perm filename F80.OUT[LET,JMC]2 blob
sn#554839 filedate 1981-01-09 generic text, type C, neo UTF8
COMMENT ⊗ VALID 00099 PAGES
C REC PAGE DESCRIPTION
C00001 00001
C00008 00002 ∂27-Nov-80 1234 JMC
C00127 00003 ∂21-Oct-80 1214 JMC
C00128 00004 ∂29-Oct-80 0921 JMC
C00129 00005 ∂27-Nov-80 1521 JMC
C00130 00006 ∂27-Nov-80 1525 JMC
C00131 00007 ∂27-Nov-80 2116 JMC
C00132 00008 ∂27-Nov-80 2118 JMC
C00133 00009 ∂27-Nov-80 2220 JMC
C00134 00010 ∂27-Nov-80 2225 JMC
C00136 00011 ∂28-Nov-80 0138 JMC
C00137 00012 ∂28-Nov-80 0153 JMC
C00138 00013 ∂28-Nov-80 1732 JMC deletion
C00139 00014 ∂28-Nov-80 1831 JMC
C00140 00015 ∂01-Dec-80 1637 JMC
C00141 00016 ∂03-Dec-80 0116 JMC
C00142 00017 ∂03-Dec-80 0135 JMC
C00144 00018 ∂03-Dec-80 0149 JMC
C00146 00019 ∂03-Dec-80 0152 JMC
C00147 00020 ∂03-Dec-80 1845 JMC
C00148 00021 ∂03-Dec-80 1929 JMC
C00149 00022 ∂04-Dec-80 0016 JMC
C00150 00023 ∂04-Dec-80 0048 JMC
C00151 00024 ∂04-Dec-80 0158 JMC
C00152 00025 ∂04-Dec-80 1454 JMC
C00153 00026 ∂05-Dec-80 1021 JMC
C00158 00027 ∂05-Dec-80 1031 JMC
C00159 00028 ∂05-Dec-80 1033 JMC
C00160 00029 ∂05-Dec-80 1053 JMC
C00162 00030 ∂05-Dec-80 1844 JMC
C00164 00031 ∂06-Dec-80 1321 JMC
C00165 00032 ∂06-Dec-80 1327 JMC
C00166 00033 ∂06-Dec-80 1354 JMC
C00168 00034 ∂06-Dec-80 1602 JMC
C00169 00035 ∂06-Dec-80 2019 JMC
C00170 00036 ∂07-Dec-80 2319 JMC Barwise is now BYY.
C00171 00037 ∂08-Dec-80 0128 JMC
C00172 00038 ∂08-Dec-80 0133 JMC
C00173 00039 ∂08-Dec-80 0213 JMC form
C00174 00040 ∂08-Dec-80 1517 JMC
C00175 00041 ∂08-Dec-80 1534 JMC mail minsky%mit-ai
C00176 00042 ∂08-Dec-80 1610 JMC FOL primer
C00177 00043 ∂08-Dec-80 1612 JMC xerox and secretarial
C00178 00044 ∂08-Dec-80 1623 JMC
C00179 00045 ∂08-Dec-80 1637 JMC
C00180 00046 ∂08-Dec-80 2331 JMC
C00182 00047 ∂09-Dec-80 1145 JMC
C00183 00048 ∂09-Dec-80 1441 JMC
C00184 00049 ∂09-Dec-80 1534 JMC
C00185 00050 ∂10-Dec-80 1613 JMC
C00186 00051 ∂10-Dec-80 1815 JMC problems
C00187 00052 ∂10-Dec-80 1822 JMC
C00188 00053 ∂11-Dec-80 0913 JMC
C00192 00054 ∂11-Dec-80 2039 JMC
C00193 00055 ∂12-Dec-80 1621 JMC
C00195 00056 ∂12-Dec-80 1621 JMC
C00197 00057 ∂12-Dec-80 2139 JMC holocaust
C00204 00058 ∂13-Dec-80 1053 JMC
C00205 00059 ∂13-Dec-80 1134 JMC helping the boss
C00209 00060 ∂13-Dec-80 1604 JMC
C00210 00061 ∂13-Dec-80 1641 JMC
C00211 00062 ∂13-Dec-80 1650 JMC
C00212 00063 ∂14-Dec-80 0031 JMC book
C00213 00064 ∂15-Dec-80 2231 JMC
C00214 00065 ∂15-Dec-80 2314 JMC proof-checking for courses
C00215 00066 ∂16-Dec-80 1609 JMC
C00216 00067 ∂16-Dec-80 1819 JMC
C00217 00068 ∂17-Dec-80 1437 JMC
C00218 00069 ∂17-Dec-80 1549 JMC using FOL in CS226
C00220 00070 ∂17-Dec-80 1645 JMC ACM
C00221 00071 ∂18-Dec-80 0840 JMC
C00222 00072 ∂18-Dec-80 0841 JMC
C00223 00073 ∂18-Dec-80 1037 JMC faculty meeting agenda
C00224 00074 ∂18-Dec-80 1143 JMC
C00226 00075 ∂18-Dec-80 1210 JMC
C00230 00076 ∂19-Dec-80 0157 JMC
C00232 00077 ∂19-Dec-80 0217 JMC
C00233 00078 ∂19-Dec-80 1001 JMC
C00234 00079 ∂19-Dec-80 1521 JMC
C00235 00080 ∂19-Dec-80 1542 JMC
C00236 00081 ∂19-Dec-80 1616 JMC
C00237 00082 ∂19-Dec-80 1634 JMC
C00238 00083 ∂20-Dec-80 0110 JMC
C00239 00084 ∂20-Dec-80 1144 JMC
C00240 00085 ∂20-Dec-80 1158 JMC
C00241 00086 ∂20-Dec-80 1634 JMC basic
C00244 00087 ∂20-Dec-80 1652 JMC
C00245 00088 ∂20-Dec-80 1709 JMC
C00246 00089 ∂20-Dec-80 1837 JMC dtn bug
C00247 00090 ∂21-Dec-80 0106 JMC
C00248 00091 ∂21-Dec-80 1139 JMC
C00249 00092 ∂21-Dec-80 1626 JMC
C00250 00093 ∂22-Dec-80 1725 JMC
C00251 00094 ∂22-Dec-80 1842 JMC book and maclisp
C00252 00095 ∂23-Dec-80 0120 JMC
C00254 00096 ∂23-Dec-80 0127 JMC
C00255 00097 ∂23-Dec-80 0132 JMC
C00256 00098 ∂23-Dec-80 0149 JMC
C00257 00099 ∂23-Dec-80 0228 JMC chat and diablo
C00258 ENDMK
C⊗;
∂27-Nov-80 1234 JMC
To: pourne at MIT-MC
Jerry: Can you get a reaction to this idea?
A U.S. FOREIGN LEGION
One way of getting military manpower is to allow foreigners
to enlist in our armed forces with eventual citizenship as a reward.
We can impose rather high qualifications and require a rather long
enlistment, say six years, which will keep the numbers down so that
we can see how the experiment is going. It would have to be decided
whether to take men with families and whether to give them some
priority in bringing in wives married during or after the enlistment.
We would get intelligent manpower for the army and other services,
and we would get better than average citizens at the end. Perhaps
the deal would be something like four years in combat arms followed
by two years in a technical job that would provide a marketable
skill. Also after four years, they might be eligible for OCS.
∂26-Nov-80 2237 JMC
To: rwg at MIT-MC
If I lent you Kettenbruche by Perron, I would like it back.
∂26-Nov-80 2235 JMC
To: *
I lent Kettenbruche by Perron to someone and would like it back. - jmc
∂26-Nov-80 1615 JMC
To: energy at MIT-MC
It would be great if the Canadians can get their Slowpoke into use. The
U.S. needs all the examples we can get of foreigners getting ahead of us,
and using nuclear energy on a relatively small scale as 2 megawatts
thermal (1/2000 of a big plant) will improve its acceptability.
I have never seen a general analysis of the difficulty of getting
projects started that meet two needs, e.g. electricity and space heat.
The requirements usually won't correspond in magnitude, and the needs
for expansion won't usually arise a the same time. Doubling the number
of authorities that can say no is another problem. There is also the
problem of not being able to use standard designs that have been
developed for either purpose. Even with all these problems, I
suppose dual purpose projects can sometimes win. Perhaps the easiest
model is when one purpose is the primary justification and the byproduct
makes the operation more profitable.
The big problem with low temperature heat is often bulk which leads
to great costs. The smaller the temperature difference between the
source of the heat and its end use, the bigger the heat exchangers
have to be and the bigger the pipes that transmit steam or hot water.
∂26-Nov-80 1546 JMC
To: TOB
When you send the new bio and bib, please send cc of the 4 papers.
∂25-Nov-80 2257 JMC
To: pourne at MIT-MC
I am now on ENERGY list so no need to send separate copies to me.
∂25-Nov-80 2254 JMC
To: LWE
I have formatted your file by letting E put a directory on it as the
first page. The simplest thing to do is to go through the successive
pages with a macro defined by
αβ∞αβdαβxdel<cr>
which deletes all the information on a page and then deletes the page
mark separating it from the next page. I would do it for you but I
it wasn't sufficiently clear which pages you needed to delete from
the file. If you look at it again and tell me what pages to delete,
I'll do it for you, and it will take but a few minutes.
∂25-Nov-80 1805 JMC
To: csd.brown at SU-SCORE
I don't have a TA in mind. If I can get a TA interested and familiar with
the material this would improve the course. Otherwise I can do without a
TA. It occurs to me that if I can't find a suitable TA, it might still
worthwhile to pay someone to hold hands a bit with students using FOL. It
would be cheaper for the department to pay piecework rates for this rather
than hiring a TA. By the way, I will need to use SAIL for the course in
order to use FOL, but it will not be a large part of the course.
∂25-Nov-80 1553 JMC
To: rem at MIT-MC
Yes. The other contributions to the latest HUMAN-NETS showed that this
is what actually happens.
∂25-Nov-80 1448 JMC
To: stan at SRI-KL
Please send an abstract to FFL by tomorrow.
∂25-Nov-80 1117 JMC
To: JRA
I would be available, but I have just committed myself to a session
on AI and philosophy at Vancouver for the week before ijcai, and I'll
also be at ijcai. Other times are still possible.
∂24-Nov-80 2324 JMC mail to bboard
To: pourne at MIT-MC
The destination for the message should have been bboard@sail.
At least that worked when I tested it from mit-ai. However,
I have put your message in the bboard file.
∂24-Nov-80 2314 JMC
To: bboard
I would very much like to see innovative ideas on ways the
administration could (1) save money and thus aid in tax cuts,
(2) provide innovative public services at reasonable costs
without increasing the burdens and regulatory functions of
government {example: use computers to give local fire
departments information on construction used in local
neighborhoods so that itr's easier to put out fires} or (3) make
real investments that would have good non-zero-sum payoffs for
future.
Interestingly enough, it's possible to get your inputs
somewhere that they might help. It's also possible that you're
stuffing them down a black hole. Be warned.
Replies to POURNE%MC
∂24-Nov-80 2307 JMC
To: REM
Please put me on space list
∂24-Nov-80 2306 JMC
To: JAK
Here is one way of implementing the function that finds position
and velocity as a function of time. It is necessary to go forward or
backward from a time when these things are known.
1. Each edge of a rectangle is represented by a pair of functions
x = x0 + u0*s, y = y0 + v0*s together with limits s0 and s1.
2. The position of the ball is represented by x = x0' + u0'*(t - t0),
y = y0' + v0'*(t - t0) from the time t0 until the ball hits something.
3. Compute the interstection of the trajectory with each edge and
eliminated those edges for which the values of s are out of range.
4. Choose that intersection with the least value of t > t0.
5. Reflect at that point and continue the process.
Linear programming techniques can probably give something more sophisticated,
but it won't be needed if we keep down the number of rectangles.
Whether the ball is visible at a given time is computed similarly,
but for that it is worthwhile to have the back edges represented by giving
y as a function of x and scanning over the ranges of x to determine the
function to compare y with.
Since it might be interesting to get someone to try to figure out
the rule of the sequence, I suggest you don't talk about the problem except
with people who agree not to talk.
∂24-Nov-80 1444 JMC
To: REG
I have made an account as follows: to be billed to my ARPA.
Kelley, John 328-6077, JAK on system
1270 Hoover, St., Menlo Park 94025
∂24-Nov-80 1155 JMC
To: RWG
I assume your landlord is RWW, and I think a protest is indicated if
RWW and offspring are interested.
∂24-Nov-80 1149 JMC
To: human-nets at MIT-AI
Two points for Zaumen:
1. If the beeps of the call-waiting are at a frequency different from
those used by the modem, there might be no problem. Moreover, many people
operate in a mode in which an occasional garble causes only a slight delay
while the garbled characters are replaced. In fact I can't recall
suffering more than minor annoyance from noise on the phone line except
when there is so much noise repairing the damage from noise has a
substantial probability of losing because of more noise.
2. A few people trying to make a lot of calls to a telephone poll won't
make much difference. What they think they avoided by the late
announcement of the poll were organized campaigns to bias the results
which could have made a large difference.
∂24-Nov-80 0205 JMC
To: FFL
I need a xerox of the following article obtainable in the math library.
The full name of the journal is Mathematics of Computation.
J. Brillhart, "Note on representing a prime as a sum of two squares","
%2Math. Comp.%1 v. 26, 1972 pp. 1011-1013.
∂23-Nov-80 2321 JMC
To: pourne at MIT-MC
Dearer to my heart than a Presidential space speech would be a good
speech on energy, including a non-pussyfooting attitude to nuclear
energy, and on technology generally. I think that Reagan
could score a coup by a speech on the wonders to come, because public
opinion, including much intellectual public opinion, is ready to flip
from its negative attitudes. I'll soon send you a draft of a statement
that Stanford SE2 is preparing for a signature campaign among California
scientists and engineers. I would like your comments on it for its
own purpose as well as eventual signature and help getting others,
and it contains some ideas that need to be pushed nationally also.
