Balancing Acts: Successes and Challenges in CS&EJune 10, 2005
Talk of the Society
Paging through this issue of SIAM News, readers will see that it's an usually focused one, devoted almost entirely to computational science and engineering. Often described as the area of scientific inquiry that lies at the intersection of applied mathematics, computer science, and an application domain, CS&E is of interest to a substantial and growing subset of the SIAM community.
In fact, SIAM's largest activity group is in CS&E. With more than a thousand members, the group (which did not even exist five years ago) held its third conference this year (Orlando, Florida, February 12–15). A variety of speakers and session organizers, on the invitation of organizers Lori Freitag Diachin (chair of SIAG/CS&E) and Omar Ghattas (SIAG/CS&E's program director), contributed short articles and images for this issue of SIAM News.
Taken together, the articles should give readers an idea of exciting advances and directions, both in the widely ranging applications that intersect CS&E and in the science of computation itself. With a program reflecting the diversity of application domains--from astrophysics ("star in a box") to computational biology to reservoir modeling (porous media flows)---and the interdisciplinary nature of the research, the Orlando conference also illustrates many of the open challenges in CS&E.
Computational science has also been a recent focus in the national science policy arena. Just about a year ago, the President's Advisory Committee on Information Technology (PITAC), which is charged with exploring the information technology needs of the U.S., convened a subcommittee to study and report on computational science and the levels at which it is funded in the U.S.
As part of its charge, the subcommittee was to respond to a series of questions, including: Is the U.S. funding the right areas? Is there a healthy balance between short- and long-term research? Is there an appropriate balance between research on underlying computational techniques and application of computational science to problems in science and engineering? Consideration of educational and training issues was also part of the subcommittee's mission. Details can be found at http://www.nitrd.gov.
The report of the computational science subcommittee, which was chaired by Dan Reed, is scheduled for release shortly after this issue of SIAM News goes to press. But on April 14, in preliminary briefings to the full PITAC, Reed offered a preview of key findings and the likely recommendations of the report. When the report is released, the recommendations should be a spur to the entire CS&E community, calling on researchers to establish scientific priorities and on agencies to develop plans and accompanying budgets.
Based on Reed's remarks, the report can be expected to highlight the interdisciplinary nature of computational science, and the increasingly important role played by computational science in solving complex problems. Because of the interdisciplinarity, new organizational structures will be required, both in government funding agencies and in academia.
In the briefing, Reed identified four key elements: high-performance computing centers, architecture and hardware, U.S. leadership in high-end computing, and integration of the science and engineering communities to work on complex problems facing the U.S. This seems to indicate a balanced focus on the issues, including hardware, software, algorithms, and policy requirements, although the report will probably be silent on the magnitude of the effort required--budget figures, in other words, are likely to be absent.
The report will likely call for a multidecade roadmap that lays out a research agenda, identifying appropriate contributions to be provided by federal agencies, industry, and universities.
Perhaps the key point for readers of SIAM News is that computational science is a focus of interest at the highest levels of the U.S. government. By directing agency attention (and, we hope, funds) to computational science, the report, which PITAC itself will issue, could be influential. The report will reinforce other recent reports, including that of the High End Computing Revitalization Task Force (HECRTF) and the DOE-supported Science Case for Large-scale Simulation (SCaLes) report (written by David Keyes). Each of these reports lays the groundwork for U.S. federal programs and activities, but action from the community will be required to make them a reality.
As mentioned at the beginning of this column, Lori Freitag Diachin and Omar Ghattas, on behalf of SIAG/CS&E and the organizing committee for this year's SIAM Conference on Computational Science and Engineering, invited a variety of conference speakers to submit illustrations and short writeups capturing some aspect of their talks. It may be that SIAG/CS&E naturally attracts influential leaders, and it may be that the field is an extremely lively one.
I believe that both are true, but I also think that the researchers who contributed the writeups deserve special thanks. These contributors (a respectably high percentage of those invited) took the time to pull together images and descriptions that (we hope) will be meaningful to the entire membership.
I'm writing this column from Stockholm, where another impressive meeting of a SIAM activity group is under way: the eighth SIAM Conference on Optimization. Based on discussions with organizing committee chairs Anders Forsgren and Henry Wolkowicz, readers can look forward to similar coverage of this field, also central to SIAM.
But what, you may be wondering, of the important and productive research areas in applied and computational mathematics that don't turn up in the names of well-attended conferences, or don't have well-connected organizers, or aren't covered by any of SIAM's 15 activity groups? Are advances in such areas destined to play out unnoticed by readers of SIAM News?
Balanced coverage of our discipline is important to us, of course. And it's not easy to achieve! Our Annual Meetings are usually a good source of input as to innovative and important research developments. This summer, SIAM News editor Gail Corbett and I will be in New Orleans at the 2005 SIAM Annual Meeting (July 11–15) and the concurrent meeting of yet another well-established SIAM activity group, the sixth SIAM Conference on Control and its Applications (July 11–14). We hope that you'll seek us out to suggest a subject, perhaps a minisymposium talk you just heard, that you consider worth an article. Even better would be volunteering to write the article!
Spreading the word about our discipline, both within and outside the community, is an important, never-ending challenge. But to return, in conclusion, to people working in areas not splashily covered in SIAM activities, it may be that SIAM needs to take additional steps. If you have ideas or suggestions about structures or mechanisms for nurturing all areas represented by the SIAM membership, I encourage you to get in touch (firstname.lastname@example.org).