perm filename E80.OUT[LET,JMC]1 blob
sn#539474 filedate 1980-10-02 generic text, type T, neo UTF8
∂02-Oct-80 0001 JMC article
To: pourne at MIT-MC
I have had a lot of difficulty with it, but I think I've got it outlined
well enough so I can finish it in the next couple days. The title is
THE FREEDOM TO INNOVATE and it discusses the importance of this freedom
and the recent attacks on it with particularly glaring examples.
∂01-Oct-80 2326 JMC
To: rwg at MIT-MC
I'm afraid I'll need some help in decrypting this.
∂01-Oct-80 1923 JMC
To: FWH
I will be interested in giving a talk on new developments in Elephant
late in the quarter.
∂01-Oct-80 1235 JMC
To: boyer at SRI-KL
Yes, I would be glad to recommend both of you.
∂01-Oct-80 0354 JMC
To: human-nets at MIT-AI
Telex and TWX
If I remember correctly TWX formerly belonged to AT&T and Telex to
Western Union. We used TWX, which used ASCII instead of Baudot for remote
time-sharing terminals before dataphones and voice couplers were
available. At some time in the 60's, Western Union was about to go out of
business (partly because their technology was obsolete compared to AT&T),
which would have made AT&T even more subject to antitrust, so the FCC
induced AT&T to give TWX to Western Union which now operates both networks
but using lines rented from AT&T. Thus monopoly is created in the name of
competition.
A few years ago I looked into the advertised Mailgrams from
computers. As might be expected, the local Western Union sales office had
never heard of what their company was nationally advertising. When we
finally made contact, they explained that they would come and program our
computer to send Mailgrams into their network. The idea that we might
program our own computer if they would tell us the protocols was too
exotic for them. Anyway, they explained that the service was intended to
send identical Mailgrams to large numbers of people and the minimum charge
was $25. Since our idea was to allow net messages by Mailgram to people
not on the net, there was no meeting of minds. Maybe things are better
now.
∂01-Oct-80 0229 JMC
To: LLW at SU-AI, minsky at MIT-AI
Here's another possible dynamic compression member.
.cb A Dynamic Casing for a Hole through the Center of the Earth
A hole through the center of the earth would have lots of
uses. It would provide a fine track for a mass driver or one could
even imagine dropping a rocket through the hole and having it fire
at the center of the earth giving it a much lower escape velocity.
However, we would have to get the exhaust gases out of
the hole before we could fire another rocket.
The main problems in maintaining the hole are pressure and
temperature. The temperature problem may be the easier of the two,
because the heat conductivity of rock is low enough so that we might
be able to pump the eat out of the hole as fast enough to keep the
temperature down.
As for the pressure, we propose exploring the idea of a
"dynamic casing". This is a casing that spins fast enough to
counteract the pressure with centrifugal force. Dimensional
computations show that the velocity of spin is independent of
the diameter of the hole. It depends on the density of the
casing, the fraction of the radius of the casing taken up
by the spinning wall and the depth in the hole.
It also depends on the radial distribution of the earth's mass.
Casual computation suggests that the casing would have to
spin at about 4000 meters per second at the center of the earth.
The immediate question is whether there is a way of making
a frictionless bearing between a spinning inner casing and a
stationary outer casing. If the heat removal took place at
ambient pressure in the outer casing, the space between the
inner and outer casings could be cold if this would help.
∂01-Oct-80 0008 JMC
To: VRP
CC: LES
Will you do mail or file transfer?
∂29-Sep-80 1435 JMC III
To: AAM
I called Fenaughty, the president, but Griffith returned the call.
You are welcome to call him. They have mild proprietary worries.
I told him that you would use any techiques you knew about but
wouldn't describe in your thesis techniques of theirs that you
don't use, and that seemed ok, and that kind of thing is acceptable
to Stanford. Griffith will discuss with Fenaughty the possibility
that they might support you during the acadmeic year and/or employ
you during the summer. Since you won't start major work till
January, there is no special rush.
∂29-Sep-80 1016 JMC
To: csd.ullman at SU-SCORE
CC: FB at SU-AI
I certainly can have no objections. It seems to me you should check
with ARPA.
∂29-Sep-80 0019 JMC
To: RMS at SU-AI
1. I'll be in late in the morning, since I have to take my car in.
2. I saw Reiter last week, and he didn't mention having any new paper.
3. You are missing the fact that both Reiter and McDermott-Doyle
treat defaults as axioms, i.e. originating form the programmer,
whereas circumscription is a rule the program can decide to apply
at any time, i.e. it can generate defaults.
4. Moreover, circumscription is semi-decideable. One can generate
the set of possible conclusions that can be reached by circumscription,
while non-monotonic logic is not r.e., since the definition depends
on fixed points among models.
5. Your conclusion about logic being suitable mainly for meta-reasoning
about what programs can do would follow if something else were guaranteed
to be uniformly better for expressing the data in the program.
6. I will be interested to hear about paradoxical aspects of circumscription,
because not all is clear to me.
∂28-Sep-80 1834 JMC
To: TW
Congratulations, but are you sure such late toilet training is advisable?
∂28-Sep-80 1443 JMC
To: SEK
What is your phone number for LOTS form?
∂28-Sep-80 1200 JMC pub and print
To: FFL
Please pub and print CBCL[F75,JMC] to be sent with a dictated
memo to Prof. William Massy, v-p for Business Affairs, Stanford.
∂28-Sep-80 0201 JMC TDD
To: human-nets at MIT-AI
Whatever is done for the deaf will have a short life, because we cannot
now predict what terminals will win the mass market among the non-deaf
home users. As soon as the number of non-deaf non-hobbyist home terminal
users passes a million or two, the deaf will want to join the club. At
that time, they will not want a special design. Poor deaf people might be
assisted to buy general purpose terminals to the extent that public
support of the poor wins politically, but by the same mechanisms that
other "needs" of the poor are subsidized. According to the social work
paternalism index of the time, they will either receive terminals or will
receive money to spend on whatever they please, be it terminals or
something else.
Therefore, assuming that anything is to be done for the deaf before
terminals are in general use, promptness is more important than getting
the best solution. As for taking it out of the phone bill rather than
general funds, among Governor Brown's thinkers, redirecting other people's
money rather than directly taxing them is now the fashion in order to get
around the constitutional limitations and political dangers of new taxes.
For example, Brown beat the drum at the Democratic convention for
redirecting pension funds into socially useful channels, and I have some
fear that they will try to redirect my TIAA retirement money out of Exxon
(of which TIAA is the largest stockholder) into the design and manufacture
of solar powered outhouses.
∂27-Sep-80 2316 JMC seminar
To: FFL
SEMINA[E80,JMC] contains an announcement of a seminar and a list of
recipients for the notice. The time and place of the first meeting
need to be determined. I suppose there should be room for about 15
to 20 people, and I would like to start in the week of October 6,
and the meeting should be at 4:15 and should not conflict with
the department colloquium, which I suppose is still on Tuesdays.
∂26-Sep-80 1729 JMC
To: FFL
I'll show you what you can do about such complaints from PUB.
∂26-Sep-80 1115 JMC
To: FFL
∂26-Sep-80 1030 JRA
price 22.50 us, can, and mexico; o.w. $25
This is for the LISP conference proceedings.
∂25-Sep-80 2103 JMC
To: TOB
It looks good.
∂25-Sep-80 1444 JMC
To: JK
Hmm. I must have forgotten that.
∂25-Sep-80 1440 JMC dover
To: CT
Until program tells you that it has done the listing, the writeup
should tell you that the program doesn't.
∂25-Sep-80 1417 JMC
To: BS at SU-AI
Who called from NSF? I didn't realize that it was in any way alive.
I have to submit a new proposal for my own support starting next June,
and I'm working on it. I want to discuss whether to include Doyle
activities in it. I intend to phone them tomorrow.
∂25-Sep-80 1123 JMC
To: gosper at MIT-MC
It is Bernard Cohen, and I'll try to bring one. I have several things I
have to do before returning to the Volterra problem. I'm puzzled by the
EIR. I assume it's Federal, but what Federal action is involved?
∂25-Sep-80 0031 JMC
To: JK
I have put a copy of Nozaki's paper on your door clip.
∂24-Sep-80 1651 JMC
To: FFL
That abstract is SIGMA[E80,JMC].
∂24-Sep-80 0059 JMC
To: FFL
I will do the abstract today, perhaps before 2pm, but please call and
find out how long an abstract they would like.
∂23-Sep-80 1913 JMC adapter
To: ROB
I need an adapter that will allow connecting a Datamedia with a four prong
outlet to a telephone jack and telephone with modular jack connection.
∂23-Sep-80 1700 JMC
To: PAT
CC: FFL, ROB
Thanks, Patte. We'll restore from that dump.
∂23-Sep-80 1657 JMC
To: CET at SU-AI
I have some good will left for Evans and Sutherland, but I really don't
know anything now that is likely to be of use to them. Therefore, someone
else would be better to tell them something of use and recruit them
to the forum.
∂23-Sep-80 1403 JMC
To: RPG
Give receipts, etc. to Frances Larson.
∂22-Sep-80 2146 JMC
To: kanerva at SUMEX-AIM
Thanks, Pentti.
∂22-Sep-80 1501 JMC
To: kanerva at SUMEX-AIM
What is proposed equipment configuration?
∂22-Sep-80 1106 JMC
To: FFL
I don't have anything right now. Maybe something short by Friday.
∂22-Sep-80 1105 JMC
To: csd.tajnai at SU-SCORE
I would be willing if the time works out.
∂22-Sep-80 1055 JMC
To: LWE
Unfortunately, transcripts are carried only on the high speed wires
(Datastream) to which we can't afford to subscribe. They usually
mention it on the regular wire when they do it, so I suppose they
didn't in this case. I don't know about machine analysis, but the
comments on who won say quite a bit about the newsman, especially
concerning the intensity of their feelings. Would you agree that
Reagan is getting substantial advantage from the fact that many
people who disagree with him can't work up either anger or fear,
and may therefore express themselves by voting for Anderson or
just stay home. I'll look into the tape question, but we have no
regular operators, so it isn't obvious how to get it done. It may
cost you extra. Actually you should inquire directly by MAILing
a message to REG.
∂22-Sep-80 0113 JMC
To: FFL
Please remember to ask me for tickets.
∂21-Sep-80 0043 JMC
To: human-nets at MIT-AI
One would naturally suppose that "The Computerized Destruction of Western
Civilization" panel was a spoof, but considering that it's SIGCAS (Special
Interest Group on Computers and Society (of ACM)), it
probably isn't. SIGCAS impresses me as one of the most blindly technologically
reactionary conglomeration of know-nothings, I have ever encountered,
based on all the issues of their newletter available a few months ago
in the Stanford library. Their specialty is horrific "what if" scenarios
about data banks, electronic funds transfer and other computer applications.
Whether it is from lack of imagination or merely ideology, they don't see
any of the positive potential of computerization and the enormous costs in
delay implementing it. For example, if we can get rid of cash, we can get
rid of crime for cash. A robber can't hold you up and demand that you
transfer money to his account - not unless he can prevent you from talking
and conceal the fact of your demise. By helping delay EFT, SIGCAS
assumes responsibility for the deaths of perhaps several thousand people
per year. Another example is that the Freedom of Information Act could
be made effective by requiring that all such information be on computer
file. Indeed the discussion of telephone company practices in HUMAN-NETS
would be much better informed if the Public Utilities Commissions were
required to keep their decisions and the tariffs they allow on file. (I
apologize having the same urge to impose a duty on government agencies
that others have to impose duties on the telephone companies).
Is anyone out there interested in forming an organization that would
devise and press for useful applications of computer technology in
society.
∂20-Sep-80 1146 JMC
To: rms at MIT-AI
He was told by someone at M.I.T. that he could reach you at my number
and that you had something to do with it. Why not call him collect?
∂18-Sep-80 1717 JMC 3n+1 function
To: DEK
Do you know a reference more recent or better than the 1974
Nievergel, Farrar and Reingold book that discusses the
termination of f(n) ← if n=1 then 1 else if 2|n then f(n/2) else f((3n+1)/2)?