∂23-Nov-80 2301 JMC
To: pourne at MIT-MC
If the Russians plug away at making landers that can survive on Venus for
longer and longer times, they will do something substantial that we can't
do with merely by having a more sophisticated generalized instrumental
technology. Their moon rover was a similar attempt to specialize, but
they didn't keep it up. Assuming we can't think of any bad reason why
they might be doing it, probably we should continue to co-operate by
supplying radar maps and maybe in other ways. After all survival at 900
degrees Celsius is a very specialized and difficult problem, and there are
enough things to do in space, so we needn't do it too unless we can
imagine a military payoff. I have thought a bit and can't imagine any,
which doesn't prove that no-one else can. If anyone has been having
gloomy thoughts on the subject, Lowell Wood is likely to know about it.
Jerry, you're a very good writer, and I think you should consider
preparing an outline and draft of one or two Presidential speeches about
space, if you think a proposal to do so would be regarded favorably or
even with interest. No doubt the official speech writers would redo it,
but since they don't know much about space, the content and message might
well stick. Some good words about "the spirit of exploration" that would
get the justification of a space program out of the science or immediate
payoff category would be the most important contribution. No doubt it
wouldn't come up very quickly on the Presidential agenda, but if something
were ready, it might be just what was wanted on a suitable occasion.
On a lighter side, your question reminded me of a "secret project" I used
to tell people about shortly after Lincoln Laboratory first detected a
radar return from Venus in the late 50s. After people agreed to keep it
confidential, they would be informed of the plan to have three large
radars spaced equally around the equator and a satellite communication
link to a central computer near Washington with a direct link to the White
House. The radars would keep a continuous watch on Venus, and if it ever
disappeared, the President would be instantly informed.
∂23-Nov-80 1825 JMC
To: BS at SU-AI
Both xeroxes have paper jams, unclearable for lack of key.
∂23-Nov-80 1534 JMC
To: marimont at SRI-KL
I haven't attended the meeting before, but I will discuss it with Tom
and then maybe with Manning to determine whether he considers it
appropriate. Perhaps there will be more information about where you
stand before a decision finally has to be made. I will surely need
to be reminded, but I have put the date in my calendar.
∂22-Nov-80 1542 JMC
To: ef at MIT-AI
What are your present co-ordinates?
∂22-Nov-80 1540 JMC
To: forward at USC-ECL
It seems to me that the work done on energy required for computation
is relevant to your question. I haven't really followed it, but I
think the expert is Rolf Landauer at IBM Research Lab, Yorktown Heights
who has published papers over a many year period. However, the results
of this work have been modified by results of Ed Fredkin (EF@AI) and
others on "conservative logic". The basic idea of conservative logic
is that if no information is lost, then the entropy need not be
increased at all. Maybe you know all this already.
∂22-Nov-80 1111 JMC
To: energy at MIT-MC
Please put me on the list.
∂22-Nov-80 1109 JMC
To: energy at MIT-AI
Please put me on the list.
∂21-Nov-80 1514 JMC odear
To: pourne at MIT-MC
CC: minsky at MIT-AI
I really liked it. Pournelle for Administrator of NASA! If one could
devise a method of getting 5 speeches made by Reagan or Bush proposing
that American manned space exploration is worthwhile in itself, I believe
the public and future administrations would support a continued large
program. If it could be started during the "honeymoon", even many of the
intellectuals would commit themselves in its favor. Your memo could serve
as the basis for one or two such speeches.
∂20-Nov-80 1912 JMC please order book
To: FFL
Please ask the bookstore to order for me "The Age of Birds" by Alan Feduccia,
Harvard University Press. My account number there is 1030.
∂18-Nov-80 2011 JMC
To: minsky at MIT-AI
n516 0008 18 Nov 80
BC-LICE-2takes-11-18
By Viola Osgood
(c) 1980 Boston Globe (Field News Service)
BOSTON-In the textbooks, its full name is Pediculus humanus capitis.
Most people, however, know it simply as the head louse.
By whichever name, the parasite is making an unwelcome appearance
these days in an unlikely spot, the schools of Massachusetts.
Brookline students and parents, for example, have been fighting the
head louse problem since school opened in September. And they feel
helpless to do anything about what many see as an invasion.
Until the last couple of years, the louse problem in Massachusetts
schools was a minor nuisance consisting of some sporadic outbreaks in
late fall or early winter. Last year, however, the problem became a
major one and in one school was so severe it threatened to reach
epidemic proportion.
Dr. Gloria Rudisch, school doctor for Brookline, said that recently
there has been an upward trend of infestation in this area of the
country. She said this followed a period in which the louse problem
appeared dormant.
Despite the unappealing nature of the beast, head lice pose no
serious threat to health.
Lice spread very rapidly in a community, and there are no sure-fire
preventive measures. In some cases, a secondary scalp infection may
result from scratching.
''It's just disgusting,'' said one Brookline woman who has had the
unenviable task of delousing the hair of both her children twice in
three weeks. ''I'm afraid we're going to have this problem for the
rest of the school year. It's scary.''
Lest Brookline parents feel that the lice burden is theirs alone,
take heart. Parents in Acton, Natick, Hingham, Cambridge, Newton,
Gloucester, Winchester, Wellesley and Belmont, just to name a few of
hundreds of other Massachusetts communities, are also stuck with the
same situation to one degree or another.
Current estimates indicate that at any given time, millions of
Americans may be infested with lice, and that the number is
increasing.
According to Dr. Nicholas J. Fiumara of the state Department of
Public Health, ''There's a lot of it around. I don't know of any
school system that doesn't have a problem. We get calls from schools
all over the state wanting to know how to deal with the lice
situation. We recommend inspection of the kids by the school nurse.
''The problem is seen primarily in grade schools, where little kids
swap hats or scarves and use each other's combs. And little girls are
forever preening each other's hair,'' said Fiumara, who directs the
communicable diseases division.
Fiumara said that the treatment of head lice is easy, although it
can take hours. The hair should be washed with a special shampoo
containing the insecticide gamma benzine. Combing the hair with a
fine-toothed comb helps remove the eggs. Heat from a good commercial
hair dryer helps kill any remaining lice. The treatment should be
repeated in about a week.
Lice are parasites. Head lice are transferred by contact or through
infected clothing, especially hats and scarves. The main symptoms are
itching caused by the bites of the feeding lice; and the oval,
pearly-gray nits (or eggs) that are firmly glued to the hair shaft.
The adult female louse lives for about a month and lays about 10 eggs
a day.
In addition to head lice, there are two other types of human lice:
the body louse (Pediculus humanus corporis) and the pubic louse
(Phtiris pubis).
Throughout the ages, lice have flourished in populations of all
social classes. In the Middle Ages, lice began to be associated
solely with poverty, filth, war and disease, a belief that persists
today.
With the development of modern hygiene and insecticides, including
DDT, louse infestation dwindled to a minor problem in most areas of
the country by the early 1950s. About 10 years ago, with the ban on
DDT, increased world travel and an increase in social contact, head
lice began to come back.
Dr. Leslie Norins, former louse specialist for the U.S. Public
Health Service, in a 1977 study for a lice remedy manufacturer, found
that in 1963 there were only 250,000 reported cases of head lice
nationwide. By 1976, the number had risen to 5 million.
Fiumara and Rudisch both said one of the main problems in dealing
with lice infestation stems from the fact that many people associate
the condition with dirt and unsanitary conditions. They said lice can
be found on children from the most meticulously cared for homes. The
Norins study found that nearly 70 percent of the reported cases of
head lice occurred in families in the middle- and upper-income
brackets.
One woman who spent six hours shampooing and combing her two
daughters' hair, said: ''Don't use my name. Lice are still socially
unacceptable. People think you aren't clean if you have lice. They
are frightening. When I saw the little egg sacs clinging to the kids'
hair follicles, it wasn't as bad as I expected, but if I had seen a
live bug, I would have died. After it was all over, I itched all
night.''
There is a range of feelings about the subject. One is the casual
attitude maintained by a few enlightened people who don't worry at
all. Another is the middle group, comprising the majority, who are
apprehensive about their children getting lice, but deal with the
problem realistically.
Then there is a third group of people, perhaps the smallest, who are
paranoid about lice, subscribe to every myth that has been uttered
and endow the mite with near-supernatural powers.
For instance, one day-care teacher in Brookline believes that one
way to stop the spread of lice is to keep the children from jumping
rope.
Cheri Gillis, of Brookline, a mother of a 6-year-old who has so far
escaped: ''Every night, I wash my kid's hair and put hair on it. A
friend told me that discourages the lice.''
A Newton mother: ''It's the kids from dirty homes who start the
problem. If the schools insisted on better hygiene, there wouldn't be
all these lice. My children had them three times, and I'm sick of
it.''
A Brookline principal: ''I had one parent blame the problem on the
children in day care. I tried to explain, that's not the reason. But
it's hard to convince people who've already made up their minds.''
''What I see in parents whose children have lice is fear and
disgust,'' said Diane Genco, director of a Brookline after-school
program. ''The parents just freak out, and the kids act as if they've
come down with leprosy.''
Fiumara said, ''It's a nuisance and an embarrassment to parents, but
there's no danger. It can cause infections of the scalp. But in terms
of disease or injury to health, there's no danger.''
ENDIT OSGOOD
ny-1118 0313est
***************
∂18-Nov-80 1101 JMC
To: CET at SU-AI
There will be a final, but I won't need a special room.
∂18-Nov-80 1100 JMC
To: csd.ullman at SU-SCORE
OK about VAX proposal
∂18-Nov-80 1058 JMC
To: shostak at SRI-F2
Many thanks for the note on Woody Bledsoe. It's just what I needed.
∂17-Nov-80 1814 JMC Cheeseman
To: DBL at SU-AI, feigenbaum at SU-SCORE, buchanan at SUMEX-AIM
Peter Cheeseman from New South Wales Institute of Technology writes that
he wants to visit for a year saying "My aim is to implement (in INTERLISP)
a new knowledge formalism I have been developing. This formalism
combines both logical and probabilistic knowledge, as well as
combining extended predicate calculus representation and set theory."
I met Cheeseman and he seems bright, we are short of space, and
he would need as desk and computer use. Do you know him and have an
opinion?
∂17-Nov-80 1813 JMC
To: minsky at MIT-AI
That's interesting. Let me try re-arrange my scheduled dinner with
Jerry Pournelle. Presently I'm scheduled to leave at 9pm and was
going to keep the Pournelle dinner early, e.g. at 6. Is this
feasible, or should I try to re-arrange my travel.
∂17-Nov-80 1718 JMC
To: FFL
daniel.1
∂17-Nov-80 1648 JMC
To: FFL
jain.1
∂17-Nov-80 1628 JMC questionnaire
To: DBL at SU-AI
CC: REG at SU-AI
I have just uncovered it, and I think it misses the point, because it
asks mainly for opinions about the solutions rather than what their
problems are. It doesn't ask how much disk storage people think they
need. It encourages everyone to ask for the grandest possible personal
computer. Why shouldn't I ask for a private Cray-1? Everyone who uses
SAIL uses the line editor if they so much as backspace to correct a
single incorrect character. The other questions are more or less
appropriate except for the last which asks for an opinion about the
solution.
∂17-Nov-80 1531 JMC
To: FFL
takaha.1
∂17-Nov-80 0243 JMC interstellar
To: pourne at MIT-MC
I gather that Lowell might have security problems with a frank article
on laser propulsion.
∂17-Nov-80 0239 JMC
To: pourne at MIT-MC
What I actually had in mind was a technical book, but a popular science
book is also interesting. I would subimit an article. Glad to hear that
Sagan says we'll go. I too have various complaints about Sagan.
∂17-Nov-80 0221 JMC
To: "@SPACE.DIS[1,JMC]" at SU-AI
1. If it is possible to scoop up matter and if the velocity to which your
laser powered unit can expel it is greater than the velocity of the
spacecraft, you win by collecting it.
2. For long enough journeys, it wins to have an accompanying spacecraft
that does not decelerate but supplies energy to the smaller decelerating
spacecraft by laser.
3. If there is an ambitious editor, it would be worthwhile to solicit and
publish a collection of scientific papers on interstellar travel. The
SETI club, perhaps in order to magnify their enterprise or perhaps because
they don't want Earthmen contaminating the galaxy, often say or imply that
interstellar travel is forever impossible.
∂17-Nov-80 0000 JMC
To: pourne at MIT-MC
I assume you've asked Lowell Wood for space ideas.
∂16-Nov-80 1752 JMC
To: FFL
metzge.1
∂16-Nov-80 1647 JMC
To: FFL
snow.1
∂16-Nov-80 1635 JMC
To: buchanan at SUMEX-AIM
It seems to me that an argument about what practical contributions
AI has made is pointless. AI is now in a period of increased
interest. The next major evaluation point is a few years away,
and what people think of it then will depend on what has happened
in the meantime - not on what arguments have or have not been
accepted for inclusion in some report.
I think you should not retreat from the point that industry has
made few contributions to AI. However, it might be worthwhile to
express some reserve about whether the current AI boom in industry
will be fruitful for industry.
However, you haven't been extreme.
As for McIlroy's preference for
PDP-11s over PDP-10s, there is a substantial probability that both
he and we will be using VAXen almost exclusively in five years.
Modifying the statement about Government assistance in language
maintenance to add the phrase "on a variety of computers" might
soothe him.
The word is "bellwether" not "bellweather".
I don't think too much of the "weak vs. strong methods" dichotomy,
because a an algorithm like tree search may require heuristics like
alpha-beta to reduce the combinatorial explosion.
