perm filename W79.OUT[LET,JMC] blob
sn#430484 filedate 1979-04-01 generic text, type T, neo UTF8
∂01-Apr-79 1219 JMC
To: MRC at SU-AI
Please don't give up easily.
∂01-Apr-79 0956 JMC
To: CLT
write dad, $620 tax.
∂31-Mar-79 1403 JMC 3 user packs
To: ME
I assume this is just for debugging. As soon as it is safe, there should
be another file pack.
∂31-Mar-79 1301 JMC
To: SSO
I don't see how we can get together without intrducing an explicit
time. Monday 1:30 at your place or any other afternoon at the
AI Lab are possible for me.
∂31-Mar-79 0815 JMC
To: *
If you would like to hear an NRC recorded message on the current 3 Mile Reactor
situation, you can call (301)492-8585 and hear a busy signal.
∂30-Mar-79 2226 JMC
To: jrobinson at SRI-KL
Sorry I wasn't clear I couldn't come.
∂30-Mar-79 1226 JMC
To: jrobinson at SRI-KL
Wednesday out for Ma. How about Tuesday?
∂30-Mar-79 1206 JMC
To: MAX
Can you come to dinner Wednesday 7pm at my house?
∂30-Mar-79 1115 JMC
To: FRG
ok about change
∂30-Mar-79 0902 JMC
To: MD
Might ONR support a small proposal to improve and expand NS?
∂30-Mar-79 0848 JMC your message
To: EAF at SU-AI
One immediate step is to order another disk channel from Foonly or other source.
Exactly what to order is complicated by the possibility of getting something
that will take 3350 technology disks which admit cheaper expansion than our
present 3330 technology disks (which must also be kept). I don't know what
the cost is yet, and we may need to guarantee a purchase order with some
of my unrestricted money and posssibly some departmental gift funds. ARPA
will be a bit slow, so we may need to borrow from the University or something
like that.
Taking some such action is necessary if the separation is to take place
promptly, and it will show Chowning that we are serious about it.
∂30-Mar-79 0845 JMC Foonly channel
To: LES
I think we should probably order a new Foonly channel, guaranteeing it
if necessary from unrestricted funds - maybe with a cancellation clause
in the agreement that will be fair both to Foonly and ourselves. Time
is passing w/o any decision from Chowning. Now that I think about it,
Feigenbaum may help with Departmental gift funds as a means of putting
pressure on Chowning.
∂30-Mar-79 0019 JMC
To: jrobinson at SRI-KL
I'm fine except for intestinal bug that will probably still prevent supper
- unless it consists entirely of Jello and the like. Moreover, I have a
concert Friday also.
How would you like do come here next Wednesday evening. We would also
invite our visiting scholar from Peking University Ma Xiwen, who will be
at the AI Lab for a year and is interested in reasoning in natural
language. I should be better by then, or else we can all have Kung Pao
Jello.
∂29-Mar-79 0856 JMC
To: PAT
I'll be staying home today.
∂28-Mar-79 0002 JMC
To: BS
Please send me application forms for sabbatical next year.
∂27-Mar-79 2007 JMC
To: DPB
I meant the logic course taught in math or philosophy.
∂21-Mar-79 0829 JMC
To: JED
I'm away till Friday. Please check whether Ma needs anything.
∂21-Mar-79 0828 JMC
To: MAX
I will be away till Friday. Jim Davidson can help with any problems you may have.
∂21-Mar-79 0050 JMC
To: CLT
alice on april 5
fix fol output line length
∂20-Mar-79 2212 JMC
To: PAT
Since I won't be teaching next year and hope to revise the subject matter
then, no copies are needed for classes. I guess I will need about
40 copies each.
∂20-Mar-79 2210 JMC
To: PAT
CONCEP and MENTAL should cite both ARPA and NSF.
∂20-Mar-79 1444 JMC
To: MRC at SU-AI
∂20-Mar-79 1440 DON
Please mail me a one-line description of DIALNET for HELP HELPER.
To MRC: Send him this with a reference to additional documentation.
∂20-Mar-79 0157 JMC
To: LLW
See CHEMEM[W79,JMC].
∂19-Mar-79 2244 JMC
To: TOB
A point that hadn't occurred to me. I'd better think about it.
∂19-Mar-79 2230 JMC
To: TW at SU-AI, DBL at SU-AI, TOB at SU-AI, DPB at SU-AI,
EAF at SU-AI, buchanan at SUMEX-AIM, lenat at SUMEX-AIM
I still favor the renaming of 206 to 122. My experience and Carolyn
Talcott's in teaching the course showed that the proving part is
well withing the capabilities of the students who take the course -
even when the proving was introduced without a warning that might
scare of the non-mathematical. The way we do it, it just isn't
that difficult, especially since it is probably somewhat more
natural than the inductive assertions method for sequential
programs. I want to teach 226 in 1980-81, since I will have
spent a year working on just that material. Therefore, I would
prefer that the course not be given with that number or my course
description next year.
∂19-Mar-79 2051 JMC maclisp at score
To: MRC at SU-AI
I succeeded in creating a file at SCORE with EMACS from my Imlac,
but lines got lost from the display of it, and I didn't succeed in
reading it into MACLISP.
∂19-Mar-79 1430 JMC
To: DPB
Tuesday afternoon is fine.
∂15-Mar-79 1459 JMC
To: bmoore at SRI-KL
I hope I can come.
∂15-Mar-79 1046 JMC
To: PAT
How are we coming on FIRST as an AI memo?
∂14-Mar-79 1751 JMC
To: amarel at SUMEX-AIM
Sure, please phone to arrange time.
∂13-Mar-79 2132 JMC
To: FML
My impression would be that they haven't all been oral, and I will look
for something, but it is unlikely that I will find anything.
∂13-Mar-79 1528 JMC
To: REP
Please come and see me about possible summer work in verification. I think
that Backus is on a wrong track. Someone with a clear firm point of view
might make something useful from their work, but I think they are unlikely
to produce something useful themselves. Backus's Turing lecture reflects
ignorance of important and relevant work in mathematical logic.
∂13-Mar-79 0953 JMC
To: PAT
Thanks for telling me where you will be.
∂12-Mar-79 1916 JMC
To: BCM
Patte will try to make it pox.
∂12-Mar-79 1914 JMC POXing an old memo
To: PAT
Ben Moszkowski would like a copy of RELATI.POX[CUR,JMC] suitably poxed.
It doesn't POX anymore, maybe because it uses an old pox macro file
of Gosper's. Could you try to make it work?
∂11-Mar-79 2111 JMC photo and biography
To: PAT
Please get them to make another publicity type photo and send it together
with my biography to Donald Michie. He'll be around today, so please get
him to say which address.
∂11-Mar-79 2002 JMC
To: *
a295 1934 11 Mar 79
Donald Power
COLUMBUS, Ohio (AP) - Donald C. Power, retired chairman of General
Telephone and Electronics Corp., died Sunday at his home in Galloway,
Ohio, after a long illness. He was 79. Power was named chairman and
chief executive of GT&E in 1959, and continued as chairman until 1971.
∂10-Mar-79 1550 JMC
To: PAT
I think I sent it but please check.