∂18-Sep-80 1156 JMC
To: *
ARK is returned. Let us give thanks. Maybe now the losing graduate
students will restock the Pony.
∂17-Sep-80 1821 JMC
To: max at PARC-MAXC
Can you (Ma Xiwen) be reached by ARPAnet mail?
∂17-Sep-80 1742 JMC
To: JD
Sorry to miss your talk Friday. I'll be in Vancouver.
∂17-Sep-80 1612 JMC
To: ROB
Bob Floyd noticed some of the famous pie on a stair rail at floor 3.5.
∂17-Sep-80 1558 JMC buyer for emacs
To: rms at MIT-AI
Dick Cogan 216 371 1306 was told to call you about buying the right
to use EMACS on a Dec-20 at I forget what university.
∂17-Sep-80 0026 JMC
To: kirkley at SRI-UNIX
I'm not sure what you mean by "present system and immediate goals".
Do you mean our computer system or my research goals or the AI Lab
or AI generally or the computer science department?
∂16-Sep-80 2303 JMC
To: csl.fb at SU-SCORE
As you probably know Peter Hart left being head of the AI Center at
SRI to head a new AI group at Fairchild but with the sponsorship
of Schlumberger, which I think bought Fairchild or something like
that. His interest is in getting good display facilities for his
group not in getting a product, so I think he will be reasonable.
My idea was merely that they would share the effort by donating
the services of an engineer. I think it isn't very highly probable,
because maybe they don't have the person, but it may be worth pursuing,
since things (from my point of view anyway) seem to be going slowly.
Probably you understand all that, but since I was in a hurry this
afternoon and therefore concise, I want to be explicit now.
∂16-Sep-80 2254 JMC
To: ROB, ME
No A.P. news today, and I want to know if Dixy won.
∂16-Sep-80 1807 JMC hurley.le1
To: FFL
The papers are "Ascribing Mental ...", "Home Computer Terminals" and
"Three reviews of Weizenbaum".
∂16-Sep-80 1746 JMC
To: buchanan at SUMEX-AIM
Have a good trip. I think the outline is basically what is wanted.
Please give my regards to Jack Schwartz if he's there.
∂16-Sep-80 1730 JMC
To: FFL
michie.le2
∂16-Sep-80 1715 JMC
To: FFL
Benefi.le1 needs to be fixed up with full and accurate names for
the books with publishers, and a request for his ai conference
speech needs to be sent to Shortliffe.
∂16-Sep-80 1645 JMC
To: FFL
todoro.le5
∂16-Sep-80 1634 JMC
To: FFL
please pub and print abrams.le2
∂16-Sep-80 1610 JMC user
To: REG
Prof. Alphonse Juilland may call you about using the machine
for typing in and printing out about 4000 pages in French. He
knows reasonably well what he is doing, and I told him he might
find SAIL cheaper than Suppes. He doesn't have the money yet,
but I suppose he can get it. I have no intereste in the project
per se, but he's a good guy.
∂16-Sep-80 1532 JMC grade change
To: FFL
I need to change Carolyn Talcott's incomplete grade in CS226. Please
get a grade change card from Carolyn Tajnai and find out when
Carolyn Talcott took the course from me.
∂16-Sep-80 1515 JMC
To: FB at SU-AI
His number (apparently a home number) is 323-3616.
∂16-Sep-80 1514 JMC
To: FB at SU-AI
I was talking to Peter Hart today, and the question of their (Fairchild's)
need for high quality terminals arose. I suggested that he might
to well to lend an engineer if he had one to the SUN terminal
project, and he is willing to consider it, but he won't even be
in his office for two weeks. He suggested that you call him after
this time if you are interested in pursuing the possibility.
∂16-Sep-80 1508 JMC
To: csd.ullman at SU-SCORE
Yes, please circulate.
∂16-Sep-80 1430 JMC
To: csd.ullman at SU-SCORE
The issue recently arose of the Department paying the
travel expenses of David Waltz to attend Dick Gabriel's thesis
oral, Waltz being one of Gabriel's readers. The Department
declined to pay, and Tom Binford and I arranged for payment,
although neither of us is involved with Gabriel's work. I think
the Department should be willing for outside reader's to attend
thesis orals, when there is no research source involved. There is
altogether too little interaction between CS faculty in different
universities as is, and interaction at the thesis committee level
is very valuable in developing a common computer science public
opinion about what constitutes a good thesis and a good state
of knowledge for PhDs in computer science.
∂16-Sep-80 1047 JMC
To: ME
Can read macros be automatically loaded on login?
∂15-Sep-80 1944 JMC
To: kirkley at SRI-UNIX
Your message came on MOnday at 4pm,, so I can't attend gathering.
DNSRI is not recognized as a host by our ARPAnet program. However,
SRI-UNIX is. It seems to me that some non-Stanford person also
complained that DNSRI wasn't recognized, so maybe it's a general
phenomenon.
∂15-Sep-80 1731 JMC
To: FFL
I gave my blackboard eraser to Ketonen. Can you get me another?
∂15-Sep-80 1039 JMC
To: REG
I trust REM isn't working for nothing.
∂14-Sep-80 1626 JMC
To: buchanan at SUMEX-AIM
Further comments on recommendations: Since one of our major expenses
is software, re-equipping with software is as important as re-equipping
with hardware. For example, support for the maintenance of MACLISP and
INTERLISP would be very helpful.
I doubt that a recommendation that would apply to all sciences is
especially helpful unless some contrast can be pointed out. For example,
there is no more obvious need to underwrite journal subscriptions in
AI than in any other field. On the other hand, there are sources of
funds for postdocs in other fields and nothing in AI or even in computer
science generally.
∂14-Sep-80 1621 JMC
To: buchanan at SUMEX-AIM
I have one comment on your outline. Most of the other reports I have seen
and the instructions are rather heavy on history. They emphasize what
developments in the field have reached practice and were they done
in universities or industry. You are probably as well qualified
as anyone to do this for AI. If you want to talk now, I am at
7-4430.
∂14-Sep-80 1550 JMC
To: VRP
No. I'm afraid my level of knowledge about Dialnet is quite low.
I confess vagueness as to what a process ID is. However, I think
MRC and Tim Eldredge would know and wouldn't mind being phoned even
today if it's urgent.
∂14-Sep-80 1451 JMC todorovich letter
To: FFL
SE2 now has an office, and Todorovich no longer runs it out of his
apartment on Riverside Drive. I think you can get the address from
some recent mail from them or if worst comes to worst by phoning.
∂14-Sep-80 1239 JMC
To: REG
me has gone mad.
∂13-Sep-80 1754 JMC
To: JK
I'll look. Incidentally if you give references to page 19
of doc.ekl on [ekl,jk] in the form doc.ekl[ekl,jk]/19p, the reader
can switch to that file just by giving the αxpo command.
I still think I need a lesson to try things out.
∂13-Sep-80 0116 JMC
To: PLS
Are you an expert on POP-2? If not, is there one around? Manual?
∂11-Sep-80 1741 JMC
To: JRA
Please state a postpaid price, so people can buy without further
dickering.
∂11-Sep-80 0301 JMC Italian eval
To: RWW
Do you remember that in 1975 I was talking about a variant of
call-by-name eval proposed by some Italian and programmed it?
Was it Prini? Can you give a clue as to the reference?
I have just figured out my old programs but don't have a writeup.
∂11-Sep-80 0226 JMC
To: pournelle at MIT-MC
Can do.
∂10-Sep-80 1200 JMC
To: human-nets at MIT-AI
Subject: What will displaced workers do
"What do do with displaced workers" is not quite the right way of putting
it. Suppose a union thinks it can get another dollar an hour out of
an employer or industry, and the workers have to decide whether to take
it in direct pay or in a job security plan whereby the employer's right
to fire redundant workers is restricted or removed. How a worker will
vote in the union bargaining committee will depend on how self-confident
he is about the economy and his own personal ability to get another job.
Union officials have as an additional consideration the size of the
union which affects their personal positions. Historically, American
workers have mostly chosen higher pay, and this has paid off.
Perhaps I have digressed, but it has been more important what workers
have done for themselves in getting new jobs than what anyone, employers
or government, have done for them.
There is an apparent contradiction in the notion that displaced workers might
take part in colonizing space if it weren't too expensive. The expense
consists precisely in that it requires a lot of work.
Let me again claim that present new technology is making little progress
in displacing most jobs. If only construction technology were getting
cheaper rather than more expensive, we could rebuild our houses
and cities. If only bureaucratic technology were getting cheaper the
fraction of their income universities spend on administration could fall
rather than rise. If only automotive technology were getting cheaper
more rather than fewer people could afford new cars. Micro-processors,
given present well developed ideas for using them, seem to be a peripheral
technology, less capable than the automobile and trucks of displacing
labor. Of course, when robots of servant level capability can be built,
everything will change, but that will require new AI ideas.
∂10-Sep-80 0230 JMC
To: JRA
Have you established a price I can tell people for the conference record?
∂09-Sep-80 2307 JMC
To: kirkley at SRI-UNIX
I wasn't looking at my terminal when you sent me the message.
If you're still there you can send me messages, and I'll reply.
∂09-Sep-80 2306 JMC
I wasn't looking at my terminal when you sent me the message.
If you're still there you can send me messages, and I'll reply.
∂09-Sep-80 2138 JMC
To: ROB
J.P. Morgan said that to someone who asked the cost of his steam yacht.
∂09-Sep-80 1841 JMC 1024 vs. 512
To: REG at SU-AI
CC: AVB at SU-AI, DPB at SU-AI
By the faculty meeting on Sept. 30, it would be good to have an estimate
of the costs of large numbers of 512x512 and 1024x1024 display systems.
Denny's current count is 185 terminal slots in 89 offices in MJH. If
there are any other considerations such as development delays that are
relevant to a decision, it would be nice to identify them. Denny, it
would be nice to know about terminal requirements for CSD outside of MJH
if any.
∂09-Sep-80 1538 JMC
To: FFL
Please send a copy of the AI Journal issue to Mints in Leningrad.
∂09-Sep-80 1451 JMC
To: FFL
I need Aug 22 Science back.
∂09-Sep-80 1344 JMC
To: FFL
Faculty club for two Thursday please noon.
∂09-Sep-80 0002 JMC
To: ROB
Both are bad, but NYT is worse.
∂08-Sep-80 2350 JMC
To: ROB, ME
NS is losing lots of characters.
∂08-Sep-80 1440 JMC office
To: JK at SU-AI, JD at SU-AI
CC: FFL at SU-AI, BS at SU-AI
The office formerly occupied by Les Earnest is now yours. It isn't
fully furnished yet, but you can get keys.
∂08-Sep-80 1436 JMC
To: RWG
Could you MAIL last night's 1st order displacement of equilibrium formulas.
∂07-Sep-80 0357 JMC
To: FFL
Don't bother with scient.xgp.
∂07-Sep-80 1625 JMC
To: RPG
I plan to be at CCRMA Tuesday at 4pm
∂07-Sep-80 1114 JMC Dialnet
To: VRP
Mark Crispin (MRC) wrote the protocols and the implementation
for SAIL and TOPS-20. He is still the keeper of the protocols.
Tim Eldredge at Hewlett-Packard, who comes around a lot, is currently
trying to make his hardware work on a TOPS-20. Ralph Gorin is
nominally in charge now and will help if need be.
∂07-Sep-80 0946 JMC
To: RWG
Tonight is good for me.
∂07-Sep-80 0944 JMC
To: ME
OK, thanks.
∂07-Sep-80 0052 JMC
To: ME
It did say it but I ran through all it showed me, and then exited again,
again said NS #, and it had forgotten all of them. I can demonstrate
the effect if you wish to peek.
∂07-Sep-80 0025 JMC ns forgets
To: ME
I just finished reading all the news stories, and left NS with the e command.
Immediately, saying NS again, I try #, and it has forgotten half the stories.
20 were remembered and 21 forgotten.