How about adding a sentence
"Another aspect of AI research is determining the facts about the
common sense world of causality, time, space and knowledge and
expressing them in a form usable by a computer program."?
In discussing strengths and weaknesses, it might be well to say something
about intellectual strengths and weaknesses as well as institutional.
For example, I think there isn't enough work in theory.
In general, I approve of your report and am willing to be counted
a co-author if you really think it worthwhile.
∂16-Nov-80 1147 JMC
To: human-nets at MIT-AI
I find it hard to understand why Senator McClure, who seemed very
intelligent to me in my one contact and who will have a big job in
reviving nuclear power as Chairman of the Senate Energy Committee, wasted
his time on such a trifle as when the networks predict election results.
Whether a few marginal voters decide not to vote because the outcome has
been predicted is one more random variable in a situation that is
influenced by dozens of random variables. Solemnly making rules as to and
can predict what when and creating an organization to enforce it and
providing a basis for new lawsuits aimed at overturning election results
can lead only to harm, though probably not much.
However, we can imagine a situation in which the voters are
treated like a jury where judges hold hearings on what can be told to
them. Letting the imagination really run wild, perhaps an election would
be postponed because a secret appeal to a decision about what someone can
say that might improperly influence the voters is on its way to the
Supreme Court.
∂15-Nov-80 1401 JMC
To: pourne at MIT-MC
1. There has been a turnaround in public and intellectual attitude
toward space. Through the sixties and seventies there was a
suspicious attitude demanding to be shown concrete benefits.
However, the long drought of manned space activity and the
wind-down of the space program made many people realize that
they really want Americans in space quite apart from any
specific benefits. Most likely this attitude will strengthen,
and a Russian space station will strengthen the belief that
America dropped the ball.
Therefore, the space program should frankly aim at exploration even more
than activities with an economic payoff. Many of the latter can be left
to industry.
2. The payoff in exploration could be enormously enhanced by one-way
manned missions. The reason is that the take-off weight is exponential
in the total velocity change required.
These aren't suicide missions, but missions in
which the astronauts commit themselves to living most of the rest
of their lives in space or on the target planet.
For example, if the astronauts had
not returned from the moon, they could have taken much more
equipment, and they have been resupplied annually for much
less than the Apollo missions cost. Perhaps the
Shuttle has enough capacity to support one-way missions to the
moon. A one-way mission to Mars should be possible by 1990.
This is a case in which a willingness to take greater risks and
make greater commitments would bring enormously greater rewards.
However, the degree of commitment is no greater than that of
explorers of the eighteenth and nineteenth centuries.
3. Basic technology needs to be supported. It was really poisonous
when NASA adopted the attitude (or was it OMB) that only technology
in support of already approved missions would be developed. It led
to a situation where a new technology couldn't be counted on
because it hadn't been developed and couldn't be developed because
no mission called for it.
4. Ion rockets and nuclear rockets should be considered for revival.
5. The Space Law Treaty should be explicitly denounced, and the
Government should have a policy of selling Shuttle launches for
any non-military purpose.
∂15-Nov-80 0129 JMC visit
To: pourne at MIT-MC
Sorry I missed Saturn. I accepted an invitation to take part in
a meeting here on that day that turned out to be NASA sponsored.
Sorry I missed Saturn. III is meeting again next Wednesday. It
would be interesting to get together either Tuesday night or
Wednesday. If the latter I would have to leave early AM, since
I teach at 9:30 on Thursday, so Tuesday probably better. I have
some ideas about technology policy, but I have to sort them out.
∂14-Nov-80 2258 JMC
To: BS at SU-AI
CC: TOB at SU-AI, FFL at SU-AI
Please check with the official that it is the grant they are referring
to.
∂14-Nov-80 1751 JMC
To: FFL
I need psa reservations to LA and back Wed for III meet. As before.
∂14-Nov-80 1731 JMC
To: TOB
Please get that NSF Final report in, so NSF will ccnsider my proposal.
∂14-Nov-80 1552 JMC
To: FFL
cartwr.re2
∂14-Nov-80 1505 JMC
To: STT
Thanks. I'm collecting different versions of this quote.
∂14-Nov-80 1458 JMC
To: LGC
Lew, it seems to me that you are trying to cover too much ground in the talk.
∂14-Nov-80 1454 JMC
To: FFL
Please pub and print cbcl[f75,jmc] and enclose it with a pubbed
version of reetz.1[let,jmc].
∂14-Nov-80 1036 JMC
To: PAM
No, it didn't show up, and it would be worthwhile to see if one of
yours works. I recall it labelled 9 volts. Thanks. John
∂13-Nov-80 1521 JMC
To: FFL
schank.1
∂13-Nov-80 1144 JMC
To: csd.tajnai at SU-SCORE
Holding it open is fine.
∂12-Nov-80 2232 JMC
To: SEK
If you log in before midnight, please phone.
∂12-Nov-80 1542 JMC
To: csd.ullman at SU-SCORE
I haven't started yet on Binford papers. Perhaps I can do them next
week.
∂12-Nov-80 1541 JMC
CC: FFL
∂12-Nov-80 1004 Nilsson at SRI-KL Chinese Scholar
Date: 12 Nov 1980 1005-PST
From: Nilsson at SRI-KL
Subject: Chinese Scholar
To: JMC at SU-AI
cc: Nilsson
John, I just received a letter from a Chinese student, Tie Cheng
Wang (working under a Lou Oi Ming). He apparently is under the
impression that I am connected with a university, and wants to come
to the US and work under me. He also needs financial assistance.
From his letter he sounds pretty bright. Might there be any
possibilities for him at Stanford? I would be glad to work with
him if he becomes a student at Stanford, or perhaps he would want
to work with you. I'm sending a copy of his letter and an abstract
of a paper to you by ID mail. Let me know if you want to follow up
on this. Otherwise, I will merely refer him to the various US
universities with good AI programs. --Nils
-------
When the letter comes, I will arrange for him to be sent
application material. We are treating applicants from China like
any others, but he will have to act quickly if he is to get in
an application by our January 15 deadline for the PhD program,
although there is more time for the MS program.
∂12-Nov-80 1541 JMC
To: nilsson at SRI-KL
When the letter comes, I will arrange for him to be sent
application material. We are treating applicants from China like
any others, but he will have to act quickly if he is to get in
an application by our January 15 deadline for the PhD program,
although there is more time for the MS program.
∂12-Nov-80 0915 JMC
To: FFL
I have changed the name of my calendar file to CAL[1,JMC].
∂12-Nov-80 0009 JMC
To: pourne at MIT-MC
Will anyone be disturbed if I phone? No-one but me here now.
∂11-Nov-80 2245 JMC
To: ROB
Many thanks for restoring lib.lst
∂11-Nov-80 1625 JMC Q and K vs. R and K
To: JJW
Professor Donald Michie
Department of Computer Science
University of Illinois
Champaign,Illinois
would like a writeup of your solution method. Actually, so would I.
∂11-Nov-80 1445 JMC
To: BIS
Can you come to 356 at 1:30 Thurs to tell Thompson people about robot work?
∂11-Nov-80 1443 JMC
To: *
Who took the power supply from the Chess Challenger in the lounge? - JMC
∂10-Nov-80 1857 JMC
To: Pool at MIT-MULTICS
CC: Solomon at MIT-MULTICS
The following includes some of the topics we discussed. Its weakest
point is whether suitable databases would be useful in an
emergency.
.cb Mobilizing Data Bases in National Emergencies
.cb by John McCarthy, Stanford University
These ideas arose in a conversation with Ithiel Pool.
concerning what might be done with computer and communication
technology in a national emergency including even a nuclear war.
The conversation started with a discussion of replacing travel
by communication, and there are many ways of doing that. However,
much of the development that will make this increasingly possible,
such as the increasing number of computer based offices, is
proceding at a pace that is unlikely to be much affected by
planning for an emergency.
The discussion did turn up one area where action now at
reasonable cost could make a large difference in an emergency.
The idea is to make information required for dealing with an
emergency available in data bases accessible anywhere in the
country from existing computer terminals. Moreover, local organizations
could provide information about what facilities were destroyed
or damaged that could be used for central decision making.
We did not discuss
much what information should be included, but here are a few
examples. The location, form and quantity of food stocks is of
little interest to anyone outside the food distribution industry
in normal times. In an emergency, many people outside the industry
might need it. Another example is the location of earth moving
equipment and people and firms qualified to operate it.
Hospitals could provide information on how many patients and
doctors they had and the current availability of beds, equipment
and medical supplies.
One form of preparedness is for FEMA (The Federal Emergency
Management Administration) to maintain a computer accessible
data base in which such information is kept. This database should
be available through the ordinary telephone network from any
terminal in the country whose user knows the telephone numbers,
accessing procedure and has appropriate passwords. The
database could be used in preparedness exercises from terminals in
local government and business offices that are ordinarily used
for other purposes.
A study is required to determine what information is wanted
and how it might be collected and kept up-to-date.
There are already hundreds of thousands of computer terminals
in the country that could be used to access the data base.
Hoever, many of them terminals do not have access to the
telephone network, because providing them with modems and telephone
connections is not required by their normal applications. However,
the additional cost of such a connection is less than α$100 and
has other uses. Therefore, it probably wouldn't require more than
minor Government encouragement to persuade most terminal owners
to equip them for such use. The modem is usually available as an
extra cost option by the manufacturer of the terminal.
Two character sets, the standard ASCII and the IBM EBCDIC are in
widespread use as are several communication speeds. The FEMA system
could easily be equipped to handle all the common variants.
In an emergency,many
thousands of inquiries and reports per hour might be required.
FEMA probably cannot afford to maintain enough lines
and enough data handling capability for the emergency. Moreover,
the database would have to be distributed to avoid overloading the
long distance telephone system.
However, there are hundreds if not thousands of database
systems in the country with tens of thousands of connected
terminals that could be mobilized in an emergency with suitable
advanced planning. Insurance companies, airlines, car rental
agencies, and local governments have database systems that could be
loaded with emergency data and programs. The loading could be
accomplished from the FEMA database, but the process would require
development and some debugging. Perhaps the debugging could be done
once for each commercially available system.
The police database systems can play a special role, because
they are accessible from police cars, and because the police will
have special responsibilities in emergencies. Moreover, in a war
emergency, the police will have more important things to do than
to check whether there are outstanding traffic warrants against
people stopped for illegal U-turns. Much time can be saved by
deciding in advance what kinds of information would be useful to
include in them in an emergency and by practicing loading the
information and interrogating it.
Developing an emergency database mobilization plan requires
considerable study. Probably there needs to be a preliminary
conference including people from Federal, state and local governments,
the companies that supply database systems, the telephone companies,
and academics interested in problems of computer science, database
design and emergency planning.
∂10-Nov-80 1356 JMC
To: REG
Jim Koerber, v-p Marketing Dastek (an IBM disk spinoff), 866-0550, Los Gatos
is interested in letting us have a 400 megabyte disk in exchange for
trying it out on D.E.C. equipment. I told him you would call him to
see if there was mutual interest.
∂10-Nov-80 1036 JMC
To: MRC at SU-AI
I was not offended by your criticism of WAITS, and I have no intention
of trying to perpetuate WAITS beyond the present hardware. I was offended
by your refusing to take your boots off the chair.
∂09-Nov-80 2052 JMC charges
To: TOB at SU-AI, REG at SU-AI, BS at SU-AI,
csd.ullman at SU-SCORE
I'm not convinced you are being overcharged - at least not by very much.
Given the department's computer facilities budget of $800K per year (part
paid by other users of SCORE) and research budget of $6.5 million, it would
seem that about ten percent should go for computing - or a bit more.
This was about what it was when SAIL was operating solely supported by
ARPA, but the non-ARPA grants were probably undercharged. It might be
that the department facilities are over-opulent. In particular, it might
be that the Altos are a gift horse that must be looked in the mouth.
You should include REG in your distribution of such messages. I'll forward
it.
∂09-Nov-80 2052 JMC
To: REG
∂09-Nov-80 2004 TOB
To: BS at SU-AI, JMC at SU-AI, DPB at SU-AI,
csd.ullman at SU-SCORE, TOB at SU-AI
I think that we are being overcharged heavily on a grant which is in
bad trouble financially. We have been charged $13,461 through 3/4 of
a year. We are almost broke now and have six months to run. I was
told that there would be refunds when I pointed out overcharges before,
but there have been new, heavy charges.
The situation is serious. There are several changes that I would like:
0. I would like refunds on our NSF contract.
1. I propose that cost center charges be paid for by direct orders signed by
those responsible for research contracts, etc. They are now charged against
contracts without being approved by the investigators. I want to sign
for charges against us.
2. I would like to have charges stabilized so that we can plan and
budget for them. It is difficult to plan for retroactive charges.
I would like to have some voice in setting charge structures.
3. I propose changing the arbitrary charging for facilities which we
don't use, i.e. ALTO and SCORE. I have not heard a convincing
justification for it.
SCORE and SAIL were research facilities. The cost center seems to be
moving from that position. I oppose making them open machines. There
is an infinite sink of potential computing use outside.
Research groups have allocated substantial money to provide adequate
facilities. Until academic computing provides both funds and facilities,
I strongly urge that we consider current realities that there aren't
enough facilities, and that we restrict access. I think that it is
better to seek separate facilities for large-scale academic computing.
∂07-Nov-80 1136 JMC
To: FFL
The Dialnet expert is Mark Crispin - MRC to the computer.
∂06-Nov-80 1617 JMC
To: FFL
schank.1[let,jmc].
∂05-Nov-80 1744 JMC
To: DEK
I'll bring it in..
∂05-Nov-80 1426 JMC
To: nilsson at SRI-KL
COMMON.[E80,JMC] Programs with Common Sense
∂05-Nov-80 1423 JMC
To: CET at SU-AI
I would be happy to give a talk in the colloquium on January 13.