∂10-Mar-79 1544 REF
John -
If you haven't already done so, could you please send a current
recommentation letter to:
Aravind K. Joshi
Chairman, Department of Computer & Information Science
Room 268
Moore School / D2
University of Pennsylvania
Philadelphia, Pennsylvania 19104
Thanks.
bob
∂09-Mar-79 1724 JMC
To: ijcai79 at SUMEX-AIM
I have received one or more papers.
∂08-Mar-79 1541 JMC
To: jrobinson at SRI-KL
$1 from me.
∂07-Mar-79 2255 JMC chowni.le1
To: PAT
Please repub it. You may have to move your .next page commands, because
I have added a bit here and there.
∂07-Mar-79 1950 JMC
To: DPB
Who is head of admissions this year?
∂07-Mar-79 1937 JMC
To: keller at CMU-10A
CC: reddy at CMU-10A
I have not been a member of the admissions committee for two years, and I am
not sure how things are done. In particular I don't know how many people
look at each folder in the course of selecting the top 50. If Raj or somone
of equal stature finds the outcome strange, then a message from him to
the current head of admissions might or might not cause re-examination.
∂07-Mar-79 1937 JMC
To: keller at CMU-10A
CC: reddy at CMU-10A
I have not been a member of the admissions committee for two years, and I am
not sure how things are done. In particular I don't know how many people
look at each folder in the course of selecting the top 50. If Raj or somone
of equal stature finds the outcome strange, then a message from him to
the current head of admissions might or might not cause re-examination.
∂07-Mar-79 1429 JMC
To: SWB
sure
∂06-Mar-79 2150 JMC
To: PAT
Please get Arthur Koestler "The Act of Creation" from the library.
∂06-Mar-79 1842 JMC foonly channel
To: LES
If you agree with Rubin, then you should talk to Poole about lead time,
and we should find out the state of the University's mind on helping
out the musicians.
∂06-Mar-79 1622 JMC
To: mcgoveran at SUMEX-AIM
I could see them Monday PM after Denny at SAIL.
∂06-Mar-79 1528 JMC
To: feigenbaum at SUMEX-AIM
I would like your opinion of the following draft letter. Who else should
get copies?
.require "let.pub" source
∂AIL Professor John Chowning∞
Dear John:
As Ed Feigenbaum has told you, the Computer Science Department
is in a desperate position with regard to space in Margaret Jacks
Hall and is eager to get the space that was allocated to computer
music in that building. This should have been apparent before,
but some were saying it would be bad and others were saying it
wouldn't, so I didn't take any position until the situation
clarified. The situation has clarified and it is indeed bad.
Specifically, as presently allocated, research associates will
have to share offices, and I think this is an untenable situation,
if we want to retain research associates for long terms, and this
is essential to the AI Lab. When Margaret Jacks Hall was first
planned, I was told that the AI Lab would have the same amount
of space there that we had in Powers, and no-one told me different.
Also no-one told me when computer music was allocated space in
the building that the statement about not squeezing the AI Lab
was no longer operative. Anyway the above is my excuse for not
biting the bullet previously.
Anyway I agree with Feigenbaum; we have got to have the
space. This leaves you with the problem of how to get your
computing done, and I am strongly motivated to help solve it.
It seems to me that while computer music has contributed financially
to the Lab, the contribution to hardware and software has not
been sufficient to amount to a whole computer facility. Nevertheless,
there is enough hardware that the AI Lab can release to make a
viable facility, even if it is insufficient to do everything you
now do.
What music now owns is the Foonly channel, a disk controller and
one disk. What I propose to do is to keep the Foonly channel for the KL
and trade for it the KA, the Suppes memory and new Ampex totalling 256K,
the Petit channel and the old terminal multiplexer. You can also have the
tape drives and any PDP-6 equipment you can use. Relatively soon you
could also get the Data Disc and maybe the III displays and the Analex
printer. Martin Frost says he could put together a system that would run
on this equipment, but I think he would need help. Qualitatively it would
be the system the Lab had before the KL came in 1975. We could help get
it going, but you would have to find and pay someone to maintain it.
Naturally if you could afford it, you might be well advised to junk the
old equipment and get a shiny new Decsystem 20, but, so far as I know, you
can't afford to do that.
This solution leaves you without a disk drive for user packs;
maybe you could use tape. It also leaves you temporarily without
terminals, and you would have to acquire some el cheapos.
Some temporary music use of the KL not involving office space
might continue. For example, Leland might continue to use the
III's and the XGP, since he doesn't use office space now.
Ethically, I find myself in the following position. I regret
letting you and Andy leave for Europe without having warned you that
this was coming up, even though I hadn't given it much thought. On
the other hand, the justification for substantial music use of the
ARPA purchased equipment was that the incremental contribution of the
musicians to the Lab was sufficient to justify their use of it. As
you know, not everyone at the Lab took that position. Now the space
situation is such that a divorce is in the best interests of the
computer science research, and this is a decisive consideration for
me.
There is the alternative of continuing the music use of
a combined system with a remote connection to where music might
be located. E.g. Serra house or a wing of Cedar might be obtained.
This is somewhat less attractive to me than a divorce, but it would
leave you with a smaller equipment maintenance problem. It involves
substantial technical problems in getting adequate bandwith between
two locations, and you would have to solve these problems.
I have seen the list of equipment that you consider necessary
for continuing with the same facilities after a divorce. The list
might be haggled with, but I think it is substantially correct. My
proposal, which essentially involves no-one buying any equipment
leaves you at a lower standard of living than you now enjoy. The
list was not accompanied by a written memo, so I don't know the context
in which it was presented. However, one could conjecture that you
thought that either the AI Lab or the University was obligated to
provide for that standard of living as a condition of divorce. I
don't know about the University, but I don't consider that the AI
Lab has any such obligation. Most likely if you want the money,
you will have to raise it. My one thought is that the University
might lend it to you on the basis of your Yamaha expectations.
It seems to me that Andy should come back and work on
reviving a viable system, because I don't think the Lab, which
has the problem of moving the KL will be able to do it entirely.
The simplest solution is to leave the KA here when the KL moves.
You would then have plenty of air conditioning and power and
wouldn't have to worry about reconnecting many cables.
If you want to get some help from the University, then you should
come back too - at least for a while. My impression is that
you have received enough good publicity recently, so that the
University is likely to be motivated to give some help.
Stanford is scheduled to get the building on April 20,
and there is some pressure to move the AI Lab promptly, because
the Department's new computer has a very shakey distant host
connection to the SUMEX Imp and will share the AI Lab TIP when
we move. Probably we won't be able to move till June in spite
of this pressure.
For these reasons, I think you should consider the matter
quite urgent.
.reg
cc: Lester Earnest, Leland Smith, Edward Feigenbaum, Gerald Lieberman
∂06-Mar-79 1513 JMC
To: PAT
please put an address on chowni.le1
∂06-Mar-79 1032 JMC
To: bmoore at SRI-KL
Thanks for the Sloman reference. It is clear that he is not a technically
educated logician - or if he is, he doesn't have his logic integrated
with his philosophy. I have gotten half way through his "Computer
Revolution in Philosophy", and I think it is very good. There is an
outside chance of getting him invited to take part in the study.
∂05-Mar-79 1440 JMC
To: PAT
speigh.le1
∂05-Mar-79 1324 JMC
To: PAT
partee.le4
∂05-Mar-79 1206 JMC
To: CLT
I spoke to him, and he said that there would be no problem.
