perm filename SAVED.MSG[ESS,JMC]1 blob
sn#134756 filedate 1974-12-12 generic text, type T, neo UTF8
∂11-DEC-74 2153 1,PLW @ AMET
Prof McCarthy:
I am a student in your cs206 class. I would like to turn in my final
project slightly late, i.e. it is done now but was not at 1:00
this afternoon when they were due. If for some reason you feel
that you cannot accept the project now, could you please give me an
incomplete in the course and I can turn it in at the beginning
of next quarter (or earlier.) I'm willing to drive up to the lab to
deliver it to you or whatever if nescessary since incompletes
are rather distastefull.
I can be reached at 329-1449 or as PLW on this system.
Thank you very much. -- Phil Wadler
∂11-DEC-74 1223 1,MG
TODAY'S MEETING IS POSPONED
AT THE REQUEST OF PEOPLE AT PARC-MAXC TODAYS MEETING ON REASONING ABOUT
PROGRAMS HAS BEEN POSPONED. THEY SUGGEST SAME TIME NEXT WEEK - IS
THIS OK?
SORRY
MIKE
∂11-DEC-74 0923 1,MG
Subject: Discussions on reasoning about programs
The first meeting is TODAY in the seminar room of the Stanford A.I. Lab
at 2.30pm.
∂11-DEC-74 0907 network site ISI
Date: 11 DEC 1974 0906-PST
From: FIELDS at USC-ISI
Subject: 16k chip
To: jmc at SU-AI
all we got from noyes was the typical graph showing a 16k chip
in n years ... n=1 to 3 ... low reliability...
sorry, but he wasn't giving anything away.
best
craig
-------
∂10-DEC-74 1004 E,ALS
Your bug has been fixed. It occurred after a CET entry because of poor house-
keeping. This label should go away after a file has been written-out and the
<control>? command does write the old CET file out even if it is empty.
aAnyway it is fixed. Thanks. ALS
∂10-DEC-74 0956 network site SRI
Date: 10 DEC 1974 0956-PST
From: COLES at SRI-AI
Subject: STATEMENT FOR SIGART
To: JMC at SU-AI
ALTHOUGH I NOTE THAT YOU ARE NO LONGER SOLICITING ENDORSEMENTS, I
WANTED YOU TO KNOW THAT I FULLY SUPPORT YOUR POSITION REGARDING
IJCAI-4.
BEST REGARDS, STEVE COLES
-------
∂09-DEC-74 1725 LCF,MAL @ UKT
this is just to thank you for effort in reading (and signing) my thesis.
∂09-DEC-74 1452 1,TW
ACHIEVEMENTS:
REPRESENTATION LANGUAGE
Theoretical basis for the representation. Several papers
published. (1974)
Knowledge base for a simple scheduling system (1974)
Design of knowledge formalism for the Diplomacy world (1974)
DISCOURSE AND SYNTACTIC STRUCTURE
Several transcripts have been collected and studied (1973-4)
A system has been written for collecting and annotating
transcripts in a form suitable for machine analysis (1974)
Development of syntactic theory (1971-74) (see references)
A small set of dialogs have been analyzed in detail, providing
a basic set of discourse phenomena to be attacked. (1974)
MILESTONES:
Initial specifications for the knowledge representation
language (Jan. 1975)
Operating prototype interpreter and deductive system (May 1975)
Working system for simple scheduling tasks (July 1975)
(using non-English input)
Working system for underlying reasoning in Diplomacy (Feb. 1976)
Knowledge base for reasoning about programs (see next section) (June 1976)
Limited story-understander (using pre-parsed input) (June 1976)
Redesign of implementation (1976-77 completed by Jan 1977)
Polished implementation complete (Summer 1977).
DISCOURSE AND SYNTACTIC STRUCTURE
Development of a base of discourse knowledge in the
representation language (Summer 1975)
Separate programs which use this knowledge to analyze
different aspects of a particular set of discourses
based on the game of Diplomacy (Feb 1976)
Limited parsing system interacting with semantic base (December 1975)
Comprehensive grammar of English written in Knowledge Representation
Language (July 1976)
Large-scale parser ready to be interfaced to other projects (Sept. 1976)
Integrated program for analyzing and discussing dialogs (Dec. 1976)
PROGRAMMING LANGUAGES AND SYSTEMS
Preliminary programming assistant system (July 1975)
Incorporation of extensive knowledge base on programming (July 1976)
Limited automatic programming system (specialized domain) (July 1976)
Incorporation of programming knowledge base into the
interpreter design and implementation (July 1977)
∂09-DEC-74 1425 1,TW
NATURAL LANGUAGE GROUP
The main research goal of this group is to develop formalisms capable
of representing a wide range of knowledge within an artificial
intelligence system. The work will concentrate on building systems
which understand some body of material based on natural language,
focussing on general theoretical issues of artificial intelligence --
how knowledge can be represented and manipulated. We feel that the
natural language tasks give us an especially valuable perspective on
understanding how people represent and manipulate knowledge, and that
the techniques and formalisms we develop will have applicability in a
wide range of "understanding" processes which do not explicitly
involve language.
Our basic methodological attitude is that it is impossible to attack
the significant issues of representation by viewing knowledge as an
abstract calculus of objects and relations. It is necessary to pick
a specific domain, involving a significant amount of knowledge, and
to develop the representation by actually trying to use it in that
domain. In line with this view, the project will consist of building
several large programs, each trying to apply the representational
ideas to a specific knowledge domain. These include directly
linguistic areas, such as the structure of discourse and the
syntactic structure of natural languages, as well as more
semantically motivated domains involving simple stories, the
knowledge needed by an office assistant
We believe that the formalisms now used for representation in AI
programs (such as predicate calculus or PLANNER-like languages) are
not sufficiently rich in allowing knowledge to be expressed in the
variety of conceptual forms which a person naturally uses. Such
systems impose a kind of one-sided uniformity, and things which we
would naturally represent quite differently (like processes and
"simple facts") must be coded into a single formalism designed to
handle everything in a uniform way. We are developing a Knowledge
Representation Language which makes it much easier to create a large
knowledge-base, and to combine procedurally oriented knowledge and
deductive methods into a single integrated system. for things like
scheduling, and the base of programming knowledge for a "programmer's
assistant".
