616 JMC
To: csd.golub at SU-SCORE
My query may be confusing. I mean why don't we allow engineering?
∂07-Jan-82 2112 JMC
To: KLC
There is no literature that I know of that mentions both except a sentence
or two in my paper. I agree that circumcscription might be used to establish
a sample space for probabilistic reasoning. I would be glad to discuss the
matter with you at a mutually convenient time - tomorrow afternoon any time
after 2:30 would be convenient for me. Since I don't remember any Wednesday
remarks, any apology is indeed moot.
∂08-Jan-82 0811 JMC
To: rwg at MIT-MC
W B
B W
won't blink according to my rules, because flipping requires a majority
including self. Of course, the rules can be as we choose, but for
the competitive game, it seems to me that my rules are more likely
to lead to a situation that can be called a victory. Experiment will
tell, however.
∂08-Jan-82 0848 JMC ekl problem
To: JK
I have been trying to use the new ekl to prove the associativity of append.
The axioms and declarations are in lispx.lsp[f81,jmc]. This file is redundant,
because it has both cons and an infix ~ that I am currently using for cons.
The proof as far as it goes is in lispx.ppr[f81,jmc]. I don't know how to
make the last step.
After you look, can I phone you?
∂08-Jan-82 0914 JMC
To: JK
I'll look.
∂08-Jan-82 1001 JMC
To: JJW
I'll be back to you. It will be some combination of the department
and overhead, but I'll have to talk to Golub.
∂08-Jan-82 1421 JMC funds for computing
To: JJW
∂08-Jan-82 1326 Gene Golub <CSD.GOLUB at SU-SCORE> funds for computing
Date: 8 Jan 1982 1321-PST
From: Gene Golub <CSD.GOLUB at SU-SCORE>
Subject: funds for computing
To: jmc at SU-AI
cc: bscott at SU-SCORE
We shall try to cover the computer costs from the residual IBM funds.
GENE
-------
∂08-Jan-82 1546 JMC Computer accts
To: JJW
∂08-Jan-82 1545 Susan Hill <CSD.HILL at SU-SCORE> Computer accts
Date: 8 Jan 1982 1539-PST
From: Susan Hill <CSD.HILL at SU-SCORE>
Subject: Computer accts
To: jmc at SU-AI
cc: csd.gotelli at SU-SCORE
Betty said something about your asking for permission for students in your
class to have computer accounts. How do you plan to handle this?
-------
∂09-Jan-82 0019 JMC lispx inelegances
To: JK
lispx.lsp[w82,jmc] is a source file, and lispx.ppr and lispx.prf are the others.
I wanted to get the sort information and the facts about append entirely
into a symbolic form. In order to do this I followed each line containing
such information by a LINENAMES that updated the line in question. It worked,
but there are the following inelegances.
1. I had to answer the question of whether I really meant it after
each use of LINENAMES for updating purposes, so there is a line containing
y after each such line.
2. SORTINFO has extra lines.
3. In the .ppr file, the command still comes out with some of the
numerical line number information instead of the purely symbolic information
that went in.
∂09-Jan-82 0020 JMC more on lispx
To: JK
It is elegant to be able to do the append proof essentially in one line.
I don't yet fully understand how it works, and I'll have to ask you some
questions.
∂09-Jan-82 0126 JMC
To: rwg at MIT-MC
Indeed!
∂09-Jan-82 0150 JMC more lispx
To: JK
I have added a definition of
flat(x,u) = if atom x then x.u else flat(car x,flat(cdr x, u))
but don't understand rewriting well enough yet to know how far I
can get toward
∀x u.listp flat(x,u)
in one ∀e step.
∂10-Jan-82 0044 JMC
To: JK
Is the following a bug? If not how do I get use an axiom ¬p to get
if p then a else b = b?
the proof BUG:
(DECL (P) |TRUTHVAL| CONSTANT)
(DECL (A B) |GROUND| CONSTANT)
(ASSUME |¬P|)
3. ¬P
ctxt: (1) deps: (3)
(TRW |IF P THEN A ELSE B| |*3*NIL|)
4. IF P THEN A ELSE B=IF P THEN A ELSE B
ctxt: (1 2) deps: (3)
(ASSUME |P|)
5. P
ctxt: (1) deps: (5)
(TRW |IF P THEN A ELSE B| |*5*NIL|)
6. IF P THEN A ELSE B=A
ctxt: (1 2) deps: (5)
∂10-Jan-82 0046 JMC
To: JK
∂09-Jan-82 0933 JK
∂09-Jan-82 0150 JMC more lispx
I have added a definition of
flat(x,u) = if atom x then x.u else flat(car x,flat(cdr x, u))
but don't understand rewriting well enough yet to know how far I
can get toward
∀x u.listp flat(x,u)
in one ∀e step.
------------
Assuming that ∀x u.sexp flat(x,u) is clear, one could just instantiate
a case of car induction for ∀u.listp flat(x,u)
and that should go in one step.
------------
∀x u.sexp flat(x,u) isn't clear, since that would also amount to the
termination of the program. Even so I don't see how it would be
used.
∂10-Jan-82 1735 JMC more flat
To: JK
CC: JJW
I now understand your proof of
∀x u.listp flat(x,u)
The next one Joe and I tried was
∀x u. flat(x,u) = flatten x * u
where
flatten x = if atom x then list x else flatten(car x) * flatten(cdr x)
The necessary definitions along with various junk (to be cleaned out)
are in lispx[w82,jmc].
We are still having difficulty in working out the technique of the new
rewrite modes, but I think it is worthwhile to debug our technique with
the present mode, even though I suspect another major change will be
wanted eventually. The present scheme is looking like a macro language,
and they usually need to be modified as they become more programmable.
∂10-Jan-82 2333 JMC
To: reid at SHASTA
BIOJMC[1,JMC] is a vita in PUB.
∂11-Jan-82 1834 JMC
To: csd.irmgild at SU-SCORE
Dr. Keith Clark
Department of Computing
Imperial College of Science and Technology
University of London
l80 Queen's Gate
London SW7 2BZ, England
∂11-Jan-82 2328 JMC bug?
To: JK
I don't know whether it's a bug or my misunderstanding, but I can't get
ekl to expand out x~nnil*v. The last line on p. 2 of lispx.lsp[w82,jmc]
is my attempt. I got it to expand flat and flatten with a simpler
mode, which I have been elaborating, but it goes no further.
∂12-Jan-82 0008 JMC
To: JK
Cancel previous message; I have got through my difficulty.
∂12-Jan-82 0135 JMC
To: JK
New bug
Step 47 in lispx.lsp[w82,jmc] or lispx.prf[w82,jmc] is produced by the
command below it. Step 51 above (not in lispx.prf) was produced by
the same command with $ in place of the &. The & was an attempt to
get it to expand flatten without again expanding flat. How is that
to be done? I want a mode that will use all applicable occurrences
of the line range referred to but won't expand the results of expansions.
Then I could expand all definitions in part of the formula with a
single reference to DEFINFO without danger of a loop.
The bug is that the term (x~y)~nil*u should be
(flatten(x)*flatten(y))*u.
∂12-Jan-82 1602 JMC proposal
To: JK
The proposal left Stanford Dec. 29. I got a postcard confirming its
reception from NSF. It has the proposal id no: MCS 8206565.
∂12-Jan-82 2134 JMC
To: ROY
Imlac conked out agin - won't load.
∂12-Jan-82 2150 JMC
To: RWW
I know of no way to relieve your nesslessness.
∂13-Jan-82 0259 JMC
To: ROY
It came back to life after temperature changes.
∂13-Jan-82 0323 JMC lisp axioms
To: YOM
My current version with working material removed is LISPAX.LSP[W82,JMC].
There are still a few changes to be made in order to maximize convenience.
∂13-Jan-82 1051 JMC
To: csd.walker at SU-SCORE
Yonatan Malachi was a high pass. Frank Yellin was a pass. Joe Weening
didn't pass but will take it again in late February.
∂13-Jan-82 1101 JMC
To: ef at MIT-AI
Alas, I can't take the time to go to the Information Mechanics Meeting.
∂13-Jan-82 1142 JMC
To: REG
∂13-Jan-82 1135 JJW SCORE account for Boyer-Moore system
I asked Lynn Gotelli for 60 aliquots, which is 3180 blocks
(Boyer said we need 3200), but she has told me that there are only
555 blocks left to be sold on SCORE. So I don't know how we will
be able to bring up the theorem-prover unless something is done to
free up disk space at SCORE.
∂13-Jan-82 1155 JMC
To: JJW
I'll start politicking.
∂13-Jan-82 1327 JMC system
To: ME
I have noticed that in the last few months, the load levels have been
somewhat lower, especially the residual load level when no-one is
doing anything. Is this a real change or just a change in the way
things are measured?
Formerly the load was (as I remember it) never beow .10, and now it
goes down to .04.
∂13-Jan-82 1417 JMC
To: FFL
Knuth, Donald E., DECIPERING A LINEAR CONGRUENTIAL ENCRYPTION, 1980. ONR
please have a copy of this sent to Whit Diffie
∂13-Jan-82 1440 JMC
To: JK
Thanks; you can expect at least a few to be my bugs.
∂13-Jan-82 1853 JMC proof done but could be improved
To: JK
lispx.ppr[w82,jmc] has the proof. The main parts are done in
single rewrites, but the loose ends are done separately. I would be
grateful if you would look at it from the point of view of style.
1. Can it be improved to reduce number of steps?
2. How should I change it as a model for CS258 students?
3. Does it suggest improvements to EKL?
In any case I plan to separate the proofs about flat from the
basic LISPX proof.
∂13-Jan-82 2136 JMC new bug
To: JK
The proof went through ok, but I decided to clean matters up by separating
the basic lisp axioms (now called lispax.lsp and lispax.prf) from the
definitions and proof peculiar to flat. The lispax part went ok and so
did flat up to the point when I did (pretty-proof flat flat).
(save-proofs flat) had worked ok. I got some kind of lisp error and then
tried (show) and got the error message you will find in flat.lsp[w82,jmc].
Running the proof again worked, but (show) failed again. In the previous
lispx context, I had no trouble with (pretty-proof).
Another subject: Part of the difficulty with proving the facts about
flat was that there is no direct way to use the premisses of an
implication in rewriting the conclusion. If you can figure out how
to make this possible, at least two of the proofs will be more
straightforward. In particular,
∀x y.(∀u.flat(x,u) = flatten(x)*u) ∧ (∀u.flat(y,u) =flatten(y)*u)
⊃ flat(x,flat(y,u)) = flat(x)*(flat(y)*u)
is obtainable as a valid formula in this way. Perhaps der could do
this or try this or let the user try it.
The proof of ∀x.listp(flatten x) would also be more straightforward
were this possible.
∂13-Jan-82 2156 JMC more on bug
To: JK
I tinkered with the file a bit more and changed the bug. I added the
command that proves ∀x u.listp flat(x,u), and this put another step
in the proof. Again the proof ran without difficulty and produced
the desired results. However, this time (save-proofs flat) got into
an apparently infinite loop.
∂14-Jan-82 1950 JMC improved lname
To: JJW
Could lname generate a (comment) line of the form (COMMENT LNAME foo baz)
so that .ppr files would have the line names attached to the axioms?
∂16-Jan-82 1705 JMC
To: RPG
Try to get LLNL to pay half.
∂16-Jan-82 2026 JMC
To: FFL
russel.3
∂17-Jan-82 2313 JMC your axioms
To: YOM
I have looked at them. In the main you have the right facts, but they
require many changes to make them suitable for ekl proofs. Mainly they
need to be put as much as possible in forms that can be used as rewrite
rules in which the left side is replaced by the right side. Those that
can be used indefinitely need to be separated from those that can't such
as recursive definitions and commutativity. Associative operations need
to be declared such. There are also some bugs, such as leaving out the
conclusion of the induction axiom.
∂18-Jan-82 1056 JMC
To: csd.hill at SU-SCORE
Gosper should still have a courtesy account, but probably of minimum
magnitude. Is this the minimum now, and if not, what are his quotas?
I don't imagine he uses anything but SAIL.
∂18-Jan-82 1110 JMC
To: reid at SHASTA, csd.schreiber at SU-SCORE,
csd.herriott at SU-SCORE
1. This course description ignores the existence of 156.
2. In my opinion, the description contains much
material, e.g. automata, formal languages and Turing machines, that
shouldn't be in a basic course. It is an obsolete carryover from the
time when computer science theory was an offshoot of mathematics.
It should be covered sketchily in advanced courses.
3. I won't vote for it as a math sciences option in its present form,
but its content is really more the business of the computer science
department than of the math. sciences committee.
∂18-Jan-82 1341 JMC
To: csd.hill at SU-SCORE
I'll find out something of what he's doing. There are other possibilities
for his disk. I'll be back to you.