∂04-Mar-79 0026 JMC
To: PAT
I now have the library copy of Aaron's Sloman's "The Computer Revolution
in Philosophy" - Harvester Press and have concluded that I need to
buy my own copy. Please order it for me.
∂03-Mar-79 2355 JMC
To: pratt at MIT-AI
I don't understand what the different effects of mailing to MODAL and
[pratt ... are intended to be.
∂03-Mar-79 1830 JMC
To: PAT
beeson.le1
∂02-Mar-79 1517 JMC agenda item
To: feigenbaum at SUMEX-AIM
Could I have a few minutes to discuss some issues of the departmental
display system? Questions of the number of terminals and the importance
of quality vs. number need at least a few minutes of faculty discussion
just to bring out opinions.
∂02-Mar-79 1511 JMC
To: pratt at MIT-AI
Our mailer refuses [pratt;modal mail]%ai.
∂02-Mar-79 1509 JMC
∂02-Mar-79 1505 PRATT at MIT-AI (Vaughan Pratt)
Date: 2 MAR 1979 1805-EST
From: PRATT at MIT-AI (Vaughan Pratt)
To: JMC at SU-AI
Ooops, I hadn't planned on anybody adopting a different strategy from
the one I had suggested, of sending mail to MODAL at MIT-AI. There was
a silent member of the group, namely a secretary file in which all
correspondence accumulates. I think if you put [PRATT;MODAL MAIL]@MIT-AI
as one of your addressees the right thing should happen (assuming your
mailer can parse it - it would depend on whether it treated [] differently
from alphamerics).
∂02-Mar-79 1451 JMC Sloman
To: bmoore at SRI-KL
Thanks for the Sloman article. It certainly clears up the misconceptions
of his earlier paper, and I think it is correct in what it covers, although
one could imagine a more detailed discussion of the uses of non-obvious
analogical representations. I am getting hold of his book today, and if
it lives up to the promise of the introduction, then I may have to campaign
for his inclusion in our study.
Can you give me the bibliographic information on the article you
sent me, i.e. where and when did it appear?
∂02-Mar-79 1443 JMC
To: ES
I may have it at home. Message me tonight if you still haven't found a copy.
∂02-Mar-79 1119 JMC
To: PAT
Wilkins letter to:
Chairmain
Dept. of Information Sciences
237 Applied Sciences Building
U.C.S.C.
Santa Cruz, California 95064
∂02-Mar-79 1032 JMC
To: REM
It is ok to browse through my unprotected directories. Not all files
with extension ESS are in ESS,JMC; some are in my quarterly files.
∂02-Mar-79 1029 JMC
To: "@MODAL.LST[LET,JMC]" at SU-AI
The answer about some modal logics being decidable isn't good enough. There
are many decidable theories in predicate calculus languages, and an AI system
that uses predicate calculus at all should be able to use them. The pure
theory of equality, the theory of list structures, the elementary theory
of real numbers, and monadic predicate calculus are useful examples.
The translation of a decidable modal theory into predicate calculus
would be decidable by the same decision procedure provided the translation
was accurate. Decidable theories are entirely usable as subtheories of
larger undecidable ones. My own opinion is that modal logic should normally
be translated into first order logic, i.e. predicate calculus + functions,
conditional expressions, etc. However, all the problems of modality that
are discussed in connection with modal logic will still exist. If one is
interested in studying modal problems in pure form, e.g. the axiomatics
of knowledge, modal logic is a good enough formalism; when one comes to
extensive applications, then a wider system is desirable.
Challenge to modal programming theorists: Express and prove the associativity
of append, and then come back for a harder problem.
∂01-Mar-79 1810 JMC
To: FB
Thanks for the reference to Lucid. It isn't the same as Elephant,
but it shares some properties. Do you have references to recent
developments in Lucid?
∂01-Mar-79 1526 JMC
To: HVA
I am puzzled about having money in the account. It seemed to me we
have been paying Mark Crispin from it, and we started soon enough to
have used it up. Do we get an automatic extension on the spending the
money? Are you taking into account that Mark is half time at CSD now?
∂01-Mar-79 1523 JMC
To: PAT
Scott is at Xerox PARC - not at Oxford.
∂01-Mar-79 1522 JMC
To: PAT
Albert Meyer is at the Labortory for Computer Science, M.I.T.
∂01-Mar-79 1005 JMC
To: mcgoveran at SUMEX-AIM
Yes, depending on the precise time.
∂01-Mar-79 0009 JMC
To: WVC
I killed your hour old dial job after paging. It cost line charges.
∂28-Feb-79 1727 JMC Dover
To: FEIGENBAUM at SUMEX-AIM, FB at SU-AI, LES at SU-AI,
admin.gorin at SU-SCORE
Well that's surely good news from Xerox. My own view agrees with that
of Peter McColough that the Dover's are of obvious use and that the
utility of the Altos depends on how research goes. In fact if I had
my druthers, I would try to trade in the Altos for the ability to put
a Dover on each floor. This would really put us into the modern world.
We shouldn't forget, however, that it won't be trivial to connect them
to the computers, but the offered Ethernet is certainly a good bet.
How are we coming with checking out the Grinnell possibility on a
display system?
∂28-Feb-79 1720 JMC
To: PAT
Please excuse previous incomplete messages. I was experimenting with
a feature of MAIL. Anyway, please finish the address, enclose and
enclose a biography with CHANDR.LE1.
∂28-Feb-79 1714 JMC
To: PAT
please decorate chandr.le1 with
∂28-Feb-79 1712 JMC
To: PAT
Please finish address and pub chandr.le1
Dr. Ashok Chandra↓Thomas J. Watson Research Center↓Yorktown Heights
↓New York∞
Dear Ashok:
As we discussed on the phone, I am accepting your offer to
come for about a month from July 15 to August 15, but I forget whether,
formally speaking, it is your move or mine.
.reg
∂28-Feb-79 1658 JMC
To: PAT
please pub and decorate gardne.le1 and include aiphil.pro.
∂28-Feb-79 1651 JMC
To: PAT
Both are for Idaho.
∂28-Feb-79 1650 JMC
To: bmoore at SRI-KL
Many thanks, but I shall be in Massachusetts on the 10th and 11th.
∂28-Feb-79 1506 JMC
To: DEW
No objections.
∂28-Feb-79 1505 JMC
To: DEK at SU-AI, feigenbaum at SUMEX-AIM
By the way, Ma Shih-wen of Peking U., whom I invited, writes that he
will leave Peking on March 9 and appear after some time in Washington.
∂28-Feb-79 1501 JMC
To: DEK
Dear Mr. Tang:
Don Knuth and I would like to know if you are interested in
an extended visit to the Computer Science Department at Stanford.
We both remember your part in the discussion following our
lectures at the Institute of Computing Machines in Peking, and
we think that such a visit might be beneficial both to you and
and our department.
If you are interested in pursuing this possibility further,
please write me including a resume and an indication of what scientific
problems you might like to work on at Stanford.
We don't have a definite plan in mind at present, but if
you want to come, we are eager to have you, and we will try to
arrange something.
