Federal Funding of Computer Science A Preliminary Report
Joel S. Yudken and Barbara Simons
CPSR/Palo Alto

Dr. Joel S. Yudken is Director of the Project on Funding Policy in Computer
Science. Barbara Simons is vice-chair of the Special Interest Group on Automata
and Computability Theory (SIGACT) of the Association of Computing Machinery
(ACM). Dr. Simons is also chair of the Committee on Scientific Freedom and Human
Rights of the ACM, and she is a member of the Palo Alto chapter of CPSR. This
article is adapted from a paper presented at the conference on Directions and
Implications of Advanced Computing, sponsored by CPSR/Seattle in July.

The General Problem Reconsidered:
Federal Funding and Science Research

Historically, defense agencies have provided the bulk of funding for research
and development (R&D) in the United States. Since World War II, the defense
portion of federal R&D funding has rarely dropped below 50 per cent. However,
the 1980s appear to represent a departure from the pattern since 1965. Beginning
in 1979 the defense portion of the federal R&D budget jumped from 48.4% ($13.6
billion) of total R&D federal obligations to 71.9% ($46.6 billion) in FY 1988
(estimated). In constant 11982) dollars this is equivalent to an annual growth
rate of 9.1 %. Meanwhile, civilian R&D obligations declined in constant dollars
at a rate of 2.3% each year. They now account for less than 30% of all federal
R&D spending.

Concern about the implications of these trends has been expressed by various
members of the scientific and engineering communities. Most recently, the
Strategic Defense initiative (SDI, or "Star Wars") has sparked criticism from
scientists, including some who are otherwise comfortable receiving DoD research
funding. In addition, a number of researchers view the acceptance of funding as
an endorsement of the goals and policies of the funding agency. Consequently,
some express moral and ethical qualms about taking military grants, even if the
actual military applications for their research appear remote.

Questions have also been raised about the effects of federal R&D funding on the
direction, quality and nature of science research in the U.S. Specifically, has
the prominent role of defense agencies in R&D funding since World War II had a
stimulating, distorting, or stunting effect on the advancement of scientific
foundations? Have certain fields suffered while others gained? Have some
branches within fields been favored while others are left underdeveloped?

Another area of concern involves the social uses and impacts of scientific
discovery and technological innovation. If most of the funding for developing
the nation's technology base is controlled by a small number of government
agencies with specific mission objectives (DoD, NASA, DoE), does this preclude
important scientific and technological advancements that serve other areas of
civilian economic and social need?

Finally, there are questions specifically related to academic research, which is
funded primarily by federal agencies. What is the impact of federal funding on
the work performed in the universities? Does the funding create an atmosphere
that advances or hinders the development of science in these institutions? Can
funding affect other areas of university life such as department sizes and
tenure decisions? Are professors being pressured by their universities to bring
in "large" grants? Are students forced to work in particular areas because most
graduate student support comes from faculty grants?

Federal Funding and Computer Science Research

All of these issues are being debated within the field of computer science (CS).
CS has historically been heavily funded by defense agencies. As a 1982 DoD
report states, 'the military power of the United States is inextricably tied to
the programmable digital computer." Computer-based technologies are also of
great importance to society. Therefore, the question of how to best fund CS is a
significant science policy issue.

The funding of CS research has increased considerably over the past decade. Yet
there has been concern expressed within the CS community that research funding
is not keeping pace with the even faster rise in demand for CS advances,
applications and educational programs within our society. A 1985 Computer
Science Board study2 and a more recent National Science Foundation report3
argue, for example, that the current pattern of federal funding may be
inadequate to support the increased levels of CS Ph.D. production necessary to
meet projected supply requirements for adequately trained CS personnel in the
1990s. This can be called the adequacy of funding problem. That is, do the
current CS funding policies of federal research agencies adequately support the
real needs of CS?

There has also been literature dealing with a second set of issues which can be
called the role of DoD problem. For example, Clark Thomborson (formerly
Thompson),4 from the University of Minnesota argues that the substantial growth
of DoD's role in CS research funding distorts the direction of the field,
inhibits advances in scientific foundations, threatens scientific and academic
freedom, and limits the potential for much needed CS research "spin-off" for
commercial and other social uses. Others have also singled out the DoD with
special focus on the Strategic Computing Program (SCP) and SDI programs.5

Proponents of DoD sponsored CS research note the pioneering role of the DoD in
the early development and use of the computer. DARPA, in particular, is viewed
as a farsighted and effective sponsor of CS research, with some of its programs
being credited for many important advances in the development of civilian
computer technologies.6,7 Yet, some advocates, as well as opponents, of DoD
funding have raised concerns both about the greater directedness and mission
orientation of DARPA's SCP program and about the potential impact of the SDI.

The Project on Funding Policy in Computer Science

In December 1986, the Science Policy Committee of SIGACT initiated a study on
the question of funding in CS. This study has subsequently been endorsed by the
chairs of all the SIGs of the ACM and has received additional endorsements from
the SIGBOARD of ACM, the executive committees of SIGACT and SIGGRAPH, the
Association for Women in Mathematics executive committee, and the national board
of Computer Professionals for Social Responsibility.

Its stated goals are:

(1) To perform a detailed study on the impacts of federal funding on the
pattern, direction and quality of current and future research in academic CS in
the U.S;

(2) To publish, widely disseminate and publicize a report on its findings. This
will be made available to the scientific community, the U.S. Congress and the
general public;

(3) To use the report to catalyze support for Congressional hearings on federal
funding of scientific research.

Information for the study is being drawn from official government funding data
series, published literature and documents, and interviews. The research
involves a review of literature, analyses of statistical funding data, and an
interview survey of academic researchers, CS department chairs, government
officials, and industry representatives. A specially designed interview format
for this study is currently being field tested.

The study is still in its early stages. Some of the preliminary findings,
presented below, center around four related problems.

The Problem of Funding Levels

To begin, we examine how federal CS funding is allocated according to types of
research (basic, applied) and subdisciplines within the field. There are
methodological difficulties, however, with interpreting the available data.

The standard source of federal funding data is the National Science Foundation
(NSF) series on federal funds for R&D based on its annual survey of federal
departments and agencies 8 In this series CS does not appear as a separate
category until 1976. Before that, funding for CS and mathematics was reported
under a single classification. Another source of confusion is presented by the
"Mathematics and Computer Sciences, Not Elsewhere Classified (NEC)" category,
which shows a sharp jump after 1984. Thomborson12 notes that this anomaly
results from the funding of the DoD's Strategic Computing Program. In our
analysis, Thomborson's convention of combining the NEC category with the
"computer sciences" category is used to obtain total CS funding figures after
1976.

Further data is being gathered from other reports1,3,10,11, the NSF's yearly
Summary of Awards, several journal and media articles, and communications with a
number of individuals. These data often do not correlate with the
classifications and magnitudes in the NSF series. There is ambiguity in the
definitions of some research categories e.g., what is basic and what is applied
research? What is CS research and what are CS applications to other fields? A
full review of these differences is not possible here, although the study will
attempt to examine and incorporate them into its final report. Unless otherwise
indicated, therefore, the figures reported below are derived from the NSF
tables.

The uncertainties in the data notwithstanding, the following trends and patterns
of federal CS research funding are readily apparent.

(1) Most obvious has been the dramatic rise in federal funding for both total
and academic CS research between FY 1976 and FY 1987.

