∂12-Oct-79  1339	JMC  
To:   LES
CC:   RPG   
Sounds plausible.

∂12-Oct-79  1337	JMC  
To:   JRA    
You can phone him at 494-8670.

∂12-Oct-79  1115	JMC  
To:   JRA    
I suggest you phone Pat Hayes at 321-2052 to ask for suggestions.

∂12-Oct-79  1114	JMC  
To:   wilkins at SRI-KL
SEARLE[E79,JMC] is the source and SEARLE.XGP[E79,JMC] the xgp version
of a reply to Searle's second draft.  If you want a paper copy, message
Patte to send you one.  I shall be away for two weeks, but after the
first of November, I will be glad to discuss these matters with you.
Perhaps you would like to visit our AI and philosophy group at the
Behavioral Science Center for lunch sometime.

∂09-Oct-79  1857	JMC  
To:   DPB at SU-AI, "@FAC.DIS[1,DPB]" at SU-AI  
My opinion is that letting Bell Lab make a pitch is a good idea.  If it
gets out of hand we can either limit it or stop it.  A more genuine
interaction between industry and the academic world is desirable.
However, it should be tried out rather gingerly with pitches in only one
or two classes where the instructors agree.  Alternatively, we could ask
them to give a seminar at a time of their convenience on work at Bell
Labs.  If we say we're trying it out, we don't have to let anyone else
do it until and unless we decide we like it and what form it should take.
This is assuming the recruiters are actually researchers and not just
random personnel men.

∂08-Oct-79  2336	JMC  
To:   CLT    
abilit.xgp again

∂08-Oct-79  2038	JMC  
To:   ME
ME - About 3500 tracks (of quotas).  Of course, we won't get it back until
the musicians have had a chance to transfer it to their new computer.
I don't understand the units.  How many tracks are there, and what
is a typical quota of 150K words in those units.

∂08-Oct-79  1843	JMC  
To:   ME
About how much disk will we recover when we separate from the musicians?

∂08-Oct-79  1519	JMC  
To:   PAT    
Thanks for the references.

∂08-Oct-79  1357	JMC  
To:   HPM    
think so but not sure

∂07-Oct-79  1327	JMC  	references    
To:   PAT    
There were two special issues of Reviews of Modern Physics containing
reports of American Physical Society studies on nuclear energy.  One
discussed reactor safety and the other the fuel cycle and waste
disposal.  Please get me the precise references including exact titles
and dates.  The physics library will have it.   I don't need the
documents themselves.

∂06-Oct-79  2050	JMC  
To:   DON    
Speaking of the wetsuit, did you know that if you put the wetsuit on when
you find it but drop all the wetsuit will be on the floor.  Nevertheless,
it will protect you in the water.  Indeed a magic wetsuit.

∂05-Oct-79  2220	JMC  	nonmonotonic conference credit    
To:   CLT
CC:   LES, HVA   
HVA will find out who paid for Drew's McDermott's
trip and I'll try to think whom to credit
if anyone.  Basically, we had no expenses except for that and so needn't
credit anyone but ourselves and SAIL as an organization.  I'll ask Les
if he can think of any benefit to crediting any AI Lab sponsor.  Probably
there will be some benefit in crediting ARPA.

 ∂05-Oct-79  2211	CLT  	??? 
I received the following query.  I can answer most of the it.  
What about the funding?
Does someone/agency get credit for bringing Drew out?  If so who?


 ∂05-Oct-79  1619	BOBROW at PARC-MAXC2 	Credit line for Nonmonotonic lgic confeence
Date:  5 OCT 1979 1616-PDT
From: BOBROW at PARC-MAXC2
Subject: Credit line for Nonmonotonic lgic confeence
To:   talcott at SAIL

I am putting together a special issue of the AI journal which
will contain all the papers presented at said conference which you
and John organized ( I have been told).  I would like to give
credit to the conference and the organizers for stimulating the
special issue.  Will you please tell me the dates, the exact title,
who to credit (You, John, funding agency, SAIL???) etc.
Thanks.
danny
-------

∂05-Oct-79  2203	JMC  
To:   DON    
I'll trade one way of lighting match (only one I know) for how to live.