.sgn
∂28-Feb-79 1451 JMC
To: DEK
I also had an excellent impression of Tang. I will draft a letter to
him asking him what are his qualifications and what would he like to
do. It seems to me that we should try to get him as a visitor, but it
may not be possible right away.
∂28-Feb-79 1449 JMC
To: levinthal at SUMEX-AIM, feigenbaum at SUMEX-AIM, LES at SU-AI,
TOB at SU-AI
I am hereby agreeing to Levinthal's current proposal, however I have
the following reservations.
1. The implied space commitment to house Sid Liebes is
for the period of this proposal. It has not yet been determined that
continuing it indefinitely is in the best interests of AI at Stanford.
2. The proposal contains no money to share the costs of maintaining or
improving the AI Lab computer facility as distinct from providing some
display hardware that may be used by other research. Making a preliminary
agreement that did not call for this was a serious oversight that
won't be repeated. Not even the cost of interfacing the new hardware
is covered.
In the long run, the Levinthal's NASA proposal and things
like it seem appropriate for the Lab, but they must pay their way
directly and not rely on their symbiotic stimulatory effect on other
AI Lab work.
∂28-Feb-79 0058 JMC photos
To: PAT
I chose a photo tuesday. Please call about them, pick up one and airmail
special it.
∂27-Feb-79 1801 JMC
To: mcgoveran at SUMEX-AIM
I don't have the time March 2 meeting begins.
∂27-Feb-79 1554 JMC
To: CET
There will not be a final in CS258.
∂27-Feb-79 1422 JMC
To: bmoore at SRI-KL
Yes, please send it. The library will send me a copy of his
Computer Revolution book.
∂27-Feb-79 1420 JMC
To: levinthal at SUMEX-AIM
We do have to meet, so please phone when convenient.
∂27-Feb-79 0903 JMC
To: RWG
The Times did. Why?
∂26-Feb-79 2039 JMC abolition of function variables
To: jonl at MIT-AI
CC: gls at MIT-AI
Besides killing a number of programs, this is an arbitrary change in
the language. Seems like bad news. - A recent victim.
∂26-Feb-79 1801 JMC
To: PAT
Patte: Please fill in the references in the last three pages of RENEW.PRO[W79.
∂26-Feb-79 1753 JMC my renewal proposal
To: LES, HVA
RENEW.PRO[W79,JMC] contains the main prose for the renewal of my NSF.
Naturally it is late and therefore urgent. The first paragraph contains
a few missing numbers which can be supplied from the previous proposal
and the award documents.
∂26-Feb-79 1642 JMC
To: bmoore at SRI-KL
suppes moravcsik perry
∂26-Feb-79 1636 JMC
To: bmoore at SRI-KL
Aaron Sloman
∂26-Feb-79 1627 JMC
To: bmoore at SRI-KL
I have been remiss in not sending out a mailing to the group with the
final composition and other information. Anyway here is the list. What
do you think about having a short meeting near the beginning of the study
and perhaps another near the end to which we would invite other people
interested in the field? Do you have ideas about who might come?
McDermott, Chomsky, Dreyfus, Doyle (M.I.T.), Suppes, Creary (Stanford),
Winograd (Stanford), Weyhrauch (Stanford) come immediately to mind, but
there must be many more.
Study group on artificial intelligence and philosophy, Center for Advanced
Study in the Behavioral Sciences - Academic year 1979-80.
∂26-Feb-79 1625 JMC
To: bmoore at SRI-KL
Names of participants
Daniel Dennett (Tufts University) Philosophy
John Haugeland (U. Pittsburgh) Philosophy
Patrick Hayes (Essex University) Computer Science
John McCarthy, (Stanford University) Computer Science
Marvin Minsky (M.I.T.) Computer Science and Electrical Engineering
Robert Moore (S.R.I.) Computer Science
Zenon W. Pylyshyn (U. Western Ontario) Psychology
∂26-Feb-79 1621 JMC
To: bmoore at SRI-KL
I would strongly recommend Mitch Marcus as a participant in any meetings
Other suggestions: David Marr, if we want to get into philosophy of perception
not really mainstream philosophy.) Julius Moravczick (probably misspelled)
also invite John Perry. He is much more personable than Julius and I think
systems. David Kaplan and possibly even Kieth Donelan (both at UCLA) also
for AI/Phil meeting. I would strongly recommend Barbara Grosz from the
Bill Woods and Brian Smith.
∂26-Feb-79 1449 JMC
To: ef at MIT-AI
SCHEINMAN, Victor
Unimation West
188 S. Whisman Rd.
Mountain View, Ca, 94041
(415) 965-0557
∂26-Feb-79 1448 JMC
To: TOB
CC: LES
Unless the source of money is one unusable for other purposes, I think
we have more urgent requirements - new xgp, new display system, modern
tapes.
∂26-Feb-79 1430 JMC
To: TOB
No protocols.
∂26-Feb-79 1429 JMC
To: CET
Thursday 2:30 ok for AI course meeting
∂26-Feb-79 1346 JMC
To: ef at MIT-AI
Vic Scheinman would know if anyone. His phone is (415)965-0557,and
his address is Unimation, 154 E. Dana Street, Mountain View, CA. He
built the hydraulic arm as well as our subsequent arms.
∂26-Feb-79 1113 JMC
To: kanerva at SUMEX-AIM
What you are doing may be reasonable for printing your thesis or
as a research project. However, it seems that the world is taking
what you call the compiling option. In particular, better and
better printers are becoming available, and better and better
document compilers are becoming available for using them.
∂26-Feb-79 1110 JMC
To: feigenbaum at SUMEX-AIM
I have been reading, and it hasn't led to a definite conclusion. I guess
the vote should be recorded as an abstention for now, but I will continue.
∂26-Feb-79 1108 JMC
To: feigenbaum at SUMEX-AIM
Yes, thank you very much. Sorry for being slow. - John
∂23-Feb-79 1829 JMC
To: bmoore at SRI-KL
Fine, how about 4pm or a bit later on Monday the 26th. I want to discuss
organizing the auxiliary meeting of the AI and philosophy working group.
∂23-Feb-79 1346 JMC
To: mcgoveran at SUMEX-AIM
No problem for me.
∂23-Feb-79 1343 JMC
To: PAT
Make a note to John Perry, chairman of the philosophy department.
You may be interested in cross-listing the following course in
philosophy. I discussed it with Pat about two years ago, but I didn't
get around to rewriting the catalog description.
226. Epistemological Problems of Artificial Intelligence - Formalisms
for representing what a general intelligent program must know about
the world including facts about causality, ability, knowledge and
action. Modes of rigorous and conjectural reasoning for computer
programs. Connections with philosophy, especially philosophical logic.
Related material in logic and elementary set theory.
∂23-Feb-79 1337 JMC
To: CET
Here are some course descriptions. Changing the number of 206 to 122
and introducing 222 was tentatively agreed by the Curriculum Committee
last year, but they or at least its chairman should review the matter.
CS226 will be considered for cross-listing in philosophy.
122. Recursive Programming and Proving - Recursive programming using
the LISP language and techniques for proving the corrrectness of
recursive programs. Computing with symbolic expressions
rather than numbers, e.g. algebraic expressions, logical expressions,
patterns, graphs, and computer programs.
