OFFICE MEMORANDUM " " OFFICE MEMORANDUM " STANFORD UNIVERSITY " OFFICE MEMORANDUM

Ln -41 z> -.1 O X DatE: March 6, 1974 o z To Distribution m< PO in -. From William F. Miller RECEIVED Vice President and Provost o "Tl Subject: . MAR 8 1974 22 o m COMPUTER SCIENCE DEPT. m o x Dear Colleagues: > Z o Attached is copy of the report of the Stanford University President's c a 3: 1973 Computer Science Advisory Committee. I am enclosing the sub-committee reports, but not the cover letter which was sent by the Chairman to the Ln President. The cover letter was very brief and contained one or two comments -4 which I believe were intended for the President alone. As stated in the z> cover letter, the Chairman expects us to distribute the reports of the o"tl x sub-committees. o c Best regards, Z m< X y 7-AA'^A Ln William F. Miller o "n ■n ri Distribution: R. Atkinson m C Dickens R. Floyd m G. Franklin O J. x Herriot Z> v W. Massy c M. Roberts H. Royden

Ln -4 z> oT. D z m< x Ln ~: -<

O Tt Tl o m

m o x > Z v c S

C

C * « Appendix A

Stanford President's Computer Science Advisory Committee Annual Meeting

October 28-30,1973

Attendees

Dr. Bruce W.Arden * Dr.Frederick P. Brooks Dr.William H.Davidow Dr. Andrew H.Eschenfelder Dr. S.Fernbach Dr. Juris Hartmanis Prof. Richard M.Karp Dr. Donald A. B. Lindberg Dr. Kay B. Magleby Mr.Richard G.Mills Prof. James McKenney Dr.Richard L.Shuey Mr.Robert Taylor Prof. Joseph F.Traub Prof. Keith W.Uncapher Dr. Ronald L.Wigington Appendix B

To: K.Uncapher and K.Magleby i From: J. F.Traub

Subject: Report ,on Sub-Cornny ttee on Computer Science

Bob Floyd feels the two biggest problems are:

I. Fragmentation

2. Over-extension » v " 1. Fragmentation - There is a long_ history of fragmentation in the Department. Floyd is trying to remedy some of it with regular lunch meetings.

At least part of the fragmentation is due to the physical scattering of faculty and students, and there was a feeling that both faculty and students pay a price for this dispersion. Space in the quad which would include everything but lab space seems desirable from several points of view (economic and good central location).

It's our understanding that this' space wi I'l not be ava i lable' for at' least five years. It is also our understanding that room may become available in Polya Hall and/or Pine Hall. We strongly urge the ad- ministration to give interim space to the Computer Science Department until their quad space is ready.

2. Over-extension - Floyd reported that the faculty is severely over- extended by both its undergraduate teaching load and its graduate students. Most of its teaching is to students outside the Department and cannot be curtailed without harm to the University. Graduate ad- t \ missions have to remain high in order to have teaching assistants for the courses and research assistants for the contracts which bring sub- stantial overhead to the University.

The Visiting Committee felt that the average faculty load was far greater than at other departments with which the Committee is familiar For example, the graduate student/faculty ratio is 11, whereas 5 is typical at other top departments. Although the Department is still one of the several best in the , Floyd and the Committee are concerned about the future. Floyd feels he's living off his re- search capital. Some members of the Visiting Committee feel that some of the recent theses have been of poor quality. 2

Appendix B (cont.)

Of course, the pressure could be relieved by cutting back on obli- gations. The Department feels, and the Visiting Committee concurs, that this would be a disservice to the University, the discipline, and society. Reasons include:

1. There is strong application pressure on the Department

2. There is heavy demand for the graduates of the Department

3. There is heavy demand for service courses

k. There is a need for additional depth and breadth in the coverage of computer science. V

5. Because of its expertise in computing, the Department could be of service to other components of the University if it were not so heavily overloaded.

The Visiting Committee recommends that the Department develop a quantitative model of how it should grow. Parameters of the model should include faculty size, undergraduate and graduate service loads, number of computer science graduate students, computer science student/ faculty ratios, and size of contract support. One advantage of the model will be that the Department will have to write down facts.

Floyd discussed which areas needed coverage. He felt, and the Committee agreed, there were needs in:

Languages and translators

Numerical analysis

Formal theory, probably computational complexity

There was a discussion of appointments in areas which the Department does not currently view as core. The Committee urged Floyd to review how de- cisions about new positions should be made by the Department in this time of l imi ted growth.

The Visiting Committee is aware that Stanford is in a period of essen- tially no growth. However, we still urge the administration to assign additional positions to the Department for the reasons given above.

It was suggested that a meeting might be arranged between certain members of the Committee and Deans Atkinson and Royston in order to bring the need for additional slots to the attention of the administration in as strong a way as possible. Appendix C University of Missouri - Columbia

-■_!

Lewis Hall SCHOOL OF MEDICINE Telephone Mo. 65201 Information Science Group 314-8.2-6*56

November 6, 1973

Mr. Keith Uncapher University of Southern California

Information Sciences Institute t Admi ral ty Way Marina Del Rey, California 90291

Dear Kei th :

Below is my summary statement of the observations and recommendations of the subcommittee on computing of the Computer Science Advisory Committee to the President of Stanford, following our meeting on October 28-30, 1973-

I. Thanks are due to the panel members who gave their time to the subcommittee effort. These were:

Sid Fern bach Dick Mi 1 Is Dick Shuey Kei th Uncapher Ron Wigington

Thanks also are due to Chuck Dickens and his staff for the excellent background documents and their patient presentation of SCIP plans.

