48

From my understanding many CS conferences are highly selective and the papers are peer-reviewed. The CS field also seems to publish traditional journal articles. I do not understand the difference between the two and in particular I am curious about the advantages of having two different publication streams. The type of things I am curious about are:

  1. What is the difference between a paper published in a top CS conference and a top specialist journal?
  2. Is one more prestigious than another and if so how do they line up?
  3. Why not just publish conference papers as a special issues of a journal?
ff524
  • 108,934
  • 49
  • 421
  • 474
StrongBad
  • 104,237
  • 30
  • 262
  • 484

2 Answers2

34

I'm afraid this answer may be unsatisfying, but at its root, it just boils down to the culture of the field, as determined by early choices in its particular history. Now that it is well established, practices have developed that cement it in place, including clear distinctions between what goes in a journal and what goes in a conferences.

Michael Ernst has a very nice write-up of the typical way a computer scientist thinks about conferences vs. journals. In short:

  • Conferences are fast, higher status, higher selectivity, and higher visibility.
  • Journals are where you put a review article or a final "extended version" of a paper (the typical threshold is "at least 30% new material").

Journals are also important for getting tenure at lesser institutions, which apply the same standards to biologists and computer scientists. This also further lowers the perceived status of journals, since the best departments are less motivated to publish there than the worst.

jakebeal
  • 187,714
  • 41
  • 655
  • 920
  • 3
    Correct, but there actually are pretty reasonable journals that approach A+ conference level of selectivity and prestige (typically Transactions, but one should not make the mistake of assuming all Transactions are necessarily great). – xLeitix Feb 02 '15 at 15:36
  • 2
    Also, computer science has grown quite a bit. Things differ between different subfields. Bioinformatics has some fun combinations (ismb/eccb conference with proceedings in oup bioinformatics -- good journal by itself). – choener Feb 02 '15 at 20:16
  • 1
    @xLeitix Nothing says the journals aren't good, just that the perception is different. – jakebeal Feb 02 '15 at 21:33
  • 2
    @choener Absolutely: I see this in synthetic biology as well, which has a lot of overlap with computer science and currently no proper peer-reviewed conferences. I really hope that these hybrids will tend to bring more conference culture into these other fields, because I think it's a crying shame to have to wait for years to have access to a result because it's doing the rounds at various journals. – jakebeal Feb 02 '15 at 21:38
  • also VLDB and ECML/PKDD conferences experiment with linking conference presentations to journal publications. But in the "publish or perish" race it might be convenient to have two publications (conference and journal) instead of one:) – al_b Feb 03 '15 at 08:01
  • 1
    While this is mere speculation, I have often wondered whether the information presented is another reason why CS-subfields frequently opt for conferences. A chemist or a physicist will not repeat an experiment live after presenting their findings, a medical scholar will not perform a live treatment in front of a conference audience, and even an engineering scholar will not construct a building or bring heavy machinery to a conference. On the other hand, a computer scientist is, in various CS subfields, quite likely to bring an interactive prototype for conference attendees to try. – O. R. Mapper Feb 17 '15 at 08:43
  • "Conferences are — — higher status, higher selectivity": I do not agree here. Consider e.g. the Journal of the ACM; it is more selective than the top conferences in theoretical computer science. The conferences are certainly faster and more visible, though. – Jukka Suomela Feb 17 '15 at 14:17
  • I note that this page states "Journals have higher acceptance rates, giving the opportunity to get your research published." -- Is this due to fewer (higher quality) submissions to journals, or is it due to lower standards when reviewing journal submissions? – kwah Sep 23 '15 at 17:24
  • 1
    "...same standards to biologists and computer scientists..", exactly, one has to publish more in journals just because of such stupid regulations. – alim Apr 21 '17 at 17:13
  • @JukkaSuomela As far as I know, most JACM papers appeared in a conference first before being republished there years later. So I don't think it makes sense to compare JACM to a conference, as it performs a vastly different role. – Thomas Jan 23 '19 at 19:04
2

A good summary of positions papers and empirical research on conferences vs journals in CS is the paper by Bowyer Mentoring Advice on “Conferences Versus Journals” for CSE Faculty

Grudin also have a nice bullet-like summary and link to some 20 papers on the subject in http://research.microsoft.com/~jgrudin/CACMviews.pdf. One of Grudin's paper is a history of how CS became so attached to conferences (I dont remember which onebut it should be among the ones mentioned in the pdf).

Jacques Wainer
  • 237
  • 1
  • 2