Reinvigorating the Field

Musings from the Chair

By Eric Grimson, CRA Board Chair

A few weeks ago, members of the computing research community assembled for the 19th biennial Conference  at Snowbird, the flagship conference for chairs of Ph.D.-granting departments of computing and allied fields and leaders from U.S. industrial and government computing research laboratories and centers. Here are some observations on trends in the field evident during the meeting.  

Throughout the conference, I was struck by a sense of cautious optimism and renewed energy shared among department chairs. Challenges remain, but the perception that key aspects of the field are slowly, yet steadily, improving was evident throughout the conference:

• Funding for computing research is increasing, and sources of funding seem to be stabilizing. Ken Gabriel, Deputy Director of DARPA, highlighted the commitment of DARPA to re-engage with the academic research community—a re-engagement that recognized the need to balance the desires of the research community to tackle difficult problems without onerous bureaucratic constraints on reporting and deliverables with the need to serve DARPA’s core mission to national security. DARPA is clearly committed to a major shift in its interactions with academic research. Other major funding agencies for computing research similarly show signs of growth, as highlighted by Peter Harsha’s plenary presentation on CRA’s government affairs committee’s interactions with Congress. Over the past 10 years, government funding of IT R&D has doubled; a very encouraging sign.
• Enrollments in computer science, computer engineering, and information science are stable and slowly rising, as documented in CRA’s most recent Taulbee report. New majors have been slowly rising since 2005; while we are still below our peak, we are on a positive trend.
• We are weathering the financial storm of the past two years. The CIFellows program, funded by NSF and administered by CCC and CRA, has helped 60 recent PhD graduates find exciting research homes for two years, while the job market stabilizes.  Listening to presentations on the effort, and hearing from selected participants, this program has clearly helped bridge a gap in the intellectual continuum of the field. More generally, discussions with many department chairs indicated that departments and institutions are cautiously emerging from hiring freezes, another encouraging sign.

There was also a growing sense of shared community among department heads. One instance emerged in the discussions and presentations of the CRA-E committees. The first version of the CRA Education group, chaired by Andy van Dam, released its final report, articulating a vision of how computation curricula have evolved: more flexible, better integrated with other disciplines, yet still providing a foundational mode of thinking that supports other intellectual disciplines. Their report, discussed in this issue of CRN, provides a valuable roadmap of alternatives being explored at multiple institutions, and articulates a vision for further evolution of curricular material. The second instantiation of CRA-E formally launched its activities at Snowbird during a packed breakfast meeting where the committee laid out plans and goals (as articulated elsewhere in this issue of CRN). Clearly there is a shared sense in the community of the need to adapt our curriculum to meet the interests of today’s students, and to develop a new generation of researchers who will tackle compelling challenges in security, energy, environmental sustainability, finance, health care, and information technology. 

A second instance of shared community emerged during a session on faculty hiring practices. While every institution is governed by local nuances, and must act to best serve its needs, what emerged from this session was a conviction that coordination among departments can improve the hiring process for everyone. It was very encouraging to hear many department heads publicly commit to common dates for applications, interviews, and decisions, and to shared methods for communicating decisions in a timely manner to applicants. 

Challenges remain, however. Despite significant efforts by many members of the CRA community, there is a clear sense that computation remains a “poor cousin” of the sciences in the eyes of national leaders—evidenced by challenges in ensuring that computation is a full partner in discussions on STEM education, in revising the CS AP exam to make it relevant to modern views on computing, and in ensuring that congressional committees include computing in their long-term visions for technological and scientific growth for the nation.

There is a concern that computing is still not fully “taken seriously” by policy makers, a view potentially exacerbated by the pending NRC rankings of doctoral programs, as articulated in the May 2010 issue of CRN. During a panel session at Snowbird, Charlotte Kuh presented current plans for the release of the NRC rankings report and data. Some progress has been made since last spring:

• For computer science, the NRC will now count publications in journals and conferences based on CVs submitted by faculty members, rather than relying on citations from the Thomson-Reuters Web of Science database. Unfortunately, the corollary is that NRC will not be able to make use of any citation data in their report.
• The NRC will print a revised methodology report highlighting the use of 90% confidence intervals for ranking ranges, and other changes. 
• Computer science rankings will be released when all the program rankings are released, now expected in fall 2010.
• Computer engineering will not be ranked, as there were not enough programs to meet the NRC criterion.

While this is an improvement in that rankings will not be based on flawed publication data, the lack of ranking of Computer Engineering departments, the lack of citation data for Computer Science publications, and the concern that even with best efforts by the CRA, the set of publications used by the NRC is incomplete—all create concern that the field may not be well presented in the final rankings. One continuing concern is that university administrators will naturally compare statistics across fields. If CS and CE are not fully and fairly reported, we may suffer in such a comparison, with artificially low productivity numbers compared to other fields.  So work remains, and the CRA Board will continue to work with the NRC and other groups to create a fair evaluation system.

Eric Grimson is the Bernard Gordon Professor of Medical Engineering and head of the Electrical Engineering and Computer Science Department at MIT.