9
$\begingroup$

I come from mathematics and, for whatever reason, am trying to publish in theoretical computer science. I'm still trying to understand the role of conference proceedings, and I have two specific question about them. I have tried to phrase the questions so that the analogous questions in mathematics would not be opinion based; if you think they are opinion based here then that is interesting in an of itself, so I would be grateful if you could comment to explain this!

In maths, a common marker of academic prestige is the specific journals someone's articles appear in, while I understand in TCS conference proceedings are more important.

Are some conferences better thought of than others? If not, then how do conference proceedings add to prestige?

In maths, there are four un-disputed "top" journals (Acta Mathematica, Annals of Mathematics, Inventiones Mathematicae, Journal of the American Mathematical Society), and then there is a large second tier of world-class journals which is too large to mention here and varies wildly between different areas of mathematics. Getting a single paper in one of these four journals is a fast track to a permanent job.

Are there a small number of top, world-class conferences where getting an article in the proceedings stands out like a gold star? If so, what are they?

Any suggestions for further reading is also appreciated :-)

$\endgroup$
  • 3
    $\begingroup$ I wouldn't say TCS conferences are necessarily "more important". The truth is that TCS people quickly get bored and want to explore new ideas instead of polishing their work for a shiny journal publication (something that I think is a problem), so many important results never wind up in journals. Having said that, it's certainly prestigious for example if your TCS paper gets published in Annals of Math, J. ACM, special issues, and such. $\endgroup$ – Mahdi Cheraghchi May 6 at 15:11
  • 2
    $\begingroup$ Are you aware that you can "publish" in a CS conference, and then publish in a journal after? (These are not considered two separate publications, but it does potentially let you try to reach two audiences with the same paper.) If you don't know about this I might write an answer w/ some more details. $\endgroup$ – Joshua Grochow May 6 at 15:41
  • 1
    $\begingroup$ @JoshuaGrochow Yes, I am aware of this. My plan is to submit to a maths journal after sorting out the conference proceedings. (How maths journals, which are not necessarily familiar with CS, deal with this is another question which I don't know the answer to though!) $\endgroup$ – user1729 May 6 at 15:52
  • 1
    $\begingroup$ @MahdiCheraghchi But while J. ACM is prestigious, it isn't of the level "a fast track to a permanent job", either. $\endgroup$ – Jukka Suomela May 6 at 16:25
  • 2
    $\begingroup$ One additional point that hasn't been made yet: while no single publication in a conference will get you a job based just on the line it adds to your CV, it can help you get jobs because it helps with visibility and networking. Many people regularly check when the list of accepted papers are announced for the top conferences, and many people attend the conferences, see the talks, meet people there, etc. $\endgroup$ – Joshua Grochow May 6 at 20:10
7
$\begingroup$

I don't think there is any TCS conference that would be a fast track to a permanent job.

STOC, FOCS, and SODA are the most competitive conferences, and papers in those will certainly look very good in your CV, but just one paper won't help that much with the career. To really stand out, you will need something more than that (say, lots of papers in STOC, FOCS, and SODA).

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks for this. I googled your acronyms, and (according to Wikipedia) noticed that the Knuth Prize is awarded every year alternately at STOC and FOCS while the Gödel Prize is similarly awarded at STOC and ICALP. Should ICALP then be mentioned here, or is it a step lower? $\endgroup$ – user1729 May 6 at 15:59
  • 3
    $\begingroup$ @user1729: While these things aren't always clear cut, I don't know anyone who would consider ICALP to be on equal footing with STOC/FOCS/SODA. $\endgroup$ – Peter May 6 at 16:03
  • 5
    $\begingroup$ It also depends on the sub-area. For example, for a paper on computational complexity that appeared in CCC, I would put that on par with an algorithms paper that appears in SODA (but I'd put both of those below FOCS/STOC). I think pretty much everyone agrees that STOC/FOCS are the top generalist TCS conferences. $\endgroup$ – Joshua Grochow May 6 at 16:09
  • 1
    $\begingroup$ @user1729 From what I can tell, ICALP carries more prestige for work closer to logic and computability theory. But, when it comes to algorithms & complexity, it is definitely below STOC/FOCS/SODA. $\endgroup$ – Thomas May 7 at 13:59
  • $\begingroup$ csrankings.org has lists of top conferences. For algo&complexity it gives STOC/FOCS/SODA, as mentioned here many times. For logic&verif it gives CAV/LICS. For PL it gives PLDI/POPL. In general, you can go explore various branches. $\endgroup$ – Radu GRIGore May 23 at 13:15
5
$\begingroup$