∂18-Jan-82 1344 JMC
To: reid at SHASTA
Perhaps I was hasty. I don't think a week-by-week outline should be asked
of you, any more than from anyone else. I'll react further later.
∂18-Jan-82 1647 JMC
To: YOM
(DECL (*) |GROUND⊗ground*→GROUND| functional NIL INFIX 840 both)
∂18-Jan-82 2205 JMC
To: JJW
Many people already have used up some of the single character e macro names.
∂18-Jan-82 2332 JMC m and b
To: JJW
I have rewritten it and checked it with ekl. I'll show you the files
tomorrow.
∂19-Jan-82 1252 JMC
To: RDG
m and b is ready
∂19-Jan-82 1252 JMC
To: RDG
essentially ready
A partial message was sent by mistake. Here's the whole message.
I have declared *, i.e. append, to be associative, and this seems to
work fine; it makes one of the proofs go without special appeal to
associativity. I have advocated to Yoram that he declare + and times
to be associative also. However, it occurs to me that declaring them
associative may cause problems in using their commutativity. Suppose,
for example, that we want to prove a+b+c+a = a+a+b+c. If EKL doesn't
allow us to group the terms freely, then there will be difficulty in
using associativity. There will also be a problem in matching
subexpressions. It seems to me that allowing operators to be
declared associative is good and even necessary, but it imposes further
requirements on interactive theorem provers such as EKL. These seem to
include
1. Parts designators should be able to pick out segments, e.g. b*c
should be a part of a*b*c*d.
2. Matching the operator should try to match segments. It is important
here whether the null segment should be considered, and this depends
on the operator. Thus the matches of x*y*z against a*b*c*d are (nil nil a*b*c*d),
(nil a b*c*d), (nil a*b c*d), (nil a*b*c d), (nil a*b*c*d nil),
(a nil b*c*d), (a b c*d), (a b*c d), (a b*c*d nil), (a*b nil c*d),
(a*b c d), (a*b c*d nil), (a*b*c nil d), (a*b*c d nil), (a*b*c*d nil nil)
in the unrestricted case and the three without nil in the restricted case.
If an operator could be declared both commutative and associative, there
are many more considerations.
1. Ideally the expression should be represented internally by a multi-set
of arguments. Arbitrary sub multi-sets should be designatable and
matched.
2. Because the ideal becomes very expensive, the user needs to be able
to restrict the potential matches that will be examined. It will require
experience to determine what restrictions should be available.
∂19-Jan-82 1551 JMC paper
To: FFL
Please send Home terminal paper to
Harold Gilliam
1747 10th ave
SF 94122
∂19-Jan-82 1650 JMC
To: JK, JJW, YOM
A partial message was sent by mistake. Here's the whole message.
I have declared *, i.e. append, to be associative, and this seems to
work fine; it makes one of the proofs go without special appeal to
associativity. I have advocated to Yoram that he declare + and times
to be associative also. However, it occurs to me that declaring them
associative may cause problems in using their commutativity. Suppose,
for example, that we want to prove a+b+c+a = a+a+b+c. If EKL doesn't
allow us to group the terms freely, then there will be difficulty in
using associativity. There will also be a problem in matching
subexpressions. It seems to me that allowing operators to be
declared associative is good and even necessary, but it imposes further
requirements on interactive theorem provers such as EKL. These seem to
include
1. Parts designators should be able to pick out segments, e.g. b*c
should be a part of a*b*c*d.
2. Matching the operator should try to match segments. It is important
here whether the null segment should be considered, and this depends
on the operator. Thus the matches of x*y*z against a*b*c*d are (nil nil a*b*c*d),
(nil a b*c*d), (nil a*b c*d), (nil a*b*c d), (nil a*b*c*d nil),
(a nil b*c*d), (a b c*d), (a b*c d), (a b*c*d nil), (a*b nil c*d),
(a*b c d), (a*b c*d nil), (a*b*c nil d), (a*b*c d nil), (a*b*c*d nil nil)
in the unrestricted case and the three without nil in the restricted case.
If an operator could be declared both commutative and associative, there
are many more considerations.
1. Ideally the expression should be represented internally by a multi-set
of arguments. Arbitrary sub multi-sets should be designatable and
matched.
2. Because the ideal becomes very expensive, the user needs to be able
to restrict the potential matches that will be examined. It will require
experience to determine what restrictions should be available.
∂19-Jan-82 1654 JMC
To: JK
Aha. You're ahead of me.
∂20-Jan-82 1254 JMC organization meeting for Porto course
To: FFL
PROLOG[W82,JMC] is an announcement of Porto's course. It lacks a place
for the organization meeting. Please get a room for (say) 15 people
and add it to the announcement. It should be mailed to bulletin boards
at SAIL, SCORE and SRI-AI and PARC. Get Arthur or RPG to tell you how.
∂20-Jan-82 1357 JMC
To: RDG
See pages 4 and 5 of APE.LSP[W82,JMC].
∂20-Jan-82 1835 JMC
To: twe.hp-labs at UDEL
Course in logic programming and Prolog starting soon
Starting in the week of February 2 and continuing through the
quarter there will be a course in logic programming and the Prolog
language taught by Antonio Porto of the University of Lisbon. Logic
programming in general and Prolog and particular have attracted wide
interest in Europe but are only getting started in the U.S.
There will be an opportunity to use Prolog.
There will be an organization meeting to fix hours on Thursday,
January 28 at 3pm. Anyone interested who can't make the meeting or who has
questions should contact John McCarthy by telephone or net message.
∂20-Jan-82 2036 JMC
To: warren at SRI-AI
He'll arrive on the 26th.
∂21-Jan-82 0051 JMC Hurd title
To: csd.golub at SU-SCORE
Is it too late to get Dr. Hurd's former title correct. I'm sure it
wasn't "project manager", and I think it was something like "Director
of Data Processing".
∂21-Jan-82 2102 JMC
To: JJW
Ten it is.
∂21-Jan-82 2318 JMC
To: rem at MIT-MC
I have no special reason to suppose that I can be much help, but if you
would like to have dinner some evening and discuss your problems, I would
be glad to do it - perhaps early next week.
∂22-Jan-82 2034 JMC
To: katz.hp-labs at UDEL
The point of the organizational meeting is to set a time for the class.
As it happens, Porto won't arrive till the day after the organizational
meeting which will be held as announced. Therefore, all it can do is
set a time. It would be best for you to try to be represented.
∂22-Jan-82 2322 JMC
To: llw at S1-A
Congratulations on Lawrence award.
∂23-Jan-82 0055 JMC Umnov and Roth
To: TOB
Please look at UMNOV.NS[W82,JMC]. I think the statement by Bernie Roth that
we learned about robotics from the Russians is a total exaggeration, and the
statement that it is impractical to put restrictions on Umnov if he visits
is essentially false. I am thinking about a letter saying so either to the
Stanford Daily or the New York Times. This is is independent of the question
of whether such restrictions are desirable. What is your opinion?
∂23-Jan-82 0125 JMC
To: TOB
See also DAILY.7[LET,JMC] for a draft.
∂23-Jan-82 1104 JMC
To: TOB
Would you be willing to take part in a joint letter or even a press conference
dissenting from the official Stanford view? We might ask a few other people,
e.g. Cannon and deBra, but we'd do something even if they weren't inclined
to participate. I guess a letter is appropriate in any case and a press
conference if there are enough, say 4 or 5.
∂23-Jan-82 1801 JMC Gosper
To: CSD.HILL at SU-SCORE
CC: RWG at SU-AI
He should be reduced to the half-an-aliquot guest level.
He will move some files elsewhere.
∂23-Jan-82 1814 JMC running out of screeen while running out of E
To: RPG, ME
My experience in running out of E convinces me that our work would be
improved by the ability to display a lot more characters. No matter
how nicely they are arranged on the screen, there just aren't enough.
Would it be feasible, for experimental purposes, to make the E-system
interface convenient for someone who had two or even four datadiscs?
I think this is worth doing soon in spite of the current shortage of
datadisc channels, because it is relevant to the question of whether
having lots of characters is more important than having beautiful
characters. My conjecture is that having two or four low resolution
displays will prove more valuable than having a high resolution display
with the same number of pixels unless the high resolution display is large
enough so that just as many characters can be put up non-microscopically.
∂23-Jan-82 2316 JMC improved string search
To: cl.boyer at UTEXAS-20, cl.moore at UTEXAS-20
Here is an "improvement" on your string search whose correctness statement
might be difficult to formalize. The algorithm is essentially that given
in your book, except for a change in the initialization to avoid updating
the table for all letters of the alphabet. Each entry in the table includes
your (DELTA1 K) in its low order bits and an integer <search number> in its
high order bits. When we start a new search, we update the entries
corresponding to the letters in the pattern string but with a search number
one larger than that of the last search. We don't bother changing the
entries for letters that don't appear in the pattern string, because a
simple comparison when the letter is encountered in the text shows that
the entry is obsolete. Only when the search number threatens to overflow
the allocated field do we go through the entire table and restart with
search number 1.
Since I don't know the literature and couldn't find a reference to this
kind of searching in Knuth vols. 1-3, this may be an old idea.
Notice that stating the correctness and the efficiency of the algorithm
requires taking into account that the algorithm may be used many times.
∂24-Jan-82 0006 JMC
To: RWG
Should your situation change to need more disk here, it will be arranged.
∂24-Jan-82 0121 JMC flap about Soviet robotics visitor
To: llw at S1-A
Page 80 of \bboard at SU-AI contains an AP dispatch and commentary by
various people including me about Stanford declining to restrict a
Soviet visitor and the State Department cancelling the visit. I made
some speculation about the bureaucratic process. Can you (dis)confirm
them, and do you know more?
∂24-Jan-82 1752 JMC
To: JPM
What I know came from Ralph. Perhaps I should not have spoken to the reporter
at all given that fact, but I felt obliged to convince her (Alice Cuneo in case
you should have the misfortune to meet her) that the computer couldn't be used
to fire missiles or write checks. If someone offers $3.5 million for the
system in Margaret Jacks, the Department should accept.
∂24-Jan-82 2333 JMC
To: feigenbaum at SUMEX-AIM
∂23-Jan-82 HPM 24-Jan-82 JMC Secure robots
a021 0047 23 Jan 82
PM-Scientist, Bjt,580
Universities Rebuff State Department Restrictions on Soviet Visitor
STANFORD, Calif. (AP) - A widening dispute over the federal
government's attempt to extend anti-Soviet sanctions to academia has
lopped two universities from the itinerary of a Soviet scientist whose
specialty is robots.
Stanford University has been declared off-limits to Nikolay Umnov
because the school refused to honor State Department restrictions on
his visit. The university said it could not successfully police a
visit, and that to make the attempt would disrupt the free environment
needed for creative work to take place.
The University of Wisconsin followed Stanford's lead Friday, saying
it was an ''open institution.''
Wisconsin Asssociate Dean Camden A. Coberly and professor Ali A.
Seireg said they plan to withdraw their acceptance of Umnov's visit.
That move will reduce the scientist's itinerary to Auburn University
and Ohio State University.
But at Ohio State University, where Umnov was to spend six weeks,
professor Robert B. McGhee said Umnov would be ''welcome for two or
three days. I'd accept restrictions for a visit of that length, but
not any longer.''
Umnov has been caught in a dispute over what visiting Soviet
scientists can see and do during a three-month scientific exchange
program that is sponsored by the National Academy of Sciences and
funded by the U.S. government.
As part of the Reagan administration's attempt to restrict Soviet
access to U.S. technology, the State Department said Umnov couldn't
see anything not already published in open literature.
It also said he should be restricted to mechanical theory of robotic
locomotion, with no industrial visits and no access to control units
or computer programming techniques that have given robots new
flexibility and made them more useful in the workplace.
The measures are justified because the government is paying for the
visit, said Sue Pittman, a State Department spokeswoman. She also
cited ''possible technology transfer risks,'' an administration term
for thefts or leaks of U.S. scientific secrets.
The restrictions reached universities hosting Soviet scholars via
letters from the national academy. As criticism of the restrictions
mounted, the NAS on Tuesday halted distribution of the letters.
The restrictions are ''absurd,'' according to Bernard Roth, a
professor in Stanford's Department of Mechanical Engineeering.
''Guys in the U.S. who know something about it (robotics) learned
from the Russians,'' he said. Potential military applications are
''very small. ... I don't think anyone will win a war on the basis of
those machines. It's certainly not of critical national interest.''
He added that nothing is being done in the field that the Soviets
''either don't already know or can easily figure out for themselves.''
In refusing to honor the restrictions, Stanford Vice-Provost Gerald
J. Liberman said, ''We believe the best interests of American science
and technology are served by open exchanges of university research
activities.''