∂06-Sep-80 2338 JMC
To: FFL
Please print SCIENT.XGP[LET,JMC].
∂06-Sep-80 1232 JMC
To: FFL
We need a white marking pen for putting titles on bindings of reports.
∂06-Sep-80 1153 JMC
To: RWG
I have some differential equations to juggle, and I wonder if you
could give me the Lisp machine - Macsyma lesson you offered in
connection with it. Almost any evening would be suitable.
∂05-Sep-80 1820 JMC
To: FFL
The paper you're typing should become COMMON[E80,JMC].
∂05-Sep-80 1608 JMC philosophical papers
To: nilsson at SRI-KL
The published papers I would like to include are listed in the
file PHILPA[E80,JMC]. I would also provide an introduction but
wouldn't mind is someone else wanted to write a preface. If
there is room, I might include a few short unpublished memos.
All the material is in the computer files listed after the
bibliogrphic reference except for the first paper which my
secretary, Fran Larsen FFL, will type in shortly.
∂05-Sep-80 1311 JMC
To: TOB
Have a good trip.
∂05-Sep-80 0739 JMC
To: REM
I'm not really satisfied with what I said in my last HN. Maybe
the point is more like this. There is a tendency to regard people in the
future the way we regard our children - requiring our guidance. They
don't need our guidance, and it's just as irrelevant as the ideas of the
people 100 years ago on what our problems will be. Moreover, if we act on
a problem about which action can be postponed, we use up our resources -
mainly our extremely limited political ability to act sensibly in several
areas at once. Since, our social knowledge is non-quantitative, and
social movements to encourage action have positive feedback from success,
society is likely to overreact and go into oscillation.
Suppose, for example, that a movement against assembly robots in
the automobile industry acquired a charismatic leader, successfully forced
GM to abandon a major project and defeated a major politician. This would
cause anti-robotism to become a successful cause, and it would attract
adherents in the U.S. and other countries. The success of the cause over
a ten or twenty year period might be entirely independent of whether it
saved any jobs. Eventually, there would be a reaction, and its success
might be independent of whether it produced economic growth.
The main cure would be to develop a successful scientific
social science. Pending that, it is best to put off almost anything
that can be put off. Saving jobs in the next ten or twenty years is
a valid goal - if not an all consuming goal. Preventing a nuclear
war in the next ten to twenty years (though maybe we can't really see
ahead more than ten) is worth almost any action that can be shown
to do any good. However, a claim that some action of no use today
will prevent nuclear war even thirty years from now needs to be scrutinized
extremely carefully.
Supposing someone had claimed six months ago that worker
discontent in Poland over the next ten years would cause a big strikes in
August 1980 and that this would cause a revolt in East Germany in 1981,
which would cause the older generation of Soviet hardliners to start a
nuclear war. This analyst might have further claimed that the best action
was to have the CIA assassinate the Lech Walesa, an obscure Polish worker,
which would postpone the Polish strikes by three years, at which time the
Politburo situation would be different and permit real liberalization
without war. The hypothesis would be conceivable, but no-one could
possibly have good enough reasons for believing it to justify an
assassination.
This is the case for concentrating on seeing clearly what is in
front of our nose. How does it justify developing science and potentially
useful but long range technology? That requires betting that our
successors are likely to have at least as clear a picture of their
interests as we have of ours and have a good chance of benefiting from the
greater variety of options provided by improved knowledge and technology.
∂05-Sep-80 0702 JMC
To: csd.friedl at SU-SCORE
Thanks Paul.
∂04-Sep-80 1647 JMC
To: VRP
C would be fine, but alas ...
∂04-Sep-80 1419 JMC
To: human-nets at MIT-AI
"While the situation
probably will remain one of increasing employment for awhile, it
must inevitably end up with fewer jobs, once the factories that
build automatic factories are built. I don't have the answer to
this problem, but it definitely calls for more than just saying,
'Well, it won't happen for awhile, and it might turn out okay.'"
There is good reason to take precisely the attitude criticized
above. The attitude of doing something now about eventual job
loss has led in England to bitter struggles against new
technology that have put England far behind France and Germany
in standard of living and with no less unemployment. The trouble
is that over-sensitivity to distant threats, whether military
or environmental or social or unemployment, leads to bureaucracies
that don't have to justify their activities by current
achievement. They can simply claim success by showing that
the disaster they purport to be averting hasn't happened. In
the case of technological unemployment, the burden of proof is
on the advocates of action to show that the situation is different
from that of ten and twenty and forty years ago when immediate
action against the same danger was also being demanded.
∂04-Sep-80 0121 JMC
To: ME, REG
That's correct assuming that's the total and not the cost per modem.
If it's the cost per modem, then it doesn't seem in accordance with
what Les negotiated with Product Associates in connection with the
inadequacy of the Product Associates modems.
∂03-Sep-80 2341 JMC
To: REG
The following should be self-explanatory.
∂03-Sep-80 2249 TOB request for display purchase
Andy Bechtolsheim has offered to build a 1024x1024
display with 8 bits per pixel combined with a
high performance display (1000 lines without interlace)
for $20k. I would like to have it for the Horton box
to make some beautiful graphics. Can we buy it from
ARPA? Do you favor it?
Tom
Is this a private venture for Bechtolsheim or is it an internal Stanford
cost? If Bechtolsheim is available to do it, I would like to persuade him
(within Stanford or otherwise) to build the large department wide display
system.
Also I thought you were broke.
John
∂03-Sep-80 2330 TOB
This is a private venture for him. As far as the department system,
that might be too large for him (a real manufacturing effort). I personally
hope that he doesn't become overcommitted before I get a display.
The NSF is broke. ARPA is a question mark. I did not make clear that
I was asking if you were feeling both rich enough and especially generous
to get it for us out of other than ARPA IU funds. Otherwise I will find
a way in time.
Also I thought you were broke.
I am feeling neither rich nor generous. However, the design effort on
the departmental system may not be so great, and Ralph has more money
than brains for the task. I think I'll tell Ralph about it, and he
can explore the matter with Andy if he has no better scheme. The
Data Disc may die soon, and Schnurle is tied up on Ethernet until
at least January according to Ralph.
John
∂03-Sep-80 2336 JMC
To: TOB
I am feeling neither rich nor generous. However, the design effort on
the departmental system may not be so great, and Ralph has more money
than brains for the task. I think I'll tell Ralph about it, and he
can explore the matter with Andy if he has no better scheme. The
Data Disc may die soon, and Schnurle is tied up on Ethernet until
at least January according to Ralph.
∂03-Sep-80 2322 JMC
To: TOB
Is this a private venture for Bechtolsheim or is it an internal Stanford
cost? If Bechtolsheim is available to do it, I would like to persuade him
(within Stanford or otherwise) to build the large department wide display
system.
Also I thought you were broke.
∂03-Sep-80 1857 JMC
To: FFL
Please be sure Carolyn's swivel chair gets to Hersche's office.
∂03-Sep-80 1728 JMC
To: JEF at SU-AI
Don't try to be too fair; I don't want to use you up as department
chairman. I don't see Gabriel as a special case. I think Princeton
should have paid, and I will argue that we should encourage outside
readers and other forms of external participation in our educational
process. Sept. 30 will be fine.
∂03-Sep-80 1411 JMC
To: JEF at SU-AI, DPB at SU-AI, RPG at SU-AI
It seems to me that the decision not to finance Waltz's trip if made
on the grounds that
"It was decided that the central department funds
should not be used for what is essentially research activity." is a
mistake. Obtaining readers for a thesis is essentially an
educational activity. I am unfamiliar with the circumstances
under which Waltz became a reader of Gabriel's thesis, and I
don't even know who is Gabriel's adviser - perhaps Cordell Green.
However, being able to get external readers for thesis projects
for which an outsider is more appropriate than additional
internal readers is a big advantage, and it is a favor to us
when someone from another university undertakes this task.
It also gives our students more freedom in choice of thesis
topic and protects the student against changes in faculty
interests or personnel during the preparation of his thesis.
Moreover, having external readers promotes inter-university
comparison of dissertation and oral committee standards. In
many British universities external readers are required and
their transportation is an accepted element of the cost of
having PhD students.
I think the next Department meeting should discuss the question
of external readers.
∂03-Sep-80 1359 JMC
To: FFL
Please pub newbor.le1 and fredki.le1. The first was sent to Newborn
A copy of the former (which I sent last year)
is to be an enclosure in the latter.
∂03-Sep-80 0141 JMC hold mail
To: FFL
If they haven't gone, please hold interdepartmental envelopes addressed
to Lieberman, Hastorf and Kennedy.
∂02-Sep-80 2051 JMC
To: JK
To tell the truth, I was a bit discouraged by difficulties in
following the writeup, and I think I'll need a demonstration
lecture. The idea of construing proofs as operators on proofs
seems promising.
∂02-Sep-80 1718 JMC
To: masinter at PARC-MAXC
I see it is mainly a function of the kinds of program people write. In
writing an algebraic simplifier, complex patterns like a sum, some
of whose terms are sums, arise all the time. Thanks for the message.
∂02-Sep-80 1623 JMC
To: FFL
There isn't an AI program separate from the general program of the
Computer Science Department. The department has both
masters and PhD programs, and inquiries should be made to the Department
Admissions Secretary. Dot Dale is the secretary still, I think, and I
think Gene Golub is still admissions chairman.
∂02-Sep-80 1400 JMC
To: teitelman at PARC-MAXC
My interest isn't in FLIP per se, but in pattern directed computing
in general. Is some other system for pattern dircted computing
in general use among INTERLISP users? I am writing the chapter
on pattern directed computing for McCarthy and Talcott "LISP: Programming
and Proving". Larry Masinter just expressed the opinion that
pattern matching compilers (and perhaps he meant pattern matching
systems in general) aren't useful. Is that your current opinion also,
and if so why?
∂02-Sep-80 1358 JMC
To: masinter at PARC-MAXC
Thanks, Larry. I would be interested in your opinions as to why pattern
matching compilers (or is it pattern matching systems in general) are
unlikely to be useful.
∂02-Sep-80 1110 JMC
To: FFL
∂02-Sep-80 1102 FFL
To: JMC, FFL
I am getting require let pub, etc. on top of everything that I start with ET.
Is that good?
I'm puzzled, and I'll be in shortly.
∂02-Sep-80 1108 JMC
To: FFL
All the papers are in printed form in the file cabinet on right as
you come into my office.
∂01-Sep-80 1820 JMC
To: RPG
Thanks for the inverse. It is something I wanted.
∂01-Sep-80 1625 JMC pattern matcher
To: RPG
I think I understand your match.>[aid,rpg] well enough to use it, but I
don't understand how it is coded, and I'll catch you some time soon
for an explanation. Thanks for the information.
∂31-Aug-80 2223 JMC
To: human-nets at MIT-AI
Peter Deutsch sees in NSA's attempt to supervise cryptography
research "as merely the latest step in a long process of subordinating
traditional democratic concerns to 'national security'". In fact, the
government had an effective monopoly on cryptographic research from when
it seriously began around World War I to the time a few years ago when
serious academic cryptography research started.
World War II experience shows that ability to crack enemy ciphers
and keep one's own effective is an important advantage. It may have saved
hundreds of thousands of lives at the time of the Normandy invasion. My
opinion is that the country is in some danger and needs all the technology
advantages we can get to offset the Russian draft, their regimented
society, and their internal control of information which keeps their
population believing that they are always right in international
conflicts. Peter Deutsch apparently doesn't consider it a relevant
consideration in the cryptography controversy, and I have no idea of what
would constitute a decisive argument.
There is also an issue of whether there are any cryptographic
secrets left worth keeping, and this might be objectively decided. A few
people trusted by the academic cryptography researchers and NSA and
clearable by NSA could be given access to information justifying NSA's
claim that restrictions on open cryptography publication would strengthen
the country's defense. They could then give the academic community an
informed opinion on whether there was such an advantage. Some people
would be influenced by this - others perhaps not.
∂31-Aug-80 1943 JMC
To: *
Has anyone been saving AP digests (other than LWE)? Reply to JMC.
