OFFICE McMORANDUM © STANFORD UNIVERSITY © OFFICE MEMORANDUM ¢ STANFORD UNIVERSITY © OFFICE MEMORANDUM

To

From

SUBJECT:

MAR 2 2 1971

Date: March 19, 1971

Chuck Dickens

T. David Philli fe 20 Ron Jamtgaard Ki
(VA-€

XS

Approach to Analysis of Campus-ACME Facilities Merger

This Wednesday, we had lunch with Drs. Lederberg, Levinthal, and Feigenbaum.
Dr. Lederberg initiated the discussion with a brief recap of the Medical
School's computing requirements. Among the requirements were: (1) Strong
support for small machine users; (2) communications among various data

bases, (3) minimum disruption to ongoing research programs currently using
PL/ACME, and (4) cost-effective computing where the effectiveness is stressed
along with the cost.

Perhaps the most significant single topic to be discussed was the purpose

of the analysis of the Campus-ACME merger which Dr. Lederberg requested at
the last Computer Facilities Committee meeting. Dr. Lederberg stressed that
this analysis should be totally divorced from the issue of outside computing
dollars on the 360/67. The analysis should be based upon our mutual con-
cern of providing the most cost effective and responsive computing to the
university community.

Dr. Lederberg indicated that NIH was likely to continue to support computing
at Stanford at approximately the current dollar rate. He also stated that
NIH wanted to support specific development tasks rather than generalized
computing subsidies for Stanford Medical School. He felt that approximately
40% of the NIH funding could be spent for systems development activity, most
of which would be accomplished presumably through SCC. He stated their

were two other types of support that he would anticipate from the grant.
Approximately 30% of the grant might be spent for applications having a heavy
computer science flavor. The remaining 30% would be devoted to more clini-
eally oriented applications. Future income from the Medical School from
computing has been estimated at $250,000 - $400,000. Dr. Lederberg felt that
$250,000 was a reasonable and conservative estimate. It was recognized that
current ACME users obtain far more service for the dollar than they are likely
to obtain on the Campus Facility or any other facility in view of the heavy
subsidy of ACME by NIH. This fact would influence the amount of work which
would transfer to any other computing system if ACME were no longer available
or if the general subsidy were no longer available.

During the luncheon Ed Feigenbaum described several facets of the ARPA network
in which Dr. McCarthy's A.I. Lab participates. The network will soon include
a 10’ bit data store and hardware links among over a dozen computers operating
at 50 kilobaud. In addition, a new hardware linking device has been developed
to permit persons with only a terminal rather than a full computer system to
participate in the entire network. The spectre of a California or National

ALISM3AINN GYOANVIS @© WNONVYOWIW 3d1IO © ALISHJAINN GYOAINVLS © WNONVYOWIW 3Dd14IO © ALISYZAINN GYOANVILS «

WNONVIOWSW F3d1ddO e@
Chuck Dickens Page 2

T. David Phillips and Ron Jamtgaard

Approach to Analysis of Campus-ACME Facilities Merger

Computer was raised as a potential development of the next five years by
Feigenbaum. Lederberg questioned the wisdom of moving to any new gener-
ation of large hardware if the potential of the "California" system
appeared realistic over a five year period.

Dr. Levinthal presented an organizational concept which would link
computing and electronic engineering and development efforts in the Medi-
eal Center. Specifically, the new "division" would include a hardware
development group, a hardware service and maintenance group, and a Medical
Center group on computing which would handle applications programming plus
hospital information systems. Large computing services would be purchased
from SCC along with systems development activity.

Those present expressed early in the meeting a feeling that such a meeting
was overdue and a need for early planning of ACME's future existed now.
Some of ACME's planning must be contained in the annual report to be com
pleted in May, 1971.

RJ/ma
ce: J. Lederberg

E. Levinthal
E. Feigenbaum