He said in a letter to the NAS, ''The campus is completely open; and
the thousands of visiting scholars to campus each year have free
access to any of the university's programs. Even if we had the means
to monitor or police the activities of visitors, such actions would
drastically disrupt the academic environment which is essential in
fostering creative research endeavors.''
Attempts to restrict visiting scientists are not new, and last
February Stanford President Donald Kennedy expressed ''grave concern''
over federal attempts to apply export controls to academic teaching
and research. He was joined in writing protests to the secretaries of
state, defense and commerce by presidents Marvin L. Goldberger of the
California Institute of Technology, Paul E. Gray of MIT, Frank H.T.
Rhodes of Cornell and David S. Saxon of the University of California.
Kennedy told the faculty senate earlier this month the universities
had done ''sporadic negotiating'' with some agencies and had had
''skirmishes'' with others.
ap-ny-01-23 0346EST
***************
kjk - Question for the robotics people: isn't Umnov in robot locomotion?
Aren't the Russians significantly ahead of us in locomotion?
What are the expected gains from Russian→US technology transfer vs.
US→USSR?
In other words, do we stand to gain more than Russia from Umnov's
visit? If so, a letter to the State Department informing them that they
are applying an import restriction rather than an export restriction
may be in order.
A brief review of the status of known Russian work in robotics might
be useful for clarifying the tradeoffs. Is anyone here willing to do that?
Lest people get the wrong impression--I'm opposed to applying
export restrictions to academic work. I think the State Dept is wedged.
However, a specific instance of the loss we incur from these restrictions
may do more than a thousand pages of moral, ethical, and philosphical
argument.
TAW - If the US Gov't doesn't want Umnov to know anything about
state-of-the-art robotics, why are they letting him in?? That seems
to be the way to handle these things, as opposed to making every
US technologist in a sensitive field take Diplomacy courses.
JMC - I think that both Roth and Lieberman have made misstatements in
order to make their case stronger. In my knowledge and Tom Binford's,
the Russians are mainly behind us in robotics. Only in walking machines
have they done anything substantial. Of course, they are very secretive
about what they do, and before they receive any visitor at an institute,
they determine what he may and may not be shown and what may be discussed.
Many institutes are secret and known only by Post Office Box numbers within
the country. When people from one such institute (the Institute of Applied
Mathematics in Moscow) wanted to talk with me in 1965, they pretended to
be from a different institute. Only later, when that institute wanted to
buy an American computer, did its existence emerge from secrecy; I never
got to visit it.
My attempts to collaborate with Ershov on MTC broke down mainly because
it was illegal for him to put scientific information in a letter except
by sending published documents. It is illegal for a Russian to give a
foreigner written scientific information except in the form of published
books or papers. In the 1960s, these rules were often bent, but exceptions
have been few since then.
Roth is also wrong in saying that robotics has no
military applications. Lieberman's statement
''We believe the best interests of American science
and technology are served by open exchanges of university research
activities.'' contains the assumption that putting no restrictions
on Umnov would achieve that result.
Finally, the statement that restrictions are impractical and would
disrupt the University is phoney. A visitor's schedule is made by
the University's "Office of Foreign Visitors", and nothing is easier
than leaving something off the schedule and notifying the Principal
Investigators of a few DoD sponsored projects that this has been done.
It doesn't prevent someone who wants to from telling all in the coffee
shop, but it would probably satisfy the State Department and have a
high probability of accomplishing what they want.
It is often argued that the Soviets can get all the information they
want by monitoring the literature. It isn't easy to tell to what
extent this is true. Certainly they are often behind in many areas,
and in many areas of computer science, their work is derivative - e.g.
putting on a Soviet computer systems done earlier in the U.S. Of course,
there is the possibility that derivative work is all we hear about.
Overall, it seems that restrictions have important but random effects.
Of course, the restrictions the Soviets place on travel by their own
people are far more effective than our restrictions, but the effects
are additive.
Stanford is imposing restrictions on access to information, because the
Center for Integrated Systems will give special access to the companies
that are paying for it. Even the Computer Forum gives special access to
its members. The CIS has been agonizing for the better part of a year
about how best to give privileges to the 12 companies.
All the above constitutes criticism of the misstatements by Roth and
Lieberman. There still remains the question of whether the particular
action of the State Department is reasonable. If I had to make a
decision myself, I'd want more information about what the Russians do,
how open they have been, and what the military applications are.
On the basis of my experience and the information contained in the
article, the State Department position seems a reasonable compromise.
In contrast to this, Stanford seems to be taking an absolutist position,
and its spokesmen are inventing facts to fit the desired conclusions.
Perhaps the position is a leftover from the 1960s.
Apart from cost-benefit considerations, there is a political issue.
The AP story will be summarized in the Soviet press
as follows: "In spite of the efforts of
the mad dog Reagan to whip up anti-Soviet hysteria over the false
issue of Poland, American scientists realize the leading role of
the Soviet Union in robotics as in other fields. They know the
Polish issue is false". For more restricted circles,
the assurance may be offered that Comrade Umnov is politically
mature and will make certain that he gets more than he gives.
Should anyone wish to be actually constructive about exchanges
with the Soviet Union, I would suggest the following:
1. Form an estimate of what the Russians are doing in certain fields,
e.g. robotics or computer science generally. Make a guess as to where
they are being open and where they are being secretive. The CIA, which
works for us you know, could be asked to help; it interviews many
emigrants from the Soviet Union and has other sources of information.
2. If we decide there is something we want to know or institutions we
think we would like to visit, the State Department or the National
Academy of Sciences can bargain on our behalf. Their bargaining will
be more effective if they get our co-operation. Our problem will be
that very few scientists will find the Soviet Union interesting enough
for long visits.
Maybe it isn't worth the trouble, however. Perhaps the effort
would be better spent on deciding what open Japanese publications to
translate.
Comments on the comments: KJK wants to be sure people won't get him
wrong. Let it be recorded that his views are orthodox.
TAW wants to know why the Government lets Umnov in at all if they want to
restrict what he learns. Often the Defense Department wants to restrict
a Soviet visitors or argues against letting him in. The
present case has the earmarks of one where the Defense Department opposed
the visit, and the State Department got a compromise. The State Department is
in the business of making agreements and generally prefers admitting them.
However, the State Department is often more aware than others of
bargaining considerations. For example, Americans were able to visit
Novosibirsk before any other foreigners including Soviet satellites in the
1960s while it was still a closed city, because the State Department had
mapped it on Pittsburgh and wouldn't let Soviets visit Pittsburgh until
Americans visited Novosibirsk. The National Academy of Sciences likes
exchanges. Individual American scientists like visits and don't much care
if the exchange is very uneven. In fact, I have found it good for the ego
to be on the informative side of such conversations. Only later do I
notice that I have learned almost nothing.
JMC - There is one consideration in favor of accepting grossly uneven
"exchanges" of information. Our openness and even naivete has a
psychological effect on Russians. There is some hope that it makes our
society look better, and some people believe this strongly. However,
emigres and dissidents who have come to dislike Soviet society are
very vehement against Western naivete. On balance, I'd bet that it's
to our disadvantage to look like suckers.
∂25-Jan-82 0201 JMC
To: LLW at SU-AI
Your remarks have been interpolated.
∂25-Jan-82 0548 JMC
To: ARK
CC: YOM
∂25-Jan-82 0300 ARK Yoram Moses request advisor change
∂25-Jan-82 0246 YOM Advisor
Hi Arthur!
Please change the listing of my advisor from Floyd to McCarthy.
Thanks, Yoram.
ARK - Is this OK with you?
JMC - Yes.
∂25-Jan-82 2120 JMC
To: ARK
ok
∂26-Jan-82 1034 JMC
To: FFL
Thanks Fran. Professor McCarthy has no further comment.
∂26-Jan-82 1309 JMC
To: atp.bledsoe at UTEXAS-20
I think we should write separate letters and will write mine.
∂27-Jan-82 1521 JMC
To: DCO at SU-AI
Please update your directory entry or plan to give present co-ordinates.
∂27-Jan-82 1526 JMC
To: llw at S1-A
Lowell:
I believe the people supported under that contract now mainly work
for LLL. What would be the possibility of either waiving that report or
getting the Livermore contract people to accept a one page report? Otherwise,
you could lend us Jeff for a few weeks.
John
∂25-Jan-82 1323 Betty Scott <CSD.BSCOTT at SU-SCORE> LLLContract 9628303
Date: 25 Jan 1982 1316-PST
From: Betty Scott <CSD.BSCOTT at SU-SCORE>
Subject: LLLContract 9628303
To: JMC at SU-AI, CSL.FB at SU-SCORE, CSL.JLH at SU-SCORE
cc: CSD.BScott at SU-SCORE
Proposal Entitled: "An Operating System and Memory Switch for the S-1
Computer," John McCarthy, P.I.
A final report on this contract was apparently due at the end of
December, 1979. It is difficult for me to determine whether John
McCarthy was actively engaged in the research, or whether Forest Baskett
and John Hennessy performed most of the research. Anyway, I am sending
each of you a copy of the proposal. The performance period was January 1,
1978, through December 31, 1979, and the funding totaled $304,472.
A final payment of approximately $3,500 is due Stanford from LLL, and they
won't pay until the report is submitted--15 copies of it. If the report
is not submitted LLL could conceivably disallow the entire contract,
leaving us holding a $304K "bag."
Would you please let me have a final report just as soon as possible. I
will have it duplicated and sent.
Thanks very much,
Betty
P.S. The LLL delay in notifying us of the above is apparently due to their
inability to locate their paperwork until recently--after hiring some
consultants to get their files in order. -- Thanks again.
-------
∂27-Jan-82 1816 JMC
To: jonl at MIT-MC
I didn't get your "Two little ...
∂29-Jan-82 1532 JMC
To: morris at PARC-MAXC
Just come to the course which will be Mon and Wed at 11 in Room 301 Margaret Jacks
Hall.
∂29-Jan-82 1533 JMC
To: oppen at PARC-MAXC
Well, I see you're still keeping your PARC location secret.
∂29-Jan-82 1535 JMC
To: BCM
mrc:<prolog>prolog at SCORE is the command.
∂29-Jan-82 1536 JMC
To: mccall at PARC-MAXC
Monday and Wed at 11 in 301 mjh.
∂29-Jan-82 1546 JMC
To: RPG
I don't know Arthur Norman, and I'm not inclined to give Denne an
account to be come acquainted with LISP and SAIL. If it were something
that exists here uniquely like EKL, I would do it, but there
are LISPs and Algol-like languages in England and presumably at
Cambridge, and the differences aren't worth fussing about.
∂29-Jan-82 2144 JMC
To: BCM
It has been removed from the disk; watch for its return.
∂29-Jan-82 2150 JMC report and next friday
To: LLW at SU-AI
I plan to spend next Friday at LLL and will try to catch Jeff and write
something then.
∂29-Jan-82 2207 JMC
To: SGF
Are you in a position to get your own terminal yet?
∂30-Jan-82 0032 JMC report to LLL
To: csd.scott at SU-SCORE
Well, it seems we have to do something, but one or two pages may do it,
and I'm going to Livermore next Friday anyway. Can you send me a list of
who was paid by the contract, so I can report what they did?
∂30-Jan-82 1711 JMC Antonio Porto
To: csd.hill at SU-SCORE
I have created a SAIL login for him AP. It should have one aliquot
and be charged to my ARPA. If he comes to see you about a SCORE
account, an aliquot can be charged also. He is a visiting scholar
from Portugal and will be teaching a course in Prolog.
∂30-Jan-82 2040 JMC
To: YOM
What progress are you making with your axioms for natural numbers?
∂30-Jan-82 2041 JMC
To: RJA
What are you up to these days?
∂31-Jan-82 1146 JMC
To: CLT
no problem
∂31-Jan-82 1722 JMC
To: feigenbaum at SUMEX-AIM
Do you think the following is ok for presentation to Lieberman along with
the AP story and the Stanford press release. In particular, can you
support it?
FACULTY STATEMENT ON SOVIET VISITORS
We are distressed by the situation described in the attached
Associated Press story. It gives the impression that Stanford takes an
attitude of total opposition towards the Government's efforts to restrict
technology transfer to the Soviet Union. Moreover, this attitude seems to
be supported by a number of untrue statements, namely (1) The Soviets are
ahead of the U.S. in robotics. (2) Robotics has no military applications.
(3) There are no possible restrictions on Soviet visitors that won't
disrupt the functioning of the University. We are further distressed that
that when the Soviet Consulate in San Francisco reported on the situation
to Moscow, this report probably re-inforced the idea in Moscow that
American protests about Poland and Afghanistan were a governmental sham
without any support in the institutions of the country such as
universities.
