COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	@make(letterhead,Phone"497-4430",Who"John McCarthy",Logo Old,Department CSD)
C00012 ENDMK
C⊗;
@make(letterhead,Phone"497-4430",Who"John McCarthy",Logo Old,Department CSD)
@style(topmargin 2 inch)
@style(indent 8)
@blankspace(10 lines)
@begin(address)
Professor Herbert Stoyan
Universitat Erlangen-Nurnberg
Institut fur Mathematische, Maschinen und Datenverarbeitung
Martenstrasse 3
D-8520 Erlangen, Germany
@end(address)
@greeting(Dear Dr. Stoyan:)
@begin(body)
	I have just read your recent 15 pages, but I haven't
read yet the earlier material.  The English in both is nowhere
near good enough for publication.  My secretary, Mrs. Larson, has corrected
the English in the early material, and we'll send it to you.

Here are some comments on the new material.

h60 - There was no agreement that I remember that maintenance and
further development would be done at Stanford.  Anyway, I didn't
undertake it except to comment on the LISP 2 efforts.  The work
by Smith and Enea on what they called LISP70 but never finished
wasn't directed by me.  They were part of Ken Colby's group at
Stanford and worked with his support and then later with someone
else's.
  Oh yes, I forget about
the improvements to the 7090 version.  Frankly, I forget how that
came about, although I guess it was largely done by people working
for me.

h60 - While I first proposed time-sharing in 1957 (orally only),
when Morse responded to my memo, perhaps having gotten some support
from NSF (National Science Foundation), he asked me if I wanted
to lead the time-sharing work myself or would it be OK to entrust
it to Herbert Teager, a new assistant professor of electrical
engineering with whom I had discussed my time-sharing ideas.  I
agreed that it should be entrusted to Teager.  The results of doing
this were mixed.  The first typewriter was connected by Computation
Center staff members (Arnold Siegel and/or Frank Helwig, I think)
before Teager took charge.  Teager undertook to replace this one
typewriter system by a three typewriter system, and this seemed to
take a long time.  A change of computers from 704 to 709 to 7090
took place.  The time-stealing mode was never used except for
demonstrations, because there were too many operational difficulties.
In particular, time-sharing requires a file system, and there wasn't
one.
@newpage
@blankspace(5 lines)

	I was never asked to be head of the Computation Center.
When I proposed to leave M.I.T. for Stanford in 1962, it was suggested
that if I stayed at M.I.T., I would be made Director of Research for
the Computation Center with some additional salary.

	There were two activities:

	1. There was the long range committee with Teager as initial
chairman.  Teager proposed a STRETCH as an interim machine, and
something vague as the ultimate machine (I don't remember it. You'd
better read the documents before you say anything.)

	2. Simultaneously, there was the plan to time-share the existing
7090.  Teager's plans for this never became concrete and were
gradually superseded by Corbato's interim scheme.

h62 -	Swapping wasn't invented only as a stopgap.  It was in my
plans for time-sharing from the beginning, because we didn't expect
large enough memories to keep all in core at once, especially since
it was already clear that AI and symbolic computation would require
single LISP programs much larger than existing core memories.
Fredkin's swapping drum was an elegant realization of the concept.

	It should be emphasized that while Hart may have introduced
macros into LISP, they already existed in symbolic assembly languages.

Digital Equipment was started by former employees of M.I.T. Lincoln
Laboratory.  This organization, which worked on air defense and
related matters, supervised IBM's work on the computers for the SAGE
air defense system, at least at first.  I don't think any of the
founders of D.E.C. had been IBM employees.

	As to time-sharing, I don't believe Fano made any technical
contribution.  He was the electrical engineering professor put in
charge of Project MAC, when Licklider (from his post in ARPA)
persuaded M.I.T. to undertake it.  The main technical person
was Corbato.  Digital Equipment Corporation gave an early PDP-1
to the M.I.T. Electrical Engineering department (1961?), and Dennis
undertook to make a time-sharing system for it with the help of
students.  Dennis's project at M.I.T. and mine at BBN were
contemporaneous, but the BBN project was working first.  On the
other hand, BBN didn't continue to use it, whereas Dennis's system
was in continuous use from the time it started working.

	CTSS was started (perhaps with that name) before Project
MAC existed, but it was not put into general use, as this would
have involved taking over the machine at the M.I.T. Computation
Center which was fully occupied with batch processing.  I'm not
sure whether it was even demonstrated.  After Project MAC was
started, a new 7090 was bought specifically for the time-sharing
work.  CTSS wasn't put into use until 1964, and the summer of
1964 was when Project MAC invited many people to spend the summer
and try out CTSS and time-sharing.

	I haven't seen the Glauthier paper.

	BASIC and JOSS were single language systems.  The others
allowed arbitrary languages.  The latter were first.
@newpage
@blankspace(5 lines)

	I'm trying to get McIntosh's address.  I believe he is still
in Mexico, and I'm writing to Adolfo Guzman, who just sent me
a paper, "A Heterarchical Multi-microprocessor LISP machine".  He
is at

@begin(verse)
Instituto de investigaciones en matematicas applicadas y en sistemas
Universidad Nacional Autonoma de Mexico
Apartado Postal 20-726
Mexico 20, D.F.
548-54-65
MEXICO 
@end(verse)

	It isn't obvious whether the 548-54-65 is part of the address
or a telephone number.
@end(body)

Best regards,




John McCarthy
Professor of Computer Science