∂05-Nov-80 1005 JMC
To: REM
The six move mate was at levels 2 and 3 of the Chess Challenger 7. At level
3 it takes about one minute and twenty seconds per move. The level 7 17
move mate (I think it was 17) takes about 3 minutes per move. The levels
aren't consecutive. The games are on the AI bulletin board.
∂05-Nov-80 1002 JMC
To: solomon at MIT-MULTICS
Yes. I would be glad to help. While I don't know the Hoover
people who are closely connected with Reagan, probably Edward Teller would
help introduce us to the right people. I'll await the call, since I want
to be sure I understand what is needed.
I have an additional concern. I think that computers are used
very inefficiently everywhere, but probably especially in the Government.
Mainly they are overstaffed. Moreover, I think new technology could save
a lot of money for the Government and for everyone who has to deal with
the Government. For example, the Government could probably save money by
offering a $5 to $10 premium for electronically submitted income tax
returns, and the companies that prepare the returns for accountants and
tax companies could easily arrange this for their clients.
∂05-Nov-80 0829 JMC
To: csd.strauss at SU-SCORE
Thanks for your comments. I will do a large number of examples
tomorrow. Some changes in the format of the proof chapter
haven't worked out as well as hoped, but since I think that a good
understanding of proof techniques is important, I'll take whatever
time is required.
∂05-Nov-80 0106 JMC account at SU-AI
To: pourne at MIT-MC
I will tell you by phone the password of the account I have created for
you, but it is the first six letters of a word connected with a matter
we have discussed.
When I see you, perhaps this weekend, I'll show you how to use it.
∂05-Nov-80 0105 JMC
To: REG at SU-AI
CC: pourne at MIT-MC
I have created a directory 1,jxp for
Jerry Pournelle
12051 Laurel Terrace
Studio City, California 91604
213 762-2256
It should get one aliquot and its costs should be charged to 2FCZ601.
∂04-Nov-80 1903 JMC
To: REM
To Huang via REM: McCarty is a common name. No relation.
∂04-Nov-80 1139 JMC
To: ARG
I am willing to be on your committee.
∂04-Nov-80 1137 JMC Datamedia
To: buchanan at SUMEX-AIM
CC: ME at SU-AI
We can lend you a Datamedia temporarily, but probably until
the end of the quarter. You can get it from Martin Frost in
030.
∂04-Nov-80 0918 JMC
To: buchanan at SUMEX-AIM
Do you still want to borrow a Datamedia?
∂04-Nov-80 0010 JMC
To: CG
If you want a Datamedia for home, it's available now from ME.
∂03-Nov-80 1830 JMC seminar
To: FFL
October 6, 4:15
Topic: General discussion on non-monotonic reasoning
∂03-Nov-80 1830 JMC
To: FFL
Please reserve for 2 faculty club also Friday.
∂03-Nov-80 1828 JMC
To: REM
What happened (re spirit of chess challenger 7), somebody mate in 4 ?
I offered a prize for the fastest mates, and the deadline was the
end of October, and the prize was won by Alan Siegel, and I'm presenting
it tomorrow - all of $10. The best mate was six moves, I believe.
∂03-Nov-80 1559 JMC →14010,14011
To: *, "#EVENT.LOG[2,2]"
Event of TUE 4-Nov-80 at 4:00pm
Presentation of Hubert Dreyfuss and Bobby Fischer memorial award for
breaking the spirit of a chess program, in this case Chess Challenger 7.
Third floor lounge.
∂03-Nov-80 1557 JMC →14010,14011
To: *, "#EVENT.LOG[2,2]"
Event of TUE 4-Nov-80
, 4:15pm
∂03-Nov-80 1427 JMC
To: FFL
Please make a reservation for two for Wednesday at 12:15 at the Faculty Club.
∂03-Nov-80 1426 JMC
To: nilsson at SRI-KL
Fine. How about Faculty club at 12:15?
∂03-Nov-80 1102 JMC
To: FFL
OK, I'll chair it. Please call and put it in my calendar.
∂31-Oct-80 1607 JMC
To: stan at SRI-KL
Since I plan to continue the seminar next quarter, there is certainly
time. However, it may be worthwhile to do it this quarter, but I
don't have the schedule yet.
∂30-Oct-80 2249 JMC imlac
To: ME, ROB, ROY
It didn't come up when the system came up this evening. Is there
anything SYSTEMatic about this?
∂30-Oct-80 1535 JMC
To: csd.tajnai at SU-SCORE
There will be no text for my winter course though there may be some
notes.
∂30-Oct-80 1450 JMC
To: LGC
Can you talk next week?
∂30-Oct-80 1443 JMC
To: konolige at SRI-KL
OK, then I take it you can give the lecture given more time. I would
like to do it before people have forgotten about non-monotonic
reasoning, but I want you to take whatever time is required to
prepare. We can discuss how much time is required this afternoon
if you plan to come.
∂29-Oct-80 1959 JMC
To: Pool at MIT-MULTICS
Fine, let's talk, but there may be some misunderstandings. I said
$500 not $50 (perhaps just a misprint) and that with a substantial
market. Also I have some doubts that what should be a commercial
product will end up having its development subsidized by ARPA. I
will be around most of the time this week and next. I can be reached
at 415 497-4430 most afternoons and 415 857-0672 in the morning
except Tuesday and Thursday and most evenings. Not today, however.
∂29-Oct-80 0923 JMC
To: TOB
Sure I'll buy it, even though I'm not rich these days.
∂29-Oct-80 0913 JMC
To: human-nets at MIT-AI
Between 1955 and 1961, IBM developed a computer called Stretch
briefly marketed as the 7030. It was an advanced computer but
was too expensive and didn't meet its speed goals. A variant
of this computer called Harvest with the capability of handling
data streams was developed by IBM for NSA. It would have been
delivered in the early 1960. The computer project was not
classified, but there can be no guarantee that the use of the
computer coincided with the rumored use - cryptanalysis. There
is also nothing to prevent the recycling of a 20 year old computer
name. However, the Progressive has so much other wrong information
that it's entirely possible that "Harvest" refers to that old
computer, which seems unlikely to still be in use.
∂28-Oct-80 1412 JMC
To: FFL
Knowledge and Action seminar, thursday Oct 30, 4:15pm
room: MJH 301
speaker: John McCarthy
Non-monotonic reasoning - especially circumscription
Because questions have arisen concerning the motivation
for developing and formalizing various systems of non-monotonic
reasoning, this talk will concentrate on motivation and examples.
Some distinctions between AI motivations and philosophical
motivations will be proposed. Some ideas will be presented
about using non-monotonic reasoning to avoid making certain
kinds of distinction until and unless necessary.
∂28-Oct-80 0113 JMC
To: pourne at MIT-MC
Well, I'm torn between coming down for the parties and coming down for
the actual show. I accepted the invitation for the 12th that somehow
appeared, so I suppose that would take priority, and I have to teach
in between, though, come to think of it, I could schedule an exam
for Tuesday the 11th, but then I probably shouldn't take that much
time. Anyway, I'll dither a while longer.
∂28-Oct-80 0110 JMC
To: nilsson at SRI-KL
How about lunch Wed, Thurs or Fri to discuss book?
∂28-Oct-80 0108 JMC
To: pourne at MIT-MC
Well, I could have made it better with more time, but considering how
difficult it was to make it jell, it's just as well there wasn't more
time considering everything else I have to do. Thanks for your patience.
∂27-Oct-80 1536 JMC
To: admin.library at SU-SCORE
OK, I'll bring it in today.
∂27-Oct-80 1407 JMC lecture in seminar
To: konolige at SRI-KL
Assuming you are taking the knowledge and action seminar for credit,
I wonder if you would be willing to give a talk on Ray Reiter's
"A Logic of Defaults" paper in the AI Journal. If you are already
somewhat familiar with the paper, we could try for next week in order
to finish of non-monotonic reasoning, but it is probably more realistic
to plan to do it later in the quarter.
∂27-Oct-80 1245 JMC reference
To: FFL
I need the full bibliographic reference to a report of the
American Physical Society on the subject of Nuclear Fuel
Cycles and Waste Disposal. This may not be the exact title,
but it appeared as a special issue of the journal Reviews of
Modern Physics. The librarian of the physics library would
be able to find it.
∂26-Oct-80 2114 JMC Solvency of computer facility
To: csd.ullman at SU-SCORE
CC: BS at SU-AI
CC: DPB at SU-AI
CC: REG at SU-AI
I am very much worried that proposals are being submitted with
budgets for computing that are too small by a factor of about five.
We have about six million per year in contracts and grants and
need to recover about $650,000 per year in computer costs, but
some of the grants and contracts may do their computing on SUMEX
and not contribute to the $650,000. Betty just prepared a budget
for me which, on the advice of Denny Brown proposed $1850 for next
year in computing costs out of a total direct costs of $78,000.
I fear that in juggling the pricing algorithm to initially make
an aliquot apply to all three systems and then allowing a project
to buy time on just one system, some error in arriving at a total
has been made.
It won't do to say that if the facility isn't paying for
itself, we will raise the price of an aliquot. If the computing
money isn't in the budgets of the grants and contracts, then CSD
won't get it. The consequence of the CSD computer facility going
bankrupt might be that the university would take it over and put
it under the control of C.I.T.
Perhaps I have somehow got it all wrong, but I would be
relieved if someone would add up the sources of money again and
verify that it comes out right.
∂26-Oct-80 1706 JMC
To: JJW
It occurs to me that I have a UDP I can lend.
∂24-Oct-80 1215 JMC
To: FFL
Please print the new von.xgp[f80 and ulam.xgp[let.
∂24-Oct-80 1119 JMC
To: FFL
University of Florida, and I'll write a cover letter.
∂23-Oct-80 2126 JMC
To: FFL
Please print von.xgp[f80,jmc] to send to Ulam.
∂23-Oct-80 1553 JMC
To: csd.ullman at SU-SCORE
No I have not been getting notices.
∂23-Oct-80 1456 JMC
To: rms at MIT-AI
Tapes were left for you on my Russian-English dictionary.
∂23-Oct-80 1258 JMC
To: pourne at MIT-MC
It was mailed Monday. If it hasn't arrived Friday, call me and
I'll send another Federal Express or something like that.
∂23-Oct-80 1257 JMC
To: REM
The SAIL computer is now part of the computer science department's
facilities, and Ralph Gorin is director of these facilities. To the
extent that the AI Lab exists apart from that, I am still the
director.
∂23-Oct-80 0037 JMC
To: FFL
Please get me another key to my office.
∂23-Oct-80 0003 JMC boxes
To: FFL
Now that I have my bookcases, I have many cardboard boxes, which
I have collapsed again. Would you find out if the Department or
some standard part of Stanford wants them? Otherwise, I'll discard
them.
∂22-Oct-80 2353 JMC
To: JD
Sorry I'll miss your talk. I'd arrive too late.
∂22-Oct-80 2257 JMC
To: FFL
Did you locate Ulam?
∂22-Oct-80 1656 JMC
To: REM
You should send the above message to Ralph. Since, as I understand it,
you are doing some work for Ralph, we could arrange an internal account
for microfiching stuff relevant to that work, and you could pay normal
rates for adding some stuff of your own. Eventually, we will be able
to ship them files for microfiching by Ethernet. I don't see anyway
that Stanford is likely to let you make a business using Stanford
facilities, because the Stanford facilities are subsidized in various
ways, tax paying competitors could rightly object. Do you have
a microfiche reader? I have one reader in the office and one at home,
and I find it a pain in the neck. If you don't have one but have
some fiches to read, I will be glad to lend you a reader, so you can
see how you like it.
I have given up being boss of SAIL, so I am not the right person to
whom you should present such propositions.
∂22-Oct-80 1526 JMC desk
To: JJW
You can use the desk that belongs to Jon Doyle in room 360 for now.
You can get a key to the room from Fran Larson in 358.
∂22-Oct-80 1520 JMC
To: gaschnig at SRI-KL
Ignore previous. I have filled in paper form.
∂22-Oct-80 1516 JMC search algorithm mailing list
To: gaschnig at SRI-KL
Please include
John McCarthy
Computer Science Department
Stanford University
Stanford, CA 94305
∂22-Oct-80 1336 JMC
To: FFL
boyerm.re1
∂22-Oct-80 1333 JMC
To: FFL
colmer.2
∂22-Oct-80 1143 JMC
To: RWG
Aren't you glad you're a tenured SAIL user rather than a faculty member?
∂22-Oct-80 0933 JMC
To: csd.jeanie at SU-SCORE
CC: DEK at SU-AI
While Bill Gosper (RWG) has indeed been around a long time as a visitor
and occasional computer user, I must have missed the meeting last year
at which he was given tenure in that position, and what did poor Bob
Floyd (RWF) do to lose his? Who wrote Gosper's long form, and what is
his exact title?
By the way, what's on the agenda for the Nov. 4 meeting?
∂21-Oct-80 2257 JMC
To: rms at MIT-AI
Would you like to talk now?
∂21-Oct-80 1831 JMC
To: FFL at SU-AI
CC: rms at MIT-AI
Stallman is a guest and can xerox courtesy the ARPA contract or
my NSF.
∂21-Oct-80 0921 JMC
To: FFL
Please print ENERGY.XGP[F80,JMC] and make 5 copies before noon.
∂21-Oct-80 0147 JMC
To: DCO at SU-AI
Is the following the paper that should be a reference in the LISP text
concerning your algorithm?
%3Oppen, Derek%1 [1978]%*: "Reasoning
about Recursively Defined Data Structures",
in %2Proceedings of the 5th ACM Symposium on Principles of Programming
Languages (POPL)%1.
∂20-Oct-80 2235 JMC
To: FFL
I need to send a biography to Colmerauer.
∂20-Oct-80 1759 JMC article and voyager
To: pourne at MIT-MC
The article is in the mail, and I have received an invitation
to view Voyager on the 12th, and I'll probably come.
