2010
06.19

I would like to suggest that to invigorate the SoCG conference, we need to expand its scope (well, the PC/community need to do it, but you get my drift). In particular, I believe that one can easily make the case that certain fundamental topics in data-structures and algorithms are as relevant for developing algorithms in computational geometry as, for example, discrete geometry is. Such topics include (IMHO):

  1. Data structures – heaps, trees, dynamic data-structures, etc.
  2. Lower bounds and hardness of approximation (on data-structures, algorithms, etc).
  3. Algorithms on graphs or metric spaces.
  4. Algorithms for other computation models (IO efficient, cache oblivious, multi-core, cloud computing [for example, Map Reduce]).
  5. And probably many others.

There is currently no good conference like SODA taking place around June (unless you count STOC, which accepts only certain kind of such papers), and I think SoCG can be a good venue for such papers.

That is all.

Update: There is also an ongoing discussion about the future of STOC. I think my suggestion is more modest – I suggest to expand the scope of SoCG and see what happens. I am less excited about colocating SoCG and STOC, because STOC is usually considerably more expensive than SoCG to attend. However, there are some clear advantages to colocating both conferences.

  1. (this is not a bad thing, but) wouldn’t this merely be a conversion of SoCG into a summer algorithms conference ?

  2. This would work well if the proposal to enlarge/kill STOC actually went anywhere. Fortunately, NOBODY seems to like that proposal.

  3. I see the expanded SoCG as a narrower conference than SODA. Without papers on quantum computing, crypto, etc. It would be strictly on topics that are directly relevant to developing geometric algorithms.

  4. It’s definitely doable if the PC wants to do it. but it would have to be a systematic effort over a few years. coupling this with a STOC colocation could also be cool.

  5. I doubt that the community/PC would be uninviting if people submit papers. The trick is to get people to submit and to some extent it has happened, I think. The main issue, however, is the fragmentation of the overall TCS/algorithms community into many conferences. We cannot attend too many so people either attend if they have a paper or to a “home” conference. Not clear that this is ideal and it can only get worse over time.

  6. Good suggestion. How do we implement it?

    We have in the past more actively sought to expand scope in various applied directions, with mixed success (from my perspective, we have caused members of our community to look in these directions, more than we have caused those working in these directions to come to our community. There are, however, several directions (GIS, radiation treatment, structural molecular biology) in which there is a great appreciation for those who do attempt to cross over.)

  7. I think in this case its easy – the call for paper mentioned more topics and emphasize that the scope had been enlarged, and people pass the word around the SoCG now accepts papers in these “new” topics. It is really not a great effort. Am I missing something?

    I think expanding the scope of SoCG in more applied direction is much harder. Here, we are trying to expand towards people that already know about SoCG and would be (hopefully) happy sending us their papers.

  8. I agree with Sariel. Expanding towards applied areas is basically a nonstarter because it’s easier to publish in those applied venues. Case in point. I had a paper rejected from SoCG because the reviewers felt it was more appropriate to publish in the relevant applied community (even though the paper itself was pure theory). Expanding towards other areas of theory is easier, because it’s a “different kind of theory” and is easier for the PC to evaluate. Also, SoCG is more “well known” in theory circles anyway.

    I recall Jeff Erickson making a concerted effort to do this when he was chair. It was hard though: there’s still a lot of PC resistance. That’s originally why I thought workshops might be a gentle way of getting people more familiar with certain topics. But maybe that’s too slow.

  9. BTW, I think having workshops in concert with the conference is a great idea.