2. Recommendations We noticed a tendency within SCIP — doubtless because of the difficulty of their task and number of their current problems -- to think in two year terms and likewise a tendency to act like a computer center. In both cases, we felt the most appropriate stance was slightly different: namely, to plan with a five year horizon and to act and plan more like "computing at Stanford" than as a computer center.

3. There are a number of operating guidelines which derive from this bas i c pos i t i on.

A. First, we recommend that SCIP explore actively the possible

Columbia, Appendix C

Mr. Uncapher November 6, 1973 i» Page Two

advantages for selected applications of an increased number of mini-computers, interconnections between these machines and larger University machines, and the extent to which SCIP might become involved in the future in the architecture of information systems as manifested in such a hierarchy . of computers.

We are not recommending any particular configuration for any particular SCIP job; just urging that their approach to network solutions be very positive and aggressive.

B. We discussed the question, should SCIP serve as broker to Stanford for computing services from the world? We agreed that SCIP should indeed try to do this function whenever possible and reasonable.

C. Within this overall planning framework, we discussed the question of SCIP's attitude toward the SLAC 91 and 2-168's. We were cognizant of the danger that computing 1/10 as expensive as campus rates would tend to draw off "business". Yet we felt that SCIP must during the coming 2-3 years pursue its present plan to facilitate the shifting of jobs between its campus 158 and the SLAC complex. Thus it would eliminate any technological bar to utilization of this excess capacity.

Whatever political bars may exist to establishing a reasonable billing basis for SLAC computing should be addressed as a reasonably high priority issue -- perhaps by Associate Provost Franklin.

k. We had some recommendations concerning the more immediate future.

A. It seems absolutely essential that SCIP develop a means to make absolutely clear the limits of its responsibility for applications design, programming, and maintenance in each of its ventures.

This might take the form of contracts and specifications, user acceptance tests, or whatever. In some sense it would be worthwhile and a great protection to offer what amounts to a tariff of services. Appendix C

Mr. Uncapher November 6, 1973 Page Three &

B. This issue becomes especially important in our next recommendation for the short term. Specifically, we urge that SCIP publish a statement announcing its intention to support the interconnection of a list of mi n i -computers , to spell out very clearly what services are to be available to these mi n i -computers , and to set up a time table, and certain other actions to be noted later.

C. SCIP should make some public demonstration that makes it - clear to Computer Science graduate students that SC IP's attitudes are now quite different from those in the past and that Computer Science students are now welcomed into SCIP facilities and that there have already been opportunities in some cases actually to study and to change operational systems, code, and procedures -- all under the direction of SCIP management, of course.

5. SCIP must do some things in the immediate 18-2^l month period which demonstrate progress, encourage support, and generally take the heat off while the major work of Operating System building is going on.

A. Above all else, SCIP should continue to view new computer systems and research applications as new opportunities to broaden the spectrum of computing at Stanford and as potential ways to broaden services to the Campus. An example is the potential TIP computer on the Feigenbaum/ Lederberg machine which will essentially purvey Stanford services to the ARPA world. SCIP could conceivably be able to participate with Ed and Josh so as to get a benefit for the campus from this connection with other institutions.

B. SCIP should consider picking some early winners, however trivial, to back. That is, announce and meet implementation schedules on some jobs — perhaps administration or hospital. C. Meet personally with proprietors of the kO mi n i -computers , perhaps on the pretext of discussing interconnections, but really because of the need to make clear to the campus public SCIP's concern for users and intention to be of help.

6. We note with regret that working interrelationships between the Computer Science Department and SCIP, while most cordial, are not intimate. There surely is much to be gained by both groups through a close working . Appendix C

Mr. Uncapher November 6, 1973 Page Four

relationship and common definition of problems and objectives, a situation which we recommend. On the other hand, vie note that both groups — as with most universities -- are quite busy as 'well as self contained. The situation unfortunately is common.

7. We note in passing that there was not time nor opportunity to study or evaluate Stanford's plans for a new network of Ik computer terminals in the Department of Computer Science and a possible extension to a total of 200 to include other student users. Certainly we did react to this news with the hope that these kinds of developments would be integrated into thinking on behalf of un i vers i ty-wi de plan for computing at Stanford. We cannot record a more cogent recommendation because of lacking information on the project. Yet we ought to note that we did not find it intuitively obvious that provisional plans to attach such a net to the existing 3&0/67 would likely be entirely reasonable.

8. In closing, we unanimously record our thanks to Chairman Uncapher for the preliminary investigation and planning which preceded this meeting. Unquestionably this is what made our efforts reasonably productive. We endorse continued vigorous surveying of activities by Kay Magleby previous to next year's meeting.

Respectfully submitted on behalf of the Subcommittee on Computing by.

7 t ] ■; __) /■J S? 3*~^ i Donald A. B. Lindberg, M.D. Director, Information Science Group Professor of Pathology

DABL/jg

cc: Doctors Magleby, Fernbach, Shuey, Wigington Mr. Mills