DARPA Assailed for Cutting Back Support of Basic Computing Research at U.S. Universities

by Barton Reppert

IEEE-USA and other major professional technical organizations, together with key members of Congress and prominent computer scientists and engineers, have sharply criticized the Defense Advanced Research Projects Agency (DARPA) for cutting back its support of basic, open-ended, “blue-sky” computing research at U.S. universities.

“At a time of growing global competition, DARPA’s disinvestment in university-based, long-term research is, in my view, a risky game for the country,” declared Wm. A. Wulf, president of the National Academy of Engineering (NAE), at a 12 May hearing before the House Committee on Science.

According to data provided by the committee staff, the amount of DARPA computer science funding awarded to universities dropped by 42.5 percent from $214 million in fiscal year 2001 to $123 million in FY 2004.

Over the same period, the National Science Foundation’s burden of supporting research in this area increased substantially — with funding provided through NSF’s Computer and Information Science and Engineering (CISE) Directorate going from $478 million in FY 2001 to $605 million in FY 2004. The number of grant proposals submitted to CISE jumped from 3,866 in FY 2001 to 6,496 in FY 2004, while the success rate for researcher grant applications declined from 24 percent down to 16 percent over that period.

Russ Lefevre, IEEE-USA vice president for technology policy activities, contends that the sharp drop-off in DARPA funding represents “a dramatic departure from the historic government support for basic research at U.S. universities and colleges, especially in information technology.”

“This support has spawned spectacular successes — including the fundamental research leading to the Internet — and remains vitally important to the nation’s competitiveness,” Lefevre says. “IEEE-USA is especially concerned about the long-term implications for cybersecurity and high-performance computing, which will suffer immensely without sustained federal investment.”

As evidence of the breadth and intensity of the computer research community’s concern, Lefevre points to an editorial published in the 6 May issue of Science, the weekly flagship journal of the American Association for the Advancement of Science (AAAS). Authoring the editorial were David A. Patterson, president of the Association for Computing Machinery (ACM) and holder of the E.H. and M.E. Pardee Chair of Computer Science at the University of California, Berkeley; and Edward D. Lazowska, who holds the Bill & Melinda Gates Chair in Computer Science & Engineering at the University of Washington. Both are members of the President’s Information Technology Advisory Committee (PITAC) and past chairs of the Computing Research Association.

Assessing the situation at DARPA, Patterson and Lazowska observed that “policy changes at the agency, including increased classification of research programs, increased restrictions on the participation of non-citizens, and ‘go/no-go’ reviews applied to research at 12- to 18-month intervals, discourage participation by university researchers and signal a shift from pushing the leading edge to ‘bridging the gap’ between fundamental research and deployable technologies. In essence, NSF is now relied on to support the long-term research needed to advance the IT field.”

The two leading computer research experts concluded: “At a time when global competitors are gaining the capacity and commitment to challenge U.S. high-tech leadership, this changed landscape threatens to derail the extraordinarily productive interplay of academia, government and industry in IT. Given the importance of IT in enabling the new economy and in opening new areas of scientific discovery, we simply cannot afford to cede leadership.”

Long-simmering worries in the academic computing research community came to a head on 2 April, when The New York Times published an article by veteran technology correspondent John Markoff, reporting that DARPA, “which has long underwritten open-ended ‘blue sky’ research by the nation’s best computer scientists, is sharply cutting such spending at universities, researchers say, in favor of financing more classified work and narrowly defined projects that promise a more immediate payoff.”

“University scientists assert that the changes go even further than DARPA has disclosed,” the Times story said. “As financing has dipped, the remaining research grants come with even more restrictions, they say, often tightly linked to specific ‘deliverables’ that discourage exploration and serendipitous discoveries. Many grants also limit the use of graduate students to those who hold American citizenship, a rule that hits hard in computer science, where many researchers are foreign.”

At the 12 May hearing, Rep. Sherwood Boehlert (R-N.Y.), chair of the House Science Committee, noted that the session was “extraordinarily important” in view of the fact that “information technology provides not just a web, it is the warp and woof of our society.”

Boehlert declared in his opening statement that in the IT field, “current federal funding is not properly balanced. It does not adequately continue our historic commitment to longer-range, more basic research in computer science, and it does not focus sufficiently on cybersecurity.”

DARPA Director Tony Tether sought to reassure the House panel that contrary to the 2 April Times story, “there has been no decision to divert resources, as the article implies. DARPA’s commitment to seek new ideas, to include ideas that support research by bringing together new communities of research scientists, is the same as it has been, dating back to the agency’s inception in 1958.” Tether also testified that “the article implies that DARPA is moving away from long-range ‘blue sky’ research. Let me assure you that also is not the case.”

Tether emphasized the need for funding more multidisciplinary research efforts, even though that could mean cutbacks in support for particular disciplines such as computer science. “Rigidly funding specific, established disciplines would severely limit the flexibility DARPA needs to be successful,” he said. “DARPA needs the ability to promote multidisciplinary work to solve important national security problems.”

During the committee hearing, however, Wulf hammered DARPA’s position relentlessly. Wulf has served as NAE president since 1997 and he is on leave from his academic position as AT&T Professor of Engineering and Applied Sciences in the Department of Computer Science at the University of Virginia.