∂31-Aug-80 1936 JMC pattern matching
To: masinter at PARC-MAXC
I'm surveying pattern matching systems preparatory to writing
a chapter in our LISP text. Is the pattern match compiler
described in the 1978 revision of the Interlisp manual still
in use? Is there a SCORE expert on it? Is FLIP in use, and
if so, what's the best way to try it at SCORE? Are there
successors to these. Some pattern match compiler features
seem to work but => got "und." as an answer, but maybe I
didn't use it right.
∂31-Aug-80 1706 JMC pattern directed computing
To: RPG
Are there any general purpose pattern matching programs available
in maclisp or accessible packages. Production systems are also of
interest. I'm currently working on a chapter for our book.
∂31-Aug-80 1700 JMC
To: JK
Please mail me a message then. I'm looking forward to trying it.
∂31-Aug-80 1654 JMC
To: JK
When can I try out ekl?
∂30-Aug-80 1322 JMC
To: human-nets at MIT-AI
Technological Unemployment
In spite of all obstacles their unions can raise, postal workers
will eventually be displaced by electronic mail. This will be good,
because they will then be employed doing other things society needs. They
are entitled to bargain about the terms of this displacement. For
example, it is negotiable demand to protect the job of a 50 year old
postal clerk with 15 years to go till retirement. Severance pay is
negotiable and is part of the cost of introducing the new technology.
The worry (at least 100 years old) of technology causing permanent
unemployment has so far proved groundless. There is no correlation
between increases in productivity and increases in unemployment. In the
last couple years in America, there has been a REDUCTION in productivity
and an increase in unemployment. On the whole countries with low
technology have more unemployment than countries with higher technology
(e.g. China and India have much unemployment), but these phenomena
probably have a common cause (lack of entrepreneurship) rather than one
causing the other.
For the last 30 years, some people have been saying that America
is just about to run out of jobs because of automation. If we were
anywhere near self-reproducing robots with human intellectual ability,
there would be something substantial to adjust to. At present, however,
the problem is the reverse. Low productivity, coupled with the increasing
cost of social and environmental programs (whether or not the increases
are justified) are causing people to feel poor and demand increases in
income that cannot be met without excessive inflation. Also, we have
neglected to develop synthetic fuels and have stalled nuclear energy. One
reason we feel poorer is that our work is going to build palaces and
mosques and giant air terminals in Saudi Arabia?
Many people hastily assume that "mindless" (do they mean manual?)
work is unnecessary. Perhaps it justifies getting out of doing any and
relieves the conscience of people who regard their own (perhaps well
paid) work as useless. Richard Lamson gave examples of assembly line
work as mindless and caring for the sick as necessary. This is stacking
the deck, because both are necessary and both are often boring.
∂29-Aug-80 2323 JMC
To: REM
REM - I suggest you or your deaf friend contact Denny Brown, the associate
chairman of CSD, about his possible CS study. He would know if Stanford
has some official way of helping the handicapped apply.
If this is unsuitable or if it doesn't work out, send me a message. Brown
is DPB@SU-AI.
∂29-Aug-80 2251 JMC photos
To: FFL
I mailed them at SFO Post Office and they should arrive Tuesday. Wouuld
you phone Sigma Xi and tell them. Of course, if I hadn't been so sluggish,
there would have been no need for haste. Maybe you should start collecting
my mail and doing the secretary thing with it, especially reminding me
of commitments.
∂29-Aug-80 2225 JMC
To: csd.barb at SU-SCORE
I have rediscovered and reread Feigenbaum message about Brandenstein. I
forgot because of all the conferences. 9am is still very early for me, but
I do want to meet him.
∂29-Aug-80 1800 JMC photos
To: FFL
They were returned for postage. As far as I can see, you stamped an
account on it, but Stanford forgot to put postage on the envelope.
I forgot to tell you that the Stanford mailing service is unreliable,
and when mailing is urgent, it is necessary to mail it at the
Post Office. Then there is only one kind of unreliability to
worry about.
∂29-Aug-80 1443 JMC Faculty interests file
To: CET at SU-AI
Mine are still current and it is worthwhile to know the interests of
other faculty members. This would be easier if the paper copy would
say what the name of the file is and their was public read access to
it.
∂29-Aug-80 1421 JMC
To: FFL
Should Barbara Roberts ask you for papers from me, please give her
the AI Journal Circumscription reprint, the Israel speech, "Concepts
as objects " and "Recursive programs as objects ...
∂29-Aug-80 1409 JMC
To: csd.barb at SU-SCORE
9am is almost never ok with me for meetings. Also no-one has even
told me the name of this ONR person. You can collect recent papers
from Fran.
∂29-Aug-80 1404 JMC
To: FFL
∂29-Aug-80 1324 FFL REMINDER
To: JMC, FFL
Do you wish to send request for hotel reservations for Sigma Xi?
Yes, for me alone for whole meeting. I guess it would be Friday
and Saturday nights.
∂29-Aug-80 0011 JMC
To: GG
Have a good trip, and come back again sometime.
∂28-Aug-80 2257 JMC
To: pournelle at MIT-MC
Jerry - This is not the article I promised you, but it may interest you.
∂28-Aug-80 1426 JMC Tahoe bomb
To: LLW
What is the effect of refrigerating a bomb with liquid nitrogen?
At some temperature detonators might become inoperative.
∂28-Aug-80 0116 JMC
To: boyer at SRI-KL
Next Tuesday at 2.
∂28-Aug-80 0000 JMC
To: REG
CC: JK
Ketonen's (JK) disk allocation has been reduced from 250K to 120K.
It should be 250K.
∂27-Aug-80 1023 JMC
To: FFL
Please send a note to his office address.
∂27-Aug-80 0046 JMC
To: TOB
Suggest you accept gracious Feigenbaum offer or let's talk.
∂26-Aug-80 2131 JMC
To: *
Confession isn't THAT good for the soul.
∂26-Aug-80 0944 JMC
To: FFL
∂26-Aug-80 0851 Nilsson at SRI-KL (Response to message)
Date: 26 Aug 1980 0849-PDT
From: Nilsson at SRI-KL
Subject: (Response to message)
To: JMC at SU-AI
cc: NILSSON
Yes, I have two cassettes (part 1 and part 2) of the Lighthill
debate (Called "Controversy: Robots" by the BBC). I can either
send them over to you or can send them directly to Karmin
(I will need his address). I forgot how I came into possession
of these tapes, but SRI will be glad to loan them free of charge
to whomever wants to borrow them. -Nils
-------
Fran - Please tell Karmin to call Nilsson at SRI.
∂26-Aug-80 0227 JMC
To: FFL
Please put in ZIP code and pub BOBROW.LE1.
∂26-Aug-80 0059 JMC
To: nilsson at SRI-KL
∂25-Aug-80 1426 FFL Call from Bennett Karmin, I.B.M. 408 256 5307
To: JMC, FFL
Understands you may have or have access to a film, video tape or written
print on the Lighthill Report, which he would like to borrow. He wants to
have an article on it in THINK, the house organ. Can I call him back for
you?
Nils: Do I remember correctly that you have the Lighthill tape?
∂26-Aug-80 0057 JMC Lighthill tape
To: FFL
I think I remember that Nils Nilsson at SRI has it. I'll send him
a message.
∂26-Aug-80 0053 JMC
To: FFL
BIO[LET,JMC] now exists.
∂24-Aug-80 2223 JMC
To: boyer at SRI-KL
2pm, Boyer here
∂22-Aug-80 2100 JMC
To: DPB at SU-AI
CC: RPG at SU-AI
I am now not quite as broke as I was and am willing to pay half of the
cost of RPG's requested trip for Dave Waltz. How about the Department
paying the other half?
∂22-Aug-80 2041 JMC purge
To: pylyshyn at CMU-10A
There will be a big file purge Aug 29 in preparation for the
new cost-center regime. You should FTP anything you value unless
some arrangement for paying for your use is being made.
∂22-Aug-80 2028 JMC phone John Haugeland
To: FFL
Please tell him that he should transfer his files by Aug 29 or they
will probably be deleted in a purge. When he has transferred them,
he should delete them here. His numbers are in my PHON file.
∂21-Aug-80 1618 JMC
To: selfridge at BBNC
Do you think the enclosed makes the whole thing to III oriented?
.require "let.pub[let,jmc]" source;
∂AIL Dr. Robert Kahn↓Defense Advanced Projects Research Agency
↓1400 Wilson Blvd.↓Arlington, VA 20009∞
Dear Bob:
Marvin Minsky and Oliver Selfridge and I have been discussing
the fact that character recognition is one of those fields which
was declared solved many years ago, but isn't. The enclosed is a
sort of proposal for a request for proposals. Oliver is in touch
with other agencies that might also have a need for good character
recognition, and he hopes to be able to discuss it with you.
To keep matters clear, all three of us are on the Board of
Directors of III. However, our belief in a need for general document
reading capability is independent of this, and some of our goals
probably require an approach different from that pursued by III.
.reg
∂21-Aug-80 1502 JMC
To: DPB at SU-AI
Denny - It seems as though you're right, but I prefer to say nothing
unless it seems required.
∂21-Aug-80 1355 JMC
To: FFL
Yes, I agreed with Suppes on the increase.
∂21-Aug-80 0745 JMC
To: JRA
No, it doesn't.
∂21-Aug-80 0148 JMC
To: csd.feigenbaum at SU-SCORE
Ed: Does the following agree with your recollection of what we
promised to do?
Feigenbaum and McCarthy promised the editorial board to propose
a scheme for using part of the journal's money for computerization.
Ideas considered included
subsidizing lecture
commissioning survey articles
on-line form of the journal
on-line bibliography
We have a year to come up with proposals.
∂20-Aug-80 2359 JMC
To: csd.feigenbaum at SU-SCORE
Where and when is our session tomorrow?
∂19-Aug-80 1553 JMC
To: FFL
Barbeque on 24th yes, by myself. Dinner on 28th no.
∂19-Aug-80 1552 JMC
To: FFL
The AP man called again, and we made the appointment.
∂19-Aug-80 1535 JMC
To: FFL
"Here" meant my office.
∂19-Aug-80 1238 JMC
To: TOB
If you declare a meeting time, I'll try to come, but it would seem that my
presence is not essential. I'll be gone most of tomorrow, though 10am
would probably be feasible. Thanks for keeping me informed with the
option of intervening.
∂19-Aug-80 1227 JMC
To: FFL
I have a breakfast tomorrow at 7:30am, which must be over by the time
sessions start at 9. I can meet the AP man at 9:15 here. I also
have a meeting of the III board of directors at 1:30 tomorrow in L.A.
and need a PSA reservation to get there on time from San Jose returning
as soon after 5:15 as possible. Please call the AP man and leave message
as to result.
∂18-Aug-80 2301 JMC
To: csd.friedl at SU-SCORE
We were too early. There was a story with IBM in it.
n094 1918 18 Aug 80
BC-AMDAHL
(BizDay)
By PETER J. SCHUYTEN
c. 1980 N.Y. Times News Service
NEW YORK - Gene M. Amdahl, who resigned from the International
Business Machines Corp. 10 years ago to start a company to produce
IBM-compatible computers, plans to take on the computer giant again.
This time, Amdahl, who stepped down last August as chairman and
chief executive of the Amdahl Corp., is forming a new venture to
compete with IBM, and specifically with its yet-to-be announced
H-series of large-scale business computers.
The new system, said Amdahl in a telephone interview Monday from
Santa Clara, Calif., could be in production as early as 1984, and
would be based on a new generation of very large-scale
integrated-circuit technology.
''This technology has the capability of providing more performance
potential for our planned product, while also offering better
economics than anything anyone has yet to introduce,'' he asserted.
The current chairman of Amdahl, Eugene R. White, is not overly
concerned about Amdahl's plans, particularly in view of the four-year
development time associated with the new computer. ''We had five
years to build the concept of software compatibility, but
increasingly large systems are sold because you can market and
service them.''
