COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Forest:
C00009 ENDMK
C⊗;
Forest:
	Would a sufficiently favorable reaction to something like the
following letter induce you to run a project?

Dear Mr. Carey:

	We have been told that when you attended tha recent Executive
Committee meeting of the Stanford Business School, you were disappointed
to learn that the School was deciding whether to buy a computer from
Hewlett-Packard or Digital Equipment and that IBM was not considered
to be in the running.  As you may remember, the requirement was for
a general-purpose interactive machine, and  that a Digital Equipment
2040 was eventually selected and installed.

	We further understand that at your initiative,
IBM is exploring the possibility of a joint study with Stanford
aimed at hardware and software systems of improved usefulness in
university environments.  We applaud this constructive approach.
To this end, IBM is exploring with the Stanford Center for
Information Processing distributed computing systems involving
a 370/145 and a number of Series 1 computers.

	In our personal opinions, however, it might be worthwhile
to explore the original problem that led the Stanford Business
School to eliminate IBM equipment from consideration early in their
search.  This is the problem of cost-effective general purpose
interactive research computing.  For this purpose, a machine must
be able to support a substantial number of terminals whose users
will carry out computations of all sizes using a large variety
of programming languages and canned programs.  Good editing and
document preparation facilities are also needed.  The terminals
must also be inexpensive, because the trend is to provide a terminal
in every office, including those of secretaries and professors
for whom computing is not a major interest.

	Above all, the personnel cost of running the installation must be
kept low, because it doesn't matter how cheap the equipment is, if an
organization can't afford the staff.  The Business School runs their
Digital Equipment 2040 with a staff of two.  The LOTS (Low Overhead
Time-Sharing) D.E.C. 2050 that serves
2500 student users every quarter has a staff of four; the similar systems
at the Artificial Intelligence Laboratory, the Institute for Mathematics
in the Social Sciences, and the Institute for Artificial Intelligence in
Medicine have similar staffs.  In contrast, the
Stanford Center for Information Processing has a staff of more than 200 to
operate its three 370/168 computers and its 360/91.  No doubt there are
reasons for the large staff, but the Business School would probably have
had to give up the idea of its own new computer system, if
even a staff of five had been required.

	We would like to explore the possibility of developing a
configuration of IBM equipment and a new operating system that
would be capable of supporting moderate scale interactive computing
using a wide variety of languages and programs.
Our own motivation is the wealth of software available for IBM
computers and our desire to make it available to medium sized
departments and laboratories in a cost-effective way.
The proposed 370/145 would be adequate unless IBM considered that
it would be better to use newer equipment.  If there is interest,
a proposal could be developed.

	To introduce ourselves, Forest Baskett is Associate Professor
of Computer Science at Stanford.
His main specialty is operating systems, and he has extensive experience
with IBM 360 and 370 equipment.  Until September 1978, he is on leave
at Los Alamos Scientific Laboratory, where he is in charge of the
development of an operating system for the Cray-1 computer.

	John McCarthy is Professor of Computer Science at Stanford,
Director of the Stanford Artificial Intelligence Laboratory, and
Director of the LOTS computer facility for student computing.  He
has extensive experience with earlier IBM equipment and Digital
Equipment PDP-10 computers.  He invented the concept of
general-purpose time-sharing in 1958 and started the original M.I.T. work
on general-purpose time-sharing computer systems.

.sgn

Forest Baskett

John McCarthy

P.S. IBM has more loyal followers at Stanford than we are, but they
are much less likely to do something different from what IBM will
do anyway.