We believe that Stanford should adopt a policy toward Soviet
visitors that balances the following considerations:
1. Preserving the orderly functioning of the academic processes of
research, publication and teaching and avoiding the creation of a security
atmosphere.
2. Showing the visitors a society more open than their own.
3. Minimizing transfer of militarily useful technology so as to
minimize our own defense costs; perhaps we can help obviate the need for a
return to the draft. Ways of implementing this without security measures
need study.
4. Increasing reciprocity so that access to Soviet research
becomes available to Americans to the same extent that our research
activities are accessible to them. This involves some degree of support
to the bargaining efforts on our behalf conducted by the State Department
and the National Academy of Sciences.
5. Maintaining good relations with the State Department, the
Defense Department and other Government agencies.
Since these goals are not entirely co-incident, some degree of
compromise is needed.
We suggest that a committee be appointed to develop a policy. If
an interim statement clarifying the impression given by the Stanford press
release could be made, it might help mitigate the unfortunate signal that
has been sent to Moscow.
∂31-Jan-82 1724 JMC
To: llw at S1-A
You may be interested that some faculty are meeting with Lieberman on
the subject (he doesn't want us to get the wrong impression). The following
is my attempt to be diplomatic.
FACULTY STATEMENT ON SOVIET VISITORS
We are distressed by the situation described in the attached
Associated Press story. It gives the impression that Stanford takes an
attitude of total opposition towards the Government's efforts to restrict
technology transfer to the Soviet Union. Moreover, this attitude seems to
be supported by a number of untrue statements, namely (1) The Soviets are
ahead of the U.S. in robotics. (2) Robotics has no military applications.
(3) There are no possible restrictions on Soviet visitors that won't
disrupt the functioning of the University. We are further distressed that
that when the Soviet Consulate in San Francisco reported on the situation
to Moscow, this report probably re-inforced the idea in Moscow that
American protests about Poland and Afghanistan were a governmental sham
without any support in the institutions of the country such as
universities.
We believe that Stanford should adopt a policy toward Soviet
visitors that balances the following considerations:
1. Preserving the orderly functioning of the academic processes of
research, publication and teaching and avoiding the creation of a security
atmosphere.
2. Showing the visitors a society more open than their own.
3. Minimizing transfer of militarily useful technology so as to
minimize our own defense costs; perhaps we can help obviate the need for a
return to the draft. Ways of implementing this without security measures
need study.
4. Increasing reciprocity so that access to Soviet research
becomes available to Americans to the same extent that our research
activities are accessible to them. This involves some degree of support
to the bargaining efforts on our behalf conducted by the State Department
and the National Academy of Sciences.
5. Maintaining good relations with the State Department, the
Defense Department and other Government agencies.
Since these goals are not entirely co-incident, some degree of
compromise is needed.
We suggest that a committee be appointed to develop a policy. If
an interim statement clarifying the impression given by the Stanford press
release could be made, it might help mitigate the unfortunate signal that
has been sent to Moscow.
∂31-Jan-82 1925 JMC
To: TOB
The current version of the statement is page 3 of STATE[W82,JMC].
∂31-Jan-82 1952 JMC
To: FFL
Please scribe STATE[F82,JMC] after centering the title, etc.
∂01-Feb-82 2320 JMC
To: INGALLS at PARC-MAXC
Nothing was distributed. I don't know how the department feels about
free loaders. The instructor, Antonio Porto, is AP@SU-AI.
∂01-Feb-82 2322 JMC
To: admin.mrc at SU-SCORE
Here are the files that should be put on the disk. OLDPROLOG should be the
file accessed as prolog. There are two copies of each file on the tape
just to be sure.
∂01-Feb-82 2316 David Warren <WARREN at SRI-AI> Tape Listing Itself
Date: 1 Feb 1982 1724-PST
From: David Warren <WARREN at SRI-AI>
Subject: Tape Listing Itself
To: JMC at SU-AI
DUMPER tape # 1, documentation, Monday, 1-Feb-82 1649
file last write size (pages) checksum
<PROLOG>READ.ME.2 12-Jul-81 2045 1
<PROLOG>PROLOG.DOC.1 12-Jul-81 2037 58
<PROLOG>GUIDE3.MEM.1 12-Jul-81 2036 36
<PROLOG>DEBUG.MEM.1 12-Jul-81 2036 22
<PROLOG>PROLOG.HLP.1 12-Jul-81 2039 1
<PROLOG>TUTORI.LPT.1 12-Jul-81 2045 8
<PROLOG>TUTORI.PL.1 12-Jul-81 2045 3
DUMPER tape # 1, oldprolog, Monday, 1-Feb-82 1650
file last write size (pages) checksum
<WARREN>PROLOG.EXE.26 31-Oct-81 1050 118
DUMPER tape # 1, newprolog, Monday, 1-Feb-82 1651
file last write size (pages) checksum
<PROLOG>PROLOG.EXE.1 12-Jul-81 2037 115
<PROLOG>PROLOG.EXE.26 16-Dec-81 0908 116
<PROLOG>PROLOG.EXE.28 21-Dec-81 1615 116
DUMPER tape # 1, documentation1, Monday, 1-Feb-82 1652
file last write size (pages) checksum
<PROLOG>READ.ME.2 12-Jul-81 2045 1
<PROLOG>PROLOG.DOC.1 12-Jul-81 2037 58
<PROLOG>GUIDE3.MEM.1 12-Jul-81 2036 36
<PROLOG>DEBUG.MEM.1 12-Jul-81 2036 22
<PROLOG>PROLOG.HLP.1 12-Jul-81 2039 1
<PROLOG>TUTORI.LPT.1 12-Jul-81 2045 8
<PROLOG>TUTORI.PL.1 12-Jul-81 2045 3
DUMPER tape # 1, oldprolog1, Monday, 1-Feb-82 1652
file last write size (pages) checksum
<WARREN>PROLOG.EXE.26 31-Oct-81 1050 118
DUMPER tape # 1, newprolog1, Monday, 1-Feb-82 1653
file last write size (pages) checksum
<PROLOG>PROLOG.EXE.1 12-Jul-81 2037 115
<PROLOG>PROLOG.EXE.26 16-Dec-81 0908 116
<PROLOG>PROLOG.EXE.28 21-Dec-81 1615 116
End of tape.
!
-------
∂01-Feb-82 2325 JMC
To: oppen at PARC-MAXC
Sorry. What set it off was that my secretary was asked by someone how to
find you, and I sent you a message without thinking about looking at
a PLAN file.
∂01-Feb-82 2329 JMC Tops-20 Prolog Tape
To: admin.mrc at SU-SCORE
∂01-Feb-82 2316 David Warren <WARREN at SRI-AI> Tops-20 Prolog Tape
Date: 1 Feb 1982 1722-PST
From: David Warren <WARREN at SRI-AI>
Subject: Tops-20 Prolog Tape
To: JMC at SU-AI
cc: Warren at SRI-AI
John, The listing of the tape will be in the following message.
Basically, you need the PROLOG.EXE from the save set "oldprolog"
for a reliable version. This version, however, only allows
Tops-10 style filenames. The version of PROLOG.EXE.28 in the
save set "newprolog" is the very latest version which permits
full Tops-20 filenames but which has a number of (minor) bugs
and snags. The first save set, "documentation", contains all
the current documentation. The tape contains a duplicate
copy of each save set for good measure. A catalog of the
various files follows. -- David.
; Files needed to use DEC-20 Prolog (see READ.ME):
<WARREN>PROLOG.EXE ; Prolog interpreter + compiler
<PROLOG>READ.ME ; Fuller description of these files
PROLOG.DOC ; User's Guide
GUIDE3.MEM ; Guide to Version 3
DEBUG.MEM ; Guide to debugging facilities
PROLOG.HLP ; Prolog help file
TUTORI.LPT ; Short tutorial on Prolog
TUTORI.PL ; Examples for the tutorial
-------
∂02-Feb-82 0036 JMC
To: ARK
Don't you have Kurt Konolige listed as my thesis advisee?
∂02-Feb-82 0144 JMC
To: ARK
He is my thesis advisee.
∂02-Feb-82 0149 JMC
To: konolige at SRI-AI
How goes the thesis?
∂02-Feb-82 1020 JMC
To: kay at PARC-MAXC
No problem. Monday and Wednesday at 11, 301 Margaret Jacks.
∂02-Feb-82 1048 JMC
To: asprey at PARC-MAXC
It meets at 11am on Mondays and Wednesdays. First meeting has occurred.
∂02-Feb-82 1620 JMC
To: konolige at SRI-AI
OK. It's "Feferman" by the way. I need to see you before Gray Tuesday,
so I can say something other than, "He seems to still exist" when they
ask, but I don't need to press you right now.
∂02-Feb-82 1717 JMC
To: FFL
Check that we are no longer paying Richard Anderson who is with HPP.
∂02-Feb-82 1753 JMC
To: JJW, AP
I have moved prolog.dis to 1,ap from 1,jjw.
∂02-Feb-82 1755 JMC
To: "@PROLOG.DIS[1,AP]" at SU-AI
<prolog>prolog enters prolog on SCORE.
∂02-Feb-82 1841 JMC
To: FFL
Please find out if Ms. Meng Lee has applied to CSD for admission to PhD program.
∂02-Feb-82 2054 JMC
To: llw at S1-A
Well, we didn't get to first base with the Vice-provost, who took an
extremely defensive attitude. We'll try the President next.
∂02-Feb-82 2056 JMC
To: RPG
I don't know how great the chances of success are. For example, I
don't know Terry's opinion. Only the names of references need be
gathered.
∂03-Feb-82 1845 JMC sign
To: BS at SU-AI
We need a sign to put on the Xerox machine saying, "Xerox down, IBM and
Kodak have been called.
∂03-Feb-82 2335 JMC
To: pourne at MIT-MC
I think you deserve considerable credit for this result.
a013 2242 03 Feb 82
PM-Space Budget,450
Reagan OKs Planet Program Money
By HOWARD BENEDICT
AP Aerospace Writer
WASHINGTON (AP) - Overriding his fiscal advisers, President Reagan
is proposing in his new budget that much of the U.S. planetary
exploration program be kept alive.
Just three months ago, the Office of Management and Budget
recommended killing most deep space exploration projects in its drive
to cut federal spending.
The effort met strong opposition from scientific organizations and
congressmen on key space committees who took their case to the White
House.
As a result, Reagan has put money for several deep-space projects in
his fiscal 1983 budget. Included is $92.6 million to continue
development of the Jupiter-orbiting Galileo satellite; $21 million to
move ahead with several European nations on a joint sun-probe
mission; and money to maintain the deep space tracking network and to
allow the Voyager 2 spacecraft to travel on to Uranus and Neptune.
If OMB had succeeded in dropping the planetary programs, it would
have meant the loss of about 1,200 jobs at NASA's Jet Propulsion
Laboratory in Pasadena, Calif., which manages the deep-space efforts.
The president is to submit his full 1983 budget to Congress on
Monday. The Associated Press on Wednesday obtained an advance copy of
the National Aeronautics and Space Administration section.
Reagan proposes total NASA spending of $6.6 billion. Factoring in
inflation, that's about equal to 1982's $5.98 billion.
A major share of the 1983 budget, $1.7 billion, is for the manned
space shuttle, which is to complete its test program with flights in
March and July and to start cargo-carrying operational missions in
November. A second shuttle, the Challenger, is to join the Columbia at
Cape Canaveral, Fla., in June.
Another $1.7 billion is earmarked for space flight operations,
mainly for the shuttle.
Reagan is a strong supporter of the reusable spaceship, primarily
because of its potential military applications.
NASA didn't get all it wanted in deep space. It lost a Venus orbiter
and a probe to Halley's comet. Other projects were scaled down or
stretched out. But, considering the bleak outlook a couple months ago,
it came out pretty well.
The proposed budget also includes $137.5 million, $61.7 million and
$34.5 million, respectively, for continued development of three major
orbiting satellites: a space telescope, an advanced Landsat Earth
Resources payload and a gamma ray observatory; $100 million for
construction of facilities; and $1.17 billion for research and program
management.
Aeronautical research dipped slightly, from $233 million in 1982, to
$232 million. Heaviest cuts were in technology for transport aircraft
and advanced propulsion.
The budget projects total NASA employment of 21,219 by the Sept. 30,
1983, the end of fiscal 1983. This would be a drop of more than 400
from the projected 1982 figure of 21,652.
ap-ny-02-04 0137EST
***************
∂03-Feb-82 2355 JMC
To: FFL
umnov[w82,jmc] needs moving the @newpage and scribing.
∂04-Feb-82 0157 JMC
To: RPG
Do you have an opinion of the 3 garbage collection papers?
∂04-Feb-82 1025 JMC
To: asprey at PARC-MAXC
If you know something or will do some reading, it may be
worthwhile to come. The class is in 301 Margaret Jacks Hall.