∂05-Oct-79  1931	JMC  
To:   MFB    
how about friday nov 30 at 2:15. that's ok

∂05-Oct-79  1928	JMC  
To:   JRA    
Stoyan should be invited to be on the committee and to give a paper.
The probability of his being allowed out is epsilon so nothing should
hang on it.

∂05-Oct-79  1926	JMC  
To:   JRA    
There is no reason that I see for Stanford LISP to be represented.  Almost
all of us have switched to Maclisp, and those that haven't don't prefer
Stanford LISP but just have large programs to maintain.  I am dubious
about the value of ACM co-operation; they have a reputation for imposing
conditions.  I think this applies only to their "sponsorship" - not to
listing it as a coming event in their publications.

∂05-Oct-79  0106	JMC  
To:   rms at MIT-AI    
Misstatement.  I meant "... couldn't be sure that learning in an ARBITRARY
model made sense ..."

∂05-Oct-79  0104	JMC  
To:   rms at MIT-AI    
I agree with the considerations of your message.  It made me nervous
that I couldn't be sure that learning in a particular model made sense,
but I don't see anything wrong with the idea that this just excludes
particular shared pointer models.  Also I don't have an actual example
in which shared pointers do harm, but my present intuition is that
it wouldn't be difficult to make one.  I'll think about it.

∂05-Oct-79  0058	JMC  
To:   HPM    
"every option ..." indeed arises after pulling oil lever.

∂05-Oct-79  0039	JMC  	haunt    
To:   HPM    
1. eating the orchid at 23:58 or at midnight has no special effect.
2. pushing normal causes the elevator to break
3. I have been unable to get anything but the sheep message from
blowing the horn which incidentally causes it to disappear - at
least if one is in the house.  I didn't check effect on score.
I thought the message might depend on the stage of the game, but
even pushing it before using special properties of elevator still
elicits only sheep message.  Do you remember under what circumstances
you got the other message?  Still, the horn may be a means of getting
hints.  Maybe I'll see what happens if I blow it after pulling the
oil lever.

∂04-Oct-79  2100	JMC  
To:   RWW    
The Aiello letter is fine except that I have changed the last phrase
which was open to the interpretation that we think she might be good
enough for the Italians even though she might not be suited to an
American univerity position.  If joint publications are in the works,
this should be mentioned even though details can be left for her resume.
Ask Patte to get it on a letter for me to sign.
Use this modified version.

Dear sirs,

Luigia Aiello has been a visitor at the Stanford Artificial Laboratory
on two occasions  -  January 1973 to August 1974, and February 1979
until the present.  During her first visit here she worked with
Richard Weyhrauch on the LCF project and they produced several papers
together.  On her currrent visit she is involved with the formal 
reasoning project and is working with the FOL system.
She has made considerable contributions to both these
projects and as a result we are currently augmenting her Italian 
financial support.  I believe that she is well suited for a
university teaching position.

∂04-Oct-79  1205	JMC  
To:   JRA    
Raj Reddy said IJCAI might sponsor LISP meet, and you should send him a message.

∂04-Oct-79  0038	JMC  
To:   rms at MIT-AI    
The Sato paper got back, thank you.

The circumscription paper is finished.  Its source is circum.new[s79,jmc]
if you are interested in FTPing it.  Otherwise, I'll send you a paper
copy if you want, or you can wait for it to appear in the non-monotonic
issue of the AI journal or as a Stanford AI Memo.

I don't remember saying that anything might not have finite models.
What I think I did say was that I wanted to axiomatize the wise men
and S and P puzzles so that there would be no merging accesibility
paths in order that my definition of learning would work properly.
That would force there to be only infinite models and I know how
to do it.  All I need is an axiom that says that a world can be
accessible only from one world other than itself.  The point is
that a merging of accessibility paths may be compatible with
an initial state of knowledge, but an inconsistency could arise
when I define learning as allowing only those accessibility relations
compatible with what is learned.  If the same world were accessible
in two ways, learning something might want to cut off one of its
outgoing accessibilities from one point of view and not from another.
I lose nothing by postulating non-merging.

∂03-Oct-79  1922	JMC  
To:   admin.gorin at SU-SCORE    
The time is ok for me, but I leave at noon.  How about the Dept. Chairman's
conference room?

∂02-Oct-79  2329	JMC  
To:   HPM    
Kicking the football broke the dome, but it was moonlight and bat escaped.