∂20-Oct-80 1757 JMC
To: FFL
Please pub and print boyerm.re1[let,jmc].
∂20-Oct-80 1616 JMC
To: RPG
Many thanks (unless there was a disaster, in which case grrrr).
∂20-Oct-80 0211 JMC
To: FFL
innova.xgp[f80,jmc] and pourne.xgp[let,jmc].
∂20-Oct-80 0155 JMC ce <filename>.>
To: BUG-e
Mostly references to <filename>.> refer to the file with the largest
numerical extension, but if there are no files <filename>.<ext>, then
ce <filename>.> creates one with extension >, but a subsequent attempt
to delete it gets the reply
">" illegal where used. I had expected that
ce <filename>.> would create <filename>.1 under these circumstances.
∂18-Oct-80 2322 JMC
To: FFL
digill.1[let,jmc],von[f80,jmc], a short biography
∂18-Oct-80 0117 JMC
To: ROY
It looks like my Imlac needs its frequencies adjusted again.
∂17-Oct-80 1800 JMC
To: FFL
NEWBOR.3[LET,JMC].
∂17-Oct-80 1156 JMC
To: FFL
2:30 is fine for Shubik.
∂16-Oct-80 1809 JMC terminal and Fred Abramson
To: JK
There is a possibility that we might be able to use some
of the $25K to pay Fred this summer, but Ullman would
prefer to spend it sooner so we can get more.
Also I just found a surplus dd line you can use, but Ralph
says he also has one for you. He said you should keep
pestering him.
∂16-Oct-80 0936 JMC
To: FFL
Please get me a faculty club reservation for two for noon.
∂15-Oct-80 2337 JMC
To: JMM
There is a FOL manual in the M box for you.
∂15-Oct-80 2336 JMC
To: JJW
There is a FOL manual in the W box for you.
∂15-Oct-80 1229 JMC
To: FFL
monthl.1[let,jmc]
∂15-Oct-80 1222 JMC
To: JD
I'll try to make it.
∂15-Oct-80 1041 JMC
To: DEK
Thanks for message. Unfortunately, I have visitor thurs but will try to
make tues.
∂14-Oct-80 2217 JMC
To: rem at MIT-MC
Back later. I'm going to drink coffee now.
∂14-Oct-80 2214 JMC
To: rem at MIT-MC
No, I didn't see it. What kind of program is it?
∂14-Oct-80 2211 JMC
To: rem at MIT-MC
I'm back at terminal.
∂14-Oct-80 1759 JMC
To: JK
Sorry. I rushed home to meet the driver with my bookcases and forgot.
∂14-Oct-80 1343 JMC tell Sue MJ211
To: FFL
There is a system message about proofs for Robert Brayton. Tell her
that he is with IBM Research Yorktown. If he is visiting someone
around here, Friedl might know about it.
∂13-Oct-80 2014 JMC
To: RWW
As to the livability of the office, I suggest you begin with the
occupants who seem reasonable, or at least polite. Barring satisfaction,
you can try Denny, but you are liable to comebacks about no being
in residence. Paul Martin is scheduled to leave at the end of the
quarter, and we can negotiate about Ann. Of course, there have been
delays before with Paul Martin.
∂13-Oct-80 1825 JMC FOL manuals
To: CJS
CC: FFL, JJW, RWW
We need about 10 quite soon. If they are packed, Joe Weening
has volunteered to help you unpack them. If they have to be
reprinted, about 40 should be printed and charged to the
NSF proof-checking grant (RWW).
∂13-Oct-80 1541 JMC
To: FFL
POSSIBLE WORLD SEMANTICS FOR KNOWLEDGE AND ACTION
In the "classical" AI planning paradigm, the planner is assumed to
have complete knowledge of all relevant aspects of the problem domain
and problem situation in which it operates. In the real world,
however, planning and acting must frequently be performed without
complete knowledge of the situation. This fact imposes at least two
additional burdens on an intelligent agent trying to act effectively.
First, when the agent entertains a plan for achieving some goal, he
must consider not only whether the physical prerequisites of the plan
are satisfied, but also whether he has all the information neccessary
to carry out the plan. Second, he must be able to reason about what
he can do to obtain the necessary information that he currently lacks.
In this talk I will outline some of the basic ideas behind a system
for reasoning about the interactions between knowledge and action.
The key point is that knowledge and action can both be described using
modal operators having possible-world semantics. This means that we
will reason not about what facts someone knows, but rather what
possible situations are compatible with what he knows, and what
situation would result from performing a given action in a given
possible situation. These ideas will be illustrated by exploring what
it means for an action to be a test.
-------
∂13-Oct-80 1206 JMC
To: Solomon.Datanet at MIT-MULTICS
No. I've never heard the term before.
∂13-Oct-80 1129 JMC
To: bmoore at SRI-KL
Title and abstract?
∂13-Oct-80 1128 JMC
To: FFL
For Moore on Thursday, we'll need a transparency projector.
∂13-Oct-80 0344 JMC
To: pourne at MIT-MC
Better call.
∂13-Oct-80 0339 JMC
To: pourne at MIT-MC
Shall I phone now or you phone (415)857-0672 before 5 min.
∂13-Oct-80 0336 JMC
To: pourne at MIT-MC
I wasn't planning to come, and it would depend on the which day, but
if I could visit you and tag along, it would be very interesting.
I now have a draft with which I am not satisfied. It has some
duplications which are partly consequences of editing through a
knothole. Would you prefer
(a) I submit it as is and you say yes or no.
(b) You ftp the file (which is INNOVA[F80,JMC]) and you
make editorial suggestions.
(c) I U.S. mail a copy for same purpose.
(d) I electronic mail you a copy.
(e) You telephone (415)321-4562 and tell the computer
TYPE INNOVA[F80,JMC].
All the electronic means will get you versions that have
a few random PUB commands. These won't confuse you about the
sense, but they might confuse Fawcett.
∂13-Oct-80 0123 JMC
To: rem at MIT-MC
Ok, but when you do want editing work, remember it's "minuscule".
∂13-Oct-80 0031 JMC
To: rem at MIT-MC
It's not so easy to get perfect accuracy if you also require conciseness.
The speaker was right that the kind of epidemic as when plague killed
half the population of Europe or smallpox killed thousands in one city
in a year or cholera killed thousands have been eliminated. This has
caused the word "epidemic" to be inflated to refer to the common cold
and venereal disease. Can you reword the statement equally concisely,
and so that it will convey just as much information to the average
person and still be literally correct? Sometimes I have condensed
a statement by consciously eliminating a qualification with the
expected side-effect that an expert would find the statement not
literally correct, but would get the information as would also the
non-expert. I could reword this message, so as to make it shorter
and with fewer misstatements if I took more time.
Are you interested in work as an editor?
∂12-Oct-80 2147 JMC
To: rem at MIT-MC
What is the ENERGY mailing list, and how can I read samples?
∂12-Oct-80 2039 JMC
To: JK
Will you be in some time tomorrow?
∂12-Oct-80 2004 JMC
To: rem at MIT-MC
How do you get there from here? Is there any way the last month's
discussion in HUMAN-NETS could get boiled down, written better,
rid of acknowledged factual errors? I suppose if there were an
editor, who could promise a larger and paying audience, this editor
could get a number of the HUMAN-NETS contributors to rewrite their
contributions and maybe those of editors. My impression is that
a factor of five compression wouldn't be hard.
∂12-Oct-80 1832 JMC
To: RPG
The previous message from me was intended for VRP. You were a misprint.
∂12-Oct-80 1830 JMC
To: VRP
I think that maybe SAIL's use of those codes antedates the final decisions
about ASCII, but I don't know. Dave Poole (Foonly) might be able to
say if you run across his path, and Steve Russell certainly would know.
It might be well to stick with S-expressions for a while; Ketonen is
doing that with his new proof-checker, and it allows more freedom for
the eventual syntax. I don't have any other suggestion than to see what
RPG will get around to doing, and since you are in good communication
with him, I doubt that I could add much. However, I'll mention it
next time I see him.
It seems that the world lost when I let Feldman push SAIL as the
"Stanford Artificial Intelligence Language" even though I did
reduce him to the acronym not regarded as an abbreviation.
∂12-Oct-80 1829 JMC
To: RPG
I think that maybe SAIL's use of those codes antedates the final decisions
about ASCII, but I don't know. Dave Poole (Foonly) might be able to
say if you run across his path, and Steve Russell certainly would know.
It might be well to stick with S-expressions for a while; Ketonen is
doing that with his new proof-checker, and it allows more freedom for
the eventual syntax. I don't have any other suggestion than to see what
RPG will get around to doing, and since you are in good communication
with him, I doubt that I could add much. However, I'll mention it
next time I see him.
It seems that the world lost when I let Feldman push SAIL as the
"Stanford Artificial Intelligence Language" even though I did
reduce him to the acronym not regarded as an abbreviation.
∂12-Oct-80 1720 JMC
To: VRP
You wrote "I've been trying to sell the SUN people on the virtues of Lisp,
but I find that Maclisp doesn't seem to work properly from datadiscs
(seems to map ..." It seems that we should talk to RPG, though I wouldn't
be surprised if he claimed it was a system problem. My query as to what
applications are recommended still stands.
∂12-Oct-80 1643 JMC
To: csd.ullman at SU-SCORE
In the past fifteen years, Les protected me from much knowledge of good
and evil. Probably Betty also knows, and the problem is as you say. I
don't believe there are any deferrable expenditures coming up like trips.
I have an various unrestricted accounts that can guarantee money, but they
don't have enough in them to pay salaries for very long.
∂12-Oct-80 1638 JMC
To: JK
I can't even understand how to begin using EKL from the manual, but
maybe starting at the beginning of it isn't right. As I mentioned,
perhaps I need a lesson, but surely we eventually need an introduction
that assumes no previous acquaintance with the project. Also Joe Weening,
a new graduate student, will need an introduction.
∂12-Oct-80 1634 JMC
To: VRP
For what purpose are you advocating LISP? I am now inclined to be more
assertive than formerly about the virtues of LISP for general purpose use.
I can talk to RPG once I understand what's at issue, but I can see that
e won't get very excited about a general claim that Maclisp doesn't work
from Datadiscs, since people have been using it from Datadiscs for years,
and presumably have been getting around minor character problems, which
I confess I haven't heard about.
∂12-Oct-80 1628 JMC
To: csd.ullman at SU-SCORE
It is my understanding that the basic AI Lab contract began in Oct 79
and expires in 1981. If this is correct, maybe it can loan money
to the other activities.
∂12-Oct-80 1515 JMC
To: REG
What about 2nd Datadisc for Ketonen-Doyle office?
∂12-Oct-80 1004 JMC
To: rem at MIT-MC
There is one further complication if your idea were to become the
dominant way of conducting discussions. The present publication
bottleneck requires editors, who have standards of writing and make
authors write well and fit into limited space. Since readers' time
is limited, this is very important. Even the present size of
HUMAN-NETS is probably causing many dropouts from reading it. I am
sending this to you rather than broadcasting it, because I may want
to say it again later only better for HUMAN-NETS. The question is
"How do editors fit in?"
∂11-Oct-80 2332 JMC
To: pourne at MIT-MC
It's coming.
∂10-Oct-80 2115 JMC
To: csd.ullman at SU-SCORE
Does this apply to all ARPA money or just that from contracts (ARPA
doesn't give grants) recently awarded or renewed?
∂10-Oct-80 2114 JMC
To: JEF at SU-AI
Does this apply to all ARPA money or just that from contracts (ARPA
doesn't give grants) recently awarded or renewed?
∂10-Oct-80 2111 JMC
To: Solomon.Datanet at MIT-MULTICS
Thursday at 11 would be fine.
∂10-Oct-80 1619 JMC
To: DPB at SU-AI
I succeeded in changing the appointment, so next Thursday is
now convenient. My title will be "The role of mathematical logic
in artificial intelligence".
∂10-Oct-80 1418 JMC
To: DPB at SU-AI
Giving CS200 talk depends on moving an appointment with NY Times.
∂10-Oct-80 1007 JMC
To: FFL
Please add me (greep@Rand-Unix) to the network mailing list for your
seminar. Thanks. - Steven Tepper
∂09-Oct-80 2309 JMC
To: REM
I don't think Reagan "would try to institute a theocracy". He does
seem to be a bit fuzzy about scientific matters, however.
I'm not even sure that the Rev. Falwell would, although I'm sure
he would try to legislate some of his morality that I don't agree
with. By theocracy I mean as in Puritan New England.
∂09-Oct-80 1339 JMC
To: csd.ullman at SU-SCORE
I was waiting hear from you that Dantzig and Golub had assented.
You might have someone send me copies of the long green forms.
∂09-Oct-80 1137 JMC
To: FFL
Please tell Dr. Gibbons' office that I haven't been receiving notices of
CIS meetings and supposed that I had been dropped from the committee.
When I accepted membership, I said I couldn't serve till Fall.
∂09-Oct-80 1134 JMC
To: katsuki at BBND
There is John J. Craig a Phd student called JJC and a Computer Science -
Computer Engineering student John Craig called JRD.
∂08-Oct-80 2041 JMC
To: REM
I dunno about any of those things. I have become resigned to the
fact that Reagan goes off half cocked unduly often. It's not bad
if its just in speeches, but it could be bad if, as President, he
got stubborn about one of his bad ideas. He doesn't now seem to be
a stubborn person - unlike Carter. I hope he has a good
science adviser. It should be possible to figure out what he reads
or from whom he gets these ideas.
∂08-Oct-80 1721 JMC loop or something
To: BUG-DOVER
DOVER FOLPRM.REF[AIM,DOC] ran for 11 minutes and had produced a
monstrous number of records when I killed it on the advice of ME.
The file is about 22K.
