Wednesday, June 03, 2009

Thoughts from STOC

I haven't been to STOC/FOCS since FCRC in San Diego two years ago. Since then I dropped thirty pounds and got a haircut last week. Amazing how many people thought something looked different about me and asked about the haircut.

I missed the Valiant celebration but it brought a large number of senior theorists who don't often attend STOC/FOCS. I enjoyed seeing and talking with them. Still, despite a relatively easy to reach location (say compared with last year's Victoria), STOC only attracted about 260 attendees, nearly half students. We just don't get draw many non-local attendees who don't have papers in the conference. Though David Johnson told me he's attended every STOC since the early 70's.

I don't attend many talks at STOC, I prefer to hang in the hallways chatting with people, about research, politics, gossip and baseball. The talks I did attend usually had about five or so good minutes of motivation and results followed by technical details for experts in that particular area. That's probably the right way to give those talks but I was rarely one of those experts.

And then I see a talk as amazing as Moser's and I realize: This is why I love theory.


  1. "We just don't get draw many non-local attendees who don't have papers in the conference."

    There was no special reg fee for (non-industrial) postdocs who often have a (small) fixed award to be used for travel/conferences. I guess they could register as students, but not advertising a lower rate for them is not very inviting. (i.e. normal NSF postdocs usually get about 1500 per year rather than the 15000 that is to be awarded to the "stimulus package" postdocs.)

    Also, I think many people feel self-conscious attending without a paper. I've heard many people/even students express this sentiment.

  2. From all I'm hearing, it appears I missed a truly great talk (Moser's)... And it makes me wonder: when we will enter the 21st century, like the theoretical physicists?

    They've been doing this kind of thing since the late 20th century!

    -Amit C

  3. That's probably the right way to give those talks

    Is it? My advisor used to say a conference talk should be an advertisement for the paper. I'm of the mind that most of a talk should be understandable by most of the audience, and the people who are experts will read the paper for the technical details. If the talk and the paper are both only understandable to the experts, how will any cross-fertilization happen?

  4. The worst talks are those that spend most of their time hashing through difficult technical material where a momentary loss of concentration is death to further understanding.

    Though they are generally better talks, I am almost equally frustrated by talks that spend all their time on motivation, background, related work, and statements of results and completely ignore any aspect of the main ideas in their proofs.

    There has to be a happy medium. Out of 20 minutes the old rule of thumb was that there should be at least 5 minutes on the main new ideas and at least 10 minutes on explaining the work to a general audience: (e.g., 10 minutes intro and statement of results, 5 minutes technical overview or sample, 5 minutes summary and open problems.)

    Maybe the two-page overviews for FOCS will encourage people to design their talks around these overviews so that talks will be accessible to more than just the experts.

  5. I am a student and I don't feel like attending stoc/focs unless I have a paper. The stoc/focs research community is very uninviting and unwelcoming as a whole.

    If you don't have a paper, you are probably unworthy of interacting with people who do.

  6. The stoc/focs research community is very uninviting and unwelcoming as a whole.

    I have felt this too; I recently moved from the stoc/focs community to a more applied one, and it is amazing how nice even very senior people in other communities are, compared to the stoc/focs community.

  7. If you don't have a paper, you are probably unworthy of interacting with people who do.

    I am probably relatively senior (with respect to the poster of this comment) and have published several times in STOC/FOCS as an outlier. I have also being amazed at the "stand-offishness" everytime I have attended with one or even multiple papers to present. My usual group of colleagues (while being no less competitive) are incredibly welcoming and friendly by comparison. I have put down this sort of behavior to the fact that most papers in TCS are based on one or two (possibly very clever but still small ideas) and people are probably always on guard to "protect their turf" so as to speak. In contrast, in my main area of research, it is not possible to write papers based on one or two ideas, and people are much more comfortable in being expansive about their work.

  8. It is due to this condescending and unwelcoming behavior of the stoc/focs community that I have moved to more applied areas despite starting off (a few years ago) as a theory PhD student.
    Now I've grown to like the more applied research areas much more (not to mention the warm and welcoming behavior of the people in the community) so I doubt I'll ever return to "hard-core" theory.
    Some day, however, I'll come back and write a stoc/focs paper or two, just to prove to myself that I am left the area because I liked the other areas more and not because I was as incapable of doing theory as the stoc/focs-regulars make others feel.