Pattern matching and syntax directed computation. Preparation for
work in artificial intelligence is emphasized.
Prerequisite: 107 or equivalent ability to program.
222. Programming languages for artificial Intelligence - Advanced LISP,
Microplanner, QA4, Scheme. Production systems and other forms of
pattern-directed computation. Applications to problem solving and
language understanding.
Prerequisite: 122, 224 or consent of the instructor.
226. Epistemological Problems of Artificial Intelligence - Formalisms
for representing what a general intelligent program must know about
the world including facts about causality, ability, knowledge and
action. Modes of rigorous and conjectural reasoning for computer
programs. Connections with philosophy, especially philosophical logic.
Related material in logic and elementary set theory.
258. Mathematical Theory of Computation - Abstract syntax and formal
semantics of programming languages. Recursively defined and algolic
programs. Proving assertions about computer programs using formalisms of
Burstall, Cartwright, Floyd, Manna, McCarthy and Scott. The emphasis
is on functional programs rather than sequential. Use of
proof-checking programs. Prerequisite: 156 with 166 recommended.
∂22-Feb-79 2312 JMC
To: csd.mccarthy at SU-SCORE
At 23:11, foo again.
∂22-Feb-79 1727 JMC
To: PAT
Please get me, from the library if possible, "The Computer Revolution in
Philosophy" by Aaron Sloman, Humanities Press.
∂22-Feb-79 0950 JMC
To: PB
Thanks. I had heard something about it, and had it occurred to me
I might have tried to add something about the tendency of the government
to reiterate statistics known to be dubious, when they support positions
the government has decided to take.
∂20-Feb-79 2227 JMC first
To: PAT
Please write Takasu asking for the title, publisher, etc.
of the volume in which
FIRST will appear. The make FIRST into an AI memo referring to the
volume in which it will be published.
∂20-Feb-79 2214 JMC
To: EAF at SU-AI
Results so far are mixed. Need more time.
Richard Gabriel showed signs of having something substantial to say,
so I quizzed him and then asked him to write you.
∂20-Feb-79 2036 JMC
To: JMC
crawford
∂20-Feb-79 2029 JMC
To: EAF at SU-AI
CC: LES at SU-AI
What is Chowning's shopping list?
∂20-Feb-79 2028 JMC
To: nilsson at SRI-KL
1:30 feb 26 would be ok for talking with ibmers
∂20-Feb-79 2023 JMC
To: feigenbaum at SUMEX-AIM
CC: LES at SU-AI, kedes at SUMEX-AIM
OK for an account for Kedes to be used mainly for XGPing. I assume he can
do his PUB or TEX at SUMEX.
∂20-Feb-79 2018 JMC
To: TOB
OK on Soroka.
∂20-Feb-79 2017 JMC
To: ME
It would be nice if the command Xreply in E looked at the current page
to find whom the message was from and MAILed the attach buffer or the
page to that person. Fancier yet. XREPLY <rest of line> would find
the sender and transmit to him <rest of line>.
∂20-Feb-79 2014 JMC
To: bmoore at SRI-KL
I would like to discuss organizing the meeting associated with the workshop.
Could you phone me Thursday or Friday afternoon.
∂20-Feb-79 1949 JMC
To: csd.mccarthy at SU-SCORE
foo
∂20-Feb-79 1754 JMC
To: csd.mccarthy at SU-SCORE
foo
∂20-Feb-79 1530 JMC
To: CET
Yes, I will be happy to introduce Dana Scott.
∂15-Feb-79 1603 JMC
To: RPG
Yes, certainly sign it.
∂14-Feb-79 2127 JMC ka system
To: ME
CC: LES
The musicians may get the KA as a separate system with one disk and
controller, Petit channel, tapes and possibly data disk but not the
switch and 256k memory. Is there a system that can readily be
resuscitated to run on such a configuration.
∂14-Feb-79 2123 JMC
To: PAT
In case I forgot to say it, FIRST is FIRST[w79,jmc].
∂14-Feb-79 1007 JMC Lecture in class
To: scott at PARC-MAXC
Could you tell my class some thoughts on MTC on Feb 27 or March 1 Tues
or Thurs 11 to 12:15? They will be reasonably well prepared and will
know about continuous functions, etc. though probably not about models
for the lambda calculus. The class is small, but I think there will be
a larger attendance for your lecture.
∂14-Feb-79 1006 JMC
To: TOB
Tuesday is ok provided it doesn't run over. My class is at 11.
∂14-Feb-79 0103 JMC
To: PAT
Please don't leave letters of recommendation on your desk.
∂13-Feb-79 2349 JMC
To: bmoore at SRI-KL
Are you back from M.I.T., or when will you be?
∂13-Feb-79 2308 JMC
To: feigenbaum at SUMEX-AIM
I hope you will take such comfort as possible from the fact that
the major space allocation fuss will end in finite time.
∂13-Feb-79 2156 JMC JS % A chess computer
To: REM
It has 6 levels of play with the 6th taking up to 3 days. I have beaten
it at levels 1 2 and 3. It suffers badly from the horizon effect, i.e. it
gives up material in order to fruitlessly postpone some lossage. It was
made in Hong Kong and the brochure is in what appears to be a Chinese
approximation to British English. It costs $100.
∂13-Feb-79 1920 JMC Levinthal memo
To: feigenbaum at SUMEX-AIM, LES at SU-AI, TOB at SU-AI
It is hard for me to get used to evaluating all propositions in
terms of space, but I now don't remember who committed what in
connection with Sid Liebes's transfer to CSD. In so far as what
the AI Lab can do in the future is limited by office space, it
isn't clear to me that this transfer is in our interest. At least,
I vaguely remember considering it as some kind of courtesy
not competing with other goals. What has actually happened?
∂13-Feb-79 1419 JMC
To: hayden at USC-ISI
Your message was forwarded to me at SU-AI. You must have the wrong name
at MIT-AI, because that is my name there. - John McCarthy
∂13-Feb-79 1028 JMC
To: DEW
Tentatively it is all right, but I might just be travelling. If so,
I propose a rescheduling in good time.
∂13-Feb-79 1027 JMC
To: RWG
CC: MLB
After you have done all that, look at the sum for some particular n. We
see that the sum is infinitely often zero in accordance with your construction,
i.e. each sum of 2↑k terms is zero, but for other values of k, the sum can
fluctuate wildly. Is that summing to zero? Alas, convergence questions
cannot be avoided in matters like this.
∂12-Feb-79 1926 JMC
To: pylyshyn at CMU-10A
I have put your notice on the bulletin board and have mailed it to CASBS.
∂12-Feb-79 1621 JMC
To: PAT
WILKIN.RE1 is about Dave Wilkins
∂12-Feb-79 1553 JMC
To: DEW
What is the exact title of your thesis?
∂09-Feb-79 1551 JMC
To: PAT
Please send minima to Dan Bobrow at PARC.
∂09-Feb-79 1440 JMC
To: PAT
abstra[w79,jmc]
∂09-Feb-79 1016 JMC account on department machine
To: CET
I would like an account with name csd.mccarthy and password 848-0898.
∂09-Feb-79 0931 JMC
To: scott at SRI-KL
A manual is available from Greenblatt at mit-ai, and I believe that one
of the machines is accessible via the Arpanet through an M.I.T. local net.