White did concede, however, that ''there may be a tendency on the
part of some people to overreact'' to the news that Amdahl was
starting a company that would compete not only with IBM but with the
Amdahl Corp. itself. In trading on the American Stock Exchange
Monday, Amdahl's shares fell 3 1/2 to close at 23 1/2.
Amdahl, who on Friday formally severed all remaining ties with the
Sunnyvale, Calif., company where he had been a consultant and
chairman emeritus, said he expected to make formal a business plan
for his new company between now and the end of the month, and that by
early September he would actively begin negotiating with financial
institutions for backing.
He put start-up costs for the new venture at around $50 million, or
about what it cost in direct investment before the Amdahl Corp.
became profitable.
''We have a pretty good idea of what we are looking for in the way
of financing, and we have already had quite a number of expressions
of interest,'' said Clifford J. Madden, who for the last six years
had been chief financial officer of the Amdahl Corp. before resigning
to become president of the new concern.
There is little doubt that Amdahl, who is not only credited with
being the architect of the revolutionary IBM 360 series but also
enjoys the distinction of being the only senior level employee ever
to quit the computer giant twice, will be able to raise the necessary
capital.
''In view of what he has done in the past, he is going to be popular
among the venture-capital people,'' predicted Ulric Weil, a vice
president of Morgan Stanley & Co., who follows the computer industry.
Nevertheless, Weil and others expressed doubts as to whether Amdahl
would be able to duplicate in the mid-1980s what he was able to
achieve so successfully in the middle of the 1970s.
For one thing the nature of the computer business is changing, from
one that was dependent on dramatic reductions in the price of
hardware to a software-driven business. For another, Amdahl's concept
of building computers that could run on IBM software was for its time
unusual. Now, however, the computer field is crowded with the
so-called plug-compatible manufacturers.
''Taking the longer view, hardware alone isn't going to be as
promising an activity as it once was, and he may be hard-pressed to
show the kind of superior profitability that was possible back in the
mid-1970s,'' Weil added.
ny-0818 2218edt
***************
∂18-Aug-80 2215 JMC
To: JK
Can you phone 857-0672?
∂18-Aug-80 2153 JMC
To: DPB at SU-AI
It would be worthwhile to find some basis for keeping Gosper on the machine.
∂18-Aug-80 2133 JMC
To: REM
She decided she didn't want to pursue it.
∂18-Aug-80 1146 JMC
To: LGC
I have fixed that bug, but the file is just a start on a draft, and
it has never been pubbed.
∂17-Aug-80 2243 JMC
To: RWG
I'll see what I can do about your continued access.
∂17-Aug-80 2212 JMC
To: RWG
Many thanks. Please persist.
∂17-Aug-80 1712 JMC
To: ME
My terminal is in the shop. ZEN is long gone, and the terminal he
was using is one of those belonging to CASBS which we are buying
and which Hersche knows about.
∂17-Aug-80 1709 JMC
To: FFL
It is to be continued with a raise. Suppes's office has details of
when raise starts, but I suppose it is Sept. 1.
∂17-Aug-80 1707 JMC
To: REM
CC: RWG
I believe Lowell Wood and Rod Hyde have analyzed stability.
∂15-Aug-80 0143 JMC
To: *
That which was whipped is now to be kissed.
∂14-Aug-80 1500 JMC
To: RWW
How about Monday 3pm here?
∂14-Aug-80 1430 JMC
To: FFL
I called Lois Shapiro.
∂14-Aug-80 1117 JMC Your outside work
To: RWW
We need to talk about whether you should take some unpaid leave.
∂13-Aug-80 1723 JMC
To: FFL
Search programs work better with a blank line between items in PSHELD, etc.
∂13-Aug-80 1629 JMC
To: JRA
∂13-Aug-80 1034 JRA conf party
carolyn,
did you get my msg about conf. party on aug 28?
john
yes, sorry to dither around so long. JMC and I will come
∂13-Aug-80 1421 JMC
To: DPB at SU-AI, REG at SU-AI
∂13-Aug-80 0843 JRA SAIL USAGE
JOHN,
I JUST GOT A MESSAGE FROM DENNY BROWN PRECIDTING MY REMOVAL FROM THE MACHINE
SINCE RUTH AND I WILL BE LOSING MONEY ON THIS CONFERENCE, HOW ABOUT BLESSING
ME AS A USER AT LEAST 'TIL THE END OF THIS YEAR?
JOHN
I'll see what I can do about regarding helping the LISP conference as a
legitimate overhead expense justifying an overhead account till the end of
the year.
∂13-Aug-80 1035 JMC
To: JRA
I'll see what I can do about regarding helping the conference as a legitimate
overhead expense justifying an overhead account till the end of the year.
∂13-Aug-80 0056 JMC august
To: LWE
I think the file august.ns[1,lwe] contains the august digests, although
I haven't gone through it. You should edit it to see it it has what you
want. It was created by the sequence
/on 1980 ; asks for all 1980 stories
supervisor*digest*laserphoto ; gets the digests
o august ; outputs the 27 stories obtained to the file august.ns
I hope it's what you want.
∂12-Aug-80 1725 JMC
To: EAF at SU-AI
The time is ok; your office, I presume?
∂12-Aug-80 1716 JMC material for CS206
To: SEK
On Carolyn's advice, I decided it would be better and give more
time to have you sell the course notes in class than to have the
Bookstore do it. Please try to find out the right number (from
DPB's estimate of class size)+ copies for ourselves + 15 extra.
Even more important, please find out about the status of MACLISP
manuals. I think I want the students to buy the complete 3 part
manual, though I might be talked into something less. Ellen Gulden
(formerly Ellen Lewis) at M.I.T. is said by the resident MACLISP
expert (Dick Gabriel) to know about its availability. I think the
Bookstore has them on order, which is ok, but please check both
with them and with Ellen. Also please be sure that the MACLISP on
LOTS is suitable and working before class. Dick Gabriel has been
hired by Ralph Gorin to polish MACLISP for a few months, but he won't
start till he finishes his thesis, but he should be willing to
answer a few questions.
∂12-Aug-80 1651 JMC
To: EAF at SU-AI
Friedl?
∂12-Aug-80 1613 JMC saving digests
To: LWE
You might try the following command to ns.
$digest*supervisor*laserphoto/ap/quiet/extend/output digest
This should cause all AP stories with the words DIGEST, LASERPHOTO and
SUPERVISOR to be put in the file DIGEST in your area. The switch
/ap restricts it to AP stories,/quiet causes it not to send you
notification messages when it finds the story and /extend causes
it to extend the file DIGEST rather than replace it when it finds
a story. Let me know if you have problems. I checked a few days
and found that the combination DIGEST*LASERPHOTO*SUPERVISOR gets
all the digests and very little more. An occasional check that
the digests still always contain these words would be worthwhile.
∂12-Aug-80 1148 JMC
To: FFL
Please pub and print BLAINE.RE1[LET,JMC].
∂11-Aug-80 1757 JMC Douglas Hofstadter
To: DPB at SU-AI
He is the author of Goedel, Escher Bach and used to be here
in Suppes group. He has a Guggenheim this next year, is visiting
Knuth (away in August) and wants to use various facilities. I
told him to call you.
∂10-Aug-80 0242 JMC
To: RPG
Thanks.
BIS is the pusher, REG will be the offerer. I will add persuasion if asked,
as I have, more or less unasked.
∂10-Aug-80 0241 JMC !BIS is the pusher, REG will be the offerer. I will add persuasion if asked,
as I have, more or less unasked.
∂10-Aug-80 0129 JMC
To: RPG
Is there a way within STEP of evaluating an expression on the side?
∂09-Aug-80 0944 JMC
To: GIO at SU-AI
Gio - I'm impressed with the start you have made on a proposal.
However, I have two reservations about the present approach.
First, the initial emphasis on Stanford's need may
jeopardize getting money for it as a research proposal. Second,
I have some doubts about whether we want to be involved with C.I.T.
I have talked to Ed, and he thinks that CSD might be able to swing
a ten gigabyte system at a cost that I estimated at $700K (and
Ralph regards as reasonable) without involving C.I.T. The meeting
left me with continued doubts about their interest and capacity
to avoid excessive costs. I hope to phone you later this weekend
when I have had a chance to think a bit more about it.
∂08-Aug-80 1303 JMC
To: rms at MIT-AI
I will see what I can do about helping find a place to stay. For how
long will you be coming and when?
Scientifically, there isn't much change. Administratively, SAIL is
being absorbed by the Computer Science to save money, although the
ARPA contract remains.
∂08-Aug-80 0116 JMC
To: pourne at MIT-MC
OK, have a good trip to the Carribbean.
∂08-Aug-80 0106 JMC
To: pourne at MIT-MC
On 20th, there is another III board meeting, and it would be nice to
see you that evening, though there is a possibility that I'll have
to scoot back to some AI conference thing. It would be interesting
to discuss technological initiatives.
∂08-Aug-80 0055 JMC
To: pourne at MIT-MC
I plan to meet your September deadline.
∂08-Aug-80 0053 JMC
To: pourne at MIT-MC
I suppose a Reagan administration is likely to be somewhat better
disposed towards NASA than the anti-technology crew that has substantial
influence among the Democrats and Carter himself. Maybe he
or his NASA nominee will be interested in a major long term goal like
a lunar colony. However, his technology advisers (and I hope they
are good) will be heavily involved in getting going again various
earthbound enterprises like nuclear energy that have been stymied by
the present administration. It can't hurt to propose, but if you
have a way of reaching him or his close advisers, beware of going
off half-cocked. Between now and the election, of course, he will
want to keep it simple, since that strategy has brought him a
tremendous lead. Most likely a lot of effort will go into trying
to win at least one house of Congress.
∂07-Aug-80 2304 JMC
To: LLW
Congratulations.
∂07-Aug-80 2008 JMC
To: JMC
Stanfor Barn tick. 327-7722 or CAPA 329-2623
∂07-Aug-80 1450 JMC
To: FFL
eiserm.le1
∂07-Aug-80 1441 JMC
To: LWE
If you get this please phone me 497-4430.
∂07-Aug-80 1417 JMC
To: BIS
CC: DCL, TOB, RPG
Many thanks for taking the initiative on a proposal for the support of
MACLISP, and I agree with your proposal entirely.
∂07-Aug-80 0035 JMC book
To: FFL
Please find out if the bookstore has the paperback edition of
"Models of Thought" by Herbert Simon.
∂07-Aug-80 0019 JMC nsf
To: CG
I talked to Suppes, and his suggestions are taken into account in
the draft letter to Kreisel in the file KREISE.LE1[E80,JMC].
∂05-Aug-80 2220 JMC
To: BUD
Sorry to hear that. Anything useful I can do?
∂05-Aug-80 2118 JMC
To: JK
I will talk to Pat, but I need to know your preferences, because
maybe I can temporize about money or maybe I can persuade Pat of something.
∂05-Aug-80 1226 JMC
To: FFL
fenaug.le2.
∂05-Aug-80 0816 JMC
To: kirkley at SRI-KL
Somewhere we have the old M.I.T. program. Dave Wilkins at SRI would
know how to find it.
∂04-Aug-80 1259 JMC
Somewhere we have the old M.I.T. program. Dave Wilkins at SRI would
know how to find it.
∂03-Aug-80 1752 JMC
To: REM
Not yet.
∂03-Aug-80 1226 JMC
To: *
Conjecture: Kennedy will be Anderson's running mate.
∂01-Aug-80 2341 JMC
To: bobrow at PARC-MAXC, hayes at SRI-KL, bmoore at SRI-KL
Dear Dan:
Your inclusion of circumscription in the list of formalisms
requiring "non-constructive" proofs is mistaken. Circumscription of a
formula generates an axiom schema, and whether a sentence can be obtained
by circumscription combined with the usual predicate calculus rules of
inference is just as semi-decidable as whether it is provable in
predicate calculus. Whether a circumscription leads to true statements is
not well defined, because this depends on facts about the world not
expressed in the sentences. This is as it should be, since
circumscription is a rule of conjecture, not a rule of inference.