∂04-Feb-82 1032 JMC
To: csd.armer at SU-SCORE
Bolles is at SRI, Cartwright is at Rice and Wagner is at Intel.
Put me on your mailing list as JMC@SAIL, although mail addressed
as you did is automatically forwarded.
∂04-Feb-82 1511 JMC
To: TOB
Well, I sicked the guy on you. Pournelle, who sponsored the meeting,
is a prolific and well known science fiction writer. He has also
written on defense policy, etc. in collaboration with a now retired
Hoover professor. He is also a Republican politician. His Citizen's
Advisory Council on Space Policy, in which I have taken part, together
with the L-5 society, deserves a good part of the credit for Reagan's
decision, reported last night, to save the planetary program of NASA
against Stockman's advice.. In general, I think Pournelle's enterprises,
which involve both scientists and space fans, are surprisingly effective.
It would be worthwhile for you to do some thinking about robotics in
space. Except that I will be in France, I would have accepted the
invitation myself.
∂04-Feb-82 1624 JMC
To: bscott at SU-SCORE
I guess I need to know the amounts paid to the different people. I didn't
realize that I was paid at all on the contract.
∂04-Feb-82 2325 JMC
To: MMS
Lists of dotted pairs are what was intended.
∂05-Feb-82 0036 JMC
To: TOB at SU-AI
CC: csd.golub at SU-SCORE
I think there is a need for "critical mass" only in a few areas. Also
the adjunct faculty in language will be justly offended by your misinformation
that they are mainly native speakers - they aren't. At least the
adjunct professor in charge of teaching French has a non-French name
and is the author of a series of textbooks.
∂05-Feb-82 0959 JMC
To: FFL
I'm going to Livermore today.
∂05-Feb-82 1719 JMC
To: JJW
CC: ZM
I will in Europe in April and May, but I don't have to be present.
Alternatively, the exam could be postponed till Zohar returns again,
which I believe is in September.
∂06-Feb-82 2340 JMC
To: CLT
Please look at BOYER[W82,JMC]. It requires extension and improvement.
∂07-Feb-82 0223 JMC
To: morris at PARC-MAXC
Can you send me a copy of your "Real programming in functional languages?
∂07-Feb-82 1047 JMC
To: CLT
Sorry.
∂07-Feb-82 2302 JMC
To: feigenbaum at SUMEX-AIM, pratt at SHASTA
In reply to Ed's question about the M8l60.
It doesn't find data as quickly as a disk, since it has to spin tape,
but of course it can provide very large file backup. I fear it's too
expensive for us, but it would certainly be nice to have in the long
run. In any case, it's best to postpone serious discussion till we have
our money.
∂08-Feb-82 0111 JMC
To: REM
I suggest you send this message to RPG who is the MACLISP guru here.
∂09-Feb-82 0404 JMC
To: gxg at SRI-KL
An account for Herbert Stoyan has been created. The account name
is HST (account names cannot have more than three letters), and
the password is the one requested.
∂09-Feb-82 0404 JMC account for Herbert Stoyan
To: FFL
I have created a login for him with the name HST. Please get it entered
in the records as a minimal guest account. His address etc. is in my
PHON file.
∂09-Feb-82 0407 JMC
To: FFL
Also please mail him "Short Waits" and "Essential E".
∂09-Feb-82 0408 JMC Stoyan account
To: GXG
An account has been created for Herbert Stoyan with the name HST (account
names cannot exceed 3 letters) and with the password requested. I have
arranged for him to be sent "Short Waits" and "Essential E".
∂09-Feb-82 0410 JMC
To: FFL
The note to Stoyan should mention that his account name is HST.
∂09-Feb-82 0944 JMC
To: FFL
A guest account.
∂09-Feb-82 1149 JMC
To: ullman at DIABLO
I'll do it if you like, but ...
∂09-Feb-82 1150 JMC consulting bill
To: pjb at S1-A
Is it correct that I bill for travel between Stanford and LLL? If so,
what is the standard mileage and rate?
∂10-Feb-82 2345 JMC
To: HST
The following two files contain some comments on your translation.
Maybe I'll have further comments later, but I've delayed so long, that
I'd better pass the papers on to Luckham now.
stoyan.6[let,jmc] comments on 2nd part of Stoyan's translation
stoyan[f81,jmc] comments on 1st part of Stoyan's translation
∂11-Feb-82 1512 JMC
To: FFL
CERCON.1
∂11-Feb-82 1704 JMC
To: RPG
I suggest you offer the following as references, suitably combining people
who will give good references and those who will be asked anyway.
David Waltz (Univ of Illinois)!
Terry Winograd (Stanford)?
John McCarthy (Stanford)?
Guy L. Steele Jr (CMU)!
Lowell Wood (LLL)!
Nils J. Nilsson (SRI)!
∂12-Feb-82 1327 JMC
To: CLT
San Jose - Armenian
The Armenian Gourmet 408 732-3910
921 E. Duane, Sunnyvale
Hours: lunch M-F 11:30-1:30, dinner W-Sat 5:00-8:30; closed Sun;
accepts MC, Visa.
Good Armenian cuisine at very reasonable prices; somewhat homey
atmosphere. Try the excellent lamb and beef saute. Good hummus, too
[BVM-9/78].
∂12-Feb-82 1455 JMC guest account for Martin Davis
To: FFL
I have created an account MDD for him. Please fill out the form for
him. His address is in PHON.
∂12-Feb-82 1457 JMC account
To: davism.acf1 at NYU
The name is MDD and the initial password is PRIZE. You can change
the password if you want to.
∂12-Feb-82 2001 JMC message files
To: ME
Is it normal for someone coming from utexas-20 to be logged into
the message file area?
∂14-Feb-82 0119 JMC
To: FFL
naraya.1
∂14-Feb-82 1326 JMC
To: REM
Both would be good, but I supposet the .xgp to .pre is more urgent.
∂14-Feb-82 1328 JMC
To: FFL
How is the spindling coming? A purge is threatened.
∂14-Feb-82 1344 JMC
To: CLT
Yes. I was merely restless. By the way, see TUNER[1,JMC].
∂14-Feb-82 1659 JMC
To: csd.ullman at SU-SCORE
I suggest you phone Kahn about proposal.
∂14-Feb-82 2039 JMC
To: JMM
Lets get together tomorrow to finish off the grades for the NRO CS206.
∂14-Feb-82 2055 JMC
To: pjb at S1-A
I actually took a third route by highway 237 which I suppose is 50
miles. As to the dates of my summer travel, I think it would be best
to make it the 12 days for which they paid me. As you may recall, I
told you that they over paid me, so I worked off the extra days during
the Fall, but I don't want to overload the accounting system. It seems
to me that my first day was September 21 and got paid for the next 12
working days. Thanks in advance for your efforts to make it all come
out even.
∂14-Feb-82 2150 JMC
To: FFL
mccart.1
∂15-Feb-82 0128 JMC
To: REM
Not that I know of.
∂15-Feb-82 1351 JMC
To: ME
Would it be difficult to modify Xpo so that one option would be to dover
or xgp the file whose name was found?
∂16-Feb-82 1128 JMC Davis address
To: FFL
We'll need to look further. Aha! He has applied for a professorship
in the Department. Irmgild will have his correct address, which should
be entered in PHON.
∂16-Feb-82 2118 JMC
To: HST
The computer has been sick in the last few days. It's usually more reliable.
∂17-Feb-82 0047 JMC
To: RPG
Late tomorrow afternoon would be a good time.
∂17-Feb-82 0103 JMC
To: ME
What is a .spi file?
∂17-Feb-82 0138 JMC
To: pourne at MIT-MC
Thanks for the quote.
∂17-Feb-82 0145 JMC
To: pourne at MIT-MC
CC: llw at S1-A
Somehow I missed your previous message. A conference call is ok with
me if it's ok with Lowell. Somehow I think we haven't done enough work
yet, but if your friends won't be put off by a preliminary discussion,
that's fine with me. Have you reached Lowell or should I try.
After noon or evening is fine with me for call.
∂17-Feb-82 0149 JMC
To: ME
There were three in 1,jmc. I deleted them, but then got curious and
restored two. They have some random characters (images of control chars
I suppose) and REM. Have a look if you like. They're unprotected.
I thought they might have something to do with SPYing.
∂17-Feb-82 0155 JMC
To: pourne at MIT-MC
CC: llw at S1-A
I have no objection to your exploring the matter discreetly as you are
planning to do. Lowell may have more information for you.
∂17-Feb-82 0241 JMC
To: ME
Now I know. They're spindle files.
∂17-Feb-82 0307 JMC
To: FFL
Can you find a paper by Alan Pasternak on reducing dependence on imported oil?
∂17-Feb-82 0312 JMC
To: JPM
Let's talk about your SAIL account some pm.
∂17-Feb-82 0348 JMC film in space
To: pourne at MIT-MC
Is anyone planning to make a feature film set in low earth orbit and
with at least part of it shot in low earth orbit?
∂17-Feb-82 0350 JMC
To: pourne at MIT-MC
It needs a suitable story, of course?
∂17-Feb-82 1553 JMC
To: AP
states.pr[e81,jmc]
∂18-Feb-82 1841 JMC
To: CLT
Sten-Ake says you should resend your message to Ken Kahn.
∂18-Feb-82 2241 JMC
To: chandrasekaran at RUTGERS
I have discussed the prospects once with Dr. Teller, and we have to talk
again. Incidentally, I have been looking at the EPRI Journal and noted
Simulation Methods for Nuclear Power Systems.
WS-81-212, $32.00
"This report constitutes the proceedings of a conference on nuclear power
plant simulation that was sponsored by NRC and EPRI in Tucson, Arizona in
January 1981. Papers from government, industry, natuional laboratories,
and universities are presented; the topics covered include simulation
needs, simulator design and performance, engineering simulations,
model development methods and verification. EPRI Project Manager
P.G. Bailey
Electric Power Research Institute, P. O. Box 10412, Palo Alto,
CA 94303.
It seems to me that your efforts to get funding would meet greater
success if your proposals mentioned such efforts and explained how
your proposals supplemented what was being done. Second, it occurs
to me that EPRI is a possible source of funding.
∂18-Feb-82 2309 JMC more on simulation
To: chandrasekaran at RUTGERS
I talked with Jerry Tiemann from General Electric, Schenectady, who
told me that he didn't know of simulations being used or proposed
in an operating environment, but he didn't claim to know much. He
asked for your name, which I gave, but I suspect you're more likely
to hear from him about his efforts to recruit PhDs than about giving
your project support.
∂20-Feb-82 1432 JMC
To: FFL
meltze.1
∂20-Feb-82 1432 JMC disjunctive knowledge
To: bmoore at SRI-AI
As you pointed out in your Master's thesis, ascribing world models
to people doesn't properly represent knowledge of disjunction.
1. Did this criticism apply also to the work of Schank and Abelson?
2. Are the formalisms generally used in AI today (not referring
to your own) adequate in this respect?
I'm replying to a letter from Bernard Meltzer citicizing my "First order
theories of individual concepts and propositions" and claiming that
Schank and Abelson solve all problems. Well he doesn't actually say that.
∂20-Feb-82 1521 JMC commute mileage for consultant
To: pjb at S1-A
I just received Livermore check 046842 for 270.00 labelled
"compute mileage for consultant". If I don't hear otherwise,
I will assume it is for my summer travel when I wasn't strictly
speaking a consultant but a summer employee. Otherwise, there is
a mistake since I have submitted only one consultant bill.
∂20-Feb-82 1603 JMC
To: FFL
scienc.1
∂20-Feb-82 1629 JMC
To: FFL
yamada.2
∂20-Feb-82 1634 JMC
To: FFL
kowals.4
∂20-Feb-82 1657 JMC letter for Teller
To: FFL
CHANDR.RE1 is a draft letter for Dr. Teller to modify and send if he
chooses. It doesn't need to be SCRIBED, but should be printed neatly.
It needs Chandrasekaran's initials in the beginning and his complete
address as the last sentence.
∂20-Feb-82 1718 JMC
To: FFL
marcon.1
∂20-Feb-82 1835 JMC
To: CLT
To print on canon ordinary, i.e. non press, files, incant as follows:
r makimp
[This asks for a file name and converts the file to .imp form].
then
do cprint[can,sys]
[There are sometimes mysterious failure messages].
∂20-Feb-82 2142 JMC
To: RPG
Come to think of it I won't be in Mon or Wed but will be in Tues.
∂20-Feb-82 2148 JMC iii trips Monday and Wednesday
To: CLT, FFL
I go to L.A. Monday (Feb 22) morning returning monday afternoon
and likewise Wednesday morning returning wednesday afternoon.
∂20-Feb-82 2154 JMC L.A. trip
To: CLT
The Monday meeting is Monday rather than Tuesday so I can meet my class.