∂08-Feb-79 1923 JMC
To: PAT
Remind me to settle travel and call Pyster when I come in.
∂05-Feb-79 2352 JMC
To: EAF at SU-AI
ok, see you at 1:30 or shortly thereafter
∂05-Feb-79 1812 JMC
To: feigenbaum at SUMEX-AIM, buchanan at SUMEX-AIM,
engelmore at SUMEX-AIM, reddy at CMU-10B
I certainly think we should try to talk him out of it. As for reciprocity,
having the conference in North America every other year more favorable treatment
than most international organizations give the United States. Better
than IFIP, for example. Of course, there should be a U.S. national conference
every other year when IJCAI doesn't meet.
∂05-Feb-79 1728 JMC
To: PAT
Please type AI memo number 1 into the computer. It's very short.
∂05-Feb-79 1358 JMC
To: PAT
.<<Please find early AI memos about Kalah and enclose copies with SHERSH.LE2>>
∂05-Feb-79 1232 JMC
To: TOB
Am I on this steering committee? If so Wed Feb 14 at 2:30 is ok.
∂05-Feb-79 1229 JMC
To: LGC
∂AIL Professor Barbara Partee↓Department of Linguistics
↓University of Massachusetts↓Amherst, MA 01003∞
∂04-Feb-79 2345 JMC your paper
To: LGC
I have read it, and I like your approach. Nevertheless, I doubt it will
be accepted by IJCAI, because it doesn't go far enough. If they do accept
it, there is time to improve it before final versions are requested. I
think it needs some axioms and a discussion of its model theoretic
semantics. You should send copies to Bob Moore if you haven't already
and to Barbara Partee, because she posed a similar problem about iterated
modalities to the one you solve. Of your proposed further research, I
think the first topic is the most likely to be productive, and the last
the least likely.
∂02-Feb-79 1409 JMC
To: PAT
Please send a copy of CONCEP to Jim Rhyne,K55/282, IBM Research,
5600 Cottle Road, San Jose CA 95193
∂01-Feb-79 1859 JMC
To: DCO
Fine. Let's do it.
∂01-Feb-79 1606 JMC
To: PAT
CC: AFB
Please get grade change card for Anne Beetem in CS390 from N to P.
∂01-Feb-79 1603 JMC
To: AFB
OK, that will do for a P. If you want a letter grade, we must talk.
∂31-Jan-79 1802 JMC
To: SAL
I will come to lunch Thursday if it isn't too late.
∂28-Jan-79 1239 JMC display system for CSD
To: feigenbaum at SUMEX-AIM, FB at SU-AI, REG at SU-AI, LES at SU-AI
CC: JBR at SU-AI
Both Jeff Rubin and Richard Greenblatt recommend
that we negotiate with Grinnell about a display
system (local company). Livermore bought a small
system for S-1 and found them flexible as to specifications.
A large order might find them even more flexible. It
would place less strain on our engineeering capability
if we could meet our needs commercially.
∂28-Jan-79 1230 JMC PUB problem
To: PAT
The Kyoto paper I am currently working on has a PUB problem.
The source is FIRST[W79,JMC], and the problem is that
on the middle of page 13 it starts indenting equation
numbers and axiom labels. The paper is unfinished as regards
content, but it would help if you could solve the PUB problem
(and others that you may notice) before I come back. There is
a PUBbed version on FIRST.XGP.
∂28-Jan-79 1226 JMC various
To: CLT
I am at the St. Anthony Hotel in San Antonio.
Please phone me Monday night if convenient.
Could you extract FOL manuals and give them
to the class. There seem to be six of them.
You might also give them copies of your LISP
axioms, although my present intention is to
use in class a set of axioms based on those
in FIRST[W79,JMC], but these may have bugs,
and I have written them in conventional logical
notation rather than FOL notation.
I may be out early Monday night, but you will
eventually get me if you try late enough.
∂27-Jan-79 1858 JMC IJCAI papers on epistemology
To: buchanan at SUMEX-AIM
CC: LGC at SU-AI, bmoore at SRI-KL
At the last IJCAI, it seemed to me that papers on epistemological
problems got rather short shrift from the referees, who
treated them like proposals for experimental work, and downgraded
them in comparison with papers that reported experimental work.
I hope that such papers can be treated better this time - maybe by
concentrating them in the hands of a few referees. I might even
be able to help a bit myself.
∂27-Jan-79 1849 JMC your paper
To: LGC
I have looked at IJC.TXT, and I think it's pretty good. However, I
think there is a strong probability it won't be accepted, because
it doesn't report experimental work, and the IJCAI referees have
a strong experimental bias. Maybe I should talk to Bruce Buchanan
about this, or maybe you could. The issue of the acceptability of
theoretical papers (or more precisely papers about epistemological
problems of AI) should be kept separate from that of the merits of
a particular paper.
∂26-Jan-79 1237 JMC
To: HVA
Happy birthday!!
∂25-Jan-79 2042 JMC
To: venkat at RUTGERS
An old version of the paper is FIRST.NEW[W77,JMC] and a version currently
being updated is FIRST[W79,JMC]. You are welcome to FTP either version,
but both are PUB source files and a bit strange for direct reading. I could
U.S. mail you a copy of the old version now or a new one in a week or so.
∂25-Jan-79 1828 JMC
To: DEW
McKay, John Montreal, (514)879-8039
∂25-Jan-79 1447 JMC
To: mcgoveran at SUMEX-AIM
Feb 5 ok with me. By the way Baskett is changed to FB from FXB.
∂24-Jan-79 1550 JMC
To: pylyshyn at CMU-10A
CC: minsky at MIT-AI, bmoore at SRI-KL
I have been remiss in not sending out a mailing to the group with the
final composition and other information. Anyway here is the list. What
do you think about having a short meeting near the beginning of the study
and perhaps another near the end to which we would invite other people
interested in the field? Do you have ideas about who might come?
McDermott, Chomsky, Dreyfus, Doyle (M.I.T.), Suppes, Creary (Stanford),
Winograd (Stanford), Weyhrauch (Stanford) come immediately to mind, but
there must be many more.
Study group on artificial intelligence and philosophy, Center for Advanced
Study in the Behavioral Sciences - Academic year 1979-80.
Names of participants
Daniel Dennett (Tufts University) Philosophy
John Haugeland (U. Pittsburgh) Philosophy
Patrick Hayes (Essex University) Computer Science
John McCarthy, (Stanford University) Computer Science
Marvin Minsky (M.I.T.) Computer Science and Electrical Engineering
Robert Moore (S.R.I.) Computer Science
Zenon W. Pylyshyn (U. Western Ontario) Psychology
∂24-Jan-79 1002 JMC Filman letters
To: PAT
Please give Filman a list of the names of the people to whom letters have
been sent, so we can be sure to send them to all the people he wants them
sent to.
∂24-Jan-79 0227 JMC
To: jrobinson at SRI-KL
I think I'll pass. I balk at the protectiveness (self or otherwise).
Besides my present schedule of running when I get up takes up much
less time than a class.
∂23-Jan-79 1702 JMC
To: PAT
Ask me tomorrow about proofs.
∂23-Jan-79 1700 JMC
To: bobrow at PARC-MAXC
I will send you my paper in its present state, but I would like to revise
it.