We anticipate a very active feedback loop in which a person trying to
use the Knowledge Representation Language for some specific task will
find ways in which it is inadequate, leading to changes in the
language which in turn can be tried out in that program and in the
others being developed.
The major issues which we feel must be tackled include:
Designing "flexible programs".
We need a programming formalism which is much less rigid than
current programming languages (indluding the new AI
languages). It should be able to carry out a process in a
"data-directed" way, doing things in an order best suited to
the particular task at hand, rather than a rigid order
predetermined by the programmer.
Representing and using partial knowledge.
Closely related to flexible programs are deductive methods
which are able to work in a context of partially specified and
partially reliable knowledge. This must include ways of
hypothesizing whole scenarios from suggestive but incomplete
evidence, and being able to recover from mistaken deductions.
Reasoning by analogy
One of the most powerful of human reasoning methods involves
recognizing a fundamental similarity between two different
situations, and using facts about one of the situations to
suggest facts about the other. Most current deductive systems
do not attempt this at all, or only the most simplistic way.
Integrating procedural and declarative knowledge.
Much of the difficulty in representing knowledge in a
computer program is in establishing the connections between
the "what" and the "how". Some knowledge seems best
expressible in the form of facts which do not in themselves
imply a thinking process, but which can be used by many
different processes. Other knowledge seems better represented
in a procedural form, explicitly determining what course the
computation should follow. Current formalisms fall short in
integrating these two facets of knowledge.
Constructing large knowledge bases
Many of the ideas being developed are aimed at making the
form in which the knowledge of a domain is represented in a
computer as close as possible to the way a person would
naturally think of it. This is necessary if we expect people
to be able to comfortably enter large bodies of knowledge into
computer systems. The emphasis on natural language is not
aimed at allowing all of the knowledge to be entered in
ordinary English, but at using our insights into natural
language as a way of getting at natural representations.
∂09-DEC-74 1214 1,MG
Subject: Discussions on reasoning about programs
Most of the people who were contacted about the discussions
said they thought it was a good idea. Also most said that
wed. 11dec.(ie next wednesday) at 2.30 in the seminar room of the A.I.lab
was a convenient time for a preliminary meeting. Thus there will definitely
be a meeting then and there.
The plan is (that is unless anyone suggests something better):
(1) For each person to say briefly what they are up to
(2) To decide on a format (if any) for future meetings (if any).
(Maybe each time someone could volunteer to organise the
next meeting and then they could decide what to do).
(3)To fix up where and when to have the next meeting. (Pehaps we could
meet at different places in rotation (eg PARC-MAXC, SRI, SU-AI)).
∂09-DEC-74 1201 1,QIB
Betty Scott called and said that the Faculty Mtg. will still be held Tuesday
at 2:15 as Prof. Floyd wants to have it before Xmas week and all the other
members are able to attend. If you like, Betty suggested you call Prof.
Floyd and talk to him about rescheduling it. Queenie
∂09-DEC-74 0906 VV,BGB
- I'VE PLACE A ONE PAGE SUMMARY ON YOUR CONSOLE...
∂09-DEC-74 0732 VV,BGB
THE SIERRA CAMERA DROPPED DEAD THIS MORNING - PLEASE PUSH TAG TO FIX IT.
∂08-DEC-74 2239 1,PDQ
Do you desire my presense for the Licklider briefing?
I had planned a presentation, but I can give you some
notes.
∂08-DEC-74 1920 2,DCL
the page i gavve you is not detailed enough?
∂07-DEC-74 0427 S,LES
The memory proposal is in CORE.PUB[D,LES]. At Al Blue's request,
I attempted to justify the memories in terms of specific research
program requirements, using MTC mainly. Feel free to fiddle it,
since it should appear over your signature.
∂5-DEC-74 0758 network site ISI
Date: 5 DEC 1974 0757-PST
From: LICKLIDER at USC-ISI
Subject: 48-Bit Maching
To: McCarthy at ISI, JMC at SU-AI
cc: Licklider
Thanks for your note about the 48-Bit Machine. I was intending
to argue for it but not fund anything. Let us talk about it.
Regards
Lick
JCRL/hcb
-------
∂5-DEC-74 0124 network site ISI
Date: 5 DEC 1974 0123-PST
From: LICKLIDER at USC-ISI
Subject: My Visits to You
To: Nilsson at SRI-AI, McCarthy, JMC at SU-AI, Green,
To: Feigenbaum, Pirtle at I4-TENEX, Ron at I4-TENEX,
To: Elkind at PARC-MAXC, Taylor at PARC-MAXC
cc: Licklider
The schedule has shaped up as follows:
SRI-AI Dec. 10, 9:00-12:45
SU-AI
JMC Not arranged
Cerf Dec. 10, 5:00-6:00
EF Dec. 10, 6:00-unspecified
Ames Dec. 11, 9:00-12:00
Xerox Dec. 11, 1:00-4:00
[Wednesday evening, Dec. 11, I'll be leaving for LA.]
If anything about the schedule is inconvenient
for any of you, please let me know asap.
John McCarthy, please let me know whether it will be possible
for me to see you and your colleagues in the slot, 1:30-5:00
Tuesday afternoon, Dec. 10. If John is out of town, Cordell, would you
indicate whether or not a visit to SU-AI can be arranged for that time?
I'll send you a copy of the message I sent John. Please check the
matter with Les Ernest, whose network address I do not have here now.
Thank you all very much for helping me set up the visits. I
look forward to seeing you.
Regards
Lick
-------
∂04-DEC-74 1412 1,MG
*********DISCUSSIONS ON REASONING ABOUT PROGRAMS*********
Would you be interested in getting together with other people
in the Palo Alto area to discuss the problems involved in reasoning about
programs?
What might be nice is if those interested could meet every now and then
(say once every two weeks) to talk about current plans, difficulties etc.
in an informal and sympathetic atmosphere.
To get things started I've booked the seminar room in the Stanford A.I. lab
on wednesday 11 dec. at 2.30 pm. I thought that at the first meeting
everyone could describe briefly what they were currently doing and hoping to do.
Then we could decide whether it was worthwhile to have further meetings and if
so fix up details of where, when etc.