I have nothing to do in L.A., and I don't mind the trips, and I want to
hear about Boyer and Moore.
∂20-Feb-82 2214 JMC
To: CLT
Oho! Well, most programs aren't tested sufficiently with input where
they aren't expected to succeed. I suppose there may be similar
weaknesses in EKL and FOL. For example, FOL was working for a year
or two before I made it do Russell's paradox.
∂21-Feb-82 0007 JMC
To: csd.golub at SU-SCORE
I hope you read Shapiro's letter to see just how ambitious he is. I had forgotten.
∂21-Feb-82 0105 JMC
To: FFL
shapir.>
∂21-Feb-82 0157 JMC
To: AP
I'll miss your class Monday and Wednesday. I must be in Los Angeles both days.
∂21-Feb-82 1409 JMC
To: RPG
If you join the faculty, you'll have to change your PLAN.
∂22-Feb-82 0041 JMC
To: FFL
I need to find the Stoyan papers and pass them on to Dave Luckham.
∂22-Feb-82 0403 JMC
To: FFL
Please scribe-decorate Shapiro letter, but I'll revise it some more.
∂23-Feb-82 1644 JMC
To: FFL
wang.1
∂23-Feb-82 2246 JMC
To: FFL
bloom.1 is a letter that refers to energy.let[w82,jmc]. Please scribe both.
∂24-Feb-82 0043 JMC
To: TOB at SU-AI, feigenbaum at SUMEX-AIM
Partly against my better judgment, I have allowed myself to
be interviewed by the Stanford Daily. I don't think I said much
wrong, and the reporter called me to check, but I still can't be
sure how it will come out. I suppose my main motivation was that
I had better let something happen, since I have been prevented by
other matters from taking any action along the lines we previously
discussed.
I told the Daily reporter to call the rest of you who
took part in the meeting with Dean Lieberman, but, of course, I
can't be sure that he will do so.
Here's hoping it turns out all right.
Sincerely,
∂24-Feb-82 0044 JMC
To: FFL
Please print FLORY.1 and take it to chemistry department.
∂25-Feb-82 1110 JMC
To: csd.armer at SU-SCORE
In the past this has been settled by the Department Chairman or Denny
together with the individual faculty members. Next year as usual, I
will have 50 percent support and propose to teach two courses, CS206
in the Fall and CS226 in the Winter. I alternate CS226 (Epistemology
of AI) and CS258 (mathemtical theory of computation) in alternate years.
Since a given faculty member doesn't always teach in the same area, I
don't see how your scheme is practical. It would be better if you or
Gene spoke individually to the faculty members. I have no idea whom
your are expecting me to call to a meeting.
∂25-Feb-82 1421 JMC
To: konolige at SRI-AI
The following variant of this puzzle that does not require an initial bound
on the numbers was given in Martin Gardner's Mathematical Games column
in %2Scientific American%1, 1980 June:
S: I see no way you can determine my sum.
P(after a delay): That didn't help me. I still don't know the sum.
S(after delay): Now I know the product.
Assuming the Goldbach conjecture, the numbers must be 5 and 6.
- due to Barry Wolk, University of Manitoba
∂25-Feb-82 1430 JMC Kuck paper
To: bobrow at PARC-MAXC
The first page is badly reproduced. Can you send me another copy of it?
∂25-Feb-82 1743 JMC
To: AP
fred.pr[e81,jmc]
∂25-Feb-82 1748 JMC
To: feigenbaum at SUMEX-AIM, csd.ullman at SU-SCORE
CC: csd.armer at SU-SCORE
Next year I plan to teach CS206 (Recursive programming and proving) in
Fall and CS226 (Epistemological problems of artificial intelligence) in
Winter and nothing in Spring (using 50% support). Should this offer any
problems for planning the AI offerings or the systems offerings (assuming
CS206 to fall under systems), please let me know.
∂25-Feb-82 1751 JMC
To: RWW
CC: FFL
Traditionally, graduate students get zapped effective the end of
a quarter, and I suppose you mean to zap Kasturia at the end of
Winter quarter.
∂25-Feb-82 1753 JMC
To: RWW
CC: FFL
Come to think of it, I guess RA's are rarely zapped except at the end of an
academic year. Perhaps you should discuss earlier zapping of Kasturia
with Gene or postpone his zapping till end of Spring.
∂25-Feb-82 2122 JMC
To: perlis.ee at UDEL-RELAY
The Department is looking for people, and I will refer your resume to
Professor Golub, the Department Chairman, when it comes. I will write
you if I have a definite reaction to the reprints.
∂25-Feb-82 2123 JMC
To: RWW
From your last two messages, it seems to me that you know what you are doing
and have consulted sufficiently with the authorities.
∂26-Feb-82 0016 JMC
To: HST
There was a Richard Watson at SRI, but he left to work for some
company, maybe an oil company. I don't think he had any relation
to LISP, but he may have had some relation to the chemical structural
questions which dendral tackles. I believe he wrote a book on
time-sharing some time in the sixties. Of course, it may be a
different Richard Watson, so you had better be sure it's the same
one. I think perhaps he was in the Computer Science Department
at Stanford for a year or two.
∂26-Feb-82 1202 JMC
To: csd.irmgild at SU-SCORE
I certainly want to talk to Shapiro. Late afternoon Thursday is best for me.
∂26-Feb-82 1337 JMC
To: CLT
ok, let's do it.
∂26-Feb-82 1637 JMC
To: csd.jake at SU-SCORE
Ressmeyer, Roger 1230 Grant Ave. #574↓San Francisco, CA 94133∞
photographer who took NYT pictures dec 1980, 956-1205
∂27-Feb-82 1536 JMC
To: llw at S1-A
Your reaction to ELSALV[W82,JMC] would be welcome.
∂27-Feb-82 1759 JMC
To: YM
It will be open, and it will be announced. I believe it will be in
Feigenbaum's SIGLUNCH.
∂28-Feb-82 1433 JMC
To: JMM
I have finished the exams. Let's go through the grading algorithm.
∂28-Feb-82 1820 JMC Hayes quote
To: csd.smith at SU-SCORE
What is the precise reference of "the way to control reasoning is to reason
about control"?
∂28-Feb-82 1832 JMC Wolfram
To: csd.golub at SU-SCORE
I have looked briefly at the SMP manual. It is rather poorly organized.
On the basis of looking at it, and having heard his lecture, I
see no reason to disagree with Tony Hearn's conclusion that it's another
algebraic computation system. I would not offer SLAC an enthusiastic
reaction but would offer him a courtesy appointment if SLAC appoints him.
My advice to them would be that tenure is grossly premature.
∂28-Feb-82 1848 JMC
To: FFL
bean.1 and yovits.1
∂01-Mar-82 0300 JMC meta-cognition: reasoning about knowledge
To: csd.lenat at SU-SCORE
I take it that your contribution to this paper was the odd numbered
pages; at least this is all I received. Could you send me a fresh
copy? The odd pages seem intriguing.
∂01-Mar-82 1156 JMC
To: pratt at SHASTA
To tell the truth, I don't see the import of the categorization either,
and we didn't have such a formal categorization previously. I suspect
it's merely an artifact of Paul Armer's retailing the task of deciding
what courses are to be given next year. The course seems like an advanced
one, and therefore should be whatever you decide. While the elementary
courses need to be fitted together, advanced courses should be at the
discretion of the professor.
∂01-Mar-82 1228 JMC
To: warren at SRI-AI
Unfortunately I will be abroad from the end of March till about the
20th of June, and I can't think of anyone else who would be interested.
If she will still be in the U.S. after June 20, I think we could pay
part of the expenses.
∂03-Mar-82 1546 JMC
To: FFL
hofst.1,hirsch.1
∂03-Mar-82 1551 JMC
To: FFL
pub ijcai.rep[f75,jmc] and include it with letter to Hirsch.
∂03-Mar-82 1700 JMC Common LISP discussion
To: RPG
Is it technically feasible to put a file on the mailing list for the
Common Lisp discussion messages. It would then be convenient for me
to check it with CKSUM rather than have them in my already large general
mail file. If it were feasible, then it might also be feasible for
several people as SAIL to share the same discussion file.
∂03-Mar-82 1812 JMC
To: RPG
Yes. Please take me off the list.
∂05-Mar-82 0112 JMC
To: cl.moore at UTEXAS-20, cl.boyer at UTEXAS-20
I forgot to answer your request for a syllabus of my course in
mathematical theory of computation. Anyway there is, alas, no
syllabus. However, covered the following:
1. representation of recursive programs by sentences in first order
logic
2. A Scott axiomatization of typed lambda calculus. (1969 CUCH, ISWIM and
OWHY paper - not the models of lambda calculus that came later.
3. Use of Ketonen's EKL interactive theorem prover.
4. The Boyer-Moore system. Because of problems in getting enough
disk space and some laziness, both on my part and on the students'
part, they never got to run any problems on the computer.
5. My Elephant formalism for representing sequential programs as
sentences in first order logic.
6. Manna's method (his PhD thesis) for proving total correctness.
7. Axiomatization of flow charts by sentences of first order logic
(multiple entries and exits).
8. Formalization of dirty LISP, including rplaca s and imbedded
setq s.
This doesn't amount to a syllabus, because I don't assert that this
is what should be covered, and it may not represent what I'll do
next month in Marseille or what I'll do when I teach the course again
in two years.
∂05-Mar-82 0114 JMC
To: csd.walker at SU-SCORE
Weening retaking the mtc qual has been until Zohar and I are next
both in town - probably in September.
∂05-Mar-82 0318 JMC dinner with Ehud Shapiro
To: csd.geneserth at SU-SCORE, csd.lenat at SU-SCORE, VRP at SU-AI,
buchanan at SUMEX-AIM, RWW at SU-AI
Are any of you interested in having dinner with him tonight (Friday)?
∂05-Mar-82 1416 JMC
To: llw at S1-A
This is a draft one page report. If it is not likely to be satisfactory,
please let me know what improvements are required.
The period covered is up through 1979, and the people paid significant
amounts on the contract were Forest Baskett, Les Earnest, Martin Frost,
John Hennessy, Marc Lebrun, Ted Panofsky, Armando Rodriguez, Jeffrey Rubin
and Arthur Samuel.
report.lll[w82,jmc] Final report for LLL contract on S-1
During this period the Stanford Artificial Intelligence Laboratory
worked on many aspects of the S-1 project. These included
1. The first architecture manual and various features of the architecture -
Jeffrey Rubin, Ted Panofsky, Martin Frost, Marc Lebrun
2. Simulator for S-1 on PDP-10, simulator for PDP-10 on S-1, simulator
for YUK-7 on S-1 -
Jeffrey Rubin
3. Design and debugging of Pascal compilers and intermediate languages
on the S-1 -
Arthur Samuel, Forest Baskett, Armando Rodriguez
4. Memory switch for S-1 -
Ted Panofsky
The results of these efforts are contained in various S-1 documentation
where they have been merged with results of efforts after the project
moved entirely to Livermore.
There follows some detail of the software work that may not have
been included in other reports include
Some specific programs developed include
UPASC - a translator for full standard Pascal to U-code (as described in
the U-code document[4]).
SOPU - U-code to S-1 code generator. Developed jointly by this group and
the Stanford group headed by Gio Wiederhold.
PASCAL* compiler - based on the UPASC compiler, it translates PASCAL to
standard U-code.
∂05-Mar-82 1418 JMC
To: FFL
Please put the part after ***** in report.lll[w82,jmc] into SCRIBE.
∂05-Mar-82 1419 JMC
To: BS at SU-AI
This is a draft one page report. If it is not likely to be satisfactory,
please let me know what improvements are required.
The period covered is up through 1979, and the people paid significant
amounts on the contract were Forest Baskett, Les Earnest, Martin Frost,
John Hennessy, Marc Lebrun, Ted Panofsky, Armando Rodriguez, Jeffrey Rubin
and Arthur Samuel.
report.lll[w82,jmc] Final report for LLL contract on S-1
*****
During this period the Stanford Artificial Intelligence Laboratory
worked on many aspects of the S-1 project. These included
1. The first architecture manual and various features of the architecture -
Jeffrey Rubin, Ted Panofsky, Martin Frost, Marc Lebrun
2. Simulator for S-1 on PDP-10, simulator for PDP-10 on S-1, simulator
for YUK-7 on S-1 -
Jeffrey Rubin
3. Design and debugging of Pascal compilers and intermediate languages
on the S-1 -
Arthur Samuel, Forest Baskett, Armando Rodriguez
4. Memory switch for S-1 -
Ted Panofsky
The results of these efforts are contained in various S-1 documentation
where they have been merged with results of efforts after the project
moved entirely to Livermore.
