Tuesday, June 10, 2008

SODA 2009 Call for Papers- and a comment on it

(Claire Mathieu requested that we post this.)

The web site for submitting papers to SODA 2009 is up and running. Please see the "Submissions" button in the right-hand side menu of the page at SIAM.

A few important differences from previous years:
  1. There is a pre-submission deadline of June 26, one week before the final submission deadline of July 3. By June 26, authors must have entered a title and short abstract of the paper which they intend to submit. (One goal of this is to speedup the assignment of submissions to program committee members.)
  2. The submission must contain complete proofs. (Part of the proofs can be relegated to the appendix if they do not fit in the requisite number of pages.)

BILL'S COMMENT ON THE POST: I wonder why they require complete proofs. Did some big result fall apart recently? Is this a good idea? Will it cut down submissions? Drastically? Will the proofs be badly written? very badly written? ignored? Why is SODA doing this and no other major theory conference has required this? Will I refuse to submit my proof that P=NP because of these requirements?


  1. Bill: maybe they are experimenting? Personally I wonder how one can accept a paper to a top conference without the proof: if it is not there and you do not need because you can imagine it, many see it as a weak paper, and if it is not there and you cannot imagine it, you will depend on the names of the authors to know whether you trust the result or not...

  2. One of the positive outcomes could be to force people to write complete proofs in the first place.
    Without this requirement many researchers will not bother writing up a journal version (or even a techreport)
    if their paper is accepted.

  3. As somebody who works in a very different part of TCS (Logics and Semantics), I am always very surprised when I hear that conference papers are not accompanied by a journal version. In my field, if an author doesn't have a journal version to go with the extended abstract, the community assumes that there's a problem with the research.

    One of my conference papers does not have an associated journal version. Why? Because a reviewer for the submitted journal version found a serious mistake that I have been unable to correct.

  4. If you haven't noticed, there are three more excellent changes :
    1) Submitting authors are encouraged to post their submission on their website.
    2) A list of accepted papers, along with the short abstracts, will be posted on the internet.
    3) Submission uses Easychair (as opposed to the old e-mailing system)

  5. The 3:00 pm June 26 date seems UNREASONABLE. The FOCS notification date is only June 25.

    Either this hurts anyone who happens to be away on the notification date or everyone who might possibly think that their FOCS paper won't be accepted will submit an abstract which will make much of the process a waste of time (though it will leak information about who had a paper accepted at FOCS who was not confident of their chances).

  6. I think this is a good positive development. I've recently seen quite a few papers that stretch the notion of "we omit the details from this extended abstract", by claiming results for which there is no written proof. In some of these cases, the conference version appeared several years ago, and no journal version exists. This seriously hinders research: no one else can publish a proof (unless its a major result), and the community is left with an unproven claim. (Often such claims are padded onto to other otherwise reasonable submissions, which would explain the acceptance.)

    So yes, I think requiring proofs is an excellent development. I wonder how easy it would be for the PC members to adapt, and enforce this requirement.

  7. The website says "Acceptance/rejection notices will be sent to authors via e-mail in September 2008." It would be nice if they gave the exact date (and preferably put it before the STACS submission deadline of September 15).

  8. Anonymous #5:

    Why don't you pre-submit your paper to SODA before knowing the outcome of FOCS, in case it gets rejected from FOCS. If it gets into FOCS, then just don't send in the full submission.

  9. I don't like the idea of sending abstract a few days before the regular deadline. I know it may help the PC a little, but it's a hassle for many people submitting the papers.

    As for the deadline for the abstract: what's the problem with having it half a day after FOCS notification? If (sadly) one's paper is rejected, then normally, the abstract will stay the same for the next conference. So, I wouldn't see it as evil. But I would prefer to have it on July 3, the same day when is the regular deadline.

  10. I think it's a great idea to require complete proofs, and I hope STOC and FOCS follow suit...

  11. I think it is a good idea to have the abstracts if it helps the PC Chair allocate the papers to the PC in a more effective way. With the PC and number of papers so huge, any measure that is helpful in this respect should be taken.

    People just don't like things to change from the way they are, which is why the above person complained.

  12. Am I the only one who thinks that it is a bad idea to have a rejected paper from FOCS almost "automatically" sent to SODA? And then to STACS?

    I know FOCS is hard. But SODA is a good conference, too. And personally, I always tend to think that if one of my papers got rejected, i have to improve it.

    I really hate that trend of preparing a sequence of submissions for one paper. People don't bother sending their paper to the appropriate conference, they aim at FOCS, cause they know it will just go unmodified to the next conf if not accepted. The impact on the community is HUGE: look at the number of papers submitted to all these confs! Increases the work for referees, also decreases the time they can spend on good papers.

    And finally, as a reviewer, if I were to reject a paper from, say, FOCS, because i think that it is correct but not strong enough, i will also give comments about the general presentation of the paper. If I see the same paper at the next conf, I expect to see the comments implemented!

  13. Sorry for entering this fray late -- one of the main problems with our community is that authors are not writing up journal papers once they are accepted to a top conference. This is not a good long term strategy for a field. I have been reading
    a book on the history of Poincare's conjecture (now theorem), and its amazing how many times it was claimed to have been solved until gaps were found in claimed proofs. Even papers in FOCS and STOC are not carefully scrutinized to verify all the details. This is inevitable if there are so many papers to be reviewed in a few weeks. The journal process is the main one I know of that works. If a result is very important perhaps many people will read the paper and bugs will be discovered. If the result is not so important that a lot of people will read it immediately, then its important for the community to
    have authors publish the complete work in journals where it is actually carefully read by a few reviewers who ask questions about gaps and the authors have a chance to respond.