A frequent dilemma of many researchers refers to the question of where to submit their work to. This question arises both with respect to conference presentations and journal publications. I will start with considering this question when applied to submission for possible conference presentations.
Commonly used criteria for deciding to which conference to submit refer to the conference's prestige and deadline. In my opinion, a much more important criteria is maximizing the correspondence (or fitting) between the work to the conference. That is, I suggest to submit the work to a conference that offers the most appropriate audience for this work. Details follow.
The above criterion is best illustrated when referring to the most common dilemma: the dilemma between submitting to a TOC-wide conference (like STOC or FOCS) or to an area-specific conference (like CCC, CRYPTO, TCC and RANDOM). In such a case, the question that I suggest to ask is whether people outside of the specific area should really care about the work in question. The answer may be positive if either the work is really a very significant contribution for the specific area (and so those wishing to have some vague feeling of where this area stand better hear about it) or if the work contains some ideas that have a clear appeal (and maybe even application) outside the specific area. If the answer to this question is negative, although the work may be very good, then one better submit the work to the specific area conference rather than to a TOC-wide conference. Note that even if such a work is accepted for presentation in STOC/FOCS, such a presentation may totally miss the point because it will be only attended (or actually absorbed) by experts in the specific area whereas more such experts are likely to attend the specific area conference and be more open to such a presentation there.
I wish to stress that even when focusing on prestige, it is not clear that presentation in STOC/FOCS is always better than presentation in a special are conference. Indeed, the line in the CV may look more impressive, but if the consensus is that the work should really not have been presented in STOC/FOCS then such a presentation may actually back-fire (because all that the experts may remember is their annoyance at the selection, while forgetting the paper's merits that would have been appreciated in the specific area conference).
Turning back to the above question, let me further illustrate my advice by considering a hypothetical work in the theory of cryptography, which may be submitted either to STOC/FOCS or to TCC or to CRYPTO/EUROCRYPT. Here, I would ask whether this specific work carries a message also for non-crypto members of the TOC community (e.g., as is often the case with work on pseudorandom generators or zero-knowledge proofs), and if so then submission to a wide-TOC conference makes sense. Likewise, one may ask whether this specific work carries a message also for non-theory members of the crypto community, and if so then submission to a wide-crypto conference makes sense. But otherwise, although the work make be very good, I would think that it should be submitted to TCC.
The dilemma regarding journal submission seems similar, except that here deadlines (or journal-backlogs) are usually irrelevant. On the other hand, here the consideration of the potential audience is somewhat less acute, but still it exists. My warning of the back-fire of being included in a too prestigious venue, remains valid, though.
Of course crypto is in scope of any general TCS venue, but my point (and generic policy) is that there is no point in submitting papers to 2nd/3rd tier general TCS venues (rather than to specialized venues, whenever they exist). In fact, in my view, the only justification for 2nd/3rd tier general TCS venues is to serve areas that do *not* have specialized venues.
Back to Oded's page of essays and opinions or to Oded's homepage.