Saturday, October 20, 2012

Short Announcements

With a shout out to the friendly folks attending FOCS this week, some short announcements.

Read the STOC CFP before you submit the paper. There are significant changes to the submission format and procedure. Deadline is November 2.

Complexity will be co-located with STOC in 2013. Submission deadline is November 30.

The new Simons Institute for the Theory of Computing has a call for workshop proposals and research fellowships.

There will be a symposium to celebrate a new professorship named after SIGACT and STOC founder Patrick Fischer at Michigan on November 5 and a celebration of the 80th birthday of Joe Traub on November 9th at Columbia.


  1. Does one have to submit a proof of inability to produce a full paper to be able to attach an appendix? (I assume "attach an appendix" is the same as "submit a paper that has an appendix".) I hope I'm not the only one who thinks requiring authors to maintain and submit multiple versions of the same paper at the same time is not an improvement. Neither for the PC nor for the authors.

  2. Bill and Lance --

    Maybe one of you can contact Joan (the STOC PC chair) to get an explanation in the change of the call for papers, then broadcast it here? I see positive aspects in the change, but I would rather hear from the committee about why they did it than speculate.

  3. Joan doesn't actually read blogs as far as I can tell, so I'll try to respond.

    Many of us (or at least I) don't like the current system where you're supposed to submit
    a) a 10 page single-column version (submission)
    b) a full version (either required with submission, or you just have to do it eventually)
    c) a 10 page double-column version (under a short-ish time limit) once your paper is accepted.

    Personally, I hate having to do 3 versions as an author. Often, if the paper is short enough, the full version and the final conference version are the same (modulo formatting), and that is not so terrible (modulo possible time pressure). I especially hate having to submit in a 10 page single-column version, which is generally too short for what I'd like to say, and requires randomly throwing things into "appendices" to meet what seems to be a rather poorly chosen length requirement.

    By the way, some people think the 10 page single-column is better for reviewers, because it's "shorter". I completely disagree. Because people meet this restriction by randomly throwing objects into appendices, it generally doesn't save me reading time-- rather, it takes longer, as I have to go through flipping back and forth to find things in appendices if I try to read in order, or go back and re-read for context when I want to read something in an appendix.

    So, speaking my own mind, I find the process that has historically developed unpleasant as both an author and a reviewer.

    Under the new call, you have to turn in
    1) a 10 page double-column version [which should be very close to what you finally turn in if your paper is accepted, modulo whatever final changes you make, based on reviews]
    2) optionally and encouraged, a full version if needed.

    The idea is we've done away with the 10 page single-column version, which, for reasons I've described above, seems like a pointless inefficiency to me. Indeed, I expect that 10 2-column pages is enough to be the full version for a majority of the papers submitted (* perhaps I'm wrong there, we'll see) so now many authors will have -- modulo their preferred formatting -- just one version.

    Further, this seems good for reviewers, in that they'll see, as closely as possible, what the version of the paper that would be accepted to the conference is supposed to look like. This should lead to better reviews and, I believe, makes for easier reading of papers.

    One complaint I have heard is some people find 2-column ACM format hard to read. I certainly accept and acknowledge this criticism. (Though, at least electronically, there are various tools to correct this.) An alternative -- turn in an "equivalent" 1-column version which would translate to something like 16 pages -- was discussed, but in the end the decision was that the benefit of seeing what the paper would approximately actually look like in its final form was a sufficient benefit for the reviewers (and, arguably, the authors, who would have less work to do once their paper was accepted) that we went with this submission approach.

    I was going to comment previously when I saw Jason's comment that I don't get at all what he's saying or what his point is. The purpose of this proposal was to reduce the maintenance of multiple versions. I think this approach will be better, and certainly don't see how it can be worse, than the historical approach that's developed. I'm very familiar with this submission style in that pretty much every other type of conference outside of theory works this way, and I've found it's much less work both as an author and a reviewer.