eee oa care cee

Miri LI [eo Wo. 4
ee a “
THE COMPUTATION LABORATORY OF HARVARD UNIVERSITY =~
CAMBRIDGE 38, MASSACHUSETTS

MAY 7 196
fae :

tisk
May 3, 1963

Professor George E, Forsythe
Director, Computer Science Division
Department of Mathematics

Stanford University

Stanford, California

Dear Professor Forsythe:

As a member in the computer area at Harvard, I am envious of the
rapidity with which you are building up your Division. I shall try to
state the different aspects of Marvin Minsky on which I happen to know
something, not worrying about the fact that most of what I have to say
will be no news to you.

Minsky, although trained as a mathematician and as something of
a logician, is best known for his work on artificial intelligence, a
highly controversial field. The "positivists" on artificial intelligence
are nearly all enthusiastic about Minsky; even the tnegativists" usually
remark that if anybody can make some breakthrough on artificial intelli~
gence, Minsky is the most likely to do it, His IRE survey of January 1961
is regarded as a standard reference. He is one of the most popular
lecturers in the field, at various conferences and at summer institutes.

For many years, beginning at least with his Ph.D, thesis of 195k,
he has been interested in models of the nervous system and plans to do
some machine simulation of nerve nets, I understand his knowledge of
neurophysiology, at least in the aspects relevant to his problems, is
quite up to the professional standard.

On the more practical side of using machines, he is consultant to
many government organizations, and, in particular, one of the active pro-
ponents of time-sharing on computers. He has worked closely with Selfridge
on certain aspects of pattern recognition. Ina report of 1956, he sug-
gested using diagrams to prove geometrical theorems on machines: an idea
which has led to quite a bit of development work at IBM. He enjoys building
things and has a sufficient working knowledge of electronics,

Ays NtW
Professor G, E. Forsythe -~2- May 3, 1963

On the more technical side of logic, Minsky is regarded as a
leading expert on Turing machines and Post production systems; his proof
of the unsolvability of Post's tag problem is widely accepted as a very
good piece of work.

He is lively, stimulating, pleasant, original, and a highly cultured
person,

At the age of 35, he is full of new ideas and future plans. I have
no doubt that he will be very productive both in his own work and in
guiding students. I consider him ready to be a full professor,

The only other possible candidates I can think of are J. Richard
Biichi, now at Michigan, and P. M. Schtttzenberger, now in Paris. Neither
of them is much concerned with the more practical side of computers. Both
are slightly older than Minsky (probably by about ) and 10 years respectively).

Bichi has a very solid reputation as a logician and has done very
interesting work in the area between automata theory and the more classical
type of logic. He has interest in the medical side, although I do not know
whether he has actually done much work in the area. He is not as lively
and outgoing as Minsky.

Schtitzenberger is both an M.D. and a recognized mathematician,
especially on semigroups and, in recent years, on the algebraic treatment
of finite state machines. I understand he has done extensive research in
genetics. Although he regularly teaches mathematics nowadays, he was a
Visiting Professor at the Harvard Medical School for 1961-62. He has a
formidable list of publications, I believe he is the oldest of the three,
probably around 45. His writings and lectures are not as clear as Minsky's.
He seems to have a very wide knowledge both in mathematics and in the
medical sciences.

If my understanding of your requirements is at all accurate, Minsky
seems to be the most suitable candidate among the three.

Yours sincerely,
“" Hao Wang )

HW: jm