(2) Although all major government agencies increased their CS funding in this
period, the DoD had the largest gains, was responsible for the greatest portion
of these increases, and currently accounts for the overwhelming majority of all
federal funds for CS.

(3) Funding has shifted away from basic and towards applied research, both in CS
federal funding as a whole, and within academic CS. Defense programs accounted
for most of this shift.

Although we are still evaluating the relevant data, in recent years there appear
to have been funding increases in almost all CS subdisciplines. However, some
subdisciplines are growing more rapidly and obtaining a larger share of funds
than others. For example, the more substantially funded areas include machine
intelligence and robotics, machine architecture, distributed computing and
software systems, while theoretical computer science and computational
mathematics are less heavily funded. Once again, defense agencies are
responsible for most of the increase in the faster growing areas.10

We now look more closely at some of the specific data. In current dollars, total
federal funding of CS research increased from $87.9 million in FY 1976 to an
estimated $544.6 million FY 1987. In constant dollars this represents an
increase of more than 300%. Similarly, total federal funding of academic CS
increased from $27.5 million (current dollars) to $227.6 million in this same
period.

Figure 1 illustrates total CS research funding from the main government sources.
We note the DoD's role in CS research has grown very substantially since 1976.
In budget dollar terms DoD's overall CS funding has grown from $60.1 million in
FY 1976 to $394.6 million (est.) in FY 1987. The DoD's share of funding
fluctuated roughly around the 70% mark. The NSF's share, however, has been cut
almost in half, dropping from 18.4% ($16.2 million) in FY 1976, to 9.8% ($53
million) in FY 1987 (est.). Interestingly, although the DoE has consistently
supplied 2-3% of CS funding, NASA's share jumped appreciably in 1982. NASA has
provided about 10% of overall CS research funding since 1982, making it a larger
funding source for overall CS than the NSF.

Looking at CS basic and applied research funding we see another important change
over the last decade. In FY 1976, basic research funds accounted for 56.1% Of
total CS research funds; in FY 1987 it dropped to 26.6%. Correspondingly,
federal funding of applied CS research jumped from 43.9% to 73.4% in that same
period.

The DoD alone accounted for 74.3% of the total increase in CS funds since 1976.
We find that 68.5% of this same total increase came from DoD's funding of
applied research, primarily for programs such as SCP, the Software Engineering
Institute, STARS, and others. The per cent of total DoD CS funding for applied
research rose from 69.1% to 85.4%, while the percentage for basic research
dropped from 30.9% to 14.6%. The DoD's share of total federal applied CS
research funding, however, has fluctuated between 7085%. DoD's portion of CS
basic research funding dropped from about half in 1976 to under 40% in 1987.

Because these numbers do not incorporate SDIO (Strategic Defense Initiative
Organization) sponsored CS research, which is officially classified as a
development program, the DoD figures may be understated. Evidence suggests that
some CS work funded by SDIO should be classified as applied or even basic
research. We also know that some research, originally sponsored by other DoD
programs, has been reclassified as SDI research, perhaps unbeknownst to some of
the researchers. Another unknown is the nature and amount of CS research funding
from the National Security Agency.

Although the NSF has increased its total CS funding levels by 171 % since 1976,
it has not kept pace with the growth in the field as a whole or with other
funding agencies. Almost all NSF funding goes to basic research. Over the past
decade, it has provided 35-42% of all federal funds for basic CS research. But,
it has accounted for only 6-7% of the increase in overall CS funding since 1976.

As Figures 2 through 4 illustrate, the funding patterns for academic CS research
also exhibit significant shifts both in types of research performed and in the
relative roles of different funding agencies. The proportion of federally funded
academic CS research to total federal funds for CS has grown from 34% in FY 1982
to 41.8% in 1987. The DoD's share has risen markedly from 44.6% ($12.3 million)
in FY 1976 to nearly 71% ($161.3 million) in FY 1987 (see Figures 2 and 3).
Meanwhile, NSF support for academic CS has dropped 50.5% to 21.2%, although in
current dollars its funding has risen from $13.9 million to $48.4 million. The
NSF's share of academic basic CS funding, which had been 68.3% in 1976, has been
fluctuating between 35 to 50 per cent since 1982 (the FY 1987 figure is 48.9%).

The most significant change in academic CS funding has been the great rise of
applied research support. The proportion of basic to applied research funding
gradually rose from 1976 to 1983. After 1983 the trend then reversed (see Figure
4). In FY 1983 academic CS funding was 78.5% basic and 21.6% applied; in 1987
42.3% was basic and 57.7% was applied. Once again DoD funding accounts for most
of this shift (see Figure 3). Its share of academic CS applied research, which
was always high, grew from about 85% to 97% between FY 1982-87. But the portion
of DoD applied CS research of total federal funds for academic CS' research
jumped from a low of 18.4% in FY 1983 to 55.8% in FY 1987. The NSF's share of
applied research is about 1%, DoE's is non-existent, and NASA recorded 2.2% in
FY 1987.

Academia's role in CS has always been important. Its share of total CS federal
funding increased by approximately 25% since 1982. But its share of federal
support for basic CS research dropped from 78.1% to 66.4%. Meanwhile, more of
the nation's applied CS research is performed in universities and colleges.
Academia's share of CS federal applied research funding has increased to 32 9%
in FY 1987, up from 11 .8% in FY 1982.

The Problem of the Direction of CS Research

The shift in funding towards applied research in CS raises important questions
regarding the future direction of CS research. Because of space limitations and
the fact that we are still in the process of collecting and evaluating data, our
findings on this and other such questions can only be outlined here.

Although this shift towards funding applied research is of serious concern to
computer scientists, its actual effect on CS as a discipline remains an open
question. As we saw, DoD funding is largely responsible for this trend. However,
many researchers feel that the DoD, especially DARPA, has been an enlightened
funder to date, allowing the scientists it funds to largely dictate the
direction of new research. On the other hand, DARPA's SCP and SDI programs have
raised concern that DoD sponsored research is beginning to be more mission- and
applications-oriented than in the past.

There is also concern that certain types of research, such as theoretical CS,
are being short-changed by the emphasis on more applied areas. It has also been
suggested that some people who should be funded are not receiving adequate
support. Evidence from our initial interviews and other sources lends some
support to these concerns.

Understanding the relationship of funding sources to how research choices and
grant selections are made is important for ascertaining the link between funding
patterns and a field's research direction over time. We are examining this
question through an interview survey of CS researchers, department chairs, and
funding agency officials. There seem to be important differences between the
funding processes of the NSF and the DoD agencies in terms of grant sizes,
application and selection processes, stability of funding, project monitoring,
and the attention paid to networking and research infrastructure. The peer-
review process of NSF versus the "old-boy" insider network of DARPA (see note
6), for example, leads to different research emphases and distribution of funds
among researchers and research institutions.

The predominance of DoD funding in CS may also have an effect in terms of
researchers not funded by the DoD. Since some researchers who feel uncomfortable
receiving DoD grants have trouble finding alternative sources, an indeterminate
number of otherwise competent or even superior CS professionals may be limited
in their work, or end up leaving the field itself. Some examples of this have
already come to our attention, though the data is still anecdotal.

Related to this is the issue of academic freedom and the creation of an
unhealthy environment of mistrust or cynicism within the academic CS research
community. For example, there are indications that some researchers misrepresent
the nature of their work in order to obtain funding from mission-oriented
sources.