∂08-Oct-80 1640 JMC
To: FFL
Please PUB and print FENAUG.3[LET,JMC]. I now use .3 instead of .LE3.
∂08-Oct-80 0218 JMC
To: LWE
CC: REG
That address might well lose you a tape. The correct address is Computer
Science Department, Stanford CA 94305. I trust that you will pay the
reasonable bill that I hope Ralph will send you, and therefore will not be
forever in debt.
∂07-Oct-80 2024 JMC
To: bmoore at SRI-KL
Thanks. That will presumably be on the 16th.
∂07-Oct-80 1439 JMC
To: FFL
Does lib.lst live yet?
∂07-Oct-80 0129 JMC
To: AAM
thanks
∂06-Oct-80 1823 JMC credit
To: konolige at SRI-KL
I haven't done anything about getting a number to give credit. Let's
talk about it at the meeting if it is important to anyone.
∂06-Oct-80 1820 JMC seminar on knowledge and action
To: bmoore at SRI-KL
I plan Thursday as an organization meeting. Would you be willing
to speak to the next meeting? I think it would be better to emphasize
formalism and technical results rather than make programmatic statements.
The Monterey Dunes meeting was so occupied with programmatic statements
that a substantial feeling of vagueness remained.
∂06-Oct-80 1734 JMC
To: AAM
I would like a copy of the CS204 problem set writeup.
∂06-Oct-80 0119 JMC
To: rms at MIT-AI
I guess I agree that it tends or perhaps tries to say the same thing. I
doubt that it comes to the same in their fixed point interpretation.
Also, the action of circumscribing certain sentences under certain
conditions might be built into a program rather than expressed as
a rule, although I naturally prefer expressing as much as possible
in the form of rules. I fear we need more extensive examples before
we can get much more out of general meta considerations.
∂06-Oct-80 0048 JMC
To: ME
I'm a bit surprised at the slowness of string substitution
in E. In order to count words, I had it subsitute space for space
in a 110 line text. It took 11 seconds to do 889 subsitutions.
A search for a character not present took only ten ticks.
∂05-Oct-80 1205 JMC
To: REG
I suspect you have got yourself into a trap. We don't have
need for an operations staff, or even one full time operations person.
However, occasional operational chores need to be someone's job other
than the manager's. Although you may fear that they will regard it
as "beneath them", you should bite the bullet and explicitly include
such chores in the jobs of maintenance people (ROB, ROY or DC),
programming (ME or MRC), your tape occasional tape dumper, or your
secretary. Otherwise you will be tempted to create a position.
∂05-Oct-80 0120 JMC von Neumann
To: newell at CMU-10A
I am writing a review for the Washington Post of "John von Neumann and
Norbert Wiener - From Mathematics to the Technologies of Life and Death"
by Steve J. Heims, M.I.T. Press. Among other things, the author treats
their interest in the relation of computers and brains. In this connection,
I wonder if von Neumann expressed any interest in your work at Rand
on the logic theory machine or chess or expressed interest in the
use of digital computers to carry out symbolic intellectual processes.
∂03-Oct-80 1457 JMC
To: RPG
Here are some file names, and some of them contain references to
the literature.
SAMEFR[F76,JMC] 21-Jan-77 SIGART article about samefringe
SAME.LSP[F76,JMC] 06-Feb-77 samefringe
SAME.PUB[F76,JMC] 13-Dec-76 samefringe
SAMEF[F76,JMC] 13-Dec-76 samefringe
SAME.FR[F76,JMC] 10-Dec-76 Someon's improved version
RESIDU.LSP[F76,JMC] 20-Jan-77 efficient version samefringe
SAMEFR[S77,JMC] 02-Jun-77 %3correctness OF %2samefringe%1
∂03-Oct-80 1340 JMC
To: JRA
It seems to me that Stanford would rightly not consider getting special
hardware for teaching LISP. It is rather episodic, and space and
maintenance would be expensive. We have so far resisted the idea
of using microcomputers even for Pascal. The one possibility is that
if the department goes in for Pascal on microcomputers, LISP might
be done on the same. To pursue this possibility, talk to Ralph Gorin,
who is pursuing the possibilities of using microcomputers on the
instructions of Gerald Lieberman, Associate Provost for Research,
who is also the administrator to whom LOTS reports,
but I should tell you that I am unlikely to favor it.
∂03-Oct-80 0154 JMC
To: LWE
I don't watch TV, so I've missed it. The Reagan strategy of turning
away wrath with soft answers has almost neutralized several very
liberal columnists, e.g. Tom Wicker and Mary McGrory, who can't resist
attacking Carter, even though they'll almost certainly end up
supporting him. There are still almost no bumper stickers out here,
except for Anderson's, and while I, a registered Republican, have
received much Republican direct mail, and no Democratic. Both sides
concentrate on turning out their supporters in ways that run as little
risk as possible of exciting the opponents enough to vote.
∂21-Oct-80 1214 JMC
To: FFL
I have my copies.
∂29-Oct-80 0921 JMC
To: human-nets at MIT-AI
Some friends repeatedly and vainly called the Reagan-won line. If the
lines are saturated, the system merely measures which call-in-line
works better rather than the number of people who try to call.
∂27-Nov-80 1521 JMC
To: FFL
suzuki.1
∂27-Nov-80 1525 JMC
To: FFL
dennet.1
∂27-Nov-80 2116 JMC
To: minsky at MIT-ML
It seems to me that you did work proportional to the number of bits
you wish to store when you cooled the container. Moreover, when you
built the insulator you do work proportional to the R-value you want,
which is in turn proportional to the length of time you can keep the
bits.
∂27-Nov-80 2118 JMC
To: JAK
I attempted to authorize you by sending a note to REG (Ralph Gorin) when
I created your directory. You might ask him, but it takes a while to
get in the official file.
∂27-Nov-80 2220 JMC
To: pourne at MIT-MC
The extreme thing to do is to divide the file into subfiles that
aren't too long. However, there is probably a better way. Let me
see if I can locate an expert who will help you.
∂27-Nov-80 2225 JMC
To: VRP
help wanted
Would you be willing to advise Jerry Pournelle (sf writer and
mc user) with the problem that follows. "here" is L.A. and I
presume he uses a TIP from his home computer system. He is
called POURNE@MC and may be logged in now.
(2) I badly need a good protocol for downloading files; some way
that files can be sent from its or sail to here, with a
character I can send to temporarily stop the sending while I
write off to disk; then continue from where left off. I don't
know of any progrtams on that end that can do that. Know any?
The problem is that the net keeps sending when we can 't be
receiving because the machine has filled the buffer and wants to
write to disk.
JEP
The extreme thing to do is to divide the file into subfiles that
aren't too long. However, there is probably a better way. Let me
see if I can locate an expert who will help you.
∂28-Nov-80 0138 JMC
To: pourne at MIT-MC
Did you get help?
∂28-Nov-80 0153 JMC
To: pourne at MIT-MC
I didn't mean a legion as an organization. Merely the idea of
taking foreigners into the armed services. Unlike the French,
Americans wouldn't be offended by having to share tents with
foreigners. Incidentally, I'll have a technological proposal
for you shortly as soon as I finish writing it up.
∂28-Nov-80 1732 JMC deletion
To: LWE
Actually it is much simpler than I made it. The command
copy foo ← digest.ns(n1:n2)
makes a new file consisting of pages n1 to n2 of digest.ns.
et foo
then asks permission to reformat foo, to which you answer
y<carriage return>.
You then check over the file foo to see that it has what you want,
exit from E, and give the system command
ren digest.ns←foo
and it asks you for permission to delete the old version of digest.ns,
to which your reply
y<carriage return>
and you're done. I did it for you.
∂28-Nov-80 1831 JMC
To: VRP
Thanks.
∂01-Dec-80 1637 JMC
To: SOL at SU-AI
I don't think so.
∂03-Dec-80 0116 JMC
To: energy at MIT-MC
Fission may not be exciting, though that may be partly a consequence
of holding the lid on ideas for new reactors for fifteen years or so,
but that isn't reason enough to use phrases like "dead end technology".
However, once enough countries without hangups about using it fully
(i.e. with breeders) generate almost all their electricity with it
and find some good electric car technology the world can push the
energy problem into the back of their minds, like say water supply.
It will be a problem at some time in some places, but it won't generate
religions any more.
∂03-Dec-80 0135 JMC
To: energy at MIT-MC
Years to pay back energy cost of nuclear plant
I was hoping someone had the references on this. The claim
that it is some large number of years has been often made and often
refuted. I think the last estimate I read from a source I regard
as respectable was something like three years. 22.5 years is implausible
no matter how you look at it, because a thousand megawatt plant at sixty
percent duty factor produces 5x10**9 kwh per year at say $.01 per kwh
makes about $50 million per year. In 22.5 years that would be over
a billion dollars. What plants cost varies widely depending on how
much delay and rework there is in the construction, but some have
come in for about a billion dollars. But energy is only a small part
of the cost of nuclear plant taking into account energy going into
materials. The over-runs are almost all labor cost. I apologize for
lack of detail, since I know that others have done this more carefully.
∂03-Dec-80 0149 JMC
To: energy at MIT-MC
Yes but:
I grant that SPS is more exciting and is one of the ways
of getting high capacity space transportation started, and I support
SPS to the extent of getting a start on the process of construction
in space so that we can see how costly it is.
I don't agree that the social cost of fission is especially
high. The fanatical anti-nukes are few, they lost the recent
election, and their attention will turn to other issues if we
persist. Moreover, fission is the most available source for the
next twenty years at least and fortunately being actively pursued
by many countries. In five years, Taiwan, with a smaller population
than California, will have more reactors and plans to have twenty
by the year 2000. The biggest problem I see isn't the anti-nuke
fanatics but the soft anti-nuke feeling among college students, professors,
etc. that being anti-nuke is a consequence of being a liberal good guy.
I call it a soft feeling, because it can be turned around by argument,
provided the individual isn't too dependent for his views on his
social milieu.
∂03-Dec-80 0152 JMC
To: JPM
I know Beckmann somewhat and can write him for you.
∂03-Dec-80 1845 JMC
To: rms at MIT-AI
I don't know the Kripke paper.
∂03-Dec-80 1929 JMC
To: ME
No nyt today
∂04-Dec-80 0016 JMC
To: RPG
I'll think a bit, and we'll talk, but some of the initiative would
have to come from you.
∂04-Dec-80 0048 JMC
To: FFL
todoro.6. Copy to Les Dugan, and both include pubbed version of energy.st2.
∂04-Dec-80 0158 JMC
To: bmoore at SRI-KL
I let your message of 24 nov fall into a crack. How about getting together
late next week, say Friday, for lunch and discussion. I could come to SRI.
∂04-Dec-80 1454 JMC
To: bmoore at SRI-KL
12th at 12
∂05-Dec-80 1021 JMC
To: human-nets at MIT-AI
MAYBE WE SHOULD MAKE THE ELECTORAL COLLEGE WORK AS INTENDED
The original idea of the electoral college was that its
delegates, chosen by the state legislatures, would meet and recruit
the best possible man for President. Maybe the idea worked the
very first time when Washington was chosen. Thereafter, parties
formed, and electoral college delegates were pledged before the
college met. This meant that they were not free to devote time
to deciding on the qualifications of the candidates and then
make up their minds.
This was unfortunate, but at least the party conventions
could do so. Eisenhower was chosen by the Republican convention
in 1952 without entering a single primary; the delegates thought
he was the best person for the party and the country. Now there
aren't even any unpledged delegates at the party conventions, so
even the candidates are chosen entirely by media campaigns.
Were the President chosen by unpledged delegates who made
up their minds after a month or two of interviewing candidates and
their supporters, it is likely that a different man would have been
chosen. It might have been someone who is not politically prominent,
but among the politically prominent people, Bush, Haig, Baker
and Anderson among Republicans, and Mondale, Moynihan and Udall
would have had a reasonable chance. Most political writers today
are urging reducing the number of primaries and making party and
government officials ex officio delegates to conventions in order
to reduce the effects of the media and give the conventions a wider
choice.
My proposal is that a jury of 200 randomly chosen citizens
should meet for three months and choose the President after hearing
arguments from the supporters of different candidates. The intent
is to approximate the effect of each citizen taking three months to
think about who should be President with really good access to
information and the ability to converse with the candidates. A
smaller jury would permit better discussion at the risk of greater
statistical fluctuation in a priori opinions.
Another possibility would be to keep the electoral college
but choose it by random selection from Congressmen and members of
state legislatures.
The anomaly of the electoral college was noted very early
in the nineteenth century. No party has ever found it important
enough to expend the political capital required to overcome the
preference of the small states for a system that gives them a
slight advantage.
∂05-Dec-80 1031 JMC
To: energy at MIT-MC
Fission plants take five years in France, Japan and Taiwan. I'm
surprised that Jerry Pournelle, of all people, should think that
America is intrinsically incapable of doing as well as Taiwan
and that the election of Reagan can't make a difference.
I believe Taiwan plants meet the American standards applicable
at the time of start of construction.
∂05-Dec-80 1033 JMC
To: REM
The California city is Pittsburg.
∂05-Dec-80 1053 JMC
To: energy at MIT-MC
REM's numbers of getting half our energy from sources preferred by
environmentalists by 1990 and 90 percent by 2020 is an eminently
reasonable compromise. Adopting it might make everyone happy and reduce
conflicts in our society.
Unfortunately, the laws of physics have not yet agreed to the
compromise. Seven years of effort have produced no assurance or even
likelihood that the sources the environmentalists prefer can produce the
energy required. Demonstration windmills are being built that will
produce energy at enormous financial and environmental cost, and they are
not prototypes of something calculated to be better. Solar energy from
SPS or from artificial photosynthetic splitting of water remain
possibilities, but a societal PLAN to use solar or fusion by 2020 or any
date is unreal until and unless engineering calculations indicate success.
