HPCC still a worthy program

By Fred W. Weingarten
CRA Staff

The federal government's High-Performance Computing and Communications initiative has been the focus of attention in the computing research community for about 13 years. Its origins date back to a report prepared by an ad hoc study panel funded by the National Science Foundation and chaired by Peter Lax, the noted applied mathematician at the Courant Institute. HPCC has been an official administration initiative for about six years. And nearly four years ago the HPCC Act (the so-called "Gore" bill) was signed into law.

At the best of times, a program that has been around that long would--and ought to--be held up to questioning. How well has it accomplished its intended goals? Is its mission still important? Assuming the overall goals are still valid, is the program as it was designed still the best way to achieve those goals?

But these are not the best of times. Budget pressures on federal R&D are seriously threatening programs in Defense and in civilian science agencies. A new Republican-dominated Congress is asking fundamental questions about the proper role of the government in funding research, particularly research that appears to have direct economic payoff to industry.

In the face of this questioning and these new pressures, it would be reasonable to be concerned about the future of HPCC. This concern is particularly acute in computer science and computer engineering because, over the years, federal funding for CS&CE has come to fall predominantly under the HPCC rubric. It would not be a long jump for congressional challenges of the appropriateness of federal HPCC support to broaden and cover all computing research support.

Some senior researchers in the computing field have been suggesting lately that we need to find a new engine, because this one is running out of gas. While advancing under the HPCC flag, computing research came of age. It moved out of the shadows and became a leading, highly visible component of federal science and technology policy. Budgets grew enormously, as did the influence and workload of the Washington computing research bureaucracy.

On the other hand

There are two problems with this argument. First, it is not clear that any politician is on the lookout for new initiatives or programs, and that reticence would particularly include anything that looks like "big science." Few, if any, real champions of science remain in Congress, and those who are there have little influence. One might think that among the newer members, some might be searching to stake out issues on which to build their reputations--somewhat in the way Al Gore made HPCC and the Information Superhighway "his." But today's political environment is not conducive to that strategy, and most new members came to town looking to make their reputations by closing down government programs, not creating new ones.

Besides, there still may be some life left in the old program. Even considering the negatives mentioned above, HPCC may turn out to be the best chance for making a persuasive case for maintaining support for computing research.

The political climate is not all negative toward HPCC. House Speaker Newt Gingrich and other Republicans are fans of the Internet and lace their talks with references to cyberspace and Alvin Toffler's "Third Wave." They may not yet have made the connection with research, but the case is there to be made. The Thomas congressional information system that the House leadership takes so much pride in connects to the Internet, is based on a Web server and uses a University of Massachusetts data query system, all of which have come from NSF and Advanced Research Projects Agency research programs.

A strong HPCC program, if properly focused, can help the nation develop and achieve the full potential of a new national information infrastructure. This argument was made persuasively in a recent report, Evolving the High-Performance Computing and Communications Initiative to Support the Nation's Information Infrastructure, by the National Research Council's Computer Science and Telecommunications Board. Although it focused on HPCC, it made a much broader argument about the important role computing research has played and the even more critical role it will play in the evolution of the United States as an information society. The report also addressed the politically tough issue of the role of basic research by tracing the flow of ideas from university-based basic research to industrial laboratories to development. It made the vital point that strategic research is not necessarily short-term or applied. There is a long-term and enormous economic and social benefit from fundamental research directed into strategically important areas.

Another important report, America in an Age of Information, was recently released by the National Science and Technology Council Committee on Information and Communications. This report developed a strategic plan and a framework to coordinate and manage computing research throughout all federal science agencies. It described computing research programs, including HPCC, in the context of broad national objectives from an applications perspective. The focus was on user needs. This approach was in stark contrast to attempts in the late 1980s to motivate HPCC narrowly in terms of its impact on the competitiveness of the US computer industry--an approach that led to criticisms that HPCC was simply a "bailout" program in disguise for Cray Research Inc. or IBM Corp.

Meanwhile, the administration appointed a new head of the National Coordination Office for HPCC. John C. Toole will replace Donald Lindberg, director of the National Library of Medicine (NLM). This appointment means several things. First, it is a full-time appointment for Toole. Lindberg served as coordinator, yet retained his position as NLM director. As HPCC grew, so did the conflicting pressures on his time and attention. Second, Toole is from ARPA. When Lindberg was appointed, the interagency pull and tug over program leadership was so great it seemed inconceivable for someone from NSF, ARPA or the Energy Department to be selected. No one thought NLM wanted to be a lead agency for HPCC, so Lindberg was viewed as neutral.

Lindberg was a senior, respected voice in science policy circles outside the computing arena. He helped legitimized HPCC to that world.

Toole has a great deal of experience in government research program management, particularly in the information technology field. However, he has less external name recognition in the science policy community. This suggests that the administration was looking for someone who had clout within the computing programs of the science agencies and who could bring about a closer coordination. It appears the administration saw less need to legitimize computing research by picking someone from outside the field.

The HPCC program is entering this dangerous time with new ammunition and tools with which to make its case. It seems far better prepared for the fight now than it was in 1994.