The main point of my Addendum is to point out that circumscription
differs from the other two in using axiom schemata and I had in mind
remarking that this therefore made it semi-decidable, but I see that
I neglected to make this point explicit. I also pointed out that
schema-based versions of McDermott-Doyle and Reiter were sometimes
possible, but again neglected to say that this would make them
semi-decidable too.
Your "clear reason for this problem" seems correct but misleading.
The true sentences of arithmetic are also not semi-decidable, but Peano
arithmetic which is semi-decidable, is enough for most purposes. Similarly,
circumscription is only an approximation to minimal models, but should
be satisfactory except in exotic cases. Semi-decidable approximations
to McDermott-Doyle and Reiter should be possible in more general
circumstances than my Addendum proposes.
Formula (3) of the Addendum is wrong. The wff on(train,tracks)
should be replaced by PHI(train,tracks), i.e. the Greek letter not "PHI".
Incidentally, while I still think that a "lack of oars" needs
to be included in the ontology of an intelligent program, it turns out
to be neater to do the circumscriptions involved in conjecturing that
one can cross the river without it.
Perhaps I am partly responsible for this misclassification,
because minimal entailment, like entailment itself, is a semantic concept
and not semi-decidable, and you probably relied on early versions of the
paper.
∂01-Aug-80 2313 JMC
To: LWE
Do you remember in which year NS was first put up?
ME - The original APE came to life in summer 1972, and NS replaced it
around May of 1974.
∂01-Aug-80 1727 JMC
To: ME
Do you remember in which year NS was first put up?
∂01-Aug-80 1723 JMC misspelling
To: REG
The last name of the person who gets the account I created is
Erbring, Lutz W. Erbring.
∂01-Aug-80 1703 JMC new account
To: REG
I have created an account for Lutz W. Ebring of the University of
Chicago, namely 1,lwe. If he only uses it a few days for looking
at News Service, we should call it a courtesy, but he may want to
use it longer in which case we should sell him an aliquot. Bureaucratic
argument: NS is a unique feature of this system, and he wants to use
it for research - not operationally.
∂01-Aug-80 1133 JMC
To: JRA
Accept party invitation with pleasure for one maybe two.
∂01-Aug-80 0309 JMC
To: human-nets at MIT-AI
JMC - Stanford
The August 1 HUMAN-NETS strikes me as a prodigy
of wrong headedness for reasons that follow:
1. Kerns takes for granted two false propositions.
First, that there is any reason to limit energy use by public
policy, especially electricity. As much as anyone is likely
to buy can be generated by safe, economical nuclear energy.
Second, there is an economy of scale in using energy, because
distributing it is still most of the cost. Third, it is oppressive
for those in charge of "public policy" to encourage or discourage
any lifestyle beyond making sure that everyone pays the costs
that he incurs.
It is cheap to call in off hours not because of policy but
because the marginal cost is low, since telephone costs are mostly
capital costs.
The British system has suffered from being government owned
by the fact that every mode of using the telephone system for data
communication has been declared illegal in Britain until it has
been proved here. This included, and it may still include, acoustic
couplers. If you think the telephone companies are arrogant, try
forcing a government agency, e.g. the postoffice, to provide a new
service or even allow anyone else to provide it.
2. As long as data service is still such a small fraction
of voice service, and as long as the technology continues to
develop rapidly, it is probably better to postpone universal data
service. It was advocated 15 years ago, but it is well that it
wasn't implemented with the technology of 15 years ago.
3. There are various techniques for getting more service
from the telephone company or any other organization
than you are willing to pay for.
4. It isn't obvious that the extra accounting costs the
telephone company much. Its all automatic. Moreover, people
who tend to complain about bills like to know what they are
being charged for.
5. I'm damned if I can see what this National Professions
Foundation would actually do. It sounds like another useless
crock like registering engineers or computer people - a way of
limiting competition in the name of "ethics".
∂01-Aug-80 0208 JMC
To: ME
Your practical objections to greater kindness to users may still be decisive,
but
1. Anyone can learn to check how long someone has been idle.
2. People are often interrupted and leave edits hanging. If they
forget to do αβ., they lose. This happens especially when the system
halts and they have to leave before it comes back to life.
3. If one can stay indefinitely logged in even when doing something
else, one sees one's messages as they appear.
4. The urge to discipline others is one that should be denied.
How about a login option that will do αβ. when autologout occurs
including when it is occasioned by SYSDOWN? It would be somewhat more
cautious if it put the hanging page in a file with a different name.
∂31-Jul-80 1720 JMC
To: *
Overheard in a record store: "Did you know that Paul McCartney was in
another group before he was in Wings?"
∂31-Jul-80 1329 JMC
To: CG
I find p-calculus hard to follow. Can you come in and explain a bit?
∂31-Jul-80 0043 JMC
To: MLB
Hopcroft and Ullman might have such a system, and Ullman probably knows
that literature.
∂30-Jul-80 2350 JMC
To: human-nets at MIT-AI
Please put JMC@su-ai on list. - John McCarthy
∂30-Jul-80 1522 JMC easing autologout
To: ME
Now that we have plenty of job slots, we could replace autologout on
Datadiscs by autodetach. There could be an autologout after (say)
eight hours or when some limit of jobs was reached. I see we have
been up to 79 jobs. Is there a good reason for the current limit
of 80? What does an inactive job cost in core?
∂30-Jul-80 1520 JMC Scott letter
To: FFL
SCOTT.LE1 lacks the departmental name of Dana Scott. The Philosophy
Department would have his accurate address.
∂30-Jul-80 1501 JMC
To: CG
Who is Harrop?
∂30-Jul-80 1131 JMC
To: FFL
Nothing for Kelly now.
∂29-Jul-80 1717 JMC
To: CJS
The notes will mainly be the same as last year's, but I am tinkering
with them. I would like as much time as possible.
∂29-Jul-80 1713 JMC
To: FFL
schnit.le1
∂29-Jul-80 1046 JMC
To: bmoore at SRI-KL
I have no idea who Alan Thompson is or what he knows about non-monotonic
logic. Must be some mistake.
∂28-Jul-80 1420 JMC
To: FFL
Tuesday at 1:30 will be ok.
∂28-Jul-80 1209 JMC
To: FFL
Late this afternoon would be ok.
∂27-Jul-80 1135 JMC
To: Solomon at MIT-MULTICS
It sounds interesting. Call when convenient to specify meeting time.
My phones are 415 497-4430 work and 857-0672 home. I'm usually home
till late morning.
∂26-Jul-80 1939 JMC
To: RWW, CLT
I went through the list of files DON is proposing to delete
from [1,3]. I will hold out to save REDUCE and RLISP, and the
others have something to do with FOL or various versions of LISP.
If you think some of them should be saved communicate with DON
or ME. The complete list is on BBOARD.TXT[2,2].
BFOL DMP 128964 23-Oct-78 1142 1CLT BFOL 4-Sep-79 1 4
BIGNUM LSP 256 15-Mar-70 0534 4-May-80 1 4
BIGNUM REL 1536 25-Jul-72 1044 2ACH 4-May-80 0 1
COMPLR DMP 30660 1-Mar-78 0145 W78RWW COMPLR 24-Jul-80 6 27
COMPLR FAP 7676 25-Nov-77 1914 NETSYS F-DEC 3-Apr-80 1 3
FOL SEG 46080 15-Jun-79 0030 S79RWW FOL 15-Jun-79 0 0
FOL2 SEG 46080 13-Jun-79 1708 1 ME FOL 13-Jun-79 0 0
FOLML DMP 47044 8-Jun-78 1025 W78RWW FOLML 19-May-79 0 0
FOLMRG DMP 10947 18-Oct-78 2134 1REF FOLMRG 29-Jul-79 0 1
LSPFNT DMP 4102 27-Feb-79 1637 1 JP COPY 13-Apr-79 0 0
MLISP UCI 41984 19-Dec-74 0004 L70DAV MLISP 10-Mar-80 1 3
MLISP DMP 24516 13-Nov-74 1743 ENGHJE MLISP 18-Jun-80 6 28
MLISP2 DMP 34756 31-May-74 2114 1HJE MLISP2 2-Jun-80 1 4
MLISPC DMP 34756 13-Nov-74 1745 ENGHJE MLISPC 10-Jun-80 3 15
REDUCE DMP 51140 17-Nov-77 1321 1ACH REDUCE 16-Jul-80 7 29
RLISP DMP 25540 22-Nov-77 0909 1ACH RLISP 17-Jul-80 1 6
∂25-Jul-80 1657 JMC ebring
To: FFL
please send NS information to
Prof Lutz erbring
NORC
University of Chicago
6030 South ellis
Chicago IL 60637
∂25-Jul-80 1646 JMC
To: sandelin at SUMEX-AIM
Yes, I got your message on Feigenbaum. I prefer the 5th to the 12th,
because there are several reasons why promptness is important in this
matter.
∂25-Jul-80 1005 JMC
To: sandelin at SUMEX-AIM
Do messages to fiegenbaum get to Feigenbaum?
∂24-Jul-80 1554 JMC
To: FFL
earnes.le1
∂24-Jul-80 1553 JMC
To: bmoore at SRI-KL
I know it's late, and I should have mentioned it sooner, but Creary
(lgc here) should be invited to the Formal Methods meeting.
He has some worthwhile if not great results.
∂24-Jul-80 1435 JMC
To: FFL
blackb.le2
∂23-Jul-80 1325 JMC Lehmann
To: csd.buchanan at SU-SCORE
I will be glad to meet with Lehmann, also to take part in AI and
philosophy panel. I was agreeable to being chairman, but Bob
Moore seems to be doing the actual work, and that suits me too.
∂23-Jul-80 1217 JMC
To: REG
Maybe we could buy some more Ampex disks used? Ampex would know who has them.
∂23-Jul-80 1026 JMC
To: FFL at SU-AI, BS at SU-AI
Jon Doyle should get Martin Brooks's desk for now, and we'll do better
on 1st Sept.
∂22-Jul-80 1338 JMC
To: SEK
When you have a chance, look at boo,jmc especially book.dir.
∂22-Jul-80 1045 JMC
To: bmoore at SRI-KL
Yes, we'll share?
∂21-Jul-80 2016 JMC
To: lauren at UCLA-SECURITY
Count one vote for measured rates.
∂21-Jul-80 1716 JMC
To: ROB
xgpsyn loses because of bad channel
∂21-Jul-80 0048 JMC
To: RPG, ROB, ME, LES
αl<space> works fine, but αk<space> kills one char at a time.
∂20-Jul-80 2206 JMC
To: WLS
See Machine Intelligence n for n≤4. It will be there or at least a reference.
∂19-Jul-80 1342 JMC
To: pourne at MIT-MC
Jerry, do you agree?
Pro-technology must appeal to the imagination
Maybe the anti-technology movement has reached its peak
and will decline. However, if the age groups in the population
retain their present opinions as they get older, and if the
new generations take their opinions from their teachers and
their immediate elders, there will be a real technological
disaster. All feasible methods of producing energy may be
paralyzed. Most likely the disaster would cause a sufficient
reversal of sentiment so that the country would recover.
If the disaster is to be avoided, youth must rejoin
the pro-technology camp. However, sober arguments about
the cost-effectiveness of nuclear energy and its good safety
record won't do the job. Neither will depicting the effects
of neglecting energy production until our noses are rubbed
in the consequences of the neglect.
It seems to me that what affects the opinions of young
people about technology is in a large measure romantic appeal.
What pro-technology sentiment now exists among young people is
not gratitude for the fruits of the technological civilization
that permits them to live in comfort and study for years before
going to work. It isn't gratitude for the marvels of medicine
that saved half of them from dying of diarrhea as infants and
small children. It is rather the romantic appeal of present
and contemplated future technology.
The pro-technology young people I have met have been
more influenced by their personal experience with technology
than by reading pro-technology literature which is almost
completely unavailable. For the forseeable future, this will
continue to contribute a substantial number of people who
will respond to appeals to defend technology. The number of
pro-technology young people who will act politically is much
smaller and depends much more on the existence of specific
pro-technology political and ideological movements.