The spin-off issue is another important problem that we only touch upon in the
study. Despite much evidence that DARPA's early work has been very beneficial
for the commercial computer industry, 6,7 others contest this view.4,11 For
example, the Department of Commerce's U.S. Industrial Outlook 1987 reports that
little of DARPA's supported supercomputer work has found commercial use to date.

The Problem of Supply

The supply of qualified CS researchers and professionals is another problem we
plan to investigate further, primarily updating the findings of the Computer
Science Board's study2 and the more recent NSF report.3 These studies argue
that, although funding levels have increased, overall Ph.D. production has not
kept pace. A third important federal reports also supports the idea that there
are too few Ph.D.'s to meet the rapidly growing demand for CS personnel.

Studies [2] and [3] couple the training of CS researchers to basic research
funding. They claim CS basic research is not sufficiently supported at this time
to meet the needs for general scientific advancement or for training new
generations of workers in the field. Therefore the current shift towards applied
rather than basic research in the universities may have negative long-range
implications, a worrisome pattern .

The Computer Science Board study also pointed to the differential impact of
funding from federal agencies. Although the DoD is the primary source of CS
support, its funding is "strongly concentrated in the four top-ranked
departments," the report states. It also argues that this is not the best
situation for producing the largest number of new qualified people. The NSF's
peer-review process leads to a different distribution of funding, which is
"concentrated outside the top four ranked departments and is fairly smoothly
distributed among academic departments, but weighted toward the highly ranked
departments. Such a distribution, the report concludes, "probably has maximal
impact for increasing Ph.D. productivity."

The Problem of Control

Our final goal is to evaluate how the different funding agencies set the
direction and priorities for CS research. Does the DoD's role as primary funder
of CS give it inordinate power over the development of the CS field, or is it a
benign force that stimulates, without dictating directions? For example, a major
CS department chairman whom we interviewed expressed concern about the potential
implications of the DoD's huge role in CS funding. On the other hand, he
qualifies his view by noting that he can not say that the DoD's role has not
been beneficial to date.

Recognizing that a wide range of people consider a balance of mission- and non-
mission-oriented funders desirable, especially with respect to basic research,
what are the implications of the recent funding patterns and policies of the NSF
for the field? We know for example that its funding levels for CS research are
increasing again, although much more emphasis is being placed on experimental
work. This and other questions will be examined at a later date by this study.

References

1. Special Analysis J. Research and Development, Office of Management and
Budget, Executive Office of the President, January 1987.

2. David Gries, Raymond Miller, Robert Ritchie, and Paul Young, Imbalance
between Growth and Funding in Academic Computer Science: Two Trends Colliding,
Computer Science Board, Committee on Research Funding in Computer Science, April
9, 1986.

3. S. Rao Kosaraju, Michael J. Fischer, and David Gries, "Meeting the Basic-
Research Needs of Computer Science,'' Study Group of the Advisory Committee for
Computer Research, National Science Foundation, December 16,1986.

4. Clark Thomborson (formerly Thompson), "Military Direction of Academic CS
Research," Communications of the ACM, 29, 7 (July 1986),583Ñ585.

5. Terry Winograd, "Strategic Computing Research and the Universities," Working
Paper No. 7, Silicon Valley Working Group, University of California, Santa Cruz,
March 1987.

6. Ronald G. Havelock and David S. Bushnell, Technology Transfer at DARPA, The
Defense Advanced Research Projects Agency: A Diagnostic Analysis, Technology
Transfer Study Center, George Mason University, Fairfax, VA, December 1985.

7. Dwight B. Davis, "Assessing the Strategic Computing Initiative," High
Technology (April 1985), 41 -49.

8. Division of Science Resources Studies, National Science Foundation: Federal
Funds for Research and Development, Federal Obligations for Research to
Universities and Colleges by Agency and Detailed field of Science/Engineering:
Fiscal Years 1973-1987, Federal Funds for Research and Development, Federal
Obligations for Research by Agency and Detailed Field of Science/Engineering:
Fiscal Years 1967- 1987, Federal Funds for Research and Development, Detailed
Historical Tables: Fiscal Years 1955-1987.

9. John R. B. Clement, "Computer Science and Engineering, Support in the FY 1988
Budget, " AAAS Report XII, Research and Development FY 1988, pp. 251-262,
Intersociety Working Group, American Association for the Advancement of Science,
Washington, D.C.,1987.

10. Report of the Federal Coordinating Council on Science Engineering, and
Technology Panel on Advanced Computer Research in the Federal Government, Robert
E. Kahn, panel chairman, June 1985 (released November 15,1985).

11. L. Bruckner and Michael Borrus, Assessing the commercial impact of the VHSIC
program, Working Paper, Berkeley Roundtable on the International Economy,
University of California at Berkeley, December 1984.

Further Conversation with Robert Kahn
Part 2: Computer Science and The Department of Defense

This is the second part of a two-part interview with Dr. Robert Kahn, currently
president and founder of the Corporation for National Research Initiatives
(NRI), and former Director of the Information Processing Techniques Office of
the Defense Advanced Research Projects Agency (DARPA). The first part of this
interview covered the mission and structure of NRI, which Dr. Kahn described as
a national effort to fund and manage research and development for Òinformation
infrastructure" in the United States. The second part of this interview, set
forth below, focuses on Dr. Kahn's tenure at DARPA, and his involvement with the
Strategic Computing Program, a ten-year, $1 billion research and development
agenda investigating applications of advanced computing for conventional weapons
and support systems. This interview was conducted by Gary Chapman, Executive
Director of CPSR, in late March, 1987, in Los Angeles, California.

Some people have speculated that you left DARPA and started your new venture,
National Research Initiatives, because you thought DARPA was, perhaps, moving in
the wrong direction. Is that a fair assessment?

DARPA is really an organization of people, and it's the people in DARPA who
determine what the organization does. Sometimes the choices within DARPA are
affected by the context in which it operatesÑthe environment determined by the
Congress, or the DoD or the Executive branch. But I would say the first order of
what DARPA does is determined by the individuals within it. It was set up to be
relatively independent within the Department of Defense, specifically to
maintain a technological vigil for the country.

I was a key player in defining what DARPA did. So, to a certain extent your
question implies that there may have been others who were influencing us from
the outside.

There have been rumors that a change in personnel in the management strata
immediately superior to DARPA personnel, at the Undersecretary of Defense for
Research and Engineering (USDRE) level, had produced views about what DARPA
should be doing that were different from those of the past.

I can only speak about the time I was there. For the last several years I was
there the Undersecretary was Richard DeLauer; prior to that it was William
Perry, and just as I was leaving Donald Hicks came in, but we only overlapped
for a very short time and I can't really comment on his tenure. During the time
I was there, l would say that the perception of DARPA by the USDRE was
relatively steady. I didn't see any major changes in perception, and the
undersecretaries were always supportive of DARPA as a lead research
organization. Now, since I have left, l have heard various reports about
changes, but I haven't had any direct personal experience other than what I've
just described.

Well, we'll ask you straight out then: why did you leave DARPA ?

I actually planned to leave DARPA, or I had made up my mind to leave DARPA, many
years before I actually did. And each time I was close to leaving I had my arm
twisted by the Director to stay on for another short period of time. In 1979
when I became the Director of the Information Processing Techniques Office, l
had suggested to the Director of DARPA that it might be time to find someone
else to come in and run the office, because I felt I had been there long enough.