Wednesday 11 dec. at 2.30 was chosen at random and can be changed.
contact MG%SU-AI.
Mike Gordon
∂02-DEC-74 2138 100,100 AT TTY15 2138: 1,ELF @ SAIL
READ GRIPES (PASSWORD TO MY AREA IS PHC
∂2-DEC-74 1304 network site CCA
Date: 2 DEC 1974 1546-EST
From: TOM at CCA
Subject: ARPANET BOOK AND DATA BASES
To: IMPORTANT-PEOPLE:
cc: TOM, JMH, DALE
YOU MAY BE AWARE OF THE FACT THAT ARPA/IPT HAS AN EFFORT
UNDERWAY TO WRITE A BOOK DEALING WITH THE ARPANET. IN SUPORT
OF THIS EFFORT I HAVE AGREED TO TRY TO FIND OUT WHAT DATA
BASES ARE CURRENTLY STORED ON THE NET.
TO ACCOMPLISH THIS TASK I NEED HELP. IF YOU'RE AWARE OF ANY
DATA BASES ON YOUR MACHINE OR ON OTHER ARPANET MACHINES,
PLEASE LET ME KNOW. ANY INFORMATION WILL BE HELPFUL.
THANK YOU.
I CAN BE REACHED IN THE FOLLOWING WAYS:
MAIL ADDRESS: THOMAS MARILL
COMPUTER CORPORATION OF AMERICA
CAMBRIDGE MASS. 02139
PHONE: 617-491-3670
ARPANET MESSAGE:TOM@CCA
-------
∂01-DEC-74 1347 AP,DBL
I'll be happy to meet and discuss what exactly has been done. Green has worked as my
advisor for this project, but Winograd has examined it more recently and might be
best acquainted with the theoretical side of BEINGs. -- Doug
∂30-NOV-74 2157 S,LES
I understand that you have some reservations about proposing
the new timesharing system in the new ARPA proposal. Clearly,
we should say something about our plans in the system software
area, since we will be proposing to support several System Programmers.
REG and I would like some guidance.
∂01-DEC-74 0213 S,LES
CHAIRMAN'S MESSAGE
The recent ACM National Conference included a lively SIGART
meeting attended by thirty-some people. The business portion
dealt with two main topics: 1) the continuing controversy over
the location of IJCAI-4$$SIGART Newsletter, Nos. 47 & 48, August-
October 1974*, and 2) the proposed new SIGART Bylaws.
The Bylaws appear to be non-contoversial. A version slightly
revised from the one published earlier$$SIGART Newsletter, No. 47,
pp.21-23, August 1974* was distributed and discussed. There was
a concensus vote by those present to loosen the nomination procedures
for SIGART offices. With this change, there was unanimous approval.
It will now be submitted to ACM Officials and to
the SIGART membership for a vote.
The IJCAI discussion, led by Jack Minker, Ed Feigenbaum, and others,
continued to draw strong feelings, all negative. There was no actual
debate inasmuch as there appeared to be no person present who favored
the Tbilisi site. The following resolution was passed (31 for, 0 against,
1 abstaining) with instructions to transmit it to the IJCAI General
Chairman.
.INDENTIT
"For a variety of logistic, scientific, and political reasons
we find the location of the Soviet Union for the IJCAI-4 unsatisfactory
and ill-considered. We recommend that the IJCAI Conference Committee
immediately seek a new location for this conference. We also request
that a referendum of the full membership of SIGART on this issue be
carried out by mail at the earliest possible time."
.TEXTIT
This message has been sent to Erik Sandewall, as requested. After
thinking about the proposed referendum of SIGART members, I have decided
not to do it for the following reasons.
.INDENTIT
1. The IJCAI Committee is already canvassing by mail the registrants
of the last IJCAI, who intersect strongly with SIGART membership, to
determine the likelihood of their attending the next conference as
scheduled. While this is not precisely the same question that was put
by the resolution above, it is more concrete. On issues as complex as
this, I believe that more is to be learned by inviting the public to vote
with their feet than with their heads.
2. It is too late to begin another survey. This Newsletter will
reach the membership around the first of the year. Meaningful
results could not be compiled before early February, which is after
the deadline for submitting papers for the conference (January 15).
There was ample time to have raised these issues earlier. The choice
of the U.S.S.R. as host country was made by the IJCAI Committee in
August 1973. It was published in the October 1973 issue of this
Newslatter (page 7), although the city was misidentified. Where were
all the concerned citizens last year at this time?
3. This is basically none of SIGART's business, inasmuch as SIGART
and IJCAI are separate organizations, tied only by common scientific
interests and common membership. In particular, the use of the
word "referendum" in the resolution above is unfortunately
presumptuous in that the outcome of such a vote would not be binding
on anyone.
.TEXTIT
Considering the amount of space that has been devoted to IJCAI issues
in these pages, the casual reader may be forgiven for thinking that
IJCAI is a subordinate element of SIGART. The writers have chosen
this forum because there is no other widespread newsletter in the
field of A. I. I believe that the Newsletter is performing a useful
expository function in this respect, but I hope that our members can
keep the relationship straight.
Speaking of relationships, since my position in this matter may be a
bit controversial, I should call to your attention a potential
conflict of interest: I was appointed Secretary-Treasurer of IJCAI-4
before I took this job. I believe that I have acted in the best
interests of SIGART in the current situation, but if you don't like it,
you can fire me (I think).
∂30-NOV-74 1603 network site AI
Personal opinions on proposed change of IJCAI location
Erik Sandewall, 1974-11-30
I distinguish two groups of considerations with respect to the location
of the conference: those which concern the success of the conference, and
those which arise from our ethical and political conscience as private
citizens. The latter group comprises two sub-questions: first, should
such ethical/political considerations be taken into account in the context
of scientific interchange, and (if the answer to that question is
affirmative), is it appropriate to wage a conference boycot against
the Soviet Union (henceforth abbrviated SU) at this time?
CONSIDERATIONS OF THE CONFERENCE AS SUCH.
The following issues concern us as organizers:
- Travel cost. I agree with your comments about the San Diego resolution
in this respect. It is not clear to me how much considerations of travel
cost influenced the decision. However, even if it were the dominant reason,
I doubt that it would make a particularly bad impression on people in
Europe: it would so obviously be one party's interest, and as such could
be shrugged off.