∂23-Jan-79 1659 JMC
To: PAT
Yes the 28th, and I want to stay for whole conference.
∂22-Jan-79 1550 JMC FILMAN.RE1
To: PAT
Please decorate this file and prepare letters. In my message file, there
is a messsage from him with a list of about ten names to which letters should go.
Besides this there are more recent messages with additional names.
∂20-Jan-79 1948 JMC film of accident
To: quam at SRI-KL
I talked to Dyanna Taylor. You have the original negatives.
However, there are also 16mm color films. She is willing to make a work
print available, but will have to be talked into making the negative
available, because she is afraid it might be damaged. I think she would
let us get the color frames from the negative into the computer, if we
could assure her that her negative, which is using to make a film for sale
to TV, wouldn't be damaged.
As I told you, Janusz Onyszkiewicz, who was Alison's husband,
hopes to go with the French Annapurna expedition this Spring, so we should
try to do something in time to help him. I was in L.A. on Friday, and III
will also help if they can.
Dyanna Taylor's telephone number (both daytime and evening) is
524-3880, and I told her you would phone to ask technical questions.
∂19-Jan-79 2004 JMC
To: DEW
I guess I must have gotten confused, but most recently I have been urging
Juan to do it, so unless you have a strong desire to do so, in which case
phone me or see me at the Lab, we'll leave it at Juan.
∂19-Jan-79 2001 JMC
To: CPP
CC: REP
You are correct about that error, and there will be an adjustment.
∂19-Jan-79 1959 JMC
To: feldman at SUMEX-AIM
CC: LES at SU-AI
The part about our having told Chern we would collaborate on SAIL-Plits
(what is it?) is a decisive consideration, as is the small number of
students. If there is some part of the work not involving it, they
should do the work at LOTS. I don't even teach my Lisp course at SAIL.
So, go ahead but take it as easy if possible.
∂18-Jan-79 1500 JMC
To: REF
Thanks to the magic of electronic computers, ∞ letters is no more trouble
than one. Ask me Monday if I've done it yet.
∂17-Jan-79 2208 JMC
To: LES
Would you take charge of getting Vadic to fix their modem or something?
∂17-Jan-79 1102 JMC
To: PAT
Would you retrieve GARDNE.XGP from the XGP when it comes off.
∂16-Jan-79 1409 JMC
To: CLT
Now I know. The file is protected, and I'll fix that.
∂16-Jan-79 1407 JMC mtcpub.pub[let,jmc]
To: CLT
I can't imagine what the difficulty was or is. Xpo from your message
or from GOEDEL[W79,JMC] in E gets to the file and DI command finds it.
I'll be in shortly.
∂14-Jan-79 2130 JMC
To: PAT
If I didn't send it, please send Corky a copy of my paper on the
Takeuchi function.
∂13-Jan-79 1547 JMC
To: bmoore at SRI-KL
In reading your thesis, I am constantly having to search for where
a notation was defined. Currently I am searching for H. I suggest
you prepare an index of symbols and notations and MAIL me a copy. It
should be included in the thesis along with the list of axioms.
It could also take the form of repeating the definitions of the
notations if this can be done concisely enough.
∂12-Jan-79 1636 JMC
To: feigenbaum at SUMEX-AIM
CC: LES at SU-AI, PAT at SU-AI, LCS at SU-AI
Dear Ed:
To commit to writing what I said about the musicians, I agree
that we should try to get their office space moved to Cedar or Serra.
However, it isn't true that we don't owe them anything. While they
have mortgaged back the KA-10, they currently own one of our two
disk controllers at $34K, one disk drive at $14K and the Foonly disk
channel at $20K. While one can imagine buying them out, justice and
our commitments require a solution in which they can continue to
make music. Their van mounted 2020, which I haven't heard of recently,
is not a solution, because it will be dedicated to a particular
application, the 100th anniversary of the Boston Symphony.
I think the best solution would be office space for them
in some compact space on campus with wired connection to the computer
together with space for the Sampson box and such like equipment in
the computer room.
John
∂12-Jan-79 1538 JMC
To: LGC
I am making good progress in reading it, and I expect to continue
this weekend. Would next week, during the week, be ok?
∂12-Jan-79 1302 JMC
To: REF
I received notice today that the chairman is Fouad Tobagi.
∂12-Jan-79 1133 JMC
To: bmoore at SRI-KL
In any case by then. If I can get done earlier, then we should meet
earlier, so you could make larger changes if there were some.
∂12-Jan-79 1131 JMC
To: AFB
In that case, please come and see me and tell me what you did for 390.
∂11-Jan-79 2239 JMC call-by-name counters
To: BCM
MOSZ.LSP[W79,JMC] contains several Lisp functions for counting, mainly
on the Takeuchi function including one constructed in the manner you
propose. Yours is shorter than mine and is perhaps more directly
constructable from the original function. Also it is more directly
"homomorphic" to the Takeuchi function. It has occurred to me to try
to extend the concept of homomorphism to include cases where the
original function occurs as part of the control structure of the
homomorph. It isn't clear yet whether this is an interesting
concept, but it might get complete recursive functions by a more
general method than the ones Corky is using.
∂11-Jan-79 1352 JMC
To: jrobinson at SRI-KL
Chomsky's lecture today at 4:15 is in Kresge Auditorium.
∂11-Jan-79 0028 JMC
To: SMG
Yes. It is the Representation of Recursive Programs in First Order Logic
that I gave at the Kyoto conference. There are copies in my office so ask
me in daytime. There will be quite a few changes in the new version including
a much better set of axioms not confused by a half-assed use of sorts.
∂10-Jan-79 1547 JMC
To: BCM
As sometimes happens, giving you the problem made me think about it,
and I solved it also last night. Moreover, Corky solved it by a
different method from mine which seems to be the same as yours.
I will illustrate by the Takeuchi function which is
tak(x,y,z) ← if x ≤ y then y else tak(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y))
The function terminates for all integer x, y, z. (It probably also
terminates for real x, y, and z, but I haven't been able to prove it).
However, the number of recursions differs depending on whether the
computation is by name or by value.
My solution works as follows:
In the call-by-value case we have
ctak(x,y,z) ← if x ≤ y then 0 else 2 + ctak(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y))
+ctak(x-1,y,z) + ctak(y-1,z,x) + ctak(z-1,x,y).
This is along the lines discussed a year ago. In the call-by-name case, we have
ctak(x,y,z) ← if x ≤ y then 0 else 1 + ctak(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y))
+ (if use1(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y)) then ctak(x-1,y,z) else 0)
+ (if use2(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y)) then ctak(y-1,z,x) else 0)
+ (if use3(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y)) then ctak(z-1,x,y) else 0)
tak(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y))
tak(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y))
where
use1(x,y,z) ← T
use2(x,y,z) ← T
use3(x,y,z) ← if x ≤ y then F else
(use1(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y)) ∧ use3(x-1,y,z))
∨ (use2(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y)) ∧ use2(y-1,z,x))
∨ (use3(tak(x-1,y,z),tak(y-1,z,x),tak(z-1,x,y)) ∧ use1(z-1,x,y))
The "usen" predicate tells whether the nth argument of the function is actually
evaluated. In the above case, use3 simplifies to
use3(x,y,z) ← if x ≤ y then F else T,
because the second disjunct is the conjuction of two use2's which are identically
true. The evaluation of the call-by-name ctak must be call-by-name in cases
when call-by-value fails for the original function. For example, we have
cadiou(x,y) ← if x = 0 then 0 else cadiou(x-1,cadiou(x+1,y))
and
ccadiou(x,y) ← if x = 0 then 0 else 1 + ccadiou(x-1,cadiou(x+1,y))
+ (if use2(x-1,cadiou(x+1,y)) then 1 + ccadiou(x+1,y) else 0)
where
use2(x,y) ← if x = 0 then F else use2(x-1,cadiou(x+1,y)) ∧ use2(x+1,y).
use2(x,y) always evaluates to F, but call-by-name must be used to insure
termination.