I agree with the others that the answer to your first question is definitely "yes", and would like to add another online resource where you can have an idea of the difference in "prestige": it's the (in)famous CORE conference ranking, which ranks CS conferences (not just TCS) on a grade scale A$^\ast$, A, B, C, with A$^\ast$ being the top tier conferences. Just type in the acronym of the conference and you'll get its rank. The ranking itself is highly debatable (and I personally try to avoid giving it too much importance) but the A$^\ast$ conferences tend to stand out as "top level" without much disagreement within their respective communities (below A$^\ast$ things get much more opinionable: there are A conferences in my field that I would never want to publish in and B conferences that I consider much more prestigious).

Concerning your second question, like other people said, a single article in an A$^\ast$ conference might not be a "gold star" on its own but consistently publishing in them definitely is.

Also, the answers given so far implicitly advertise the equation "TCS = algorithms and complexity" (which I guess is the dominant North-American perception) but the situation is essentially identical in every other field of TCS that I know of. A few examples: logic in computer science (where the flagship conference is LICS), the theory of programming languages (where the top conference is POPL and, to some extent, ICFP), or even machine learning (I am less familiar with this field but definitely NIPS/NeurIPS is considered a top conference).

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks. Just to clarify your final paragraph: in these different sub-fields, FOCS, STOC and SODA are not really applicable? (Which is my interpretation from looking at their scopes, but I'm not confident here.) $\endgroup$ – user1729 May 7 at 11:56
  • $\begingroup$ The sub-areas of TCS have historically often been divided into "volume A" and "volume B" work. ICALP, and the journal TCS, have such divisions, as well as the handbook of TCS. Compare this answer: cstheory.stackexchange.com/a/19661/2367 Possibly FOCS/STOC/SODA focus mainly on volume A work, but I might well be wrong. And LICS and POPL lean towards volume B. $\endgroup$ – Hermann Gruber May 8 at 22:16
  • $\begingroup$ Hermann is correct about the "A/B" divide. Actually, until the late 70s/early 80s, people working in logic in computer science would publish at STOC (and I guess FOCS too), but for whatever reason (I honestly don't know) those topics were increasingly marginalized, resulting in the creation of a specific conference, named LICS. Today, I think that the overwhelming majority of LICS papers would be considered off topic at STOC and FOCS (and even more at SODA). Same for POPL, which focuses on programming languages. $\endgroup$ – Damiano Mazza May 9 at 5:03
  • $\begingroup$ @HermannGruber and DamianoMazza Okay, that's interesting. I do know about the divide (and that my work is in the "B" camp), but didn't know how important it is. Is it then the case that my question splits into two questions: what are the "top" A conferences, and what are the "top" B conferences? $\endgroup$ – user1729 May 11 at 16:13
  • $\begingroup$ Well, as long as you know that TCS has several subfields, and that the existence of "top" conferences is acknowledged in every subfield, I don't think it's necessary to split the question. If you want a list of "top" conferences, I think that that references you got in the answers are pretty good. Also, as I said, the CORE ranking is fairly accurate with A$^\ast$ conferences, you will hardly get a more exhaustive list that that... of course with the proviso that every ranking is debatable. $\endgroup$ – Damiano Mazza May 13 at 8:45
4
$\begingroup$

In Italy we don't have "fast tracks to permanent jobs" (especially in Academia), but we have a complicated algorithm that ranks CS conferences (we have to take it into account for the National Scientific Qualification):

http://www.consorzio-cini.it/gii-grin-scie-rating.html

According to this one, STOC, FOCS and SODA are in Class 1 (top class), CCC and ICALP are in Class 2, etc.

| cite | improve this answer | |
$\endgroup$
0
$\begingroup$

I guess SODA: dl.acm.org/conference/soda and FOCS: ieee-focs.org will be it

| cite | improve this answer | |
$\endgroup$
  • 12
    $\begingroup$ Uh...what about STOC? STOC and FOCS are isomorphic... $\endgroup$ – Joshua Grochow May 6 at 13:33
  • 1
    $\begingroup$ I like the use of the word "isomorphic" here, my head's spinning now with possible definitions :) $\endgroup$ – Mahdi Cheraghchi May 6 at 15:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.