- Local AI milieu. Chosing a conference location close to an existing AI
center has some advantages: it facilitates the work of the conference
committee, and it may also imply better access to the right kind of
computing power and to the ARPA-net, which is an advantage for demos of
programs. I fail to see any other significant advantages, and conclude
that Tbilisi is none inferior in this respect to the other proposed
locations for the conference (W Germany and Japan).
- Local arrangements. SU is perhaps not the most service minded country in
the world, but Tbilisi is generally reported to be good or excellent.
- Access for foreign delegates. After Zohar Manna's letter, I do not
see that we could have any objections against SU policy, from the limited
viewpoint as conference organizers.
- Access for dissident Soviet scientists. Holding the conference
outside the SU certainly would not make it easier for them to come.
SHOULD OUR ETHICAL AND POLITICAL CONSCIENCE AS PRIVATE CITIZENS INFLUENCE
OR RESTRICT SCIENTIFIC INTERCHANGE?
I believe not. The common reason (do not mix X and politics, where x ranges
over sports, wheat trade, science, etc.) is not completely convincing to
me, but there are also certain ethical questions which are intrinsic in
science, and which require open communication. The recent decision by some
biologists to refrain from certain research involving synthesis of new
genes, until it had been proven that such experiments could not result in
new and very contagious organisms, is a prime example. If scientists let
their business be influenced by general-purpose ethical/political
questions, such as waging a conference boycot against the SU in order
to force the government there to improve civil liberties, then we will
rout bridges that will be needed for discussing ethical problems
within science. There are some such problems in computer science as
well. Government use of computers for controlling the citizens could
of course develop to much more frightening extremes in totalitarian
societies. This has not yet happened in the SU, probably mostly for lack
of resources. We should retain the very narrow channels we have for
indirectly influencing the continued development.
A second reason for my negative answer to the question is that, if such
a political discussion starts, it could quickly get out of hand. There
are many possible reasons for questioning a country's policies. Countries
with apartheid policies would be prime targets. Countries involved in
unsolvable conflicts, such as Israel and the neighboring arab countries,
could all be charged on ethical/political grounds by their opponents.
'Safe banking' countries (with Switzerland as the prime example) could
be blasted by people who are concerned that financial criminals put their
money there. Countries which have or are supposed to have very lenient
attitudes to sex, such as my own, would of course be a threat to the
delegates and even more to their families, and therefore be unfit as
conference locations. A few years ago, many would have objected to the
US as a conference location because of the Vietnam war. Having some
acquaintance with people in various European countries, I believe that when
Western Europe becomes less dependent on American military support and
therefore on American benevolence, criticism of the US for other foreign
policies will become more overt. Finally, on the assumption that
scientists would stick together in defense of common interests, one
might argue agains holding a conference in any country which gives
insufficient research funding to its own scientists. After all, why
should their government obtain scientific knowledge at a very low
price, if they have not helped to pay the expense of developing it?
IF GENERAL ETHICAL AND POLITICAL CONSIDERATIONS SHALL BE MADE, IS IT
APPROPRIATE TO WAGE A CONFERENCE BOYCOT ON THE SOVIET UNION AT THIS
TIME?
The idea with the proposed boycot (as formulated by Minker) is that the
SU government is very interested in Western technology, and that by denying
them access to such, we could force them to respect civil liberties in
their country. The argument assumes that one could make a deal: if the
civil liberties are restored, the conferences will come back.
First, of course, we do not have an organization which could make such a
deal. If, miraculously, the SU government would make it known that they
agreed to the terms, and improved civil liberties, then this would
merely show that they were susceptible to pressure, and new protests
would arise in order to extract bigger concessions. It is easy to
reschedule a conference, and it is an infinitely bigger undertaking to
change the ideological climate in a country. We can easily shift into
reverse; they can not. Therefore they can not negotiate with us, all
other reasons aside.
They could of course negotiate with another government, but not even
the US government controls where conferences take place. However, even
if one could arrange the proper forum for such a deal, I do not think
the SU government would agree to it. The question of "free interchange of
ideas" was brought up by Western countries, including neutrals, as a key
issue during the European security conference in Helsinki. The SU
had taken the initiative for that conference, and really wanted it to
succeed, and yet they would not yield an inch on the issue. Also, the
recent trade deals between the SU and the USA stipulates that the SU
should allow a small minority, already considered as partly hostile, to
emigrate more freely, which does not strike me as such a big concession
from the Soviets. The agreement does not say anything about civil liberties
and related issues. Since increased dissent in the SU would clearly be in
the American interest, I do not believe the Americans were not interested
in buying it. They were not allowed to.
From such observations, and also from some knowledge of the country, I
conclude that the full control of debate and opinions within their country
is considered as a vital interests by the present SU government. At the
same time, the proposed price for a liberalization, namely getting
some conferences, is ridiculously
small. If the SU government wants the scientific benefits of the conferences,
it can buy them in other ways, although perhaps at a slightly higher
price. Nobody has proposed (and hopefully will not propose) to bar
Soviet scientists from visiting conferences here, or to put an embargo
on scientific books, journals, or informal reports.
However, a widespread conference boycot could grow into an image problem
for the Soviet government. If at some point they would want to react to
it, the sensible thing (from their point of view) is of course to further
restrict foreign travel for Soviet scientists, and perhaps also to
restrict visits by foreign scientists. This would serve the double purpose
of being a reprisal against our friends, and of protection: the Soviet
government must think that if we are so zealous in our political opinions,
we are not good company for their scientists anyway.
Such restrictions would be a setback with respect to the ethical issues
within science, which I argued above. But I believe it would also be a loss
for the person who seriously wants help pull the SU in the direction of our
political ideals. The best channel is then to relate to their people
during an extended time. Contacts at a conference are of relatively little
importance, due to the ease of supervision and the big difference in
culture and language, but I believe with reason that Soviet citizens who
go abroad for a longer time sometimes bring home an understanding for new
political ideals.