“I am deeply concerned about what has happened at DARPA,” Wulf testified. “On top of a many-year drift toward the less ambitious and more incremental, the Iraq war has been described as a reason to dramatically accelerate this — to focus on reaping the successes of the past, to focus on … industrial development over university research, and to shift the balance strongly toward near-term topics. While I can agree that reaping, developing and focusing on the near term are needed, so is long-term investing. Without current investment there won’t be anything to reap next time.”

Contrasting this situation with DARPA’s past successes, Wulf declared that “there was only one old-style DARPA, and it is gone … You can only wonder at what the world would be like today if the immediacy of the Vietnam War had diverted ARPA from funding crazy ideas like networking, timesharing, VLSI, graphics, RISC architectures, RAID disk systems, parallel computing — or any number of other technologies that are essential to today’s computer industry and whose results pay off daily to industry, government and the consumer, as well as the military.”

Also voicing pointed criticism of DARPA at the hearing was F. Thomson Leighton, co-founder and chief scientist of Akamai Technologies Inc,. and a professor of applied mathematics at MIT. Leighton is a member of the President’s Information Technology Advisory Committee (PITAC) and chair of the PITAC Subcommittee on Cybersecurity.

Leighton testified that “if DARPA’s current practices had been in effect in the mid-1990s, it is unlikely that the development of Akamai’s technology, to improve the distribution of content and applications over the Internet, would have taken place. That is because no other agency has stepped in to fill the gap created by the shift at DARPA. This is particularly evident in the area of cybersecurity.”

He charged that “as a result of the changes in government funding for basic research, we are now facing a serious lag in our nation’s ability to continue to innovate, at a time when innovation is most needed.”

Additional criticism of DARPA came in a written joint statement submitted to the House Science Committee, drafted by six organizations — the American Society for Information Science and Technology (ASIS&T), the Coalition for Academic Scientific Computing (CASC), the Computing Research Association (CRA), the Electrical and Computer Engineering Department Heads Association (ECEDHA), the Society for Industrial and Applied Mathematics (SIAM), and the U.S. Public Policy Committee of the Association for Computing Machinery (ACM).

“We are concerned about DARPA’s diminished role in supporting computing research and the impact that it will have on the field, DARPA’s mission, and the nation as a whole,” the statement said. “Central to these concerns is the idea that the field —  and hence, the nation — benefited greatly by having different approaches to funding computing research represented by the NSF model and the DARPA model.”

The joint statement observed that “the United States still has the world’s strongest capability in fundamental research in IT … This is a robust system that can take stresses from decreased funding for a short time as we determine our strategy. But we run a grave risk in letting the uncertainty about funding for fundamental IT research go on too long.”

Peter A. Freeman, NSF assistant director for the CISE Directorate, said in a telephone interview after the House Science Committee hearing that “in the case of computer science and engineering, there’s no question that a number of researchers who previously had been funded, for many years in some cases, by DARPA, are now coming to us as first choice. Also, there appears to be no question that a number of researchers who had been seeking and getting DARPA support, but maybe not the stars … are turning away from DARPA and don’t even bother to apply there anymore.”

Although Freeman said he is concerned about the low success rate for computer research grant applications submitted to CISE, he added: “Success rate isn’t really the issue. The issue is, is this nation funding enough fundamental computer science, computer engineering research to keep the new ideas entering the pipeline, which everybody agrees takes, in the best of circumstances, five years, 10 usually, often 15 years to get out and start to be turned into services, products and so forth.”

“You have to keep the pipeline full,” Freeman said. “If you stopped all fundamental research today, you wouldn’t see any difference for a while in what’s happening in industry. But in five years, 10 years, we’re not going to have any new ideas coming along. And where are they going to come from? They’re going to come from other countries that are investing in fundamental research.”

Jerry Engel, president of the IEEE Computer Society and professor of computer science and engineering at the University of Connecticut, Stamford, said he has “some very mixed feelings” about the controversy over federal support for basic computing research. Engel noted that he himself had worked at NSF from 1991 to 1995, for two years as a program director and subsequently as acting director of the Division of Computer and Computational Research within CISE.

“What strikes me most about the issue is that the two agencies [DARPA and NSF] have different missions. As such, it’s not clear to me that [their funding priorities] are interchangeable,” Engel said in a phone interview. He added: “It seems to me that if research falls within the NSF mission, it makes sense that NSF fund it. If it is within the DOD mission, then it should be supported there.”

Engel said his own view is that the overall level of federal support for basic computing research is “not keeping up with the expansion of the field. It is not adequate. The question becomes, how does one balance these needs within the whole set of national priorities.”

The IEEE Computer Society has not adopted any position statement on U.S. government funding for computing research, Engel said, because “we are, in fact, not a national society, we’re an international organization, so we tend to be a little more careful in taking positions on national issues.”

Barton Reppert is a freelance science and technology writer specializing in S&T policy coverage. He previously worked for 18 years as a reporter and editor with the Associated Press in Washington, New York and Moscow. He can be contacted at barton.reppert@verizon.net.