Wednesday, July 09, 2008

Attendence at CCC: We have no edge people

This summer the Computational Geometry Conference (SoCG) and the Conference on Computational Complexity (CCC) were both held in College Park Maryland. Hence a direct comparison is possible. Here are some numbers:
  1. SoCG drew 140 people. Recent America SoCG confs that were not to FCRC or co-located with anything: 2006-AZ: 125 people, 2004-NY: 180 people.
  2. CCC drew 80 people. Recent America CCC confs that were not FCRC or co-located with anything: 2005-San Jose: 67 people, 2004-Amherst: 82 people, 2001-Chicago: 96 people. For more complete data see this post
Some thoughts on these numbers:
  1. SoCG has more people then CCC. Why? (1) more people are working in it, and (2) more people on the edge--- people in graphics or vision or Comp Biology or algorithms who, IF SoCG is local then these edge-people might go. CCC doesn't really have this.
  2. We do not have edge people. Who should our edge people be? Crypto? Algorithms? Combintorists? Quantum People? Philosophers/Historians/Sociologists of Science? For Crypto and Algorithms there have been some results (though not alot) of interest to them. For Quantum People they should care about Quantum Computing , but do they? Phil/Hist/Soc of science might be interested in seeing a young field where they can still interview some of the founders, but that is not the same as going to the conference. Besides, they probably don't have grant money.
  3. I can count about 5 people who often go to CCC who missed CCC08 and 10 more who I think should go (who am I to say they should go?) who didn't. Some of those who missed it had pretty lame reasons (e.g., a daughters wedding). Before trying to outreach to Crypto, Algorithms, Combinatorists, Quantum Physicists, P/H/S of science to go, we should get our own people to go.
  4. Are there othere potential edge-people I have overlooked?

21 comments:

  1. Not going to a conference so you can help prepare for and go to your daughter's wedding, is not a lame excuse.

    ReplyDelete
  2. So how many edge people (i.e., people not "core" CG but interested in graphics, vision, comp bio, etc.) actually attended this year?

    Regarding edge people for CCC, it would be nice to have more outreach to crypto. There is (I think), on average, 1 crypto paper at CCC per year and it would be nice to see more than that.

    ReplyDelete
  3. (Regarding my first question: I meant how many attended SoCG.)

    ReplyDelete
  4. SoCG is the premier conference for computational geometry, because we essentially kicked them out of mainstream CS.

    CCC is the conference for complexity losers who can't publish in STOC/FOCS. I know tons of complexity people who don't even consider going to CCC, unless they have to give a talk or something lie that.

    ReplyDelete
  5. Anonymous #1: The comment about a daughter's wedding being a lame excuse was a joke.

    ReplyDelete
  6. 1) There were 5 definite
    edge-people at CoSG, and
    a few more who are borderline
    edge (egde-edge?).

    2) In past years they've gotten more- about 10 plus.

    3) CCC is for losers commenter: rather than agree or disagree, this could be looked at objectively: do papers
    in CCC have less citations
    than complexity papers in
    STOC/FOCS. How much less?
    Or do a survey of all people in complexity theory about the topic.
    My point is, rather than
    state this as being
    obviously true, one
    should pin it down to
    an answerable question
    and answer it. Commenter:
    If you do this, you can
    do a guest post.
    (this comment is NOT
    sarcastic or anything of the sort).

    ReplyDelete
  7. The simple (and straightforward) answer:

    CCC'08: 80 attendees
    SoCG'08: 140 attendees

    140/80 ~ 1.75

    CCC'08: 74 authors
    SoCG'08: 124 authors

    124/74 ~ 1.68

    ReplyDelete
  8. CCC is for losers commenter: rather than agree or disagree, this could be looked at objectively: do papers in CCC have less citations than complexity papers in STOC/FOCS. How much less?

    http://arxiv.org/abs/physics/0604130

    ReplyDelete
  9. As a I mentioned in an earlier commnet, the big issues are travel money and competing venues. Unlike SOCG, Complexity people still attend FOCS and STOC, not to mention ICALP (the lure of Iceland), RANDOM/APPROX, COLT, in addition to complexity workshops in Banff and Dagstuhl this summer.

    For computational geometry, there is no competition to SOCG in the summer months (SODA being the other pole).

    The issue is not the absolute attractiveness of a location but the attractiveness relative to others.

    Paul

    ReplyDelete

  10. CCC is the conference for complexity losers who can't publish in STOC/FOCS. I know tons of complexity people who don't even consider going to CCC, unless they have to give a talk or something lie that.


    The above comment is immature and
    silly. Just look at the list of
    authors in CCC this year. Most of
    them publish quite frequently and
    happily in both STOC/FOCS and CCC.
    Repeated comments that STOC/FOCS is
    the only place for good research
    are boring and show insecurity.