∂05-Dec-80 1844 JMC
To: rem at MIT-MC
I'll try to find a reference to the information you want. Many of them
aren't 10 percent here and 10 percent there, but 1 percent here and
3 percent there. Moreover, even in a country as large as the U.S.,
there are extra expenses in keeping many technologies going simultaneously
for the same basic function. However, I must admit a bias. I would like
the energy problem to disappear from the forefront of public consciousness,
because I there are many other more interesting technological opportunities
to think about, and I think the breeder reactor will do it. Also I have
always had a negative reaction to environmentalism, i.e. to their
idea of a good society, their pessimism and their liking for scapegoasts.
Whenever they attack someone, my immediate reaction is to defend the
victim. Unfortunately, they aren't always totally wrong.
∂06-Dec-80 1321 JMC
To: rem at MIT-MC
You misparsed what I said, so I suppose some others did also. It would have been
clearer had that phrase been "someone not politically prominent. However,
among the politically prominent, "
∂06-Dec-80 1327 JMC
To: rem at MIT-MC
Agnew complained that the press was unfair to the Nixon Administration,
and further complained that much of the press deliberately misinterpreted
these complaints as advocacy of suppression of the press. In my opinion,
he was substantially correct in both complaints. He also took bribes when
he was Baltimore county executive. On many other occasions newspapers have
interpreted expressions of disagreement with their positions as attacks on
freedom of the press.
∂06-Dec-80 1354 JMC
To: human-nets at MIT-AI
ESCAPING FROM FIRES
A major cause of smoke inhalation death is carbon monoxide
poisoning. The CO binds to haemoglobin so it can't carry oxygen,
and it is very insidous, as shown by the accidents in which several
people running the engine of a stalled car to keep warm all die
without any of them noticing. I would like comments on the following
method of escape which has not been proposed in print to my knowledge
or properly tested.
Put a large plastic bag over your head filled with air and
tie it around your neck. An experiment indicates that you then
have at least five minutes of walking before even becoming
uncomfortable from your exhaled CO2. If you need more time than
that to escape maybe you can find a place where it can be refilled
with fresh air.
If it works, safety rules could require that suitable plastic
bags be available. Failing that, a traveler can carry one as I do for
laundry, and many hotel laundry bags are suitable. Dry cleaning
bags now almost always have holes in them because of the opposite
danger.
Experimenters to determine how far one can go with a bag of given
size are solicited.
∂06-Dec-80 1602 JMC
To: rem at MIT-MC
The person selected mightn't have been prominent at all. However, among
the present politically prominent people, Bush, Mondale, Haig, etc. would
have had a better chance of being selected by a body that interviewed
candidates than they did under the present system.
∂06-Dec-80 2019 JMC
To: REM
I wasn't arguing for Haig. He has been chosen for high positions
at various times, and his name was floated early as a possible
Republican candidate, but got nowhere, because he is little known
to the general public. Perhaps he would impress an electoral college
that interviewed many candidates; perhaps not. I was not even claiming
that the people mentioned were better than Carter and Reagan, though
I would bet that at least some of them are.
∂07-Dec-80 2319 JMC Barwise is now BYY.
To: FFL
Since he now has a SAIL account, BYY should be in the list of
people to be notified about the seminar, and he need not receive
paper notification.
∂08-Dec-80 0128 JMC
To: pourne at MIT-MC
Doesn't the Athenian system amount to choosing Congress by lot. I
didn't and don't advocate that.
∂08-Dec-80 0133 JMC
To: pourne at MIT-MC
OK. It would be better if you withdrew or modified your remark to
HUMAN-NETS rather than have me answer it and you agree.
∂08-Dec-80 0213 JMC form
To: BYY
It might be best to get that form to Susan Hill by Tuesday and remind her
to legitimize you, so your files won't get zapped in the purge threatened
for Tuesday. If this is difficult, I can help.
∂08-Dec-80 1517 JMC
To: FFL
JOURNA.1
∂08-Dec-80 1534 JMC mail minsky%mit-ai
To: BYY
dover <filename>
or
xs <filename>
mail minsky%mit-ai
∂08-Dec-80 1610 JMC FOL primer
To: BYY
There is also a FOL primer, and my secretary, Fran Larson, known to
the computer as FFL, can get one if you MAIL her a message.
∂08-Dec-80 1612 JMC xerox and secretarial
To: ALS
Ralph Gorin has promised you an auditron charged to the computer facility,
and my secretary, Fran Larson, FFL, can do secretarial work for you.
∂08-Dec-80 1623 JMC
To: FFL
I want to keep 1,jmc small. Letters like perm.1 should be on let.
∂08-Dec-80 1637 JMC
To: csd.ullman at SU-SCORE
I think the department should practice equal opportunity but not
affirmative action. - JMC
∂08-Dec-80 2331 JMC
To: csd.ullman at SU-SCORE
Since there aren't enough cases of genuine discrimination to reduce the
alternatives available to qualified women and negroes, Stanford affirmative
action policies merely stir the pot, no matter how they are carried out.
Alas, I have no suggestions worthy of acceptance. I have only the remark
that some years ago we accepted some graduate student applicants far down
in the list of applicants - with results regretted by us and probably by them.
The qualified minority candidates for teaching positions and graduates study
are quite likely to be from CMU and MIT. I hope that Kennedy's new burst
of activity won't cause that to happen again.
∂09-Dec-80 1145 JMC
To: FFL
Please telex reply to Rose Michaelson. "Glad to sponsor you. Hope you
will be coming to this area sometime. Regards also from Susie."
If there was a return address, use that, otherwise via RCA Communications to
10, Salisbury Road
Edinburgh
∂09-Dec-80 1441 JMC
To: BYY
I don't notice Tuesday at 4:15 as a change of day in logic seminar.
∂09-Dec-80 1534 JMC
To: FFL
Please print and then delete bernst.ns[f80,jmc] and send it to Todorovich.
∂10-Dec-80 1613 JMC
To: FFL
rosenh.1
∂10-Dec-80 1815 JMC problems
To: ROY
CC: REG
1. You were going to do something about my Datadisc.
2. Weyhrauch's Imlac has been down for many months, and I understand
that the problem is with the line or at this end. Should I give up
and give him a Datamedia?
∂10-Dec-80 1822 JMC
To: FFL
fenaug.4
∂11-Dec-80 0913 JMC
To: pourne at MIT-MC
I read the Commentary article last night, and I thought about
mentioning it to you and decided against it, because it is written for the
convinced, merely telling them who their opponents are. I would suppose
that some of the books about the holocaust mentioned in the article would
be more convincing than the article itself. The author of the article
mentions a 30 year old, but cannot bring herself to believe that it is
necessary to start from square one again.
Having taken it for granted since 1945 (Note that unlike World War
I atrocity stories, it didn't appear until the end of the war.), I need to
re-examine my sources of information. I find several reasonably
convincing. (1) My mother telling me that all our relatives from Lithuania
were killed. (2) Four other people telling me that their relatives were
killed. One of them was in Russia, where the holocaust against the Jews
is officially minimized. (3) One of them describing how her parents got
her accepted as a Catholic orphan in order to save her. (4) The
acceptance of the contention by the Germans. (5) The defenses at trials
being of the form: I was just an underling and wasn't the person who
actually gave the orders to kill some large number of people. (6) The
Commentary article of (last?) year about the behavior of certain specific
Jews from Warsaw, and the controversy about whether a certain Jew was
killed by another for collaborating. This article excited responses from
several persons who were inmates of that camp and who knew something about
the matter, all saying that the people who were actually present had all
been killed.
I believe there is actually a list of names in Jerusalem as part
of the memorial.
Admittedly all the above could be somehow explained away, but of
the common opinions I presently accept without a personal investigation,
this seems one of the best supported. The existence of such books as
Rassinier's doesn't impress me in itself, because much less important
matters have excited expose books advancing all sorts of theories.
∂11-Dec-80 2039 JMC
To: FFL
Please send a Yumyum to Feferman in Math and collect from me.
∂12-Dec-80 1621 JMC
Escaping from fire
It is my impression that many people die from CO
and other gas poisoning who are never close to flames or great heat.
The primary utility of the plastic bag would be in this case, but
I don't know what fraction of the cases are covered. If the bag
could withstand heat as well, this would improve survival chances.
This can be aided by adding as much cold water as possible using
a saturated towel or pillowcase, and keeping a wet cloth to wipe
hot spots with.
I see it more as a way in which people who know about it can save
themselves than as a candidate for a safety regulation. Of course,
an oxygen apparatus would be much better, but equipping all homes
and hotel rooms is an expense larger than is likely to be undertaken.
I had hoped that someone would be interested in more experiments on
how long one could breathe at what level of activity with a bag of
known size. It would also be important to know what fraction of
cases this method of escape is applicable to.
∂12-Dec-80 1621 JMC
To: human-nets at MIT-AI
Escaping from fire
It is my impression that many people die from CO
and other gas poisoning who are never close to flames or great heat.
The primary utility of the plastic bag would be in this case, but
I don't know what fraction of the cases are covered. If the bag
could withstand heat as well, this would improve survival chances.
This can be aided by adding as much cold water as possible using
a saturated towel or pillowcase, and keeping a wet cloth to wipe
hot spots with.
I see it more as a way in which people who know about it can save
themselves than as a candidate for a safety regulation. Of course,
an oxygen apparatus would be much better, but equipping all homes
and hotel rooms is an expense larger than is likely to be undertaken.
I had hoped that someone would be interested in more experiments on
how long one could breathe at what level of activity with a bag of
known size. It would also be important to know what fraction of
cases this method of escape is applicable to.
∂12-Dec-80 2139 JMC holocaust
To: pourne at MIT-MC
I am sending you a xerox of the chapter "The Second World War and the
Holocaust" from "A History of the Jewish People" edited by H. H. Ben-Sasson.
The authors are from Hebrew University and the book was published by
Harvard University Press, so I suppose that it represents the accepted
Israeli view of the events. Note the following:
1. They also say that the main extermination camps were in Poland but
also mention Lithuania and Russia. Belsen and Buchenwald are better
known to Americans, because they were liberated by Americans.
2. Their estimate of the number killed is between 4.5 million and
6 million. One might suspect that the author of the chapter (Ettinger)
had more definite views but wanted to avoid involvement in
disputes.
3. Many were killed in many ways including shooting, burning in buildings,
and death marches. No definite estimates are given of the number
gassed.
4. The precision of some of the numbers suggests that they come from
German sources. These sources might have exaggerated to impress their
superiors or in some cases for humanitarian reasons.
3. More than half the number was from Poland and the next largest
from Russia. The number from Russia must be especially difficult
to verify, but the Poles didn't co-operate too well either. I once
had a meeting in Warsaw, and we went to the memorial for the
resistance of the Warsaw Ghetto. It was very badly kept indicating
that the Polish Government wasn't much interested.
4. There are many uncertainties stemming about the prewar and postwar
Jewish populations stemming from questions about whether an individual
or family was classified as Jewish, since this was partly under the
control of the individual and partly not.
My former student Ruzena Bajcsy, who was sent as a graduate student
by Czechoslovakia in 1967 but didn't go home after the Russian
occupation, told me that she was placed by her parents in a Catholic
orphanage in Bratislava, which saved her when parents deported to the
camps. Younger children saved by such mechanisms are not easily
counted, because they might not later have been identified as Jews.
Suppose one is interested in independently verifying some alleged
history that has been challenged as a fake. What means are available?
That inconsistencies will exist is certain even when we have
eyewitness accounts of a single event. Estimating a charge of
fakery involves asking how many people would have to be involved
in a conspiracy - in this case a very large number.
It seems to me that there are many sources of evidence that the
Germans rounded up large numbers of Jews and deported them with
the claim that they were being resettled. Since this was known
to the population of many countries, it could hardly be a fake.
No-one at all admits to having been resettled, and the non-return
of most (if not almost all) of those rounded up is also known
to their non-Jewish neighbors. The various methods the Germans
are reputed to have used to kill them are reported by a much
smaller but still large number of people. Those tried for
crimes in connection with the extermination, so far as I know,
didn't deny the facts of extermination but merely (sometimes)
denied or minimized their personal roles.
The biggest uncertainty concerns numbers. Estimates of the size
of a crowd vary by a factor of ten, and many people don't know
the population of the U.S. within a factor of ten. No doubt
more specialized books have more detailed estimates of numbers
than the chapter I'm sending you, but there are probably
methodological questions about avoiding double counting that
leave a considerable residue of uncertainty. Probably it
would take a rare person to be motivated enough by a desire
to get an accurate count to take on people who would regard
any such attempt as an attempt to whitewash the Nazis.
There may, however, be reviews of Rassinier et. al. based on
equal knowledge and less passion than Lucy Davidowicz's.
As you say, there are more productive things to think about.
∂13-Dec-80 1053 JMC
To: REM
Not yet.
∂13-Dec-80 1134 JMC helping the boss
To: pourne at MIT-MC
I read your BOSS file. While none of the writers in it knows exactly what
the White House now has, the discussion was quite reasonable. Evidently,
the idea has caught on to the point that much larger amounts of money have
already been spent than you were originally contemplating. There is
always the possibility that whoever is doing the White House job now is
behind the state of the art, but otherwise one would expect things to be
in a reasonable state.
Here are some issues:
1. Is the system used by policy people or merely by secretaries.
If the latter, something isn't being done right.
2. Secure use by remote terminals is feasible. Any documents that
can be entrusted to a safe at someone's home could also be entrusted with
somewhat greater security by a secure terminal at home. The terminal
would need its own editor and would communicate cryptographically.
Ability to use the facilities of the central computer interactively might
be severely limited in order to avoid an interaction style that can be
cryptographically attacked. Actually, I think that can be provided also.