Such reasoning is sometimes countered by saying that only "safe" people are
allowed to go, and that one should help the true opposition instead. Now,
there are certainly good humanitarian reasons for supporting dissidents in
the Soviet union, but one need only consider the minute influence of the
U.S. Communist party to realize that small groups whose opinions are
extreme relative to the group's environment, have a very hard
time against a concerted counter-campaign by a strong modern government.
It is much more important that people with some influence upwards in the SU
get a chance to study and accept some ideas and ideals here, even if they
react negatively to some others and bring back a balanced view.
Also, of course, no amount of screening will guarantee that a person is
"safe" in the sense that he does not pick up any new ideas. A prolonged
stay on one's own in a foreign and very new environment usually has a
strong impact on a person, and may break through solid psychological
defense.
For these reasons I believe that a conference blockade against the SU
would fail to achieve its intended purpose, but would have very undesirable
side-effects. Realistically, I do not see how we could change the location
of this conference now, and I have not heard anybody in the conference
committee argue for that. The present campaign may however achieve its
purpose in another way: if it raises enough embarassment, then organizers
of conferences in other fields will not consider Soviet locations.
This may still have adverse effects on Soviet scientist's opportunities
to travel, and I sincerely regret it.
Concerning European reactions, my general impression after talking with
people e.g. at IFIP was that this boycot sentiment was not present; the
general reaction was the disillusioned one: everyone has his faults.
However, I have just received a letter from Pat Hayes saying that there
are some protests in Britain also against the chosen location. No
details available.
ββββ
∂30-NOV-74 0855 1,MM AT TTY121 0855 @ AI
do you have text of acm proposal?
∂30-NOV-74 0853 network site AI
From: MINSKY@MIT-AI 11/30/74 11:52:29
I do not have text of the SIGART resolution.
If you have a copy I will put it and your letter up on the board here and collect reactions.
∂26-NOV-74 2024 CHS,REF
Very random thoughts on Berliner problem in BERLIN.PRB[CHS,REF]. You get
to look at it and tell me I don't understand what's going on.
Bob
∂29-NOV-74 1333 network site ISDT
Date: 29 NOV 1974 1333-PST
From: LONDON at USC-ISIB
Subject: IJCAI4 and Rosh Hashana
To: SIGART at CMU-10A, JMC at SU-AI, MM at MIT-AI,
To: RAPHAEL at SRI-AI, HARTHART at SRI-AI
Are you aware that Rosh Hashana (Jewish New Year) falls in 1975 on Sept. 6.
This is the Saturday of the new dates for IJCAI4 (Sept. 3-8). There
should thus be more controversy over IJCAI4.--Ralph London (LONDON@ISIB)
-------
∂28-NOV-74 2104 network site ISI
Date: 28 NOV 1974 2104-PST
From: LEDERBERG at USC-ISI
Subject: ADVANCED MEMORY CONCEPTS
To: JMC at SU-AI
I NOTE YOUR REMARKS. BUT VON NEUMANN WAS ALSO INTERESTED IN CELLULAR
AUTOMATA. I GATHER YOU DON'T THINK VERY MUCH OF WHAT HAS EMERGED IN
THAT AREA AS LIKELY TO BE CONTRIBUTORY TO PRESENT AGENDA.
JOSH
-------
∂27-NOV-74 0931 network site ISI
Date: 27 NOV 1974 0931-PST
From: LEDERBERG at USC-ISI
Subject: ADVANCED MEMORY TEVHNOLOGY
To: JMC at SU-AI
John -- I need your help in fleshing out the following concerns for
the ARPA committee you talked to
I am still troubled that we have not gone very deeply into radical
re-considerations of basic memory design, viz., to think through the inter-
relationships of software principles to hardware development. Now it
is certainlybtrue that we can simulate any other design (including, e.g.
associative memories and vector processing on the (now-)conventional
direct address memory systems. But I fear we are in a vicious cycle
where our imagination about software-potentialities, and hence long range
requirements is configured by presdently availavle hardware. The fact
that we still try to work in LISP despite the incredible clumsiness of
its implementation on contemporary machines is the illustrative counter-
example.
It is not going to be easy to collect all we need to know in the time
frame of our initial report; but it seems obvious that for a program of
this scope there needs to be some ongoing process of review and reinjection
of new insights. I hope this will be one of our specific recommendations.
Josh
⊗
-------
∂27-NOV-74 0423 S,LES
My impression is that we will not get the memory from DEC within
the next six months or so. The current plan for using it involves
the memory multiplexer, which will have to be designed and built.
∂26-NOV-74 1916 network site CMUA
***** FTP mail from [C300PM12] (MCCORDUCK)
DEAR JOHN,
I'D LIKE YOUR ADVICE AND HELP. I'VE SEEN (OR PLAN TO SEE SHORTLY)
THE FIRST GROUP OF MY SUBJECTS IN AI. BUT NOW I NEED SOME HELP DECIDING
WHO ELSE TO SEE. HERE'S A LIST OF THE MEN I'VE SEEN OR WILL SEE IN THE
NEAR FUTURE:
NEWELL, SIMON, FEIGENBAUM, GELERNTER, MINSKY, MCCARTHY, SOLOMONOFF,
SHANNON, BERNSTEIN, SELFRIDGE, ROCHESTER, SAMUEL, ARMER
HERE'S A LIST OF POSSIBLES, GENERATED FROM NILSSON'S PAPER (I.E., THOSE
WHO PUBLISHED IN AI BEFORE 1966, MY CUTOFF DATE) AND FROM
COMPUTERS AND THOUGHT, AND FROM MY HEAD. I'D BE VERY GRATEFUL IF YOU'D
HELP ME PRUNE THIS LIST (OR ADD TO IT) BY SAYING WHO, IN YOUR OPINION,
IS VITAL, WHO'S NOT SO VITAL, AND WHERE I'D PROBABBLY BE WASTING MY TIME.
SHAW, SLAGLE, TRENCH MORE, EDWARDS & HART, KOTOK, RUSSELL,
BERLEKAMP, TONGE, GERE, ROBINSON, ERNST, ROBERTS (BOTH FOR SCENE
ANALYSIS AND ARPA ACTIVITIES) HARRIS, CHOMSKY, KUNO & OETTINGER,
GREEN, BOBROW, RAPHAEL, SIMMONS, WEIZENBAUM, HUNT, LICKLIDER, LINDSAY
NEISSER, UHR & VOSSLER, AMAREL, BANERJI, BLEDSOE & BROWNING, COLBY,
MICHIE, PAPERT, ROSENBLATT, LLOYD MORISSETT
ANY HELP YOU CAN GIVE ME WILL BE GRATEFULLY RECEIVVED. SEE YOU SOON.
