Tuesday, February 24, 2009

The FOCS Call

The 2009 FOCS Call for Papers has a few changes from previous calls.

"Optimization" is added to the list of topics.

The 2009 Call requires full proofs in the papers with no promise those proofs will be looked at.

Authors are expected to include full proofs of the central claims in the paper. If necessary, these proofs may appear in a clearly marked appendix. However, any material not included in the body may be ignored at the discretion of the Program Committee.
I have nothing against full proofs but I hope authors still worry more about giving the right intuition in their submissions than about dotting every i.

The Call also asks for a brief description due a week after the submission deadline.

The brief description should be no more than two pages, using the same font size, margins and spacing as the extended abstract. It should contain an informal description of the paper. It may contain an overview of the main results, concepts, or ideas introduced in the paper. It should provide the same understanding conveyed in a brief conversation or presentation. It may replicate material from the extended abstract, or even be a copy of its first two pages. But, it must not contain any technical material not present in the extended abstract.
This one puzzled me so I asked Dan about it.
It will have a few uses:
  1. To make it easier for us to get more opinions about the contribution of a paper.
  2. To make it easy for committee members to get a feel for papers they have not been assigned.
  3. Ideally, to reveal the essence of a paper, unencumbered by the need for formal presentation.
We will soon produce and post examples of what we have in mind.
So the Call asks for both details and essence. Interesting.


  1. This is more work to the authors but I must say I like the changes. The proof requirement will reduce false positives, and in the long run will strengthen the publication status of a conference. The two-page requirement will reduce false negatives by allowing PC members to structure their work-overload better (the board should require a high read-ratio of the two-page descriptions).

  2. The submission requirement states that the page limit is 10 pages, not including bibliography, figures, and technical appendix.

    I'm conflicted on this, because my paper is easily under 10 pages if all figures move to the end and are not counted against the page limit. But the figures form the heart of the intuitive understanding of the paper. In the final submission, to be recorded in the 10 pages of proceedings, I would surely include the figures and leave out the proofs. But why are figures singled out specially; why not just, "Any material, figures or otherwise, can be placed in a technical appendix"? For that matter, where is the bibliography supposed to fit in the final submission if it is not counted in the 10 pages?

    As far as the brief description, I'm still confused and look forward to seeing examples. Good introductions already satisfy needs (1)-(3), and are usually under 2 pages. My interpretation is that the brief description is my chance to brag about the proof techniques (which I usually leave out of an introduction), at a high level. Introductions often have something like this, so maybe this is a way for the PC to force good introductions without actually forcing; the brief description is the introduction that the PC wishes you would write.

  3. For that matter, where is the bibliography supposed to fit in the final submission if it is not counted in the 10 pages?

    The final submission format will squish your paper to nearly half its size. From my understanding, this "10 pages" business for submission has nothing to do with whether the paper will fit in the final proceedings. It has to do with you saving the reviewers' time: they want you to highlight 10 pages' worth of material for them to read. From this perspective, it makes sense why they don't count figures and bibliography; a reviewer will not spend much time on them anyway.

  4. I agree that the intuition behind a proof is more important in a presentation than "dotting all the i's." But I've seen plenty of conference papers that say something along the lines of "The proof is not difficult and is omitted due to space constraints, but will appear in a full version." This type of statement can have several problems, which are what I'd guess the committee is trying to avoid:

    A) The full version sometimes (often?) never appears.
    B) In the ones I've read, the proofs have in fact been easy to reconstruct, but I imagine there are plenty of cases where this is not true.
    C) I also imagine there are plenty of cases where the "easy" proof the authors had in mind does not in fact work, because some "i" can't be "dotted."