MASSACHUSETTS INSTITUTE OF TECHNOLOGY
LINCOLN LABORATORY

LEXINGTON 73, MASSACHUSETTS

VOlunteer 2-3370

27 February 1962

Dr. Joshua Lederberg

Department of Genetics

Stanford University Medical Center
Palo Alto, California

Dear Dr. Lederberg:

Thank you very much for your kind letter of February 20,
1962. It was somewhat delayed because you will note that the
Lincoln Laboratory is in Lexington, Massachusetts, and not in
Cambridge.

The two reports that I sent you are all that I have published
on this subject. The rest of the data is still in the form of com-
puter print-outs.

At your suggestion I sent copies of my report to Dr. Robinson
in Washington.

Our work has been carried on at a very slow and part-time
pace at the Lincoln Laboratory, financed from our general research
funds. We believe that a more systematic attack is called for, but
we cannot do it without some serious support from a responsible
government agency. Since November 1961 we have had a proposal
with the National Science Foundation for a $50, 000 grant that will
enable us to do a serious engineering design of a model information
center. As of now, this proposal is still under consideration.

We may eventually come up against tough technical and logical
problems, but as of now the problem is almost sociological in nature.
How does one get a government agency to provide speculative "seed
money" for research into a new technological problem that is hardly
even defined. It is fairly easy to champion this or that specific device
or program and get some support for it. But if one states frankly that
one intends to start with only the broadest appreciation of function and
then let the components and programs evolve as the need arises, the
proposal is considered too vague. And yet those of us who have had
experience with complex systems design in other fields know that no
other approach has worked in the early stages of development. One
Dr. Joshua Lederberg -2- 27 February 1962

simply cannot design these things analytically and expect them to work
at the turn of a switch. One must first build a viable model, place it

in a realistic environment and through careful observation and monitor-
ing encourage an evolutionary process of growth.

If your panel is still in session, and if it accepts presentations
from nonmembers, I would welcome the opportunity to discuss my ideas
with you.

Sincerely,

JM: In. Qesslar__

MMK/jes M. M. Kessler