PAMELA
∂26-NOV-74 1452 network site CMU
***** FTP mail from [X180LE03] (SIGART)
John,
Yes, we would very much like a statement from you for the Newsletter.
We hope to get the issue out within a week, so would appreciate
getting the copy from you by this Friday. If this is not sufficient
time, please let us know and perhaps we could wait a bit longer.
If it is convenient for you, we would appreciate getting it over
the net (to save us from having to re-type it in). I simple
SOS or E format, with blank lines between paragraphs and without
justification, would be just fine for us.
Please let us know if there are any problems,
Thanks,
Lee Erman (SIGART @ CMU-10B)
∂25-NOV-74 1420 network site AI
Date: 25 NOV 1974 1718-EST
From: ERIK at MIT-AI
To: jmc@SU-AI
The December issue of the SIGART Newsletter will contain the resolution
from the recent SIGART meeting, and a questionnaire whether the
conference should be moved or not. In order to get a fair response
from the referendum, there should also be a statement in that issue
of the reasons for not moving the conference. Would you be willing
to write such a statement? Your message some days ago covers most of
the issues, but not the political considerations nor the question
whether political considerations should influence the choice of
location. - I have definite opinions on the topic myself, of course,
but would prefer to stay away from the discussion except when the
adequacy of the location from the scientific viewpoint is
questioned.
∂25-NOV-74 0945 1,BGB
On your console there is an envelope containing my draft of
six specific admistrative steps that you may wish to take
in order to get the PDP-11 thing done and to improve conditions
for doing vision/robotics research. I think Quam will second me
on all my points when he returns - perhaps we will talk about it
then. Also, I would be willing to give oral arguments on each item
at any time (day or night). I will finish a draft for ARPA vision
research and will then give up on protesting and planning until
January - curious to relate, I believe Earnest rather than Binford
will oppose the things I want to see done; consequently, it is up
to you to act. Bruce G. Baumgart
∂23-NOV-74 1707 S,WD
Crypt takes 3ms per word which is three times as long as my
programs and 10 to 20 times what I concider desireable.
∂23-NOV-74 1227 1,DEK
john, I think bill gosper has some revolutionary ideas about
continued fractions that are certainly worthy of support.
I'm out of money, not being in charge of a Project. I don't
know about the finances of your project, and you may well be
out f money too. However, if there is a reasonable possiblity
of supporting gosper I think it would be more than worth the
investment.
Bill needs a bit of bugging to get him to write things down and
to keep moving forward. I would volunteer to be responsible for
that.
He says you will support him to look into some topological problem
that he could force himself to get interested in if absolutely
necessary. Maybe that would be the best thing, I don't know the
problem or the potential it has. All I know is that (as a sort
of lower bound on what Bill can do) the continued fraction thing
is certain to be of widespread interest. everybody to whom i
mentioned the possiblity at oberwolfach was anxious to see his
ideas. --don knuth
∂23-NOV-74 1124 1,DEK
re your question about van der waerden's theorem...i'm not sure
you're not baiting me intentionally! the three visitors i'm supporting
this year (erdos, szemeredi, graham) are the leading authorities
on such questions. Szemeredi has obtained the best bounds, in a truly
deep mathematical paper about to be
published. he lectured on it at the vancouver
convgress and I understand was near to winning the field medal for it.
szemeredi is here for the full year, ron graham will be here for four
or five one-month shots, erdos returns next spring. phyllis knows how
to reach erdos. --don
∂22-NOV-74 0050 S,LES
Subject: ARPA PROPOSAL, 1975 + 2
As previously advertised, it is now time to write our ARPA proposal
for the two year period beginning 1 July 1975.
Your section should include a summary of recent accomplishments, with
suitable bragging about significant work. There should be a
bibliography of books, articles, reports, and films that are relevant
to an understanding of this work, including publications of outside
people where appropriate. You will probably save time by building on
our recent annual report [AIM-252], which is available in the
document room and on the disk in RECENT.PUB[D,LES].
I would welcome suggested changes and additions to our list of
"bullets" given at the beginning of Section 2 [page 3 of the printed
version, page 6 on the disk].
You should attempt to identify new directions that our research
should take and the reasons why you think the work is important. The
treatment should be aimed at readers who are knowledgeable in
computer science, but not necessarily in your speciality. Try to
convey the objectives of your work and the general approach, but keep
it informal and don't try to cover everything.
If you forsee any special hardware or system development
requirements, lay them out and we will try to estimate the costs.
I need your input no later than Monday, December 2. Happy Thanksgiving.
␈ CC: @LEADER.LST:CCG,DCL,JMC,PDQ,RWW,TOB,TW,REG,TED
∂22-NOV-74 0020 1,PDQ
please look at CART.BAD[1,PDQ] and send me responses.
∂21-NOV-74 1756 1,PDQ
When is the deadline for any material for the new ARPA proposal?
∂19-NOV-74 2011 ACT,REG
What we need in a document compiler is a set of programs (an assembler and a
loader) that manipulate paragraphs in the same way that we now manipulate
subroutines. My idea is that the "assembler" produce "relocatable paragraphs"
which contain "external fixups" and "internal symbols". An external fixup is
a cross-reference outside of this paragraph. An internal symbol is what a
cross-reference refers to. The loader assigns the relocation factor (i.e.,
page and line) to each paragraph and to each of the internal symbols of a
paragraph. The loader also resolves the value of external requests. Since
externals might change the length of a paragraph, the loader must iterate
until an assignment of relocation factors to each paragraph is found that
satisfies all constraints (I haven't worked out the details of that part
yet).
The basic idea is to confine the slow text munching to the assembler, which can
assemble separate pieces of a document to facilitate debugging of small sections
and to eliminate the need to reassemble the entire manuscript when an incremental
change is made.
Have you any interest in pursuing these ideas?