3. Managing the electronic mail of a big shot to whom very many
people want to communicate and who may want to have his mail filtered may
not be an optimally solved problem. There are better solutions than
having a secretary print all the mail and sort it, but I don't know if any
have been implemented. I have noted that the ARPA people are not very
promptly responsive to electronic mail.
4. It would be very important for the White House or Congress
to have publically available computers that can store material for
which public comment is being solicited. The Freedom of Information
Actt can be made much more effective if there is terminal access
to the information that it makes available to the public. There
are already enough terminals so that such facilities would be
extensively used. Perhaps the White House press office would be
a good place to start.
∂13-Dec-80 1604 JMC
To: FFL
Please pub cbcl[f75,jmc] and send 4 copies to Nilsson at SRI by campus mail.
∂13-Dec-80 1641 JMC
To: FFL
Please send a pub of energy.st2[f80,jmc] to Elliott Bloom at SLAC.
∂13-Dec-80 1650 JMC
To: ME
The xpoint <filename>/c should leave you in readwrite mode rather
than readonly.
∂14-Dec-80 0031 JMC book
To: FFL
Please have the bookstore order for me "Logic and Databases" by Gallaire
and Minker - Plenum Press.
∂15-Dec-80 2231 JMC
To: FFL
Please send out seminar message announcing Stallman for this Thurs.
∂15-Dec-80 2314 JMC proof-checking for courses
To: JK
I found having students prove programs without using a proof-checker rather
unsatisfactory this year. As it happens there is some IBM money that the
department gets that might be used to develop a proof-checking system for
classes. In my present state of mind even an unmodified FOL would be usable,
but the ekl has projected features that might also be useful. Can you come
to a meeting Friday at 2pm to discuss it with me and RWW and CG and CLT?
If the time is unsuitable, it can be changed.
∂16-Dec-80 1609 JMC
To: csd.hansen at SU-SCORE
I see your difficulty, but I don't know a standard solution. One way out
might be to subst the gensym into the array statement before executing
it. Since array evidently quotes its arguments, this will give it the
gensym.
∂16-Dec-80 1819 JMC
To: FFL
davis.re1
∂17-Dec-80 1437 JMC
To: FFL
The program is called SUDS, and you should refer him to Ralph Gorin.
∂17-Dec-80 1549 JMC using FOL in CS226
To: csd.ullman at SU-SCORE
This is a request for financial support from the department or other part
of the University to support the use of the FOL interactive theorem prover
by students in CS226 (Epistemological Problems of Artificial
Intelligence). There were about 10 students in all the last time I taught
the course, although I hope for a few more this time, since the subject
has come more to the forefront of AI, and the course is cross-listed in
philosophy for the first time. FOL is available only on the SAIL computer
and takes advantage of the character set of the SAIL computer which
contains the symbols used in mathematical logic. It is not readily
transferred to LOTS for this reason, and also because there is no-one
available to do the work.
Throughout the University, there are courses that should use programs
created for particular machines in connection with research. Where this
use is warranted and LOTS is unsuitable, and the machines are cost
centers, I believe the University should use its instructional computing
funds for the purpose. There is no justification for dedicating these
funds to C.I.T. use.
∂17-Dec-80 1645 JMC ACM
To: FFL
I seem to have let my membership in ACM (Association for Computing
Machinery) lapse, probably by neglecting a dues notice last year some
time. At least I haven't received any publications recently. I guess
the quickest way to renew it is to check whether the Department office
has the forms and the list of publications. If the office doesn't
have it, then many of the faculty will.
∂18-Dec-80 0840 JMC
The best is the enemy of the good.
While the best plastic bag for escaping fires was 20 years
in development, perhaps many died who would have been saved had the
possibilities of ordinary plastic bags been publicized.
∂18-Dec-80 0841 JMC
To: human-nets at MIT-AI
The best is the enemy of the good.
While the best plastic bag for escaping fires was 20 years
in development, perhaps many died who would have been saved had the
possibilities of ordinary plastic bags been publicized.
∂18-Dec-80 1037 JMC faculty meeting agenda
To: csd.ullman at SU-SCORE
I would like to raise the issue of the fluctuating content of the
comprehensive. Every time I am on the committee, there are questions
on proving program correctness, and then the subject gets zapped.
∂18-Dec-80 1143 JMC
To: energy at MIT-MC
Unfortunately, there is every reason to believe that this Kopka is
a wild-eyed visionary. No doubt he is entirely capable of drawing
artist's conceptions of "levacars" and inventing such names and
collecting visionary ideas from a variety of magazines. However,
there is little evidence that Ford has serious projects in the
areas mentioned. The article is a mixture of entirely practical
ideas and things like the solar-collector paint job, which is
probably merely a wish if it doesn't actually lose for lack of
collector area. 50 years is just far enough off, so that there
need be no actual research projects. The statement "If all this
sounds wild, consider today's futuristic technology and compare
1981 autos with those built 40 years ago". Alas, cars today are
better than those of 40 years ago only in the refinement of
technology of cars that were in production at that time.
∂18-Dec-80 1210 JMC
To: pourne at MIT-MC
The committee members referred to by STEF are certified experts.
Possibly they would suffer from pomposity and the common tendency
to advocate fancy terminals rather than making sure that everyone
expected to communicate had a terminal on his desk and another at
home. Perhaps there would also be a tendency to impose restrictions
based on a demand for security greater than that achieved by
non-computer systems. I have just begun to think about the possibilities
for Trojan horses. As I mentioned last night on the phone, one
can mail people programs that if they run them will do all kinds
of strange things. There is also the question of whether the
officially delivered hardware and software might have had Trojan
horses built in at some point. What about substituting a terminal
that would remember some of what was typed and later telephone the
Soviet Embassy and transmit it? No doubt many more possibilities
have been thought of and the experts have scared themselves into
paralysis. However, I have been thinking about non-electronic
Trojan horses. For example, how does one know that the safe that
was delivered to the White House by truck was in fact the one
shipped by the Mosler safe company and hasn't been replaced by
a model with an additional combination and also a remotely
interrogable voice recorder.
Did you read about the Politbureau members talking to one
another on their car radiotelephones as they drove around Moscow
and the transmissions being picked up in our embassy. They only
stopped when Jack Anderson's column exercised the American public's
right to be informed about such matters. You probably wouldn't be
able to get a story published with a silly plot about a secret
institution developing all sorts of sophisticated counter-espionage
devices and methodologies, while their bosses blabbed all the
secrets in private dining rooms in restaurants, where the waiter
recorded them on a twenty dollar casette recorder and sold them
to the highest bidder. (This last is imaginary).
∂19-Dec-80 0157 JMC
To: energy at MIT-AI
Answer to Schauble query about TMI 1.
It has not gone into operation. The nominal reason was that some of its
facilities are being used on a standby basis just in case Unit 2 needed
it. General Public Utilities has been trying to get permission to operate
it for a long time. The real reason may be that NRC doesn't want to let
GPU off the hook yet. Two of the commissioners are basically
anti-nuclear, and a third (the chairman) likes to take a "balanced"
position. No such "major" action seems likely to me before Reagan
appoints a chairman, he is confirmed, and takes office. Presumably he
will be unabashedly pro-nuclear, and then we'll see how the votes line up.
∂19-Dec-80 0217 JMC
To: FFL
The address for the MACLISP, etc. inquiry is PUBLICATIONS%AI.
∂19-Dec-80 1001 JMC
To: csd.strauss at SU-SCORE
I can give you more time.
∂19-Dec-80 1521 JMC
To: FFL
hamilt.1
∂19-Dec-80 1542 JMC
To: energy at MIT-MC
Other observers feel that TMI should be given the opportunity to
prove, if it can, that a substantial part of the costs are a
consequence of the actions and inactions of the NRC. With authority
to give orders should come some part of the responsibility for
carrying them out. The above "some observers" refers to me. I
wonder if David's "some observers" aren't spokesmen for anti-nuclear
organizations who prefer being called "some observers" in order to
avoid having bias attributed to them by people remembering their
official positions and their past statements.
∂19-Dec-80 1616 JMC
To: nilsson at SRI-KL
I believe Fran sent them recently, but she'll answer Monday.
∂19-Dec-80 1634 JMC
To: FFL
I sold a Yumyum and the dollar is in your drawer.
∂20-Dec-80 0110 JMC
To: MRC at SU-AI
Is there some way to use EMACS on SCORE telnetting from SAIL?
∂20-Dec-80 1144 JMC
To: FFL
∂20-Dec-80 0141 RMS My mail
Can you make sure it all gets forwarded to MIT?
∂20-Dec-80 1158 JMC
To: human-nets at MIT-AI
According to the Encyclopedia Britannica, inhalation of a one percent
concentration of CO can be fatal in 10 to 20 minutes. Therefore, there is
a substantial margin between CO danger and flash fire danger from an oxygen
apparatus. If you had the apparatus, your best bet would be to use it.
∂20-Dec-80 1634 JMC basic
To: BYY
In order to find whether there was a basic, I did
di basic.*[1,3]
This directory command asked whether there were any files on area 1,3 where
public programs are kept having BASIC as the first name. The replay was
BASIC.DMP and BASIC.TEX. A file whose second name is DMP can be run by
the RU command giving its full name or by the R command if it is on 1,3,
the afore-mentioned public area. R BASIC elicited the question of whether
I wanted the new version or the old. Since I don't know how to use BASIC,
I stopped there. I thought that BASIC.TEX might be the writeup for the
above, but it turns out to be a collection of macros for the TEX document
compiler. BASIC.DMP hasn't been changed since 1973, suggesting that it
has no active users.
I would like to persuade you to use LISP, especially MACLISP, for any
computing you do. It can do symbolic computations better than any other
language, and it can do numerical computations as well, sometimes better,
since it handles arbitrary precision integers. Also there is substantial
logical interest in proving that LISP programs meet their specifications.
Documentation is the MACLISP manual and a book by myself and Carolyn Talcott,
copies of the latter being available.
The command DI BASIC.*[UP,DOC], as in "What's up, Doc", did not turn up
a file whose first name is BASIC, which is the logical place to look for
a documentation file.
LISP also has plenty of local experts.
∂20-Dec-80 1652 JMC
To: FB at SU-AI
Now that the VAXen can be reached from SAIL, I'd like an account for education.
∂20-Dec-80 1709 JMC
To: BYY
Yes, the book assumes no previous programming, and I'll be glad to
help you get started, and I am interested in applications to logic.
I can get you a copy Monday or sooner if you want and lend you a Maclisp.
There are many LISPs, but MACLISP is the one used on this machine, and
one of the two most used, the other being INTERLISP.
∂20-Dec-80 1837 JMC dtn bug
To: MRC at SU-AI, TVR at SU-AI
The bug is more complicated than I thought. Even setting
the number of lines at 20, I can still excite it by M-20 C-n,
if this will bring it beyond the number of lines that exist.
The symptom of the bug is a loop in dtn at SAIL.
∂21-Dec-80 0106 JMC
To: TVR at SU-AI
CC: MRC at SU-AI
The bug is as follows:
DTN SCORE
TER DATAMEDIA
login
EMACS <some short file but more than one screenful>
When I do M-20 C-N under certain circumstances, the screen blacks out
and my WHO line shows that SUPDUP is in a loop at SAIL.
It is somewhat sensitive to the declared line length, but I haven't
found any line length that prevents it. It is reliable enough to demonstrate.
It occurred when I was using an Imlac from home, but I don't suppose
that SUPDUP knows about that.
∂21-Dec-80 1139 JMC
To: human-nets at MIT-AI
I don't think Hofstadter proved any point by telling stories about it,
and I don't think anyone has proved that there can't be penetration
proof operating systems. I rather think there can provided physical
security can also be maintained. They may not be necessary though.
∂21-Dec-80 1626 JMC
To: csd.bradford at SU-SCORE
FINGER speaks truth. 857-0672.
∂22-Dec-80 1725 JMC
To: FFL
Any mail for Stallman should be forwarded to him at M.I.T. AI Lab.
∂22-Dec-80 1842 JMC book and maclisp
To: BYY
I have left them in my box in MJH 2nd floor. The Maclisp is a loan.
∂23-Dec-80 0120 JMC
To: pourne at MIT-MC
Yes. I would certainly come to the space meeting if invited.
As to what you do about having said things about LISP, that the
M.I.T. people disagree with, and I would probably disagree with,
just keep saying what you think. I suppose you haven't written
the kind of program that LISP is good for. Perhaps if I look at
the "flame", I'll have some rebuttals or offer some education.
I find controversies about LISP and also about AI rarely worth
getting involved in except as a vehicle for making points that
I want to make anyway. The same effort in put into actual work
changes more minds than any amount of argument.
This depends on the fact that most of the relevant Government
funding decisions are not based on public opinion.
∂23-Dec-80 0127 JMC
To: boyer at SRI-F2
No, I didn't receive a message referring to a complete proof. Also I was
too far behind at the end of the quarter to ask for a lecture from you,
but I hope you'll still be willing next year if still here.
∂23-Dec-80 0132 JMC
To: pourne at MIT-MC
Well, Jerry, you can expect uninhibited replies from the M.I.T.
people to uninhibited remarks. LISP is well taught at many other
places than Stanford and M.I.T. - CMU, Texas, Yale, and Indiana
come immediately to mind. Winston's Addison-Wesley text is ok,
and I hope mine with Carolyn Talcott will be better, although it
will emphasize proving programs correct, which requires some
mathematical logic.
∂23-Dec-80 0149 JMC
To: boyer at SRI-F2
Got it this time. September.
∂23-Dec-80 0228 JMC chat and diablo
To: FB at SU-AI
I called Diablo with chat, logged in and tried to use learn. However,
whenever the required response was <cr> without anything before it,
chat hung, and the only thing to do was to <call> out and try again.
LEARN worked ok in the Diablo room, however. It has a certain "peek
through a knothole" approach, but I learned how to rename all files
whose 3rd letter is "w".