4

This question is primarily for the Math/TCS/Systems community. I expect that similar situations (theory vs practice/applied journals) arises in other context as well so feel free to comment what is going on in your field. Also, it is not exclusive to journals.

I have what is in my opinion a good theoretical/algorithmic result that has many applications to system stuff like cloud computing/fault tolerant stuff etc. The paper is theoretical nevertheless, without experiments etc. I have two options:

  1. submit it to a good TCS journal (DAM/Algorithmica etc) without caring too much for applications and focus purely on the math stuff.

  2. Submit it to a more "system" journal (for example IEEE/ACM TON/ToPDS or Journal of Par. Dist. Comput. etc). Such journals could accept, on principle, more algorithmic results provided they have clear real world applications relevant to the scope of the journal.

My question to the TCS community: How bad/good looks on the CV a theoretical result on a -nevertheless good- system journal but not so visible to the TCS/Math community?

My question to the System community: How do you react when you see such a pure algorithmic result in such a journal? Even if it is motivated by real life system applications.

My question to the Industry people: Does it make more sense for my CV to demonstrate that I can publish to more system journals (like the above). How do you act when you receive application from TCS people?

The last question is in purpose general. Depends the position. But do some publications in more practical journal give extra credit to the applicant?

Update: For what is worth, in case anyone is interested I can share my mild experience. I submitted to a (prestigious, if it matters) system journal my work (which is rather heavy from technical/math point of view). The paper got accepted, but it strikes me that none of the reviewers spent time on the technical part (at least not very carefully). They rather cared on the applications part (and they were happy with it). I am a little disappointed since I expected people to check the math part, and this makes me a bit reluctant to try similar journals in the future.

Still, my question stands: How does it look on the CV from a TCS/Math point of view? Does it make any difference (positive or negative)? Does it look "suspicious" that I have chosen a system journal when I could perfectly choose an algorithmic/applied math one? The reason I did that, is because I wanted to show a little bit of "diversity" in my CV (plus the fact that I like systems in general). Does it make any sense?

  • Any particular reason for phrasing your question in terms of journals? Conferences are generally better vehicles of visibility and impact in computer science. – JeffE Jun 7 '16 at 12:41
  • @JeffE The only reason is that I would like to submit it directly to a Journal at the current state. But also the question makes also sense for conferences. Would someone go for a pure TCS conference like STACS/SODA/FST&TCS (I mention only upcoming conferences) or INFOCOM etc? – PsySp Jun 7 '16 at 12:45
  • 1
    My two cents are that I would view theoretical results published in a systems journal or conference with some caution, because, as you mention, it's likely that the mathematics have not been carefully refereed. If you already have a track record of publishing strong theoretical work, I would worry a bit less. – Sasho Nikolov Mar 12 '17 at 19:07
  • @SashoNikolov In retrospect, I completely agree with you. The reason that I chosen a (prestigious) system journal rather than a (decent-strong) theory journal was that I wanted to show some diversity also in case I apply for some industry job at the future and that work fitted in such context. All my pubs so far are in pure TCS venues. But I doubt I will do it again. – PsySp Mar 13 '17 at 11:03
3

This is not a direct answer, as I do not have that much experience with work crossing the theory/systems divide. I do have experience with the theory/bioinformatics divide, which is quite similar.

First, the publication venue should be chosen according to the intended target audience. If you believe your work is directly relevant to a community, you should publish it for that community. If it is only potentially relevant, you should probably publish it elsewhere.

I frequently read algorithmic papers published in theoretical algorithms, algorithm engineering, and bioinformatics. The fields have very different standards for what constitutes a good algorithmic paper. Proofs and analysis are the primary content in a theoretical paper. Experimental evaluation, if present, is considered secondary. The situation is reversed in algorithm engineering. As a result, a good algorithm in one field is often not that good in the other.

Things get a bit weird in bioinformatics. The algorithm itself is secondary content, as people are more interested in the problem you study and whether you have biologically interesting results. An algorithm is good, if it solves a relevant problem in practice. There are some bioinformatics papers with plenty of mathematical content, but they are exceptions. Proofs are seen more as technical details better suited to the supplementary material than as scientific contributions. Hence the reviewers rarely pay very much attention to them.

  • Thanks for the answer. Yes, the question was more broader. In the systems community is more less as you describe in bioinformatics: the focus is on applications rather than on the theory itself, although many theory papers appear in such journals as well. I wonder for example in your field, how a pure theory paper (motivated by bioinformatics applications) is perceived and vice versa. – PsySp Mar 14 '17 at 11:11
  • @PsySp The bioinformatics community values the conceptual content of a theory paper over its technical content. If the problem is relevant and the solution looks promising, the paper can serve as a basis for further work. I have no experience in how the theory community values work published in bioinformatics. – Jouni Sirén Mar 14 '17 at 11:48

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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