Computing Leadership Summit, October 20, 1997

CRA organized its fourth Computing Leadership Summit on October 20, 1997, in Washington, DC. Presidents, executive directors and other key volunteers from AAAI, ACM, CRA, CSTB, IEEE-CS, SIAM, and USENIX met to discuss policy issues affecting computing research.

The day began with a breakfast meeting with Representative Vern Ehlers (R-Michigan), Vice Chair of the House Committee on Science, plus Science Committee staff members Tom Weimer and Jim Wilson, and AAAS congressional fellow Sharon Hays. The discussion centered on a congressional study of the nation's science policy, which Ehlers is chairing and which he had kicked off only four days before the Summit.

Juris Hartmanis, Assistant Director of CISE in the National Science Foundation, discussed three cross cutting NSF initiatives: Knowledge and Distributed Intelligence, Life in Earth's Environment, and Education for the Future, which are being funded with $10 million from the Director's Opportunity Fund. The group discussed the need to communicate the economic and social impact of computing research, and the intellectual excitement of the field. One comment was that we are entering a "new era of tool-building" that will drive a new understanding of human intellectual activities.

John C. Davis, who is Director of the National Computer Security Center at the National Security Agency, discussed the report of the Presidential Commission on Critical Infrastructure Protection. This report was delivered to President Clinton as a classified document on October 20. The report concludes that about 60% of the nation's "critical infrastructure jeopardy" is related to information technology: the Internet, banking systems, telephone switching systems, even power grids that rely on computer control (see related article). The report recommends a research program that would grow to $1 billion per year focusing on "the science of complex, interdependent systems." Joseph DeBlasi, executive director of ACM, suggested that perhaps our field needs to produce a document that alerts the public to the potential threats from expanded use of information technology without expanded investment in understanding how to build, operate and evolve large mission-critical software systems; perhaps the Commission's report (or the parts of it that are declassified) will provide some grist.

Albert Teich, Director of AAAS Science and Policy Programs, provided an overview of trends in government funding of scientific research and development. Overall, the situation looks better than it did five years ago, but not entirely rosy. Comparing 1994 and 1997 budgets, NIH and NSF are the only agencies with more to spend after inflation; DoD, NASA and others are down 5 to 25%. President Clinton's proposed budget for 1998 improves the downward turn only slightly. Looking over the longer period since 1986, defense support for research is down, while nondefense support grew in the 1980s, then leveled off, and declined slightly in the 1990s. NIH has been treated much better by Congress than other nondefense agencies in their R&D budgets.

Given that R&D funding comes from the one-third of the federal budget that is for discretionary spending and that the discretionary portion is where cuts are being made to balance the budget, Teich reviewed the budget-balancing activities. Over the next few years, discretionary funding will drop slightly while nondiscretionary funding will continue to grow. The budget is scheduled to be balanced by 2001, but the continuation of an economy that is more robust than planned and the willingness of Congress to make unpalatable cuts in programs and entitlements in out years will affect this schedule. There is recent talk - resolutions, but not appropriations - on the Hill to increase R&D spending overall, for example, to double these budgets over 10 years. Charts are available at www.aaas.org.

Teich provided advice on playing the funding game. He noted the important but difficult task of documenting social return on R&D investment and the irrationality of setting research priorities, in that federal funding for science is not a tradeoff system. Teich observed that the post-Cold War environment makes it harder to justify a large role for government support of science, and that it is likely that Congress increasingly will look to industry for this support.

John Guckenheimer, president of SIAM, noted a series of disturbing trends affecting computational science: the failure of the effort to establish university degree programs in computational science, the difficulties of finding a home for computational science in existing departments of computer science or mathematics, the problems of sustaining interdisciplinary research relating to computational science in university center structures, the lack of attention in industry, and the amount of poor software being written because of lack of training in, or attention to, the basics of computational science. Some solutions were posed by the attendees, including: 1) increasing the scope of computational science so it seems important enough to get attention; 2) a CSTB study on applied mathematics and computational science; 3) a discussion of these problems with NSF, inasmuch as NSF has a great interest in interdisciplinary initiatives; and 4) bringing the problems to the attention of the CS community by writing an article for Computing Research News.

Ed Lazowska, chair of the CRA board, discussed issues related to women in computing research. He circulated a document prepared by the Committee on the Status of Women in Computer Science and Computer Engineering Research, outlining its programs. He then encouraged the professional societies to examine the areas in which it makes sense to work together, and those in which multiple efforts are required. It was agreed that the professional societies would find a way to get the heads of their women's committees, or other suitable representatives, together to examine these issues, as well as ways in which industry and the federal government could be encouraged to provide financial support for these programs. To that end, it was suggested that the meeting include representatives from companies that have an interest in women in the profession.

Peter Freeman, chair of the CRA government affairs committee, made a brief presentation of the work CRA is doing in government affairs on behalf of its members. He categorized these into three areas: 1) being available in Washington to respond rapidly to the crisis of the day; 2) building up a stronger presence in Washington that can be proactive on behalf of the computing research community; and 3) establishing closer ties with industry and trade associations to leverage our efforts and have a more united front in Washington.

A roundtable discussion about the Leadership Summit reaffirmed that it should be held annually, with an emphasis on issues rather than operations.

This article was first printed in the January 1998 issue of Computing Research News.