∂19-NOV-74 0845 2,JH
John,
I will meet you in front of the Main Interior Building (18th and I Streets
I think) at 8:30 AM on Monday Nov. 25. If that's not acceptable or you want
to meet the night before the meeting leave me a message on the system and I'll
login from BBN.
Jim
∂19-NOV-74 0239 EXP,JMG
Thanks. I would like to show you one of the interactive displays I have
been working on for looking at 3-dimensional configurations when it's
a little more along the way.
∂19-NOV-74 0209 EXP,JMG
About 2/3 of the month has passed, and I have (I believe as of now)
used the 2 hours of CPU allotted. This is a petition for more CPU
(say another 2 hours). I have finished the crunching, for sure, and
am now trying to 1) display the analyses interactively, and then
2) write up the results (the dreaded document, etc.). I could well
stop for the rest of the month, but am forced to petition on the
following grounds: starvation. I now have no financial support,
and yet, have a postdoctoral fellowship waiting upon first draft of
the aformentioned document. The catch-22 here is the existence of
a deadline (also dreaded) early next year (i.e. 1975). Can we talk
it over soon? - john grey.
∂18-NOV-74 1654 FOL,AJT
an FOL meeting Friday at 1:30 will clash with Winograd's seminar. can
we have it in the morning (after 10:00) instead?
∂17-NOV-74 1631 1,JAF @ NCCT
ah yes, kicha is very moody and is about the worst student i know to
be involved in a change-of-adviser squeeze. i will try to talk with
you, forrest, and kicha when i come (wed-fri) and come up with
something. would like general talk with oyou anyway. jerry
∂18-NOV-74 0917 network site BBN
Date: 18 NOV 1974 1217-EST
From: CLEMENTS at BBN-TENEX
Subject: CURRENT TENEX SOURCES
To: JMC at SU-AI
cc: CLEMENTS, SUTHERLAND, ALLEN
HI JOHN, LONG TIME NO SEE. NOT SINCE A VOLLEYBALL GAME WHEN I
WAS DECIDING TO COME WORK AT BBN.
THE CURRENT TENEX SOURCES FOR RELEASE 1.32 ARE AVAILABLE TO THE
SITE DISTRIBUTION GROUP FROM THE DIRECTORY <TENEX-132> ON HOST
69 (OR 105 OCTAL), BBN-TENEX. EACH SITE INCLUDING YOURS HAS A USER
NAME IN THAT GROUP. YOURS IS USER "GORIN". HE CAN GET THE STUFF VIA
FTP IN THE USUAL WAY, USING NAME GORIN AND PASSWORD WHATEVER IT IS,
AND ACCOUNT 102. IF ANY DIFFICULTIES COME UP DOING THIS, SEND MAIL
TO ME AND I WILL TRY TO HELP.
A REMINDER, WE SENT YOU THE 1.31 RELEASE SOME TIME AGO. I THINK
RUBIN HAS THAT, BUT I MAY BE WRONG. TO MINIMIZE THE FTP TIME, YOU
MIGHT WANT TO JUST READ THE DOCUMENTATION FILES IN <TENEX-132>
AND IN <DOCUMENTATION>, RECOGNIZABLE BY FILE NAMES, AND JUST TAKE THE
SOURCES WHICH HAVE OBVIOUSLY CHANGED FROM THE PREVIOUS RELEASE.
YOU ARE WELCOME TO TAKE THE WHOLE THING, OF COURSE, FOR ANY ARPA
PURPOSE.
AGAIN, LET ME KNOW IF I CAN HELP.
/RCC
-------
∂18-NOV-74 0911 1,MG
I'm looking forward to the FOL meeting at 1.30. I invited waldinger (as well as
boyer) along - I hope this is OK.
∂18-NOV-74 0638 network site AI
Date: 18 NOV 1974 0935-EST
From: PHW at MIT-AI
To: jmc@SU-AI
Oops. Mistyped. ERIK is ERIK.
∂16-NOV-74 1459 PC,RWW
↑MONDAY IS OK
∂16-NOV-74 1414 THE,AJT
Monday 1:30 fine with me. arthu.
∂16-NOV-74 0401 S,DWP
Sure.
∂16-NOV-74 0000 1,DBX
OK Monday 1:30. (Still waiting for you ...)
∂14-NOV-74 0243 105,SGK
Why aren't we fighting this random move to campus more? Why are we
not threatening to leave stanford if they force us to move? Is it
true that CS thinks there is no reason to have the computers and the
people in the same place? Do they really plan to have the machine, and a few
people to maintaine it in one place, and the researchers as a separate clearly
defined class of people in another? If they do it sounds like not only
will this lab cease to attract new, good blood [as opposed to the typical
computer freakish, space war playing cretin univs. seem to have plenty of these
days], but lose everyone it has already. Stanford is headed towards losing
some of its greatness in CS. Making a move similar to the one I understand
is going to take place will only speed this process.
∂13-NOV-74 1113 1,JB
Since Jorge Morales is doing his qualifying in MTC in January, I'd like to take
mine at the same time. JB
∂12-NOV-74 1818 network site MAXC
Date: 12 NOV 1974 1819-PST
From: ELLENBY at PARC-MAXC
Subject: Uncrossing of toes.
To: jmc at SU-AI
John,
We have managed to move into part of our house:
1030 Harker Avenue,
Palo Alto, 94301. Phone: 328-6082.
Drop in sometime. Best wishes Gillian and John.
-------
∂12-NOV-74 0816 1,PDQ
Fri 1:30 is ok.
∂11-NOV-74 1936 1,PDQ
Can we meet early Tuesday sometime. Agenda: Vision, Cart, Personnel, Hardware.
Please advise of convenient time.
∂11-NOV-74 1345 network site AI
Date: 11 NOV 1974 1644-EST
From: PHW at MIT-AI
To: jmc@SU-AI
PLEASE PHONE MARVIN TO TALK ABOUT POSSIBILITIES FOR GOSPER.