When did you start at DARPA?

I started there in October, 1972.

And you left inÑ

Toward the end of 1985. That's a long time for any person to be at DARPA. My
original plan was to be there only two or three years.

In any case, the main purpose of my leaving was to start this new organization,
NRI. You see, we had just started to develop the Strategic Computing Program
(SCP)Ñ

This was started before you left DARPA.

Yes, several years before I left. I had agreed to stay on to help get that
program started. As the program was developing, the budget for research work on
directed energy weapons was transferred to the SDI program. This put the
information processing program budget on a track towards a third of the total FY
1986 DARPA budget. My prognosis was that this was about as large as it was ever
likely to get. Even though I felt that more funds could be allocated
productively to the program, it seemed unlikely that the Congress would
appropriate large increases to the DARPA budget, and, furthermore, working with
a fixed budget, DARPA had many other things to fund. It just seemed to me that
having developed this program to that stage, the remaining requirement was
effective management, and there were other people in the country capable of
doing that quite well, while there was no one who had attended to this problem
of infrastructure that we have discussed. So those were largely my reasons for
leaving DARPA.

So you were carrying around this idea about the infrastructure problem for some
years while you were at DARPA.

I'm not sure it was the infrastructure idea, but the idea of some kind of
organization in the private sector that could attend to the funding of long-term
research and development, that could address some of the longer-term issues that
could not be addressed as effectively through military research channels. While
military investment in R&D is absolutely essential to DoD and the country, there
are other areas that are also important to address from a national perspective.
The infrastructure idea came out of a number of conversations I had with a
variety of people around the country, and it seemed to supply some necessary
context to this concern for long-term funding of research and development.

One of the concerns of CPSR and many other people in the computer science
community is that the percentage for computer science research that comes from
the Department of Defense has steadily risen. Current figures suggest that about
60 per cent of computer science research funding comes from DoD, and that at
some of the top schools like Stanford, MIT, and Carnegie-Mellon, the military
funding levels can be as high as 75 to 80 per cent. In addition to this trend is
the concern that the research priorities of DARPA have shifted toward
development and applications and away from basic research. Many people are
worried about the distortion of computer science research by explicit military
goals reflected in funding priorities. Is this something that should be a
concern, in your view?

I think we've entered a period where there really are some serious concerns. And
I would lay them out as follows: The purpose of the Strategic Computing Program
was to address a set of issues in computing technology having to do with
developing an industrial base for artificial intelligence, getting the kind of
multiprocessor support that would enable that technology to come to fruition,
supporting work in VLSI that might be necessary, and so forth. What we were
trying to do was to get industry involved in a significant way, and especially
to get industry coupled with the good work that was being done in the
universities. On the other hand, supporting industrial R&D is a good deal more
expensive than supporting university research. But DoD needs industrial
participation in order to get real systems and products.

All of this required additional funding. At the same time we were trying to
generate significant industrial participation, we knew that university research
funding would have to increase because industry would need trained people in
order to participate effectively over the long run. My own view was that we
needed to increase academic funding, and we also needed to broaden the base of
the academic research community supported by DARPA. We funded critical mass
efforts based on the quality of proposals submitted, and this process naturally
weighed in favor of the most capable schools in the field. I felt we needed
additional funding to make other schools become as good as the best, and, at the
same time, to support the best schools we were already funding.

Unfortunately, because of the growth of the SCP and the budget constraints
imposed by the Gramm-Rudman cuts, there hasnÕt been enough money to do all these
things simultaneously. Funding priorities have been judgment calls on the part
of people involved with the program. As I said before, DARPA is an organization
of people. If the people in charge are more interested in industrial R&D than in
university research, the academic percentage will decrease. My own personal
strategy would be to fight for enough money so that no one will be
disadvantaged, but that only works when money is available. When budgets are
flat, hard choices have to be made.

When did the idea for the Strategic Computing Program begin?

Well, it's a little hard to say precisely, because it sort of evolved. I had
been pressing for a number of years to increase the scope of the information
processing program, because I felt the need, first, to restore the university
base to where it was a strong and viable component, and, second, to expand the
industrial sector to the point where it could leverage and take advantage of the
results produced in the universities. I saw various ways to do that. I think in
the time frame of 1979 to 1980 there was some reluctance to this level of
funding. Although, l should point out that the budget for information processing
did grow substantially during this periodÑit went from about $50 million to
almost $100 million in three or four years.

It often takes some kind of catalyst to cause a major change. I think the
catalyst for the SCP may have come from a confluence of several factors: first,
the emergence of artificial intelligence technologies; second, the maturation of
VLSI integrated circuit technology; and third, the need to transfer results from
the university laboratories into industry. And another factor, while it wasn't a
direct effect, was the Fifth Generation Project in Japan, and similar efforts
elsewhere.

These things changed the computing field so that by 1982 there was sufficient
motivation to put the SCP together. The President's budget for fiscal year 1984
was presented to Congress in January of 1983, and the SCP was included. Over the
summer of 1983, a lot of the details were fleshed out, and, by October, 1983,
the full report was issued to Congress at their request.

Why were the specific projects of the SCP selected, and how were they selected?
[The SCP consists of, among other things, research projects on an autonomous
land vehicle, a ''pilot's associate,Ó and two "battle management'' systems.-ed]
Were these things that the services had said they were interested in having, or
were these things that people in DARPA thought were good ideas for the
technology?

I think it was largely the case that people within DARPA thought they were good
applications that would stress the underlying technology. There was the issue
that came up along the way about whether the SCP should be a pure technology
program, or whether it should have some applications built into it. I would have
preferred to focus on the technology first and the applications later, but there
were many reasons why it looked like having some applications as part of the
program would be desirable. One, it would make the technology more demonstrative
to the military services. Two, it would act as a focus for research. Three, it
could get other parts of the agency involved. There were a lot of discussions
about this issue, and I supported the result.

But there was an environment created by the military services that made it
apparent that they were interested in certain applications of new-generation
computing technology. These ideas didn't come out of the blue sky.

There had been a Defense Science Board report about a year or two earlier that
basically recommended a major investment by DoD in artificial intelligence. The
services were ready for this kind of initiative, and many of their best people
were very eager to get involved. The way the program has worked out, the
services are involved in a very direct way, managing contracts and helping to
shape some of the details of the program. That was the way DARPA was set up to
work But the decision about the choice of specific projects was internal to
DARPA.

A lot of people, including people in CPSR, as you know, have commented on the
language used in the original Strategic Computing report. The report seems
unduly optimistic and in some cases almost melodramatic about its claims for
these technologies. And the timetable in the back of the report, which suggests,
for example, that natural-language processing will be solved within three years,
seems extraordinarily optimistic. Were these kinds of thing the product of
DARPA?

My role was strongest in the early phases of the program when we were
formulating what the components of the program should be. The timetables were
generated at the very end of the process to reflect a plausible set of
schedules. Many people inside DARPA and without, myself included, contributed
information to this process. It reflected as good a schedule as could be
developed in that time frame. I might add that we planned on applications that
were simply demonstrations of technology. DARPA would not be building fieldable
equipment, or anything the military could use operationally, as is. So even if
we talked about battle management, as the report does, this did not mean that
DARPA was proposing to manage any battles, but, rather, to show how the concepts
and the technology could apply to the problem of battle management.

Or presumably not apply.

