2

I was taught that conference papers in computer science count as publications. Therefore, a journal paper cannot be presented in a conference because of plagiarism reasons, and vice versa.

I confirmed this information:

The weird part for me is understanding which authority (IEEE?, ACM?) regulates this? Is there any document that support this? It seem that this policy in CS emerged from the community (bottom-up; see for example this) – sort of like hand shaking or bowing in Japan – and is therefore self-regulated by people.

Related to this question (EDIT: I will ask them separatedly later as suggested in the comments but I feel I need to clarify certain issues I seem to be confused with):

  1. I have had multiple arguments with colleagues of other fields that disagree with me in that conferences papers/presentations don't count as publications (I am in a multidisciplinary field). I feel I tend to loss these argumentations because it is one CS student against the rest. In a way, it would be easier if I could refer people to some authority.

  2. If I simply don’t want to follow the rule for conferences that apparently don't count papers as publications, what are the real consequences apart from the social lynching? To be clear, the work is done but not published - just somehow presented in a conference (*). Is there any way to confirm that a conference or proceeding is formally registered as an (or to an) editorial/publisher?

  3. Is there any other way of presenting published results in a conference as, for example, simply abstracts (*)?

(*) In some areas people just present abstracts, and depending on the area, the conference accepts almost everything. The main goal of this seems to be networking - presenting the work and having the opportunity to interact with people that assist to your presentation (if any).

  • 5
    I am no computer scientist, but there may be too many distinct (yet valid) questions in your question. I suggest to ask your questions separately, one by one (so you can build upon the answers from your previous questions). As for your question 2b (“How can I substantiate that a publication at a conference is ‘valid’?”), you might specify the background for which you need this. – Wrzlprmft Jan 12 '16 at 21:32
  • 1
    SIGMOD and ICDE (premier CS database conferences) now allow the presentation of papers now published in the TODS and TKDE journals, as posters in the conference. Note, that this is not a separate publication but instead it is used for promoting the journal publications to the conference audience. – Alexandros Jan 13 '16 at 12:14
  • I was taught that conference papers in computer science count as publications. Counting publications is fundamentally silly. What matters is the quality of the research you do. It's not a sport where we tally up a score. – Ben Crowell Jun 15 '16 at 23:37
  • @BenCrowell: As far as I know, "to count something as a publication" is synonymous to "to consider something a publication". It has little to do with "counting publications". – O. R. Mapper Jun 16 '16 at 6:30
7

There is no central authority that regulates anything. Each journal and each conference have their own policies about such things. In any case, it is not in general true that in computer science "a journal paper cannot be presented in a conference because of plagiarism reasons, and viceversa". My answer to another question explains more about the difference between conference and journal publications and how they interact with each other (the answer is not area-specific, but it was informed by my specific knowledge of computer science publication conventions).

As for your other questions, they are a bit vague but I'll try to answer them:

I have had multiple arguments with colleagues of other fields that disagree with me (I am in a multidisciplinary field). I feel I tend to lose these argumentations; and basically nobody really believes me because I would need a bunch of people in my field as proof.

If I understand correctly what it is that you are claiming in your arguments with your colleagues, it sounds like they may be in the right. However, it's a bit unclear what your position is precisely, so perhaps you should try to clarify that.

If I simply don’t want to follow the rule, what are the real consequences apart from the social lynching. In a multidisciplinary field we are already kind of being punished when we hear that conference papers don’t count as publications. Is there any way to confirm that a conference or proceeding is formally registered as an (or to an) editorial/publisher?

Again, I'm not sure what rule you are referring to. As I said above, there is no centrally regulated system of rules. Each journal or conference will have their own publication policies which you should comply with. These are not "rules" in the sense that disobeying them will get you arrested, but not complying with these policies will quickly get your papers rejected and/or get you labelled as a dishonest or unethical researcher. If you are after an academic career, it is completely counterproductive to do this, and if you don't want an academic career, I don't really understand what would be your goal in trying to work around the publication conventions in such a way.

Is there any other way of presenting published results in a conference as, for example, simply abstracts?

Again, I disagree with your premise (at least in the generality in which it is stated) that one cannot present published results in a conference. I'm in math rather than CS, but I have done precisely that many times (Edit: to clarify: in math, conferences usually don't publish proceedings, so "presenting" does not equate to a publication). Even in many CS conferences, you could do that by taking care to first submit an extended abstract version of your paper to the conference, presenting it there, and following that up with a more detailed paper which you will submit to a journal. The order does matter, as the answer I linked to above discusses, but that's the logical order of doing it anyway, so I don't really understand what the source of your frustration is. If you add more details to your question I'll be happy to discuss them.


Edit: Here are some more thoughts addressing the very helpful clarification comments you left below my answer.

