( Df ‘ 9 R
i? Le i a y
SCHOOL OF MEDICINE

STANFORD UNIVERSITY, STANFORD, CALIFORNIA 94305

JOSHUA LEDERBERG . September 19 » 1975
JOSEPH D. GRANT PROFESSOR .
oF GEXNETICS

Dr. Robert W. Floyd
Chairman

Computer Science Department
Stanford University
Stanford, California 94305

Dear Dr. Floyd,

I am havpy to have this occasion to record my unqualified endors sement of
Dr. Bruce Buchanan's reappointment in the rank of Adjunct Professor of Computer
Science. This status is a fairly new one at Stanford, and may be unfamiliar to
some members of the department; it may be recalled that the title hes replaced
that of Senior Research Scientist (old-style), that it is a term eppointment
and also coextensive with available project funding, but that it hes faculty
prerogatives in a number of areas and was intended to be reserved for candidates
of high distinction and national reputation. I am familiar with a number of
holders of this title in the Genetics Department and elsewhere in the School
of Medicine, end it is cleer that Dr. Buchanan compares favorably with any of them.

My association with Dr. Buchanan has been a close one for alrost ten years,
in the framework of the DENDRAL project and in association with Dr. Feigenbaum.
Anyone who has observed this effort has been impressed, if by nothing else, with
its success as an interdisciplinary undertaking -- it involves particivation
by several people each from the departments of Chemistry, Genetics and Computer
Science, and within the latter of a range of contributions from LISP-hecking to graph
theory. While Dr. Djerassi, Feigenbaum and I will wish to take some credit for
motivating this cooperation, it is but fair to say that it simply could not have
worked without Bruce's breadth of understanding clarity and steadiness of thought.

These are indeed Dr. Buchanan's special qualifications, and they ere
particularly notable in a field like AI which may suffer as much from the
exuberance and shortfall of critical judgment of some of its proponents as it has
gained from the brilliance and inspiration of its pathfinders. Bruce does not
display the level of innovative flash that one might seek as one of the criteria
for appointment to the reguler faculty and autonomous leadership of a distinctive
area of research, although I suspect he would not be lacking in this mode were
he to move to a different environment. As it is, he has been a perfectfit to the needs
of our ongoing effort, and he has evidently found this also to be en excellent
milieu for his own creative productivity. His fundamental interests are, after all,
in the philosophy of the scientific method; and the DENDRAL project has been a
unique context in which to put philosophical ideas to practical application as we
see them tested by working scientists in a variety of fields. I heave acauired the
deepest respect for his wisdom and oversight over the entire project, end believe
Dr. Robert W. Floyd -2- 9/19/75

he will continue to make important contributions both to the substance of
our efforts and to our understanding of the underlying issues involved in
the mechanization of scientific discovery and inference.

Useful results from AI work are usually relegated to engineering, once
they have been achieved, and academic eminence in the field is usually founded
on appreciation for contributions to mathematics, logic, or linguistics.
Bruce's forte along these lines is philosophical analysis, but coupled, as I
have indicated, with an extraordinary ability to relate to a wide range of disciplines
in surefooted, persuasive (and non-controversial) fashion. Consider how his
published work has related to mass spectrometry, to graph theory, and to the law.

I am at a loss whom to look to for plausible comparison. N.J. Nielsen's name
comes to mind as someone who has exhibited a comparable level of critical over-
sight of the field of AI, but has been undoubtedly less productive in terms of
explicit applications (in a field that badly needs them to sustain its
credibility). Bruce's role in the actual work of the DENDRAL group in the
automation of chemical reasoning is faithfully documented by his place in the
authorship of innumerable publications, and his reputation among philosophers
by other letters in his file.

For the programmatic task which justifies an appointment to be filled
by this rank, I am unable to suggest anyone of competing stature -- in a few
years, perhaps, some of the junior collaborators now connected with the project
might be thought of (Dennis Smith, Nancy Martin) -- but none of them would be
nearly competitive at this time.

Sincerely yours,

 
   
  

Shua Lederberg
ofessor of Genetics and
(by courtesy) of Computer Science