    ReplyDelete
  11. Just look at the list of
    authors in CCC this year. Most of
    them publish quite frequently and
    happily in both STOC/FOCS and CCC.


    That's beside the point: nobody writes papers of a perfectly consistent quality level. Of course the original comment was immaturely and provocatively phrased, but there's a little truth to it. The SOCG authors frequently consider these papers to be their best work, and they want to be sure to attend so as to learn about the most important new work by others in their field. For CCC, this is less true. The papers there are certainly good, and sometimes great, but most people who publish in both CCC and FOCS/STOC consider the latter more prestigious and important. I bet rejected papers from STOC get submitted to CCC more often than those rejected from CCC get submitted to FOCS.

    ReplyDelete
  12. I think Paul's argument makes sense. I go to SODA, SoCG, and would like to go to FOCS/STOC everyyear, but my pub profile drives me towards DB/data mining conferences and I have to make a choice.

    For CCC people, I imagine FOCS/STOC *is* the "applied" area :)

    ReplyDelete
  13. Maybe because the SOCG people have more money to go to conferences because they do more applied stuff?

    ReplyDelete
  14. I think we have enough `edgy' people in our community as it is...

    ReplyDelete
  15. This being the complexity blog, it has a large following among bored anonymous commenters who want to write aggressive comments. The biggest surprise for me is that people still take the time to point out that comment X is "immature" or whatever :)

    That said, I agree with the intelligent side of "CCC is for losers" comment. The complexity people that I talk to (whatever "complexity people" means) define STOC/FOCS to be the premier complexity conferences. Of course they sometimes publish in CCC, but it's like algorithms people publishing in SWAT (nobody will claim it's the top conference when sending a paper there).

    For geometers, the top conferences seem to be SoCG and SODA. It's sad that STOC/FOCS are nowhere on the list.

    @James: The number of authors seems like a circular explanation IMHO. My peronal impresson is that these conferences grew as large as possible, with the only restraint that the papers should still be "good."

    1. I remember a SoCG PC chair (who shall remain anonymous) declaring at the business meeting something like: "We accepted all papers that people who are not authors might have any bit of interest in."

    2. For CCC, I distinctly remember a complexity person telling me that "My paper is so pathological, they might even reject it at CCC."

    ReplyDelete
  16. Bill Gasarch asks: "We do not have edge people. Who should our edge people be?"

    As an "edge person" (quantum system engineer) relative to complexity theory, it was interesting to me that no-one has attempted to answer Bill's question.

    Three changes that would attract more edge folks like me would be increase the weighting of: (1) upper bounds relative to lower bounds, (2) non-rigorous design rules relative to rigorous proofs, and (3) tools relative to results.

    But I can appreciate that from the point of view of many academicians, all three would be changes for the worse! So people are bound to feel passionately about this question.

    Is CC all about proofs and truth? Or is CC all about tools and federation? ("federation" being the new "application")

    A strong case can be made either way.

    ReplyDelete
  17. Mihai,

    I think that your perception is one of ignorance. I just served as PC chair for this year's CCC conference and so got quite a good sense of the quality of the both the submissions and accepted papers. The comparison with SWAT is way off base. From what I have seen of recent SODA conferences I would say that the quality range of this year's CCC papers is not that different from those SODA conferences.

    Also, historically, CCC has had some classic results: the first communication complexity lower bounds for disjointness, the first time-space tradeoff lower bounds for SAT, and a host of other significant results that are easily of the level of strong FOCS/STOC papers (including a number of strong results this year).

    It is true that with the overlap with STOC, authors often have had to choose between STOC and CCC submission. However, CCC is a sufficiently strong venue that STOC PC members have chosen to publish their good (or even their best) work at CCC rather than waiting for FOCS.

    ReplyDelete
  18. "That said, I agree with the intelligent side of "CCC is for losers" comment."

    As unintentionally funny comments go, this is a pretty good one.

    ReplyDelete
  19. Lets all bow to Mihai's knowledge
    of conferences and their quality.
    Comparing CCC to SWAT???

    ReplyDelete
  20. I wonder where is the SWAT chair to defend his conference to the hilt - SWAT is way way better than ELGAR - and then the ELGAR chair is notified about this discussion, and so on.

    The egotistical attitude of people in this community is ridiculous. As for mip, the flavour of comments he/she makes should be done anonymously for the sake of his/her own good.

    ReplyDelete
  21. Estimated impact of publication venues in Computer Science (higher is better) - May 2003 (CiteSeer)


    110. SODA: 1.51 (top 9.00%)

    229. IEEE Conference on Computational Complexity: 1.11 (top 18.75%)

    ReplyDelete