Much of the future pro-technology political effort in
the U.S. depends on two sources - the L-5 society and the
science fiction group typified by and perhaps centered around
Jerry Pournelle.
The L-5 society is an extremely interesting phenomenon.
It was founded to advocate implementing a proposal by Princeton
Physicist Gerald O'Neill that the government establish habitats
for people in orbit around the earth. O'Neill's proposal differed
from previous such proposals in that it was more detailed and
that his cylindrical space stations were designed to enable people
to live in earthlike conditions with earthlike gravity (from
centrifugal force), a 24 hour day (from a combination of rotation
and precession of the cylinders. Internally the cylinders were
to be lined with soil so that the inhabitants could lead ecologically
sound agricultural lives. Thus O'Neill's proposals combined
the idea of space travel with the passion for ecological living
that was popular at the time of the proposal. The proposal also
involved a huge government program to plan and administer these
ecological lives.
The above may be a bit of a caricature, but all the mentioned
ideological elements were present.
Subsequently the L-5 society evolved away from the O'Neill
concept. The ecological lives have been de-emphasized and the
government planning has been de-emphasized. In fact, the L-5
Society has become a promoter of private enterprise in space
and the main and most effective opponent of the proposed treaty
on outer space on the grounds that declaring space the "common
heritage of mankind" would outlaw or at least discourage private
enterprise in space. The L-5 society has also become the most
effective constituency for a large American space program apart
from companies involved in the program. Its appeal is to the
imagination and the spirit of exploration and adventure and
not merely to the rather narrow concept of scientific
investigation.
Pournelle and like-thinkers have had a major impact on
the ideological content of science fiction. Until after World
War II, science fiction was not much affected by ideological
currents among American intellectuals. Its writers and editors
either had a technological background and a simple engineer's
enthusiasm for technology and acceptance of the present social
system or they regarded science fiction as genre with these
presuppositions. After World War II, however, liberal ideological
themes became increasingly prevalent. For example, robots changed
from being a tool or an alien menace to an oppressed minority
requiring tolerance. Sociological themes speculating on a world
dominated by the advertising industry or the Mafia appeared.
In the late 1960s, anti-technological themes became important
and so did counter-cultural themes.
The new event is the "class consciousness" of science fiction fans
and writers who are often visionaries who really would like to explore
space and welcome new inventions. It finally sunk in after Lord knows how
many stories about the evils of technology and renunciation of technology
that their hopes were the victims of this ideology. They began to notice
that there was a connection between the anti-technological propaganda and
the decline of the space program.
∂19-Jul-80 0342 JMC SUN display
To: FB at SU-AI, REG at SU-AI
It would be worthwhile to test piping signals for high resolution
through existing coax.
∂19-Jul-80 0323 JMC
To: pourne at MIT-MC
Listing my computer files, but I had just decided to give up and
go to bed.
∂19-Jul-80 0224 JMC
To: minsky at MIT-AI
LAINGSBURG, Mich. (AP) - Big Al from Laingsburg may be in line to
help decontaminate the crippled Three Mile Island nuclear reactor, but
he's not worried about it.
Big Al is a robot.
''We've designed a robot we expect will enter contaminated
environments and perform dose rate surveys and lower the levels of
contamination,'' said Jim Hatt, vice president of Hodges Anitronics,
which designed Big Al. ''If successful, it would save human exposure
to low dose rates of radiation.''
Hatt said the robot, which is 3 feet high with a 15-pound mechanical
claw, will be ready for tests next month.
''We tested one of our smaller units in the bottom level of TMI's
turbine building,'' Hatt said. ''It functioned perfectly, and we
didn't lose (radio) signal with it despite it being in a room with a
lot of pipes and concrete walls.''
Dave Delzingaro, a spokesman for TMI, confirmed Hodges Anitronics
made the robot proposal, and said ''it will be looked at seriously.''
The plant was contaminated in March 1979 in the worst commercial
nuclear power accident in the nation's history.
ap-ny-05-24 1430EST
∂19-Jul-80 0216 JMC
To: pourne at MIT-MC
I said, "I often wonder ...". I think there is something in the idea,
but I don't think four years is long enough to implement both phases
of the process. However, it may be feasible to create new organizations
for certain purposes rather than expand old ones that happen to have the
right names. This might be most applicable to government agencies, but
it appears that Reagan hasn't (at least so far) the political strength
to actually propose wiping out non-functional agencies. Anyway he must
first be elected, and there's many a slip ...
∂18-Jul-80 1751 JMC
To: HUMAN-NET at MIT-AI
for HUMAN NET
A major threat to reasonable rates for long hold times is
likely to come from populist attacks on terminal use as a luxury
of the rich. Some utilities commissioner building a record
of defending the public interest preparatory to seeking higher office
will denounce the selfish luxury of
wealthy terminal users and will soak them in order to keep
residential voice rates down. The response to arguments that terminals
will eventually be as ubiquitous as color TV's will be the classical,
"What has posterity ever done for me?". Such populist reasoning
is as likely a threat as Lauren Weinstein's fear of the telephone
companies' simple desire for profits.
There is likely to be a strong push for more phone company
income, because the nominal return on investment
of any regulated capital intensive business
is illusory in a time of high inflation. He who sold beans ten years ago
and bought telephone stock, sold the stock this year and bought
more beans with proceeds and the dividends probably has fewer beans
than when he started even though he has a nominal profit on which
he gets to pay tax. Technology improvements seem to have kept the
heads of the telephone companies above water, even though electric
many utility bonds have gone down from AAA to B, and they can't sell
stock to finance new plants. There is probably strong pressure
in the phone companies
to make money in any possible way, and the terminal users may get zapped.
In any case, it seems very unlikely that terminal users
will be able to get eight hour local calls for the same price as
five minutes for very much longer. The danger is that they will
be charged more than marginal costs in order to avoid a increases
in residential rates. The defense must be reasonable, and it must
emphasize the need to encourage the future mass use of terminals.
∂17-Jul-80 2306 JMC computer prepared manuscripts
To: DEK
I just received a brochure advertising the Journal of Algorithms
with you as editor. Its "Instructions to Authors" assume that
all manuscripts will be typewritten rather than prepared on a computer.
Shouldn't the possibility of using PUB or some other typesetting
program be admitted and regulated?
∂17-Jul-80 2223 JMC
To: BIS
Now that you mention it, that part of the myth is correct.
∂17-Jul-80 2141 JMC
To: BIS
Perhaps I misunderstood your message. The original design of our
keyboards was the work of Steve Russell and Dave Poole and others whom I
don't remember. Poole is still in the neighborhood. I don't recall
caring very much.
I was not consulted in the present change. It was discussed in
the BBOARD file, but I wasn't reading it at the time or I would have
vetoed the change. Many people including me who touch time have been
incommoded by moving not only the numbers but also the special characters
above them. I have the further problem of also using Imlac and Datamedia
keyboards that haven't been modified.
The change doesn't make our keyboards conventional, so I think it
is somewhat harmful on the whole, but I don't want to involve myself in
such matters.
∂17-Jul-80 2009 JMC
To: BIS
I was not consulted at any time.
∂17-Jul-80 1602 JMC
To: FFL
Please get library copy of G.E. Hutchinson, Introduction to Population Ecology
∂17-Jul-80 1302 JMC dired or the like
To: DON at SU-AI
Is dired suitable for the following purpose. I keep lists of
my files in categories, i.e. lisp programs, drafts of papers, notes
on mtc, etc. The names of the files are an insufficient indication
of the content, so my file FILES has a description of each file, and
I often use E search commands for key words in these descriptions.
The file names are written as <name>.{<ext>}[prj,prg] so I can get
to them with Xpoint commands. At the end of each quarter, I like
edit the directory of that quarter's [prj,prg] - I start a new
directory each quarter - and enter the files into FILES. DIRED doesn't
seem to be directly usable for my purpose because (1) it doesn't
list the files in the <name>.<ext> form and (2) it doesn't provide
for the file descriptions (3) it doesn't help with the merging.
Is the present DIRED usable for this somehow or are you
motivated to modify it in that direction?
∂16-Jul-80 1110 JMC
To: pourne at MIT-MC
John McCarthy, Computer Science Dept.,Stanford CA 94305
I remember about Survival of Freedom and hope to have some
alternatives for you in a few days. Have you seen Knowledge
and Decisions by Thomas Sowell? It's great.
∂16-Jul-80 0040 JMC
To: CG
Please phone.
∂16-Jul-80 0040 JMC
To: RWW
Is Chris back? Please ask him to phone.
∂15-Jul-80 1302 JMC
To: ME, ROB
The changed top row is really a pain.
∂14-Jul-80 1551 JMC
To: FFL
See FUSION[E80,JMC] for some comments on Fusion, for which thanks.
∂12-Jul-80 2048 JMC
To: AAM
Even though I am a member of the Board of Directors of III, I will be quite
interested in the information you obtain. III put big hopes into selling
Graphics-1 to the Social Security Administration but eventually lost out to
much more primitive methods that do less. Graphics-1 is very expensive
and based on obsolete technology, and III isn't motivated to modernize it
without customers. However, I believe that some of the ideas in Graphics-1
together with new ideas might lead to a cost-effective system, but I am
not personally knowledgable about such technologies.
I will be interested in discussing it when you come back.
∂12-Jul-80 1231 JMC
To: REM
June 30 Computer World (math library) has articles on commercial file compression.
∂10-Jul-80 1843 JMC
To: RPG
If I understand what you are proposing - some maintenance in exchange
for good access, there should be no problem. Scott Kim, who will TA
CS206 in the Fall, may have some questions and ask some advice. We have
no current plan for volume II. We will not modify the book much before
Fall, but we will work on it then and hopefully finish it off.
∂10-Jul-80 1819 JMC
To: EAF at SU-AI
Accepted.
∂10-Jul-80 1442 JMC
To: DLW at SU-AI
Is there a later LISP Machine Manual than Novermber 1978?
∂10-Jul-80 1346 JMC
To: TOB
I forget. Do you get Assembly Engineering and do you want mine?
∂09-Jul-80 1508 JMC
To: RPG
What you say sounds right, but I didn't see the original "malignment".
∂09-Jul-80 1057 JMC
To: buchanan at SUMEX-AIM
CC: bmoore at SRI-KL, balzer at USC-ISI
I would be happy to participate.
∂08-Jul-80 2102 JMC
To: ROB
Do I need to modify my Imlac keyboard?
∂08-Jul-80 1440 JMC
To: CLT
OK, around 5
∂08-Jul-80 1334 JMC a file
To: FFL
I renamed the file EXPSE2[1,JMC] to be SE2.LE2[LET,JMC]. I want
to reserve [1,jmc] for files of permanent use.
∂08-Jul-80 1329 JMC phone number for Lenat
To: FFL
I received a letter from Paul Gilmore about Reiter. If it has
Gilmore's phone number, please send it to Lenat (DBL on the machine).
∂07-Jul-80 2050 JMC
To: JK
Ok. See how he reacts.
∂07-Jul-80 1703 JMC
To: rms at MIT-AI
No, and now it isn't clear that I can. ARPA lured me into offering
someone a job and then didn't come up with the money. This
disappointment is only a few days old, so maybe I'll be able
to do something about it, but if you can find alternate financing,
it would be good.
∂07-Jul-80 1628 JMC
To: FFL
Yes, the manual.
∂07-Jul-80 1543 JMC
To: FFL
Also D. Kahnemann and A. Tversky, Psychol. Rev. 80,237 (1973).
∂07-Jul-80 1506 JMC
To: FFL
Please get from some library "Attitudes and Opinions" by S. Oscamp.
∂07-Jul-80 1401 JMC materials to Carolyn Henson
To: FFL
Please send her NS.NS[E80,JMC] and a writeup of the News Service
program which Les or Martin Frost can tell you how to get.
∂05-Jul-80 1649 JMC
To: SEK
Let's get together some afternoon and talk about CS206 preparations.
∂03-Jul-80 1644 JMC
To: LES
%3Cartwright, R.S. (1976)%1:
%2A Practical Formal Semantic Definition and Verification System
for Typed Lisp%1,
Ph.D. Thesis, Computer Science Department, Stanford University,
Stanford, California.