∂11-NOV-74 1248 ACT,REG
FYI, Don Farrow called and said that he's bringing Weiss and Kotok at 1330 Wednesday
∂08-NOV-74 2240 1,PDQ
I WOULD LIKE TO DISCUSS VISION PLANS AND PROBLEMS -- TELL ME WHEN
∂08-NOV-74 2022 1,DBX
(Still (Still waiting to your answer on my PLANS of Oct 12))
∂08-NOV-74 1609 1,MG
I've put a rather vague outline of what I'd like to do, am doing, and have
been doing on PLAN.MG[1,MG]
∂07-NOV-74 1002 LCF,MAL @ UKT
DEAR JOHN, VERY SHORTLY DAVID BARSTOW WILL GIVE YOU A COPY OF MY THESIS
TO EXAMINE. COPIES ARE ALSO ON THEIR WAY TO THE OTHER READING
COMMITTEE MEMBERS (HOARE, WEYHRAUCH, GREEN). I AM HOPEFUL, FOLLOWING
ROBIN MILNER'S APPROVAL OF THE CURRENT VERSION, THAT THIS WILL BE THE
FINAL DRAFT BECAUSE I AM SUDDENLY UNDER SOME PRESSURE TO FINISH.
AN OFFER OF A POSITION (AT AUST. NAT. UNIV., CANBERRA) THAT I WANT TO
ACCEPT, IS CONDITIONAL ON MY OBTAINING MY PH.D. SOON. HENCE, IF YOU
CAN FIND THE TIME SOON TO LOOK AT THE MANUSCRIPT I WOULD BE MOST
GRATEFUL. OF COURSE, I EXPECT THERE ARE TYPOSTHAT I HAVE MISSED AND
I WOULD BE HAPPY IF YOU COULD POINT OUT THOSE THAT YOU FIND. THEY WILL
BE CORRECTED IN THE FINALVERSION. YOU MAY MAIL YOUR COMMENTS TO ME AT
[LCF,MAL] OR GIVE THEM VERBALLY TO DAVID BARSTOW. (HE WILL RELAY THEM.)
THANK YOU VERY MUCH - MALCOLM NEWEY.
∂06-NOV-74 1258 CAR,HPM
New CRYPT ready!
Source is CRYPT.FAI[HAK,HPM]
Writeup is CRYPT.WRU[HAK,HPM]
∂04-NOV-74 1547 PAR,BLF
FOR THE MONTH OF OCTOBER, PARRY USED 374 MINUTES OF CPU TIME, WHICH
IS APPROXIMATELY 26,180 K MINUTES. THIS IS APPROXIMATELY 1 MINUTE OF CPU
TIME PER CONVERSATION. BILL
∂04-NOV-74 1035 DOC,TOB
I called Dave Grossman today. He has cleared
with IBM management that a Research Associate
appointment would be ok with them; it is ok with
him. I am interested in him coming here.
Tom
∂06-NOV-74 0923 1,BH AT TTY11 0923
↑B was disabled last night intentionally by request of LES
as the easiest way to solve a complicated and temporary problem
and is now restored. Sorry about that.
∂05-NOV-74 0842 AMP,ELM
IF YOU ARE SERIOUSLY INTERESTED IN ESTABLISHING
A MICROWAVE LINK TO CAMPUS, I HAVE LOCATED
FOUR USED TRANSMITTERS AND RECEIVERS. THE PRICE
WILL BE LOW ENOUGH TO BE UNIMPORTANT. A DECISION
WILL BE NEEDED DURING THIS WEEK.
∂03-NOV-74 1233 1,BH
They seem to be forging ahead with this TIP login stuff. The way they
are implementing it is that when you call up a TIP it will connect you
to an RSEXEC, which will log you in. Thus BBN-TENEX gets to do the
work instead of the TIP itself. (One more flaky component in the chain
needed to use SAIL from AMES-TIP, by the way.) Are we (i.e., you) going
to try harder on this scheme of letting each host log in its own users?
␈ CC: jmc;les
∂01-NOV-74 0959 ACT,REG
I talked with Kotok. He says we've got three options:
1. accept a production 1080
2. wait 9 months or more for the version we want
3. send people to marlborough to check out a customized version.
(option 3 has the Kotok stamp of approval, but his management never
heard of it, yet)
∂01-NOV-74 0851 ACT,REG
Well, I talked with Harvey Weiss again today. Not too much transpired. It looks
like we get to talk at length with Kiesewetter at DECUS, and perhaps lure him
here over the weekend.
∂31-OCT-74 1944 1,DBX
Still waiting for your reaction on my PLANS of Oct 12.
∂31-OCT-74 0710 network site ISI
Date: 31 OCT 1974 0710-PST
From: LICKLIDER at USC-ISI
Subject: SEEING YOU FRIDAY
To: JMC at SU-AI, MCCARTHY at ISI
cc: LICKLIDER
The free times on my schedule tomorrow are 12:15 to 3:00
and 5:00 to 6:00. I have to leave at 6:00 because Louise and I
have a dinner engagement. I hope that the time I have free
concides with your available time. Please let me know.
Regards
Lick
-------
∂31-OCT-74 0429 network site ISI
Date: 31 OCT 1974 0429-PST
From: LICKLIDER at USC-ISI
Subject: Seeing You Friday
To: McCarthy, JMC at SU-AI
cc: Licklider
Looking forward to seeing you. As soon as I get to the
office and get a look at the schedule, I'll SNDMSG what times are
available.
Regards
Lick
-------
∂28-OCT-74 0007 100,100: jmc @ SAIL
write bowen
∂29-OCT-74 1320 MAP,RF
I have renamed all the .PUB files for the memo, splitting them
into new pieces to facilitate reordering. Each has been given
a name matching the mask "M*.PUB". The list and cross-reference
to old names is on page 6 of MEMO.PUB.
∂28-OCT-74 0224 S,LES
I believe that it would be worthwhile to establish, for each research project,
a list of important references. We already have a file area allocated to
bibliographies, namely [BIB,DOC]. A hand-eye bibliography exists there in
the file HAND.EYE, but it is a bit out-of-date.
I further propose that any bibliographies that are created use the following
format:
<author names, separated by commas, with lead author listed last-name-first>
"<title, with any internal quotes represented by ` and ' >"
<journal or other source -- if more than one, separate with ";">
<if in a book, give publisher and city>
<month, if known, and year of publication>.
<blank line>
∂27-OCT-74 2338 105,SGK
Have you spoken with Sarah yet? and how do you feel about it?