There follows some detail of the software work that may not have
been included in other reports include
Some specific programs developed include
UPASC - a translator for full standard Pascal to U-code (as described in
the U-code document[4]).
SOPU - U-code to S-1 code generator. Developed jointly by this group and
the Stanford group headed by Gio Wiederhold.
PASCAL* compiler - based on the UPASC compiler, it translates PASCAL to
standard U-code.
∂05-Mar-82 1424 JMC
To: AP
Antonio Dias 313 994-1200x523 will call again today or Monday.
∂05-Mar-82 1449 JMC
To: csd.smith at SU-SCORE
Thanks for the references.
∂05-Mar-82 1449 JMC reprint
To: FFL
Please send a copy of "Circumscription: A from of non-monotonic reasoning"
to Francoise Harrois
19 Laguna Place
Long Beach, CA 90803
∂07-Mar-82 1517 JMC Golux
To: PJH
I have just gotten around to reading your 1973 "Computation and Deduction",
some of the contentions of which I am now prepared to agree with. Was
Golux ever programmed or is there a paper that updates the ideas of the
1973 paper?
∂07-Mar-82 1520 JMC
To: PJH
Also, what is the correct bibliographic reference to that paper?
∂08-Mar-82 0131 JMC
To: csd.armer at SU-SCORE
My course listings are as I want them.
∂08-Mar-82 1814 JMC Hurd invitation
To: REG
Cuthbert Hurd has invited us to eat pie Wednesday at 8pm, and he
suggested I mention it if I see you before he reaches you. His
number is 854-1901.
∂08-Mar-82 1955 JMC
To: kanerva at SUMEX-AIM
I don't know about VisiCal; I suppose it simply depends on whether
there is a market for it. I don't know whether C.I.T. is a healthy
place to work; I wouldn't have assumed so. Of course, they shouldn't
be, and presumably aren't, primarily research oriented.
I will be glad to talk with you, and I suggest you phone some morning
after 10 at my home number 857-0672 or some afternoon at my office
number.
∂09-Mar-82 0049 JMC coming out Saturday
To: llw at S1-A
If it is convenient, I would come out Saturday for Shackleton and S-1.
It is a convenient time for Rod, and I would come out around 1pm and
stay into the evening.
∂09-Mar-82 0118 JMC meetings this summer
To: FFL
I keep forgetting to put the LISP and Artificial Intelligence meetings
in my calendar. Please remind me or do it. The announcements are on
my desk somewhere.
∂09-Mar-82 0131 JMC
To: FFL
Please decorate boyer[w82,jmc] as a letter and scribe it.
∂09-Mar-82 1557 JMC
To: rah at S1-A
I have mislaid dates of L.A. Space Industry meeting. Can you supply?
∂10-Mar-82 1229 JMC
To: csd.golub at SU-SCORE
Donald Perlis, whose resume I am forwarding, wishes to apply for a
faculty position.
∂10-Mar-82 1543 JMC
To: GHG at SU-AI
I have no idea, but since he's 38, maybe he's too old to be a son.
As you may remember, there was a numerical analyst named Sam Perlis,
I believe also at CMU, perhaps a relative of that one.
∂11-Mar-82 0006 JMC
To: csd.golub at SU-SCORE
In my opinion, the student report on Shapiro partly reflects conservatism.
Indeed logic programming does not fit into the Stanford curriculum as in
now exists. However, I think it is an important new direction for computer
science. I am not too enthusiastic about Shapiro's idea of concentrating
on distributed computing, however. But then he made something better out
of debugging than I thought possible.
∂11-Mar-82 1923 JMC
To: feigenbaum at SUMEX-AIM
Gene has eliminated Shapiro, and I think this is a big mistake. The lack
of student enthusiasm seems to reflect a conservative attitude. Shapiro's
interests do seem special given what is currently being taught at Stanford,
but this is how any new field looks at the beginning from the outside. I
think Shapiro has a remarkable number of very wide ranging and rather clear
ideas. I expressed some doubts to him about his interest in distributed
logic programs and got a very cogent reply which I can send you if you
like. I won't be at the March 31 meeting, so I wonder if you are inclined
to urge a reconsideration. I would bet that Stanford is going to look very
silly a few years from now if we have missed a chance to get Shapiro.
∂10-Mar-82 2114 Gene Golub <CSD.GOLUB at SU-SCORE> [Tom Dietterich <CSD.DIETTERICH at SU-SCORE>: Ehud Y. Shapiro student interview]
Date: 10 Mar 1982 2111-PST
From: Gene Golub <CSD.GOLUB at SU-SCORE>
Subject: [Tom Dietterich <CSD.DIETTERICH at SU-SCORE>: Ehud Y. Shapiro student interview]
To: jmc at SU-AI
Mail-From: CSD.DIETTERICH created at 6-Mar-82 18:21:42
Date: 6 Mar 1982 1821-PST
From: Tom Dietterich <CSD.DIETTERICH at SU-SCORE>
Subject: Ehud Y. Shapiro student interview
To: Shapiro-interviewers: ;, csd.golub at SU-SCORE, csd.ullman at SU-SCORE,
csl.lantz at SU-SCORE, csl.jlh at SU-SCORE, tob at SU-AI, rwf at SU-AI
cc: CSD.DIETTERICH at SU-SCORE, csd.jf at SU-SCORE
Group student interview with Ehud Shapiro:
People present:
Rich Pattis (REP)
Michael Kenniston (csd.msk)
Yonatan Malachi (YM)
Ken Clarkson (KLC)
Dave Smith (csd.smith)
Tom Spencer (THS)
Tom Dietterich (csd.dietterich)
Shapiro conducted himself well in answering some pretty tough
questions including many detailed questions about his work. The
students present were generally luke warm about him, however. There
was a sense that he was a quick-minded person who would do well in
whatever problems he attacked. However, his area of interest seemed
to be quite narrowly defined to be PROLOG and logic programming. In
response to the question about what he would do if he came to
Stanford, he answered that he would leave debugging for a while and
instead concentrate on parallel execution of PROLOG programs and on
applications of PROLOG to distributed computing. He also expressed a
desire to collaborate with the Japanese Fifth Generation project
(which emphasizes logic programming). This and subsequent answers
gave us the impression that he was more interested in PROLOG than in
any "computer science" problems. In fact, we were not sure exactly
where he would fit within the department. He stated that he is not an
AI person and disagrees with the current AI methodology. On the other
hand, he is not really an MTC person. Nor could we really consider
him a programming languages person. His primary asset for this
department would be that he would be able to teach PROLOG.
One area of particular concern was his answer to the question about
teaching. He indicated that he has never taught and that he was not
very enthusiastic about teaching, although he would "probably try and
do well" on any teaching assignment he was given. He said he would
not want to teach a "core AI" class.
Summary: no one felt particularly strongly that we should (or should
not) hire him. Our main question was what he would bring to Stanford.
We were somewhat concerned that he would not put much energy into
teaching.
--Tom
-------
-------
In my opinion, the student report on Shapiro partly reflects conservatism.
Indeed logic programming does not fit into the Stanford curriculum as in
now exists. However, I think it is an important new direction for computer
science. I am not too enthusiastic about Shapiro's idea of concentrating
on distributed computing, however. But then he made something better out
of debugging than I thought possible.
∂11-Mar-82 1924 JMC
To: csd.golub at SU-SCORE
I have asked Ed Feigenbaum to help me urge reconsideration on Shapiro with
the following:
Gene has eliminated Shapiro, and I think this is a big mistake. The lack
of student enthusiasm seems to reflect a conservative attitude. Shapiro's
interests do seem special given what is currently being taught at Stanford,
but this is how any new field looks at the beginning from the outside. I
think Shapiro has a remarkable number of very wide ranging and rather clear
ideas. I expressed some doubts to him about his interest in distributed
logic programs and got a very cogent reply which I can send you if you
like. I won't be at the March 31 meeting, so I wonder if you are inclined
to urge a reconsideration. I would bet that Stanford is going to look very
silly a few years from now if we have missed a chance to get Shapiro.
∂12-Mar-82 0117 JMC my trips
To: FFL
For a Livermore personnel security questionnaire, I need to list my
foreign trips. Almost all of them have been paid through Stanford,
and so expense reports exist. Please find out what records exist. I
need dates and countries visited. I also am using the notations in
my passports, but one passport was in a stolen jacket. The trips date
back to 1965. The most important are trips to Russia, because the
Russians don't usually stamp the passport and trips between Sept. 1967
and July 1972 for which the passport is missing. You might ask Patty
Wood what records she thinks might be available.
∂12-Mar-82 0331 JMC
To: FFL
Please print a copy of umnov.pre[w82,jmc] and make 5 copies.
∂12-Mar-82 1257 JMC
To: RPG, LGC
ADVICE[W80,JMC] contains some 1980 ideas.
∂12-Mar-82 1349 JMC
To: csd.golub at SU-SCORE
As you know, he gives an excellent lecture. I will call him and explore
further the teaching question. In my opinion, he will in fact be one of
our better teachers, because he is so well organized. If his teaching
proves unsatisfactory, then we don't have to re-appoint.
∂13-Mar-82 0231 JMC
To: rah at S1-A
It looks like I'll have to be at Hoover all day tomorrow also - with
Teller. Perhaps I'll see him there.
∂13-Mar-82 1655 JMC Thanks and parallel Prolog
To: TOB
∂11-Mar-82 1124 Ehud Shapiro <Shapiro at YALE> Thanks and parallel Prolog
Date: 11-Mar-82 1422-EST
From: Ehud Shapiro <Shapiro at YALE>
Subject: Thanks and parallel Prolog
To: Jmc at SU-AI
cc: Shapiro at YALE
Thanks for your hospitality. I found Stanford an exiting place, but
also quite intense. I still don't know if I what I felt was the tension
around the job interview, or the "objective" nature of the place itself.
I received your letter as I came back to Yale. There are several
reasons why to investigate distributed/parallel Prolog (I do not
distinguish between the two, because I think that at the right level
of abstraction they do not differ). One is that parallel Prolog can
be executed faster on a multi-processor machine, such as the one the
Japanese are planning to build. The other is that I think that a
variant of Prolog may be a very useful for implementing distributed
applications. It took may a while to realize this, but the more I
think about it the more excited I become: in distributed/parallel
Prolog a goal corresponds to a process, and a shared variable between
two goals is a communication channel between processes.
For example, here is how you write a multiplexer, merge(Xs,Ys,Zs), that
merges the two input streams Xs and Ys into one output stream Zs:
merge([],[],[]).
merge([X|Xs],Ys,[X|Zs]) :- merge(Xs,Ys,Zs).
merge(Xs,[Y|Ys],[Y|Zs]) :- merge(Xs,Ys,Zs).
Clearly, Prolog's execution mechanism won't do here: this program is
truly nondeterministic, and the right clause should be invoked as soon
as the input arrives on the channel it reads from. Also, more stringent
assignment of "modes" to arguments should be done, to avoid
multiple-writers conflicts.
Clark and Gregory, in their paper in the ACM conf. on functional
language and machine architecture (from several months ago), give an
example of a simple operating system, with two screens, two keyboards
and one monitor, based on this style of handling streams.
It seems that that the style of programming that results from this
approach is (I think) called object oriented. Consider, for example,
the following programs for stack and queue manipulation. The stack
program receives messages of stack operations, performs the operation
and return the result.
% stack(L,Xs) :- evaluate the stream of messages L on stack Xs.
stack([],←).
stack([pop(X)|L],[X|Xs]) :- stack(L,Xs).
stack([push(X)|L],Xs) :- stack(L,[X|Xs]).
The initialization is: stack(L,[]), where L is the stream of messages.
Although stack is a "global" data structure, the program is functional,
with no explicit side-effects. The side-effect are done to the
variable owened by the process "stack".
The answers to the messages are returned in the variable X for pop(X).
push(X) is just executed, with no response. This can easily be modified
if there is a need for confirmation. For example, one can add the
clause
stack([push(X,ok)|L],Xs) :- stack(L,[X|Xs]).
Which will respond 'ok' in X to the message push(a,X), when the message
received and successfully processed.
It may seem forbiding to invoke a new process for every stack operation
(and, theoretically that's what the stack program does), but here a
new concept comes to our help: tail-recursion optimization of
sequential Prolog maps into "process-optimization" in distributed
Prolog: instead of invoking a new process with the new stack, process
optimization will retain the old process, but will modify its variables,
i.e. its "state", and thus induce an implict side-effect.
Although the approach needs synchronization primitives not available
in current Prolog-10, simple "distributed" programs, such as the one
for stack and queue (below) may still be executed directly in it. For
example:
| ?- stack([push(a),push(b),pop(X),push(c),pop(Y),pop(Z),push(d),pop(W)],[]).
W = d,
X = b,
Y = c,
Z = a
yes
Queues can be implemented with the same ease. Below an example of a queue,
simulated with a "lazy list".