%3Cartwright, Robert and John McCarthy (1979)%1: "Recursive Programs
as Functions in a First Order Theory", in E. Blum and S. Takasu (eds.),
%2Proceedings of the International
Conference on Mathematical Studies of Information Processing%1, Springer
Publishers.
∂03-Jul-80 1639 JMC
To: LES
∂29-Apr-80 1309 CG
Here is the description of my thesis work for the ARPA report.
Chris Goad is currently completing his PhD thesis, titled "Computational
uses of the manipulation of formal proofs". The thesis concerns the use
of proofs as descriptions of algorithms. In particular, a proof in an
appropriate system of a formula of the form "for all x there exists a y
such that R(x,y)" can be used express an algorithm A which satisfies the
"specification" R in the sense that, for each input n, R(n,A(n)) holds.
The proof can be "executed" by use of any of several methods derived from
proof theory, for example, by a proof normalization procedure, or by
extraction of "code" from the proof by means of one of the functional
interpretations.
A proof differs from a conventional program expressing the same algorithm
in that it formalizes more information about the algorithm than does the
program. This additional information expands the class of transformations
on the algorithm which are amenable to automation. For example, there is
a class of "pruning" transformations which improve the computational
efficiency of a natural deduction proof by removing unneeded case
analyses. These transformations make essential use of dependency
information which finds formal expression in a natural deduction proof,
but not in a conventional program. Pruning is particularly useful in
adapting general purpose algorithms to special situations.
Goad has developed efficient methods for the execution and transformation
of proofs, and has implemented these methods on the Stanford Artificial
Intelligence Laboratory's PDP-10 computer. He has done a series of
experiments on the specialization of a bin-packing algorithm. In these
experiments, substantial increases of efficiency were obtained by use of
the additional "logical" information contained in proofs.
The general objective of Goad's work has been the development of improved
technology for manipulating methods of computation. In his thesis, he has
exhibited new automatic operations on algorithms - operations which are
made possible by the use of formal proofs to describe computation.
∂03-Jul-80 1525 JMC
To: REM
I, for one, would be interested in seeing a more detailed exposition.
∂03-Jul-80 1137 JMC giant file
To: EAF at SU-AI
The best way to make a research project of it is to involve Gio, because
he has research activity in data bases. Unless you think he is about
to disappear, I would suggest that he and I work out something together,
but I would expect him to handle the technology on a day-to-day basis.
I would help think of things to do, and I have quite a few in mind already,
and I would help with the politics.
The immediate major problem with C.I.T. is that they have traditionally
charged so much for disk space that they have excluded many applications.
Also I think it might be easier to call it a research project if it
didn't involve them. However, these obstacles might not be insuperable.
It occurs to me that the big deal is to involve the library.
The goal is to make a start on having the documents themselves on-line.
We could begin with documents that are already available in digital
form. We could start with CSD's terminals and the library could acquire
terminals gradually. Incidentally, we have been slow enough in going
to a terminal system that the price of memory may have come down enough
to go for 1000 by 1000 right away.
If you agree that Gio can play an important role, I'll try to
call a meeting of you, me, Forest and Gio - with Don invited.
∂03-Jul-80 0035 JMC
To: JK, LGC
I have sent a version to Engelmore. It may or may not require revision.
∂03-Jul-80 0035 JMC
To: engelmore at USC-ISI
.require "memo.pub[let,jmc]" source;
.cb FISCAL YEAR 80 ACCOMPLISHMENTS AND FISCAL YEAR 81 PLANS
.CB FORMAL REASONING GROUP
.CB STANFORD ARTIFICIAL INTELLIGENCE LABORATORY
.CB Computer Science Department, Stanford University
Professor John McCarthy continued his work on formalization
of common sense knowledge and common sense reasoning. The goal is
to obtain a formal language for use by AI
programs with the precision of mathematical logic
and the expressiveness of natural language.
McCarthy related his circumscription approach to non-monotonic
reasoning to the approaches of McDermott-Doyle and Reiter. They cover
distinct AI requirements, but McCarthy's axiom schemata and circumscription
rule seem more suitable for computer use in both cases.
McCarthy has a new approach to using first order logic
for expressing concepts that should be vaguely defined, because the
users of the concepts are referring to entities only a few of whose
properties they know. This is related to recent work by philosophers
on "natural kinds".
Thus a person hasn't a definition of lemon to which he is irrevocably
committed; we could accept blue lemons if there were good enough reasons
why they should be included, i.e. a person regards the concept of
a lemon as a natural kind existing outside himself and about whose
criteria he could learn more. AI programs must use terms like
⊗bachelor whose definitions are conventional and natural kinds like
⊗lemon and treat them appropriately. The technical consequence of
this common sense fact seems to be that the formal language doesn't
have a known preferred "standard model".
During Fiscal 81, McCarthy will continue this research and
expects to prepare a paper on why AI programs need vague concepts
and how to make them work. An AI memo will shortly be published
on McCarthy's Elephant formalism for proving facts about programs
in first order logic.
Jussi Ketonen has been working on fragments of predicate calculus
admitting fast decision procedures. Predicate logic has been traditionally
viewed as the place to formalize the notion of a "logical inference";
intuitively it is clear that any "trivial" inference should be easily
formalizable and decidable in predicate logic. As is well known,
the full system of predicate calculus is not decidable; thus we one is led
to study suitable fragments of predicate logic. Ketonen has isolated a
fragment that seems particularly promising: In some intuitive sense it
captures the notion of a "direct" inference. The corresponding decision
procedure works in exponential time and is tailored to situations that
occur naturally in actual mathematical practice.
This work will be described in detail in a forthcoming paper
by Ketonen and Weyhrauch. Ketonen has written a program that implements the
above-mentioned decision procedure in the FOL proof checking program at SAIL.
The results were encouraging though somewhat inconclusive due
to the technical difficulties encountered in the FOL environment.
Ketonen is currently in the process of designing and implementing
a new mathematically oriented proof-checker in which to test the above
mentioned ideas. The proposed proof-checker would be radically different
from all its predecessors in its expressive power and its ability to
manipulate proofs. For example, we can then implement procedures to
unwind proofs of well known mathematical theorems, say, the
recent non-constructive proofs of Van der Waerden's theorem. This kind
of a task is purely mechanical in character yet quite impossible to do
by hand. The interest in this endeavor lies in finding new upper
bounds for the numerical function occurring in Van der Waerden's theorem;
the previously known bounds are very bad.
Ketonen hopes to have a small subset of his system running by the
end of the summer. The top level language has not been decided yet and
will be the subject of intense thought.
Lewis Creary continued his work on the epistemology of artificial
intelligence, and on the design of an advice-taking problem solver.
On the epistemological side, new ideas were developed for dealing with:
a) causal reasoning and the frame and qualification problems,
b) the formal representation of actions,
c) epistemic feasibility of actions and the need to plan for the
availability of knowledge, and
d) the structuring, classification, and use of domain-specific
knowledge within a general problem solving system.
Work on the advice-taking problem solver was concentrated for the most
part on high-level design issues and associated theoretical questions.
A tentative high-level design and long term plan were constructed,
incorporating the epistemological ideas mentioned above, as well as new
theoretical ideas concerning:
e) problematic goal seeking, goal refinement, and achievement planning
f) advice seeking and taking as a means of program development and selective
knowledge acquisition,
g) complex motivational structures for use in problem solving.
Also, work was begun on a detailed design for a modest initial implementation
of the problem solver as an advice-taking travel and communications planner.
During Fiscal 81, Creary plans to finish the initial detailed design of the
problem solver and to construct an experimental implementation of it as a
basis for further development. He also expects to produce some papers
expounding the theoretical ideas on epistemology and on problem solving that
were developed during the past year.
∂03-Jul-80 0026 JMC bullets
To: engelmore at USC-ISI
Your request for bullets about fiscal 80 and 81 work came at a difficult
time. Two people, Chris Goad and Luigia Aiello, who either have been
or will be supported are in Europe at a meeting and won't be back till
the middle of July. The next message is an approximation to your needs.
It is too long, because it includes explanation, (I fear too technical)
of the background of the accomplishments and plans. I'm willing to work
to prepare a condensed version if you consider it important.
Your message about inability to assure additional support
leaves me in a difficult position, since my offer to Goad was based
on your March encouragement to make the offer. I will have to use
the last of my unrestricted funds and may have to cut off some work
early if help isn't forthcoming in January.
∂03-Jul-80 0015 JMC datadiscs
To: REG
I understood you to say that there are some uninstalled datadisks. The
offices now occupied by Les and Hersche will be occupied by more than
one research associate each and at least Les's office has the connection
for another datadisk, and I would like to install at least that other one.
∂03-Jul-80 0003 JMC
To: raj.reddy at CMU-10B
I sent "To whom ..." letter on Monday to Beverly Howell.
∂03-Jul-80 0001 JMC
To: EAF at SU-AI
I have asked Frances to be sure Doyle's mail is held.
∂02-Jul-80 2359 JMC
To: LES
CC: TOB
Please talk to me before + or - action on Levinthal trip paying.
∂02-Jul-80 2354 JMC form from Fariello, manager of conferences
To: JRA
I received a form dated June 27 on RESERVATIONS FOR HOUSING ...
addressed to conference sponsers (sic). It looks like something
you would have already covered, but maybe you should phone her
to be sure.
∂02-Jul-80 1338 JMC
To: LGC
Difficult as it may seem, the new ideas that were developed should be
stated in single sentences. One or two will be enough.
∂02-Jul-80 1205 JMC
To: FFL
∂02-Jul-80 1040 doyle at MIT-AI (Jon Doyle) Mail coming
Date: 2 JUL 1980 1338-EDT
From: doyle at MIT-AI (Jon Doyle)
Sent-by: TK at MIT-AI
Subject: Mail coming
To: jmc at SU-AI, eaf at SUMEX-AIM
CC: DOYLE at MIT-AI
Dear John and Ed,
I'm not sure who to tell this to, but I have begun to
tell people of my new address in the computer science
department. I would greatly appreciate it if your mail-
handling personnel could be warned to hold any mail that
arrives for me before I get there.
Thank you very much.
Jon Doyle
∂01-Jul-80 1840 JMC
To: JK
Thanks for notes; I haven't really got to them yet.
∂13-Aug-80 0059 JMC august
To: LWE
You had best edit the stories by
the command EDIT AUGUST.NS and EDIT DIGEST.NS. The editor called by
that command is the same as the one promoted by D.E.C. for the
DEC 20. It was in fact written years ago in our lab. We have a better
display editor, but it isn't much better than EDIT on a teletype terminal,
and you are probably already familiar with EDIT. Incidentally, we may
eventually begin to begrudge the file space for these digests and they
might be moved to tape and by tape to your DEC 20.
∂26-Aug-80 0232 JMC
To: FFL
I found ZIP, so please just XS BOBROW.XGP[LET,JMC].
∂01-Sep-80 2354 JMC
To: RPG
A simple simplifier for sums and products sits in match.lsp[e80,jmc].
Among other things it takes (plus 2 x 3 y) into (plus 5 x y). The
match part uses %match ok, but maybe instantiate requires extension
to handle the instantiate part. The relevant case is noted in the
file.
∂07-Sep-80 1118 JMC
To: VRP
Eldredge is TWE, and Gorin is REG.
∂14-Sep-80 1454 JMC
To: FFL
I forgot that the address is already in my file under SE2.
∂14-Sep-80 1626 JMC
To: buchanan at SUMEX-AIM
Further comments on recommendations: Since one of our major expenses
is software, re-equipping with software is as important as re-equipping
with hardware. For example, support for the maintenance of MACLISP and
INTERLISP would be very helpful.
I doubt that a recommendation that would apply to all sciences is
especially helpful unless some contrast can be pointed out. For example,
there is no more obvious need to underwrite journal subscriptions in
AI than in any other field. On the other hand, there are sources of
funds for postdocs in other fields and nothing in AI or even in computer
science generally.