Or not apply, that's right. There was a Packard Commission report issued last
year, I believe, that said DARPA now wants to play a more active role in
prototyping. That wasn't its charter, and I don't know what the status is of the
Congressional authorization for prototyping funding for DARPA. It may or may not
change, depending on the needs of the agency and the Department of Defense.

Why was the language of the Strategic Computing report so emphatic and dramatic
about the promise of the technology, if in fact the program was simply a set of
demonstration projects? There's one sentence in the report about how these
applications of computer technology "could fundamentally change the nature of
human conflict."

Many people were involved in writing that report. It was a product of the
agency. I was quite involved in writing the first draft, which was largely about
the technology itself. By the time the report was actually issued, many others
within DARPA had been involved in writing it. Somewhere along the line, a
decision was made to produce a punched-up version that could give a more cogent
view of the possibilities. The original version probably made for less exciting
reading. In any case, l was not the final editor of the report, and it was an
agency-wide decision as to what should be in it.

As you know, within CPSR we have dealt with issues concerning the role of the
computer professional in an environment characterized by research and
development dominated by military funding, an increasing military mission for
computer science in general, rising military budgets, and growing concern about
computer reliability and the prudence of using computers in certain military
applications. What kind of attitude should computer scientists have about these
trends?

Well, I think national security is really a concern of all of us, at some level.
The Department of Defense for as long as I can remember has been the main
benefactor of computer science research in this country. I think researchers
have really benefitted from DoD support of computer science, in a first order
way. I think the most serious question remaining is whether that support is
sufficient for the country. Defense funding for computer research is basically a
good thing for the research community, but is it sufficient? What does this
country really need, on a long-term and sustained basis, to meet its own
requirements? That's the real question. I believe we need some sources of long-
term R&D funding that can address issues that are not directly within the
province of the DoD.

As computing technology matures, l think it's the obligation of the Defense
Department to understand how to make use of it more effectively. The concern
that your organization has been raising about responsible use of the
technologyÑwell, my own view is that the Department of Defense exhibits an
extraordinary degree of responsibility. That's part of the reason it is
entrusted with some of the most capable and controversial systems the world has
ever known, including nuclear weapons.

You must admit that there are strong pressures within the Pentagon, and outside
the Pentagon in the defense industry, for the promotion of various weapons
systems, usually involving large contracts. There's a tremendous amount of money
involved in our current system of providing for national security. And over the
years, there has been a growing concern about the vast sums of money spent on
defense, and the unreliable and potentially dangerous weapons systems being
developed, such as the Sergeant York Air Defense Gun, the B-1 bomber, the M-1
tank, and so on. Many people feel that after spending trillions of dollars on
defense, we are hardly any more secure than before.

At the same time, computer science has become the number one strategic
technological resource in the United States. Computer scientists are, in effect,
in the same position as nuclear physicists were in the late 1940s and early
1950s. And many of them are asking, as those scientists did, what is my personal
role in the arms race? What are your feelings about these kinds of things?

Well, you can't roll the clock back on the human intellectual process. That's
going to go forward, no matter what. And the essential element of a good
national defense is to understand what the capabilities of a potential opponent
could be, and what ours could be.

I've interpreted the questions up to now to be about the research community. I
think you really have to distinguish between the development of an idea, and the
decision to deploy it as part of an operational system. There are several
reasons for exploring new technology. One reason is to have the option to use it
at some future time. It is also essential to protect the country by
understanding the military potential or the threat of a new technology. Once the
potential is understood, it's another thing to put that technology into an
operational environment. Those choices are usually made through completely
different processes than those involved in research. During my tenure in the
Defense Department, my office had no role in making decisions about deployment,
and properly so, because that was not our function. We have a mechanism in this
country to debate what military systems should be produced and deployed. The
Congress is deeply involved in that, every step of the way.

I don't think you should have an inhibition on what the potential for a
technology might be. A technology may prove to be absolutely critical to
national security. So I think you need to make a clear distinction between the
process that explores ideas and technological capabilities, and the process that
puts these into operational military systems. To a certain extent, this has been
what the SDI controversy is all about.

As a nation, we must understand the potential threats and approaches the country
might face or have at its disposal in the future. A decision to produce and
deploy should be a separate process, and, in certain cases, should involve the
apparatus of decisionmaking in the widest possible sense.

But there are numerous examples of a defense contractor getting a contract to
research some application of technology, then becoming a lobbyist for protecting
that contract through the deployment stage. Even if the technology is dubious,
they think they can pull it off, or give it enough of a glimmer to sell it to
the Congress and their stockholders long enough to make the program too costly
to kill.

I come back to this dichotomy between the existence of capabilities that one can
draw on, and the decision to actually draw upon them. It cannot be the case that
the Department of Defense is charged with the defense of the country and yet at
the same time cannot be aware of all the potential threats it might face, or the
capabilities at its disposal. The decision to deal with a perceived threat, or
deploy a certain capability, can be a managed process. It's a process that
involves the Executive branch, the Congress, and occasionally the public. It
seems to me that it's this process you should be concerned with.

If Computer Professionals for Social Responsibility wants to put its view about
the SDI across to a congressional committee, we have to pay for our chosen
representatives to travel to Washington and testify. The Department of Defense
and defense contractors also pay for their witnesses, but the resources are
unbalanced to say the least. It's a David and Goliath story.

Sort of like Gulliver and the Lilliputians?

Exactly.

But sometimes Lilliput can win. There are challenges in fighting successfully
for what you believe in. In our society, if you were the only person who
objected to some policy, and everyone else in the United States was for it, it
might not matter how much money you could command. The will of the people would
be there. I'm a strong believer in the whole democratic process. If people in
this country are not up in arms about something, it's probably because they
don't perceive a threat. If they do, they have the power to speak up.

But a lot of people don 't realize what's going on. Perhaps most people,
especially on such technical issues as the use of computer technology in weapons
systems.

When you say people don't know what's happening, I would simply come back to
ask, what really is happening? To the extent there are projects that are
underway, they've usually been budgeted, appropriated, and this is a matter of
public record. The Defense Department doesn't act in secrecy, for the most part.
There are some projects which are not public because of national security
reasons, but almost everything you talk about is part of the public record.

We take it you would recommend that more computer scientists with technical
expertise get involved in public policy issues, as you've done.

I'm certainly in favor of public awareness and participation. If people are not
aware of what's going on, it's probably because they've chosen not to spend a
lot of time thinking and worrying about some subject, because they may not see
that it affects their daily lives very much. And there are people who are over-
concern, in my opinion.

Perhaps you could talk about your opinions about the development of CPSRÑnot its
internal development, of course, but the fact that it's on the scene and
growing.

I don't know all the activities of the organization now, but my impression is
that the basis of the organization has changed somewhat over the years. As it
was originally constituted, my impression was that its goal was to stop nuclear
war, or something to that effect. There's obviously a lot of concern about the
potential for nuclear war. Perhaps the greatest deterrent to nuclear war is the
weapons themselves. I don't think anyone's particularly happy about the large
deployments of nuclear weapons, and most of us would prefer they not exist. But
if they do exist, you have to find a way to deal with them prudently on all
sides. I'm very hopeful we'll find some way to resolve this problem in the
future.

I think what CPSR has gotten into in recent years is more of an educational
role, at least that's my understanding. Perhaps partly because of your work
people are now more aware of what I would call the "downside" of computers. And
I would hope that you would take a little more balanced position and also point
out the useful and "upside" potential of this technology. It's not evil that's
fueling it. It's concerns, natural concerns, about the well-being of the
country.