  1. I sympathize with your money issues, but honestly you are mixing two completely unrelated things here. The fact of the matter is, publishing in a journal first and then in a conference is less logical and rather defeats the entire purpose of the system of dual conference-journal publications (and, to be clear, I rather dislike this system myself and think it has some serious flaws, but it does have a certain logic to it, which you are trying to work against because of the unrelated money issues).

    What I think would make sense for you to do is to determine what is the optimal way to publish your research from a scientific point of view, ignoring the funding issues, and then go to your advisor and/or department, explain the problem to them and ask them to help you achieve that goal. If they are running an honest graduate program, it is their responsibility to provide grad students with the resources they need to succeed in their work. So, I think that problem ought to take care of itself if you go about things in a mature way instead of trying to work around the system, flawed though it might be, and ignore well-established conventions in your field.

  2. As for "conference papers don't count as publications", that statement simply does not have a well-defined truth value, and in fact, I think asking questions like whether a conference papers "counts" or does not "count" as a "publication" (whatever that is) is misguided and unhelpful. Here's what in my opinion is a more helpful way of thinking about this topic: a conference paper is precisely what it is, nothing more or less - it is a form of publication that has certain characteristics in terms of the rigor with which it was evaluated, the level of certainty that it is correct, etc. When you are evaluated by others in connection with job applications, promotions, etc., it is important that they understand the context for your achievements so that they can reach an informed opinion about how good your work is. When working in a multidisciplinary field, you are correct that there can be a risk that the people evaluating you are not sufficiently informed about the conventions in some of the fields you are working on to have that context. I agree that this can be frustrating and worrying, but I think there is only one correct way to address this concern, and that is to provide that context in an honest, clear, concise and readable way in your CV and other documents (research statement, publication list). If your department is unfamiliar with CS conference proceedings papers and how much they "count" for, explain to them what such papers are about; e.g., have two sections in your publication lists, one for journal publications and another for conference publication, and include a small footnote to explain the distinction.

    On the other hand, do not be tempted to try to game the system in any way, by, for example, withholding that context, trying to pass off a conference paper as a journal paper, ignoring editorial policies of a journal or conference you submit to, or any other means that can be viewed as dishonest or unethical. That will only lead to trouble.

To summarize, I understand from your comments that you have some legitimate concerns about funding travel to conferences and about having your work be evaluated by people who do not understand publication conventions in CS, but it seems to me that you are contemplating some rather ill-advised means for addressing those concerns. Academia is not perfect and one often encounters such less than ideal situations in almost every discipline, but I find that a little bit of creative thinking, maturity and honesty will almost always overcome them.

  • In my 1st question, I was refering to colleagues that always insist that conference papers don't count as publications (in CS they do, at least most of the time and I am not sure to tell when not because of the community assumption problem). In my 2nd question, I am just refering to conferences that don't count the papers as publications in which case the social punnishmment still exist. – toto_tico Jan 12 '16 at 22:43
  • Regardles the 3rd question, my frustration starts because I have to apply to travel founding (international PhD student here) prior to trips and acceptances. The award deadlines are usually way before the submission deadlines and it requires a lot of anticipation and, in terms of timing. Basically, it is just better (safer in terms of money) to publish in a journal first. Probably that would be different if I would be in a CS department, but things get fuzzy when doing working in a multidisciplinary area. Conferences should be for divulgation, journals for publication. – toto_tico Jan 12 '16 at 22:49
  • to be clear, the source of the order problem is that conferences usually require an author to assist triggering travelling expenses. It is comparatively easy to get founding for a journal paper once accepted. So, it is better to publish in a journal first, and then attempt to divulgate the results afterward (without having to produce a slightly different work to avoid the plagiarism - consider being PhD students again). – toto_tico Jan 12 '16 at 22:56
  • 1
    I had a lot to say about your comments so I've added more thoughts to my answer. Hope you find this useful. – Dan Romik Jan 13 '16 at 8:48
  • 1
    If there are no proceedings then there is no paper, so I'm not sure what you mean about whether it counts as a publication. And if there are proceedings, then as I said I don't consider the question "does the paper count as a publication" a meaningful or helpful one, so I can't say anything more about that, sorry. As for the money issues, yes, it is unfair, and as I said I sympathize, but sadly I can't solve all your problems. All I can do is provide some feedback that the approach of trying to manipulate the system to address those issues comes across to me as highly problematic. ... – Dan Romik Jan 13 '16 at 19:28
8

As is, a journal paper cannot be presented in a conference in many fields, not only in computer science, whether the conferences are counted as publications or not. Of course, you can always present an improved version of a previous work, if the improvement is significant.

It seem that this policy in CS emerged from the community (bottom-up) (e.g. see this), sort of hand shaking or bowing in Japan - and therefore self regulated by people.

