The Chronicle of Higher Ed today has coverage (free until 6/2 apparently) of the May 12th House Science Committee hearing on “The Future of Computer Science Research in the U.S.” that’s generally pretty good. But it makes an odd point at the end that doesn’t accurately represent what went on at the hearing. Here’s the paragraph:
[DARPA Director Tony] Tether challenged Mr. [Tom] Leighton [, co-founder and Chief Scientist at Akamai Industries] and Mr. [Bill] Wulf [, President of the National Academies of Engineering] to supply examples of important projects that the agency has refused to support, and they did not immediately offer any. That shows, Mr. Tether said, that the agency’s priorities are properly placed.
At the end of the 2 hour, 19 minute hearing, in response the committee’s very last question, Tether told the panel that in dealing with the university computer science community he saw “a lot of hand-wringing” but didn’t get many “actionable ideas” from the community. Science Committee Chairman Sherwood Boehlert then turned to Wulf and Leighton and asked if they could take that as a challenge and provide a list to the committee and to Tether. Both responded that they’d be happy to and Boehlert noted that he’d make that part of the post-hearing questions that will be put to the witnesses (and noted the challenge in his press release).
I understand both Wulf and Leighton are eager to respond to the challenge. Leighton told me after the hearing that he was getting ready to wave the PITAC report on Cyber Security R&D as a start (the focus of much of his testimony), which contains specific recommendations in 10 areas of cyber security research currently under-supported. Both Leighton and Wulf will be reaching out to the community to craft a list that will be most useful to DARPA and DOD and most responsive to the committee’s request (which hasn’t yet been received, as far as I know). There are plenty of resources from which to draw — PITAC’s Cyber Report, Defense Science Board, CRA’s Grand Challenges conferences, National Academies reports, etc.
The idea that either Wulf or Leighton were dumbstruck by the question is just wrong, and the idea that the community lacks an adequate response to the committee’s challenge is equally wrong.
Otherwise, the article does a decent job of summarizing the hearing. From my perspective, the hearing was incredibly useful. I could spend a lot of space here dissecting the testimony of Marburger and Tether — though frequent readers of the blog won’t need my dissection to spot the points of contention in both sets of testimony. Tether essentially argued in his oral testimony (and half of his written testimony) that DARPA has reduced its funding for university-led computer science research because maybe it’s focusing on multi-disciplinary research now; something Tether apparently deduced by looking at university web pages, he says. But in the appendix to his testimony, he provides the response to the same question he gave to the Senate Armed Services Committee, compiled by the DARPA comptroller, which includes these five reasons for the shift:
1. A change in emphasis in the high performance computing program from pure research to supercomputer construction;
2. Significant drop in unclassified information security research;
3. End of TIA-related programs in FY 2004 due to congressional decree, a move that cost universities “a consistent $11-12 million per year” in research funding;
4. Research into intelligent software had matured beyond the research stage into integration;
5. Classified funding for computer science-related programs increased markedly between FY 2001 and FY 2004, but Universities received none of this funding.
From my perspective, having the DARPA director stand before the committee (literally) and affirm that the agency has significantly reduced its support for university-led, long-range computing research was very useful. The community can raise concerns about DARPA’s priorities, but ultimately it’s up to the Director and the Administration to set them as they see fit. What’s more important to me is that the impact of DARPA’s (now undisputed) withdrawal on the overall IT R&D enterprise be adequately assessed and addressed. The gap that DARPA leaves is substantial — both in terms of monetary support and in losing a funding model that has contributed so much to the extraordinarily productive environment for innovation that is the computing research community. NSF is great at what it does — funding individual investigators and research infrastructure at universities — but there was substantial value from DARPA’s approach of focusing on particular problems and nourishing communities of researchers to address them. Without DARPA, that approach is largely absent in the federal IT R&D portfolio.
It was also useful for the Science Committee to get exposure to the concerns the community has had with DARPA over the last several years. Tether’s performance — literally standing before the committee (I staffed a lot of hearings for the House Science Committee under two different chairmen and never once saw a witness rise before the committee and wander around the hearing room while testifying…), delivering remarks 15 minutes over the 5 minute time limit imposed by the committee, and most importantly, being largely unresponsive to the three questions the committee posed to him prior to the hearing — confirmed to the committee Chair and staff that the concerns the community had shared with them had merit. The result is that the committee intends to remain engaged on this issue, which is to the community’s great benefit, I think.
The committee plans to proceed with the issue in the coming months in non-hearing venues. I’ll bring you developments as this moves forward during the summer and fall.