Wednesday, June 29, 2005

New Theory Task Force

A growing concern in computer science, and in theoryCS, has been the dwindling of research funding and support from the government (NSF/DARPA/etc) at a time when interest in computer science among undergraduates is dwindling, and the US appears to be losing its edge in basic science.

Especially in theoryCS, there is a growing sense that we need to articulate a vision of the field that is more about CS as a way of thinking, than solely as a field that develops IT infrastructure. Lance had mentioned TheoryMatters earlier: it's a new clearing house for information on why theoretical computer science matters, and how the current funding crisis affects the work that we do.

In the next phase of this project, SIGACT (the theory arm of the ACM), has set up a committee that will deal with funding, outreach and advocacy issues. Here is the original missive from Sanjeev Arora:
SIGACT has set up a committee that will deal with funding, outreach, and advocacy issues. It will also explore and promote funding opportunities for TCS in NSF and other funding agencies. The members are: Richard Karp (chair), Christos Papadimitriou, Avi Wigderson, Richard Lipton, Bob Sloan, Madhu Sudan, Sanjeev Arora, Michael Mitzenmacher, and 3 ex officio members from the Sigact EC: Hal Gabow, √Čva Tardos, Joan Feigenbaum.

One of the things we plan to work on is preparing a list of "research directions" for TCS in the next decade or two. Unlike a similar list from 2000 this one will be kept brief. It will have a few compelling items that will be comprehensible to nonspecialists and congressional aides. These will show that funding TCS is in the national interest.

We welcome your suggestions. Let's all put some thought into this. We should not be afraid to take on big challenges. Obviously, a brief list will not cover all of TCS but it is hoped that a long-term focus by funding agencies on some of the list items will benefit many or even most TCS researchers.

This is your time to speak up. If you have thoughts on this matter, (and if you are a theoretician, you should!), you may comment here, or even better, at the comments section of the TheoryMatters page (the password is the three first letters of "theoretical computer science", all in small case). Think big ! And don't be afraid to be speculative.

Here is my (first) contribution, on (what else!) computational geometry:
For the past many years now, biologists have been accumulating X-ray crystallographic structures of proteins, and developing a rich understanding of the crucial role that protein shape plays in determining its function. Today, the Protein Data Bank holds over 30,000 protein structures, and this number is constantly growing.

Understanding protein shapes, performing structural comparisons of molecules, and manipulating large molecular structures all require a sophisticated understanding of topology, geometry, and shape theory. In fact, one of the most active areas of research in geometry has been the area of "computational topology", which includes a large body of work on the use of topology in modelling shapes, especially proteins.

Being able to manipulate shapes efficiently is a core tool in doing the kinds of large scale structural biology that protein biologists have long dreamed about. Fundamentally, if I can classify the shape of a protein by comparison with others, I have learnt a lot about its function.

One goal of computational geometry should be to continue explorations in shape modelling and manipulation, with particular emphasis on the kinds of tools needed to compare and process large protein structures.

Update: There were some complaints about the make up of the committee. As Sanjeev Arora points out in the comments, Hal Gabow sent out a general invite to all SIGACT members, and not many people responded. It's important to note there are many ways to contribute to this effort.

Disqus for The Geomblog