The entire academic world is somehow self-regulated by people, and most people agree on the fact that trying to resell the same paper twice is unethical because it can lead to an artificial inflation of the number of publications, which is one of the many indicators employed to hire people in academia.

And even if you don't find that unethical, you don't want to read the same publication again and again, do you?

  • I understand the double publication problem (which I do consider unethical). It is a matter of presenting results to promote your own work and to establish with people in your area. My third question addressed that, how do you present results without entering in ethical dilemma? and my second question adressed the way around, how to present in a conference that won't count the acceptance as a publication? – toto_tico Jan 12 '16 at 22:26
  • In any case, I have also found in many instances very related work published in a journal (long version) and then in a conference (short version). It might or might not be differently enough but I found myself citing the two papers because I am not able to tell which is the most relevant (specially if it is recent work) so the problem of the same publication exists already. – toto_tico Jan 12 '16 at 22:34
  • 3
    @toto_tico Most conferences allow the publication of extended papers after the conference, once the proceedings have been published. For the extended papers to be accepted, there should be at least some 30% of new material and the proceedings should be duly cited. Thus, a short answer to your third point can be: plan your publications according to the above requirements. In fact, this is exactly what I do: I know what the important conferences in my field are, and I plan first a short paper and then an extended paper to be published later on. – Massimo Ortolano Jan 12 '16 at 22:43
  • 1
    @toto_tico Anyway, I agree with Wrzlprmft that you've posed many questions that would be better asked separately, because answering all of them at once might be overwhelming. – Massimo Ortolano Jan 12 '16 at 22:43
4

One reason conference proceedings are accepted as publications in computer science is likely because CS moved much faster than physics or mathematics. In the latter a multi-year review process was acceptable, but not in the former. So the latter accepted conferences only as a presentation platform, but not as final verification/peer reviewing opportunity.

It may be that in future, with accelerated publication channels, this will change and even CS conferences may lose the stamp of publication (except for the top ones), but cultural inertia is usually slower.

However, as long as proceedings count as publications, they should not overlap significantly with a journal publication - publications are the currency of the scientific community. And while there is no central committee enforcing that the currency is valid (i.e. publications are valuable and original), the members of your narrower community carry out social control. Compare that to pre-state social control in a small village. You didn't need a central police force, the decentralised village community was perfectly sufficient for this. In fact, historically, this would be the earlier stage and science mostly still works like this, because of expertise horizon.

  • 2
    When you speculate about speed of the field as a reason for the popularity of conferences in CS, you should probably also consider that at least in some subfields of CS, the demonstration of the subject of research is a crucial part of what is being presented. Even an applied chemist will usually not set up a small lab on stage to show an experiment live, in front of the audience, nor will an applied biologist bring any flora or fauna into the conference hall - but in applied CS, anything that can be shown on a screen is usually shown on a screen, as a running demo. – O. R. Mapper Jan 13 '16 at 13:47
  • @CaptainEmacs, I thought I was starting to get clear with the answer below of Dan Romik, in the sense that if there is proceedings, then the paper will count as a publication regardless the field. However according to your answer this is not the case in other fields, i.e. even with proceedings there is no publication. – toto_tico Jan 13 '16 at 17:24
  • @CaptainEmacs, I guess the peer review aspect is an essential way to decide on my question 2. Is it conclusive? At least, I will add (Peer-Reviewed) on the side of my conference papers in my CV. – toto_tico Jan 13 '16 at 17:27
  • 1
    Talking about peer-reviewed publications is like saying pesetas are money; they are, but so is the US dollar. The value depends on the community and the conference in which the papers are read. There are forgiving communities, but in other, you can lose reputation. In computer graphics, there is SIGGRAPH, everybody submits there. NIPS/ICML deadlines are fixtures of Machine Learners' calendars like Christmas and Easter. But other conferences are just meeting spots, and, again others, are considered 2nd/3rd tier. Depending on your field, you may want to avoid the latter, peer-review or not. – Captain Emacs Jan 13 '16 at 22:07
0

The main reason why conferences with formally published proceedings do not accept work which has been already published or accepted elsewhere is to avoid duplicate publication (https://en.wikipedia.org/wiki/Duplicate_publication). It is just the same as Journal B will not accept/publish a paper already accepted/published in Journal A. Such rules are regulated by publishers, conference organizers, and the whole ecosystem is coordinated by the Committee of Publication Ethics (COPE) - http://publicationethics.org/

In other fields, conferences do not have formally published proceedings, just sth like a book of abstracts. That's why you can present the already published work there.

While you are asking about presenting journal work at a conference, the opposite case is quite common in CS. For instance, you publish sth at a conference, then extend this with new results, more details, state the increment and voila - you can submit this new work in a journal.

  • Ah, what will happen if you do duplicate publication - one or both papers will be retracted and you can get some shame in your Uni/community... – al_b Jan 14 '16 at 9:41

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.