I think the educational role of CPSR is an absolutely essential ones and I'm not
only encouraged by it, I'm supportive of it, provided it's done in an effective
and balanced way

Parting Thoughts From Former CPSR Chairman Severo M. Ornstein

CPSR co-founder and immediate past chairman of the Board of Directors of CPSR,
Severo Ornstein, passed on the position of chairman to Steve Zilles on July 1,
1987. The editor of the CPSR Newsletter asked him to share his parting thoughts
with readers. Severo will continue to be an important part of CPSR as a new
member of the National Advisory Board.

When you set out to form something like CPSR, you are so busy making it happen
that it never occurs to you that there might come a day when you will wish to
back off and let it develop further on its own. But indeed that day has arrived
for me, and now as I retire from the chairmanship. I'd like to share a few
thoughts with all of you who have today become CPSR.

Although I am sometimes credited with being CPSR's founder or, in any case, one
of its several founders, that is not really an accurate characterization. It was
simply that watching the news in 1981, I observed a declining respect for the
possibility of nuclear war, and consequently became very concerned about the
future---of mankind in general and of my own family in particular. It seemed
that if we were going to survive the swashbuckling that was then revealing
itself within the government, something had to be done to arouse more general
awareness and concern about the possible consequences of such behavior. Unable
any longer to shirk a growing sense of responsibility, I formed an electronic
discussion group, a subset of which later began to meet weekly in Palo Alto. I
also tried to reach out to a much wider audience by appealing to the management
of the Xerox Corporation (for which I then worked) which responded with an hour-
long television special exploring public concern about possible nuclear war.

But while some of my concerns and actions thus initially pulled together those
of us who eventually formed CPSR, it was not originally my intention to create
an organization embodying those concerns. In fact in our early discussions it
was others who first felt the need for organizing the professional community. In
particular Dan Ingalls (of Smalltalk fame) was the first person to suggest that
we form an organization, and, later, tired of endless "what should we do"
discussions, Terry Winograd arranged for an initial public gatheringÑthus
simultaneously spawning CPSR and inadvertently institutionalizing such
discussions.

In looking back now, l think that one of the principal contributions I made to
CPSR came in those early days when the depth of my own conviction helped to weld
people together and, later on, to convince funders that we were in earnest. Of
course my wife Laura Gould and I also did a great deal of scurrying around, and
for the first year or so the two of us handled most of the actual work of the
organizationÑ putting together the Newsletter, establishing a membership and an
office, handling correspondence, making public pronouncements, importuning
funders, etc. But I have to confess that from the very outset I myself felt some
ambivalence about building an organization. This was a matter of both personal
taste and belief. Much as I cherish the ideals of democracy, l often find
history to be propelled by a relatively small number of determined, sometimes
fanatic, individuals. But of course groups of dedicated people can also have an
impact, if thoughtfully organized with a clear purpose in mind. There is a story
Pete McCloskey (former Congressman from California) likes to tell about himself.
One day he received a visit from a "little old lady" pressing him to vote for
some bird-related conservation measure She was quite insistent, and when he
asked her why she was pushing him so hard, she replied that it was because next
day she had to report back to her thirty thousand-member organization. That
reply, he said, cast the conversation in a whole new light.

So, accepting the importance of organizations, I have fought consistently to
prevent CPSR from becoming just one more grassroots peace organization. I am
convinced that professionalism and technical credibility are the cornerstones of
our effectiveness and that we must continue to speak as professionalsÑno matter
how outraged we may be as individuals. Serious people will listen only if we
stick to those areas involving misuse and misunderstanding of computers. So far
we have generally been quite good about this, but one must work constantly to
combat the many forces tending to push one into more general positions and into
a more stereotypical mold. I admonish each of you to fight these tendencies like
the plague to avoid mindless acceptance of standard liberal viewpoints and
instead to help preserve CPSR's fresh, critical outlook and special professional
character.

Now a couple of comments about the substance of CPSR's program. I was originally
one of those most vocal about the need to stick to the narrow goal of trying to
avert nuclear war through education about technological risk. I felt that that
constituted a unifying topic that few people could contest. But I was naive
about the general level of apathy and about the greed of those segments of the
profession with a vested interest in the arms race and/or DoD funding of
computer science. In time, l became convinced that our viability and credibility
would be enhanced if we embraced a more general program rather than a single
(albeit overwhelmingly important) issue. I am therefore pleased that today CPSR
has broadened itself, and I hope that we can make a showing in the civil
liberties area as significant as that which we have made in the area of military
misuse.

But there is also another thing that CPSR is doing that is dear to my heart. For
many years I've found the arrogance of people in the computer field extremely
distasteful. And as computers have spread throughout society, the contagion of
our arrogance has spread as well. Most appallingly of course we, or more
specifically greedy contractors, have enticed the military with absurd promises,
leading technically gullible and beleaguered politicians into such fantasy-lands
as the SDI.

But I don't think that everything can be blamed on greed alone. I fear that many
of us in the computer field have too often allowed our enthusiasm for technical
innovation to override a cautious and objective consideration of possible
consequences, many of which are often subtle and even obscure. Indeed, it would
seem that in retrospect it would have been virtually impossible to predict in
advance some of the consequences that have occurred. We are hardly alone in our
technological hubris, but that fact doesn't reduce our own need to be as
responsible as possible. It has been highly gratifying to me to find that such a
large number of people within the profession are finally becoming concerned and
are beginning to try to mend matters. I am delighted that just as I retire from
the board, the first professional symposium on these topics is taking place
under the CPSR banner in conjunction with the AAAI meeting in Seattle. I offer
CPSR/Seattle hearty congratulations and best wishes, and I hope that this marks
the beginning of a new tradition. Such events indicate a widening acceptance of
our concerns by the professional community which, after all, constitutes a major
step toward their broader recognition within society.

A final thing I want to comment on is CPSR's "overhead." All organizations bring
with them substantial overhead which must be justified by the additional power
that union brings to individuals. I feel that by and large CPSR wastes very
little motion. Since we haven't got an endowment, and since without funds there
would be no CPSR, we simply have to spend time in fundraisingÑimportuning our
membership, foundations, wealthy individuals, and industry. No one likes to
spend valuable energy in that way, but it is absolutely essential. I doubt that
most of you realize how extraordinarily fortunate we are in having an office
staff capable of maintaining the organization with one hand while doing
important substantive work with the other. We owe a great deal to these people's
unusual combination of skillsÑnot to mention their boundless dedication and
energy.

Let me close on a more personal note. The unifying force that I provided in the
early days, which helped bind together a fragile and ill-defined group, demanded
strong convictions and opinions and a forceful personality, all of which I
possess. Today, in a burgeoning national organization with a representative
Board of Directors, my strongly personal style is increasingly out of place. So,
having had a major hand in helping to get CPSR off the ground, l have decided to
move into the background and let those with greater organizational and
diplomatic skills take over the leadership. In 1982, when I announced that I was
going to retire from thirty years as a computer scientist, friends asked me what
I was going to do instead. At the time I said I was going to devote myself to
becoming a serious pianist. Little did I realize then that CPSR was on the verge
of blossoming into a full time undertaking that would occupy the next five years
of my life. But now that CPSR is reaching maturity, l intend to return to my
earlier plan. In my capacity as a member of the National Advisory Board I will,
of course, continue to watch over and participate in CPSR, but henceforth in a
much less visible role.