% queue(L,Qhead,Qtail) :- evaluate the list of messages L on queue
% whose head is Qhead and tail is Qtail.
queue([],←,←).
queue([add(X)|L],Qhead,[X|Qtail]) :- queue(L,Qhead,Qtail).
queue([remove(X)|L],[X|Qhead],Qtail) :- queue(L,Qhead,Qtail).
The initial call is queue(L,Qhead,Qhead). For example:
| ?- queue([add(a),add(b),remove(X),remove(Y),add(c),remove(Z)],Q,Q).
X = a,
Y = b,
Z = c,
Q = [a,b,c|←371]
The mode should be queue(+,-,+), so trying to remove an element
from an empty queue will violate the mode and hence fail, or raise an
exception.
However, without modes, remove(X) on an empty queue
will unify X with the appropriate element of the queue when
enough add operations will be done. For example:
| ?- queue([add(a),remove(X),remove(Y),add(c),remove(Z)],Q,Q).
X = a,
Y = c,
Z = ←110,
Q = [a,c,←110|←345]
This is a curious behavior. It means that there is no need to
synchronize add's and remove's, and any sequence of them will unify
the n'th element removed with the n'th element added, even if the
remove occured before the corresponding add.
In summary, I think that all will benefit from distributed/parallel
Prolog: it will enable a richer programming style, faster execution
of programs, and an increased range of applications that can be solved
with a higher level language.
-- Udi.
-------
∂13-Mar-82 2042 JMC aluminum power to the people?
To: llw at S1-A
Can this be??
Date: Thursday, 4 February 1982 16:46-PST
From: KING at KESTREL
Subject: aluminum power to the people?
To: energy at mit-mc
cc: King at KESTREL
I have a half-baked idea.
Lawrence Livermore Labs claims to be on the verge of an
aluminum-air battery that would have a range of 1500 miles in an
electric car, on 71 lb. of aluminum. At current prices, this aluminum
would cost about $55, and it would provide about 500 KWH using
reasonable estimates for the energy efficiency of the car.
My idea is to consider an aluminum battery for the power
source for a house. Peak demand for a house is certainly less then
∂14-Mar-82 2200 JMC
To: jmc at S1-A
Spirulina
∂14-Mar-82 2304 JMC
To: LES
What's your business address, so I can use you as clearance reference?
∂16-Mar-82 0010 JMC via S1-A
To: FFL
Please remind me to send Livermore a bill for today.
∂16-Mar-82 0232 JMC
To: rah at S1-A
Ted Anderson volunteers to help with Shackleton.
∂16-Mar-82 1647 JMC
To: pjb at S1-A
Paula, I don't want to officially submit the fingerprints yet, because I
haven't completed the other papework. However, ask them if they are
willing to look at the fingerprints and say whether they are technically
satisfactory or whether I have to have them done again. They have to be
clear enough for classification. If they are technically satisfactory,
then please keep them for me until I have the paperwork done. This may
take a while longer, because I have never applied for a clearance before,
and I have made many foreign trips, and it isn't easy to be sure the list
is complete.
∂16-Mar-82 1724 JMC
To: llw at S1-A
I the draft report I sent you on Stanford S1 contract ok?
∂17-Mar-82 0137 JMC
To: TOB
What if he asks me what OTV stands for?
∂17-Mar-82 0140 JMC
To: LGC, RPG
Let's make it Monday at 2pm to talk about the Advice Taker.
∂17-Mar-82 0237 JMC
To: csd.churd at SU-SCORE
I hope to borrow Computer Establishment after I return.
∂17-Mar-82 0238 JMC afterthought on NYT
To: REG
I don't think the added clause in the NYT invoice or letter
was aimed specifically at us. I think it was put in because of the increased
computer services provided generally and because the NYT is now offering
full news story services itself. If they had been concerned about our use
they would have written a letter addressed to us specifically inquiring
about it. Therefore, I reaffirm my opinion that you should just sign their
agreement and send it back.
∂17-Mar-82 1144 JMC
To: TOB
I phoned Mock. He said that the proposal would have to be resubmitted,
so I explained about the success of the long campaign to make you an
Adjunct Professor, so we agreed that it when it is resubmitted, you should
be PI. No action is necessary until you get something in the mail, but
I think it might be worth your while to talk to him on the phone.
∂17-Mar-82 1508 JMC
To: LGC
rpg now agrees to
03-22 Monday 4pm, rpg and lgc, advice taker
∂17-Mar-82 2159 JMC
To: FFL
eller.1
∂18-Mar-82 0222 JMC
To: FFL
I need to think a bit more before I speak to Mrs. Kuo.
∂18-Mar-82 1058 JMC
To: Colmerauer at MIT-MULTICS
test
∂18-Mar-82 1244 JMC
To: Colmerauer at MIT-MULTICS
Your test message also received ok. I have begun to think about alternatives,
and I'll send you another message soon.
∂18-Mar-82 1600 JMC Colmerauer is on ARPAnet.
To: AP at SU-AI, warren at SRI-AI
His net address is Colmerauer@multics where the capitalization of just
the initial letter is essential.
∂18-Mar-82 1718 JMC
To: warren at SRI-AI
It would be interesting to be associated with the VAX Prolog project.
∂18-Mar-82 1720 JMC
To: warren at SRI-KL
I didn't read carefully enough, logic programming architecture is even
more interesting.
∂18-Mar-82 1720 JMC
To: warren at SRI-AI
I didn't read carefully enough, logic programming architecture is even
more interesting.
∂19-Mar-82 1238 JMC
To: CLT
We have an appointment with Okner at 11:30 on Monday.
∂19-Mar-82 1256 JMC
To: warren at SRI-AI
Do you have home phones for Keith Clark or Bob Kowalski?
∂19-Mar-82 1411 JMC delay of trip
To: rah at S1-A, llw at S1-A
My European trip is delayed one and possibly two weeks. Therefore, I
plan to go to the Space Industrialization meeting and also visit the
S-1 project at least once more.
∂22-Mar-82 0325 JMC
To: FFL
Please find me letters from Keith Clark or Robert Kowalski. I need phone numbers.
∂22-Mar-82 0419 JMC
To: jbr at S1-A
Try "help login" on the s1 foonly.
∂22-Mar-82 1751 JMC one way mission report
To: llw at S1-A, rah at S1-A, minsky at MIT-AI,
danny at MIT-AI
The last paper we wanted has finally appeared. It is entitled
"The one-way manned space mission" by John M. Cord and Leonard M. Seale
and was issued in 1962. It is for a one man mission. Some of the
life-support computations may still be relevant. I'll bring it in
to LLNL.
∂22-Mar-82 2106 JMC
To: FFL
That was Pressey if you haven't sent it yet.
∂23-Mar-82 1110 JMC
To: RPG
Two is fine.
∂23-Mar-82 2131 JMC
To: FFL
golub.1
∂24-Mar-82 0309 JMC trip
To: Colmerauer at MIT-MULTICS
I now propose to arrive on Saturday, April 10 and leave on Saturday
May 8. Will that be ok?
∂24-Mar-82 1750 JMC
To: FFL
It would be good if I could leave for Livermore tomorrow at 11am.
∂25-Mar-82 0053 JMC questionnaire
To: FFL
I have added quite a few trips to the form and added my job history.
Please find any trips that may be listed in the file HISTORY[W82,JMC]
that weren't transferred to SECURE[1,JMC]. See if you can make the
whole thing neater than I left it. I'll be in to go over it with you.
∂25-Mar-82 0135 JMC your visit
To: shapiro at YALE
I should have got back to you sooner. You didn't make a hit with Golub
by talking about getting out of teaching. The standard load is one course
per quarter of the academic year. A few old hands like me teach only two
out of the three quarters (buying out with research money), but I think
the younger faculty all teach the three quarters, although leaves aren't
hard to get for almost any reason.
If you are still interested in Stanford, Golub can readily be reassured
that you plan to do your share of teaching. I suppose that Stanford
differs from the University of Illinois in this matter.
The students were impressed with some of the things you have done and
with your interview, but they "don't see how you fit in", because they
consider Prolog a narrow subject. In my opinion, they're mistaken about
this, and I will carry the ball on arguing that logic programming is
a subject for which Stanford has to make room. However, you could help
by mailing me a note about how you see the subject of logic programming
as a whole and how your work fits in to it.
If you want to phone, I can be reached at 497-4430 office or 857-0672
home. Thursday afternoon and evening, I'll be consulting and Lawrence
Livermore Laboratory at 422-0758 - all area code 415.
∂26-Mar-82 0056 JMC
To: csl.reid at SU-SCORE
jmc - One hopes he's the student who was admitted. The obvious translator
is one of the other PRC students at Stanford.
∂26-Mar-82 0100 JMC via S1-A
To: ME
at 10:24 Wednesday the NYT wire turned to garbage.
∂26-Mar-82 0312 JMC
To: csd.ullman at SU-SCORE
We might wait for it to be signed by ARPA director.
∂26-Mar-82 0327 JMC
To: RPG
(if nil 'a 'b) works at SAIL but not at S1.
∂26-Mar-82 1729 JMC binder envy
To: RWG
I have succeeded in inducing a case of binder envy in Betty Scott. She
wants to know the name of the binder at PARC and the name of the person
at PARC who can tell her about it.
Preliminary remark about flavors: Having flavors may be independent of
whether we are talking about objects or quantities. My example of vectors
and smallness referred to quantities.
∂26-Mar-82 1743 JMC
To: CLT
04-06 tues. 7:30pm, Jordan 100, Dov Gabbay lecture
∂27-Mar-82 2111 JMC
To: FFL
kahn.1, and please make a Livermore form - $11.00 food this time.
∂27-Mar-82 2328 JMC
To: MLB
I like to try out the Symbolics machine when convenient
∂27-Mar-82 2355 JMC
To: BS at SU-AI
∂27-Mar-82 2333 RWG Cheshire binders
To: BS at SU-AI
CC: JMC at SU-AI
come in two flavors, hairy and simple. The hairy one takes binder tape
cassettes in three widths and cuts them to the length you specify.
The simple one uses precut 11" tapes and cannot bind books more than
about 1.5 times the thickness of JMC's flavor papers. But it is
simpler to operate, harder to jam, and has colored tapes. We have
both flavors at PARC. (Xerox bought the company.) My phone
(494 4491) will eventually ring at Linda Williams's switchboard,
which is right next to a hairy Cheshire. I suspect she could tell
you a lot.
There's also a marvelous paper joggler into which you can put your
Dover output prior to Cheshiring, which Symbolics Inc just had to
have, until they learned it cost $600. I think it's by FMC. Call
the main Xerox number (494 4000) and ask for Jeannie Treichel or
Janet Moreland in Bldg 32.
∂28-Mar-82 0059 JMC
To: FFL
Did I send the letter on Boyer and Moore? Please send a copy to Bledsoe.
∂28-Mar-82 1454 JMC
To: FFL, CLT
I have changed my password, because I think the old one has been used by strangers.
∂28-Mar-82 1545 JMC
To: RWW
Gosper 494 4491, at Xerox
∂28-Mar-82 1848 JMC
To: gosper at PARC-MAXC
Are you interested in eating later? Call 857-0672.
∂29-Mar-82 1108 JMC
To: FFL
Yes, the letter to Bachman.
∂29-Mar-82 1159 JMC
To: warren at SRI-KL
I forgot to ask whether they can be reached on ARPAnet.
∂29-Mar-82 1255 JMC File server
To: REG
1.I think we have sufficient word from ARPA to justify proceeding
on the file server.
2. I talked to Poole today who will call you. He isn't making the
F5 and will propose an F4 but claims it will be cost-effective. He
can have a proposal in about a week.
3. Are there other plausible candidates? Should we get out an RFP?
Perhaps someone should phone D.E.C. to see if they can do better
now than before.
∂29-Mar-82 1949 JMC
To: chandrasekaran at RUTGERS
Unfortunately, I'll be in France May 3-7.
∂30-Mar-82 0100 JMC
To: FFL
geabst[w82,jmc] should be scribed. It is the abstract I'll send to GE for June talk.
∂01-Apr-82 0924 JMC
To: MDD
I'm glad to be a reference, but I'll be going to France on the 10th. If
you tell me to whom the reference should be addressed, I'll do the letter
before I leave, and my secretary can send it when the inquiry arrives.
∂01-Apr-82 1025 JMC
To: ota at S1-A
I can't leave Stanford till 11:30 Friday, because I have a meeting at 10
concerned with using our ARPA computer money. If it turns out you have
to leave early, perhaps I'd better go separately. If I leave at 11:30
from Stanford, I can be at the Lab a few minutes after 12:30 depending
on traffic. Judging by our previous trips, this should get us there
in time for the 7pm ceremonies.