Corky's solution to the same problem, which I think is the
same as yours, would write
ccadiou(x,y) = ccadiou1(x,0,y,0)
where
ccadiou1(x,y) ← if x = 0 then c else ccadiou(x-1,c,cadiou(x+1,y),ccadiou(x+1,c,y,d))
Well I'm not sure this is quite correct, although it gives the right answer,
and I'm also not sure what you intend with two or more variables.
Could you work out yours for the Takeuchi and Cadiou functions?
∂10-Jan-79 1430 JMC
To: TOB
CC: JB
I notice that I am listed as a speaker in your session at the computer forum.
I will be in Texas, and the speaker will be Juan Bulnes. I assume he
is in touch with you, but I want to be sure.
∂10-Jan-79 1042 JMC
To: PAT
If Volume 1 of S-1 software is too big to move, paint it.
∂09-Jan-79 2326 JMC 91 function
To: BCM
I now am convinced the derived function for number of recursions of the
91-function is correct. Incidentally, it gives a new proof, which I think
is shorter than any other I have seen of the termination and correctness of
the 91-function. Namely the function cff(x) = if x>100 then 0 else 2(101-x)
is such that when the 91-function recurs, cff(x+11) < cff(x) and
cff(if x>100 then x-10 else 91) < x. It would be interesting to see if
the such a mechanical process of getting the recursion equation for
a suitable derived function and guessing its value would replace the
cleverness in proving the termination of the Boyer-Moore reduction of
conditional expressions to normal form by turning a crank.
The problem of getting a derived function that gets the number of recursions
in a call-by-name recursion remains open.
∂09-Jan-79 2308 JMC
To: LES
Can you look at the Hedrick criticism of Dialnet?
∂09-Jan-79 1550 JMC
To: RPG
Is the matcher suitable for general use as part of a lisp program,
and is it worth studying from the point of view of including material
on matching in CS206 and the book?
∂09-Jan-79 1545 JMC
To: RPG
What is the basic reference on MATCH?
∂08-Jan-79 1721 JMC
To: DCL
.require "let.pub" source
∂AIL Dr. William Carlson↓Office of Information Processing Techniques
Defense Advanced Research Agency↓1400 Wilson Blvd.↓Arlington, VA 20009∞
Dear Bill:
As I said, I thought your summary of the sense of the meeting
was excellent. The meeting itself was useful in making known what
everyone was up to. I think people's plans will be affected in the
long run, and a certain amount of duplicated effort will be avoided.
Of course, it was a meeting of the tops of a lot of icebergs, and
the bottom parts will be affected more slowly.
As I mentioned, I would much like the opportunity to try out
one or more LISP machines at Stanford. It seems to me that it is now
demonstrated that the machine can be an effective tool for developing
and running large symbolic computation programs and that it willl
provide an effective environment for a user who has sufficient access
to one. It needs to be seen how many people can effectively share one,
since they are still quite expensive.
Here are some comments on the contents of the meeting:
1. I do not share the view that environment is everything,
and language is unimportant. Xerox PARC has created an extremely
pleasant programming environment but at great expense per user.
This environment has influenced many attempts at emulation, prominently
at M.I.T. It remains to be seen how important quality of facilities
is compared to the number of reasonable terminals and amount of
computer time. An experiment to determine the effect of super facilities
on rate of program development would be worthwhile.
Moreover, I think that LISP as a language for computing with
symbolic information has many advantages that have not been realizedd
in any other languages so far. An attempt at formulating these was
included in my %2History of LISP%1 paper which I enclose.
2. I think it would be worthwhile to try to bring up ADA in
a time-sharing environment and provide it with analogous facilities
to LISP. These facilities should include
a. A standard representation for the ADA language in
ADA data structures. Moreover, lists of ADA expressions must also
be representable and computable with, so that the language
essentially has to contain LISP. Primitives for composing and decomposing ADA
expressions must also be available and standard.
b. It must be possible to run single ADA statements
in analogy with the ability to evaluate single just defined LISP
functions. Since ADA has declarations and block structure (I presume)
while LISP is single level, there will be additional problems in
making a convenient system. This problem is one reason why the
outcome of a "friendly ADA environment" experiment cannot be predicted
a priori.
c. The full ability to transfer AI applications
will require the ability for ADA programs to create new ADA code,
attach it to whatever is used for property lists and to run it
from inside the program. However, this facility of LISP is not used
in all AI programs.
d. These facilities need to be a separable ADA package,
because their associated large run-time routines will not be wanted in
ordinary uses of ADA that often have to be space efficient.
Thanks for inviting me to the meeting.
.reg
∂06-Jan-79 2151 JMC book
To: PAT
Please order Computability and Logic by Boolos and Jeffrey, Cambridge University
Press, charging an unrestricted account.
∂06-Jan-79 1534 JMC
To: ME
NS hasn't picked up any digests since the 2nd.
∂06-Jan-79 0017 JMC induction to omega↑omega
To: RSC, RWW
I now think my proof is not first order, since it uses a sentence
with an implicit universal set or predicate quantifier as an inductin
hypothesis. Moreover, I now doubt that the induction to omega↑omega
as a schema can be proved from ordinary induction. Maybe one can
only prove instances of it for particular predicates.
∂05-Jan-79 1819 JMC
To: stever at MIT-MC
Ask MRC@SU-AI.
∂03-Jan-79 1403 JMC
To: jrobinson at SRI-KL
Carolyn forgot that she would be in Idaho, but anyway 5:30 at your place
sounds fine.
∂03-Jan-79 1335 JMC Chomsky lecture
To: jrobinson at SRI-KL
Sounds like fun to go together. Carolyn Talcott will come with us. Do
you want to eat together first?
∂03-Jan-79 0215 JMC
To: JB
Fine. You're on. Please talk to whoever is in charge.
∂02-Jan-79 1308 JMC
To: PAT
Also takasu.le6
∂02-Jan-79 1304 JMC
To: PAT
Please pub and print kyoto[w79,jmc].
∂02-Jan-79 1236 JMC
To: PAT
please decorate shersh.le1. His letter is on your desk.
∂02-Jan-79 1046 JMC
To: gaschnig at SRI-KL
Michie 31 Royal Terrace 031
42 Great Cumberland Place, London W.1., mother
9 Steele's Road, London NW3, 01 722-3193, Ann
15 Canal Street, Jericho, Oxford, Oxford 56273
851-3746,
Machine Intelligence Research Unit
Hope Park Square
Edinburgh EH89NW, 031 667-1001 x6693, Penny Montgomery