My best wishes to all of you, and may CPSR continue to flourish.ÑSevero M.
Ornstei

New Chairman Steve Zilles Outlines CPSR Priorities

With the installation of a new board and officers on July 1, CPSR is entering a
new period. In truth, this period began some time ago as the natural evolution
from an organization with a few members and a particular focus to an
organization of over 2000 members with a much broader focus. CPSR began because
a number of people (initially centered in and around Palo Alto, California) were
concerned about the dangers of nuclear annihilation and were seeking ways to
combat that threat. When CPSR was being formed, there was talk of 'winning
nuclear wars' and of technological solutions to political problems. Like many of
the founders, l was well aware of the strengths and limitations of our computer
technology. We knew that we could not trust it with decisions that could lead to
a nuclear mistake. We wanted to be sure that people making policy, in our
country and internationally, also understood those limitations and acted
accordingly. Therefore, CPSR started out as an organization to educate people
and decisionmakers on the realistic use of computer systems and to attract other
computer professionals with similar concerns.

A concern for the role that computers were and are playing in facilitating
nuclear war was the genesis of much of the early activities of CPSR. We began by
speaking out against policies, such as reducing the time between the launch of a
missile and its arrival at a target, that increases reliance on computer systems
to initiate a response to a presumed attack. We produced a study paper on the
Strategic Computing Initiative. When Star Wars/SDI was proposed we commissioned
a paper on the impossibility of ever building a reliable computer control system
for SDI. We organized and supported debates on the technical infeasibility of
SDI. We produced an excellent slide show called Reliability and Risk with a
primary focus on the use of computers in defense systems. This work has been
successful beyond our expectations.

Although the nuclear arms race, with the global threat that it implies, was the
focus that started CPSR, we recognized that it was only one area in which the
balance between the promise and the limitations of computer technology is an
important area of concern. The critical nature of the problem of nuclear war and
the desire to make a significant impact with limited resources caused the
founders to limit our initial endeavors to this area, but we planned that
someday we would broaden the focus of CPSR to other areas of computer usage. We
were careful to word our purpose with that broad intent in mind. In light of our
past effectiveness and the rapid growth of our membership we are in a position
to broaden our focus sooner than many of us had initially expected.

With success comes change. Some of the people that have done so much to get CPSR
started are retiring from the CPSR Board and new members are taking their
places. We owe a great deal of thanks to these members for their contributions
to the organization. Severo Ornstein and Laura Gould devoted much of their first
years of "retirement" to running and then helping out in the first CPSR National
Office. Together with Brian Smith and others, they were the authors of a number
of the initial CPSR statements. Deborah Estrin and Alan Burning were
instrumental in helping form and aiding active chapters in Boston, Los Angeles
and Seattle. For these contributions and for their sound advice and action we
who remain are truly grateful. These short comments on their contributions
hardly do justice to the real significance of their work.

In a way, the new Board reflects the broadening of scope that is taking place
with CPSR. Although only five people are retiring, the new Board has eight new
faces, in part because it is the first Board which has representatives from all
six geographic regions. This new breadth of representation and experience will
help broaden our focus and strengthen our activities.

The outgoing Board has already taken action to broaden our program. We
established a program in computers and civil liberties as a major new focus area
for CPSR. In April, we adopted a report of the program committee chaired by
Deborah Estrin to put emphasis in the following areas

Reliability and risk
Civil liberties and privacy
Chapter support and aid
Funding for computer science research
Autonomous weapons

The ÒReliability and Risk'' activity is a broadened form of the original focus
on computers and nuclear war. The topic of Òautonomous weapons" is, in turn, a
generalization to the whole question of the use of computers in weapons systems,
especially where the computer is used to decide who will die. This raises both
questions of reliability and questions of ethics and morality.

I expect that the recent changes in CPSR reflect a pattern that will continue in
the coming years. We must continue to grow if we are to improve our
effectiveness and the support for our programs. As we grow, l believe we will
continue to broaden our focus as long as we can increase our base of resources
and attract members with new suggestions and areas in which to work. What ties
us together is the overall goal of "improving the level of responsibility with
which society deploys computer systems."

CPSR provides its members with an amplified voice to speak their concerns and
with a chance to meet and work with peers who are similarly inclined. By working
together we all can have a greater effect. We have established a national
presence that is represented by our National Office. The National Office
monitors and responds to national events, including new programs such as SCP or
SDI, the proposed upgrade of the National Crime Information Center (NCIC), and
requests to testify before Congress. We work to educate our members, national
decision makers and the public about the promise and limitations of computer
technology. We gather information on the use or proposed use of computers and
pass it to our members. We have an excellent full-time staff to coordinate and
support projects undertaken by part-time volunteers. All these mean a more
informative and effective voice.

As an organization, CPSR speaks through the voices of its members. We count on
you to carry our message and to develop it as well. We need your help to educate
people using such tools as the Reliability and Risk slide show. We need your
help in both creating and carrying out chapter projects. These may be as large
as the Reliability and Risk show or more local and specific such as presenting
the

show to the colleges in your area. If you don't have a local chapter, the
national office can provide some support and help in forming a chapter.

We also need your financial help in supporting the National Office and programs.
Our dues are intentionally kept quite low. With the change in the U.S. income
tax law, society membership is not deductible, but contributions are. We are
therefore offering you the opportunity to contribute to our program support as
you are able Our expenses are about $200,000 per year. Today much of that
support comes from foundations. We cannot count on that support as we mature; we
must develop our own base of support. That means you and anyone else you can
recruit to help.

Our promise is to spend your money as effectively as possible.

We stand today with a strong heritage built by our first Board and officers. We
are beginning to broaden our focus. But we have much work to do. We must develop
a solid program in civil liberties. We must develop a sound continuing financial
base. We have tools, such as the Reliability and Risk slide show, that will help
us articulate our concerns to as many people as we can possibly reach.

I look forward to working with you to solve the problems, to listen to your
ideas and to continue to broaden our work to the extent that our finances and
resources allow. I believe we can continue the record of impact and growth that
has been the legacy of our founders and make CPSR the premier computer
educational society from the viewpoint of ethical and social concerns. --Steve
Zille

CPSR/Seattle Sponsors Conference on Directions and Implications of Advanced
Computing

On Sunday, July 12, CPSR/Seattle hosted the symposium "Directions and
Implications of Advanced Computing" on the University of Washington campus. The
event was the first research conference to be sponsored by CPSR, and one of the
largest projects ever attempted by a CPSR chapter. More than 140 people attended
the all-day conference, which included talks by two invited speakers and
fourteen contributed presentations. The topics ranged from the history of
computing to software testing, from the funding of computing research to the
implications of artificial intelligence for medicine and law.

The conference opened with a plenary session on the funding and management of
computer science research, with two invited speakers, Robert Kahn and Terry
Winograd. Robert Kahn is the founder of the nonprofit Corporation for National
Research Initiatives in Washington, D.C. Until 1985, Kahn was Director of the
Information Processing Techniques Office at the Defense Advanced Research
Projects Agency. Terry Winograd is associate professor of computer science and
linguistics at Stanford University. He is the author of several books, most
recently (with co-author Fernando Flores), Understanding Computers and
Cognition. He is national president of CPSR. The session was moderated by Gary
Chapman, executive director of CPSR.