  9. I am a recent theory graduate and I feel that the senior people in the STOC/FOCS community are for the most part, not really unfriendly, but rather unwelcoming and unapproachable at conferences. If there is someone you know after meeting them once or twice, they are just not going to act like they want to talk to you at the coffee break or sit near you at lunch. I get a very strong feeling that you are supposed to "know your place".

  10. I have a few STOC/focs papers but not in the mainstream areas. I feel that senior people are not as approachable as other research area like pure math. I donot understand that. They have tenured positions in good universities, but why do they still act so defensively and make sure that their cliques have everything?

  11. The aspect that I found most intimidating about attending these conferences as a grad student was how shockingly young so many of the big names in the field were compared with the math faculty I knew.

    However, one of the best opportunities in attending as a grad student was the chance to meet and network with grad students from other departments (who sometimes felt just as I did). These students have become friends as a result of our repeated meetings each year.

    I also realized that many people use the conferences as unique opportunities to work on research problems together face-to-face, taking many hours of the day away from the talks to meet in corners. This level of intensity and drive was not something that I had expected. It took a little bit to learn to distinguish when their conversations were social (when I always found that people were very open and friendly) from when their conversations were part of their joint research (when there were not so welcoming of interruption).

  12. Regarding the best way of giving a talk, I think it should be possible for a focused non-expert to follow the entire talk, or a distracted expert. The true experts will already know your paper, so you can't just focus on them.

    The best talks are those that have a simple motivating example, that is then extended into a simple explanation of a special case of the proof. Not every result can be presented in this way, though.

    The really bad talks are always those where the presenter has not given any thought to non-experts. Aside from this one case, though, most talks are okay.

  13. Senior researchers being unwelcoming is least of the problems at conferences. When I attended my very first complexity conference as a graduate student, a senior researcher came to me after my talk and said very rudely "I do not like it" and walked away. I had to ask somebody who that was. Let us call him S. S continued to be nasty to me in future conferences whenever we met.

    The irony is that I had never heard of S till my first conference and, to date, several years later, I have never read a single paper by him nor have I seen him cited in any meaningful fashion. He has published a lot of papers and has been a full professor for many many years. I have heard that he did a lot of service to the complexity community during its nascent years.

    My personal view is that some people who are able to advance their careers by playing the publication game without really advancing the field have deep-seated complexes and expect young people to curtsy to them. Fortunately, I think S is an exception to the rule and most people in complexity theory are not like him.

  14. THe real problem with our TCS is not the rudness of the senior memebers or some such stuff. It is the total dependence on conference as a publication media. In mature fields like maths and physis the emphasis is on journal publication.Conference
    and workshops are ment for newcommers and there will usually be tutorial sessions and youngsters will very well find it welcome. And dont let the old excuse "rapid" publication be repeated. We should learn something from those physicsts.

  15. THe real problem with our TCS ... is the total dependence on conference as a publication media. In mature fields like maths and physis the emphasis is on journal publication.

    We all recognize that this is a (possibly even the) major problem facing TCS. But what can we do to change it i.e. who wants to bell the cat ?

  16. I was just comparing the quality of review
    vis a vis the time taken for acceptance in physics (journal) and tcs (conference).

    I am comparing an reasonable
    APS journal with a high
    class TCS conference


    Time for notification (2-3 months). If
    a refree sits on stuff a new refree is sought. Review is detailed. In depth
    comments even when stuff is accepted
    help a great deal in improving the quality of exposition. Often Authors are
    blinded to their sloppy writing because
    it is their work.


    time for notification (1-2) months:
    review is often

    "The paper is terrible not interesting
    reject" or

    "Interesing accept"

    Both these reviews in my opinion
    is totally useless.

    Who looses in the long run. In my opinion every one.

  17. I am now a fourth year graduate student, and I find the community very friendly. I am thankful our community is so friendly, especially after talking to other graduate students at my university about their communities.

    The first few times I was at Stoc/Focs, a colleague and I made a practice of going up to researchers we wanted to meet(/know better/hear about their research) and asking them if they would go to dinner with us. It worked great! (though, admittedly, it felt terrifically awkward).

    Some researchers go to the conferences to work with collaborators and are thus protective of their time. They will politely tell you they already have plans. But most people are more than happy to dine with you. If you are intimidated by the senior professors, try the other graduate students.

    Sometimes our community (myself included) lacks good social skills and situations turn out stickier (or at least more awkward) than necessary. But the vast majority of people in TCS want the community to thrive and are eager to help out new comers to the field.