Kahn reviewed recent trends in federally funded computing research. He noted
that support is growing, and that a larger proportion of funds is being
allocated to applied, rather than basic research. Kahn said that this is
happening in part because the products of decades of basic research are now
sufficiently mature that they can serve as the basis for worthwhile
applications. He observed that through most of the twentieth century, housing
and automobiles were the engines that drove the U.S. economy and provided most
of the growth, either directly or by stimulating other industries like steel and
petroleum. He suggested that those industries will not be able to play that role
any longer, and that future growth must come from other sources, like the
"information industries," computing and data communications. Kahn said that just
as the country required an infrastructure of highways to allow people to
effectively use the automobile, we now require an "information infrastructure"
consisting of computer networks, databases, and software utilities that permit
people to effectively use computers.

Terry Winograd argued that the predominance of the Department of Defense in
computer science funding affects the content and direction of the research, and
thus influences the character of the world in which we must live and work. He
noted that some researchers assert that the source of funding does not affect
their choice of research topics. Winograd rejected this view as naive. He noted
that the officials of the Department of Defense say that they fund research in
order to promote work on topics they feel are important. By providing large sums
of money for work on these topics, and by building infrastructure that makes it
easier to do such work, defense funding encourages scientists to choose those
topics. Researchers are not entirely free to "take it or leave it," Winograd
pointed out, because careers and university budgets often depend upon defense
funding, and similar sums are not available from other sources. Winograd
suggested that this pattern of dependency contributes to an unduly large
influence of military institutions and values in public life.

Kahn said that the task of providing for national security obligates the
Department of Defense to research all technologies that might be relevant to
defense. He said that the appropriate stage for criticism of these technologies
is when applications are planned, not when basic research is performed. Winograd
said that the selection of research topics is an important factor in determining
the nature of the world we come to inhabit, and that projects gather
considerable momentum even during the research stage, making them difficult to
cancel later on.

During the contributed presentations, David Bushnell, Joel Yudken and Barbara
Simons described the institutional context in which computer science is done and
speculated how a variety of factors may influence its content. Richard Hamlet,
David Bella, and Erik Nilsson presented software engineering analyses of costs,
reliability and safety related to defense policy. Juergen Koeneman, Matthew
Lewis and Seth Chaiklin. Rolf Engelbrecht, Carole Hafner and Donald Berman, and
Jack Beusmans and Karen Weickert each explored the implications of applying
artificial intelligence technology in particular areas from education to combat.
Reinhard Keil-Slawic, Douglas Schuler, Susan Landau, and K. Eric Drexler
confronted present and future dilemmas in computing from philosophical and
ethical perspectives.

After the conference, the participants, families and friends attended a
reception, hosted in part by publishing company Springer-Verlag to celebrate its
new journal, Artificial Intelligence and Society.

The DIAC conference was largely the idea of Doug Schuler, president of the
Seattle chapter. Scott Rose was responsible for local arrangements. Ken Berkun
maintained and enforced the budget, and many other chapter members worked on the
project. Tyler Folsom handled publicity for the event. Contributed papers were
reviewed by a program committee including Andrew Black, Alan Borning, Jonathan
Jacky, Nancy Leveson, Abbe Mowshowitz, Herbert Simon, Terry Winograd, and
several anonymous reviewers. The papers are collected in a proceedings volume
which is available from the CPSR National Office

From the Secretary's Desk
Eric Roberts
CPSR National Secretary

As Steve Zilles noted in his statement on page 10, this issue of the CPSR
Newsletter marks a new era in the history of the organization. As of July 1, I
took over the position of National Secretary from Laura Gould who had held that
office since CPSR was founded. First of all, I would like to thank Laura for all
of the work she has done for CPSR over the years. It will be hard to live up to
the standards she has set, but I intend to do my best.

The big news item for CPSR this summer is, of course, the DIAC-87 symposium in
Seattle. The details of the symposium are presented elsewhere in this issue, but
I would like to add my own note of appreciation to the organizers for what I
felt was an extremely important and interesting event. I enjoyed the opportunity
to meet CPSR members from different parts of the country, and I hope to have
many more such chances in the future. Plans are already underway for DIAC-88 in
Minneapolis/St. Paul. Anyone interested in volunteering to work with local
arrangements or the program committee should contact Doug Schuler. c/o CPSR/
Seattle, P.O. Box 85481, Seattle, WA 98105.

The other major item on the calendar is the upcoming Annual Meeting in
Cambridge, Massachusetts, on October 17 and 18 (see program on inside back
cover). Along with all of the events listed in the program, we hope to provide
opportunities for members with similar interests to get together. If you have
any suggestions for topics (particularly if you would be willing to act as the
session facilitator), please send them to me at the National Office.

Responses are beginning to arrive in response to the announcement of Reliability
and Risk that Katy Elliot sent out to 5600 peace groups in July. The response
from the Nuclear Age Peace Foundation may be of particular interest to CPSR
members. During the week of November 9-15, 1987, the Nuclear Age Peace
Foundation is sponsoring the "Second International Peace Week of Scientists"
intended to "further arms control and disarmament and the application of science
for peace and human betterment." Suggested topics include "Eliminating Risks of
Nuclear War by Accident" and "The Economic Impact of Military Spending on
Scientists and Scientific Research"Ñboth of which are areas in which CPSR
members have already done considerable work. As part of the program, the Nuclear
Age Peace Foundation is offering a $50,000 prize for "the best proposal by a
scientist or scientific organization offering an action plan to provide
sufficient initiatives that can be used for peaceful rather than destructive
purposes." For more information, contact the Nuclear Age Peace Foundation, 1187
Coast Village Road, Suite 123, Santa Barbara, CA 93108.

As noted on this page, Mary Karen Dahl continues her work with civil liberties
and the National Crime Information Center (NCIC) project. This project offers an
important opportunity for chapter involvement, since setting up a dialogue with
congressional representatives is best handled with a locally-oriented media
campaign. For more details, contact Mary Karen at the National Office

CPSR Team to Consult With Congressional Committee Studying NCIC
Mary Karen Dahl
National Program Associate

The CPSR Computing and Civil Liberties Project has been asked by Congressman Don
Edwards (D-CA) to help form a panel of experts to provide an independent
critique of current proposals for updating and expanding the National Crime
Information Center. Edwards, who chairs the House Judiciary Committee's
Subcommittee on Civil and Constitutional Rights, explained the need for the
undertaking in his letter of invitation:

The FBI's National Crime Information Center (NCIC) is the largest computerized
criminal justice information system in the country. Its 18 million records are
accessible to 64,000 criminal justice agencies throughout the country. It poses
significant social policy issues regarding the application of information
technology to law enforcement and the impact of computing on privacy and civil
liberties.

Edwards charges the panel with studying plans for NCIC 2000 and advising the
Subcommittee on the following questions:

1. Have individual rights and society's needs, in addition to law enforcement's
needs, been fully identified and properly balanced . . .?
2. What types of information should be stored, indexed and disseminated in a
redesigned NCIC?
3. Who should have access to the system and for what purposes?
4. How can the system be designed to control access and dissemination and to
protect against unauthorized use?
5. How can the system be designed to ensure
record accuracy?
6. Does the system adequately use technology to protect privacy and civil
liberties?

Drs. David Redell and Peter Neumann (both of CPSR/ Palo Alto) will serve on the
panel. In addition, CPSR chapters are encouraged to form study groups to develop
and share research on this issue.