16

I'm currently a post doc in theoretical computer science and I have a double digit number of conference publications, about half of which are at so called "tier 1" conferences (SODA,STOC,...). Unfortunately I only have 1 journal publication so far, mainly due to coauthors dragging their feet.

Will the lack of journal publications have a great impact when applying for a faculty job in cs-theory at some point, and is it advisable to try to up this number before applying? Or is this fully compensated by having sufficiently many conference papers?

6
  • 5
    no one cares about journal pubs for hiring. Now for later, that's a different story.
    – Suresh
    Dec 12, 2012 at 19:40
  • And while the OP is asking about computer science, the answer is likely the same in most fields.
    – Ben Norris
    Dec 12, 2012 at 21:12
  • 4
    I suspect that disciplines like math and bio care a lot more about journal pubs, since they don't have a culture of peer-reviewed conference papers.
    – Suresh
    Dec 13, 2012 at 1:04
  • 1
    @Suresh: I wouldn't say no one. The big question that hiring committees ask is "Will this person get tenure?" A track record of pushing papers through the journal pipeline is one of the requirements for tenure at most places. It's at most a minor third-order concern, but it's not nothing.
    – JeffE
    Dec 13, 2012 at 3:56
  • 4
    Or is this fully compensated by having sufficiently many conference papers? — This is the wrong question. If you have sufficiently good conference papers, then the lack of journal papers won't matter. Quality is way more important than quantity.
    – JeffE
    Dec 13, 2012 at 3:59

3 Answers 3

14

CS people know that CS conferences count for more than conferences do in other fields, and although they count a little less than journals in CS, this should not be a problem. Of course there are still rankings between the various forums.

The fact that CS rank conferences higher than other disciplines only becomes a problem if you are compared against non-CS people, which often happens when competing for funding and promotions, in my experience.

6
  • 1
    Actually meant "more than conferences in other fields". I should just stop for today. Dec 12, 2012 at 19:58
  • Everbody I've talked to agrees that (even high-profile) conferences don't review properly and accept based on interest rather than rigor and substance. It's quite confusing that they are valued more than journal articles, which are (apparently) usually better in this regard.
    – Raphael
    Apr 25, 2014 at 9:06
  • 2
    @Raphael: Having just been on the ECOOP PC, I would have to disagree. Interest is certainly an element, but as there are loads of papers choose from, rigor and substance generally won out. No paper was rejected for being uninteresting (to the people in the room). Apr 25, 2014 at 9:11
  • Afaik, PC member don't read all the papars closely? You (w|c)ould have been the victim of sloppy reviews without even knowing.
    – Raphael
    Apr 25, 2014 at 9:14
  • 2
    @Raphael: PC members do not generally read all of the papers carefully, but it would be untrue in general to say that they don't read all of their assigned papers closely. There are always cases of sloppy reviews, but top conferences try to avoid these, even if it means assigning someone to review a paper while the PC meeting is going on. Overall, there is some truth to your sweeping statements, but in general, at top venues, they are false. Apr 25, 2014 at 9:27
6

Since this was a point of discussion on my answer here, I felt this might be a good place to offer an extended analysis (since it also serves as an answer to the current question).

I had stated:

While in some fields, conference papers are akin to talk abstracts, in areas like computer science, conference papers can be very meaty and there is a high churn of papers in conferences. Top conferences can have acceptance rates around 10%, and as such, A+ conference papers are often held in high regard within the community: these venues are far more competitive than many of the good journals. Still, even in the CS area, metric-wise (for hiring, positions, funding, etc.), journals will often still count for more than a conference following the norm in other academic fields.

To which Suresh countered:

the last line of the answer is just not true at all.

and later:

It's based on my experiences (at different levels) evaluating work and people, and watching how others do it, and how people talk about research work. It's not an absolute statement, but CS departments have spent many years in battles with deans convincing them that conferences count for more than journals.

And JeffE:

"even in the CS area, metric-wise (for hiring, positions, funding, etc.), journals will often still count for more than a conference" — Speaking as the chair of the faculty recruiting committee in a top-5 CS department: This is simply incorrect. I don't recall anyone on my committee ever pointing out a CS journal paper in any junior candidate's CV. (For interdisciplinary folks, it's important for research to be published in journals in the other area: Biology research in biology journals, for example.)

This is counter to my own experience in applying for jobs, and in my current department, where journal papers are held in higher esteem. I know there are Computer Science Departments with their own by-laws and policies whereby conferences can be given as much weight as journals, but this is why I said "often", not "always". It is also my experience that such departments tend to be the more famous ones: the ones with enough clout to have their own policies. Smaller/less-well-known departments (which are by their nature more numerous) often have to abide by wider faculty/collegial policy.

In any case, to try to put some meat into the answer, I tried Googling for some tenure-track and hiring criteria in Computer Science. I'll pick out relevant quotes:

1) Department of Computer Science at the University of Vermont: Guidelines for Tenure-Track and Tenured Faculty

Publication of refereed articles in both journals and conferences is very important; in many areas of computer science, publication in top-tier conferences is considered as prestigious as publication in top journals.

2) Department of Computer Science & Engineering, University of Minnesota: Criteria for Promotion and Tenure

A 1994 NRC Committee on Academic Careers for Experimental Computer Science stated “The requirements for good research and engineering in experimental computer science and engineering (ECSE) are different from those of many other academic disciplines” and then added “Because conferences are the vehicle of choice in ECSE for the dissemination of research, well-refereed conference proceedings (as well as work published in refereed private journals) should be given as much weight as archival journal articles in evaluating a candidate's research portfolio for promotion and tenure”. Certain proceedings articles should be weighted equally or even more heavily than archival journal articles when evaluating the candidate's research contributions

3) Duke University, Computer Science Department, By-Laws

[...] recognising the important role played by conferences, book chapters, and other non-journal research documents.

4) University of Stockholm: Tenure-Track Position, Department of Computer Science

The quality of the research should be documented by presentations at well-es­ta­blished conferences and publications in high quality international journals and conferences.

5) Computer Science Department: Appalachian State University: Criteria for Promotions and Tenure

Scholarly Activity

Examples of Indicators of Excellent Performance

  1. Publishes an article in a refereed journal.
  2. Publishes a textbook. [nothing about conferences]

6) Wright State University, Department of Computer Science By-Laws

Primary indications of quality normally include, but are not limited to, the following:

  • publication of research results and of extended scientific and engineering reviews in peer-refereed journals of acknowledge stature (particularly those of scholarly professional societies such as the ACM and IEEE);
  • ...

Secondary indications of quality include, but are not limited to, the following:

  • refereed conference proceedings;
  • ...

The first two departments explicitly state that conferences can be considered as highly as journals, the next two state that conferences will be considered alongside journals without any explicit ranking, and the last two state that conferences count less than journals.

These are six examples, and are biased due to having "Computer Science" in the title, meaning that they are CS-specific criteria (whereas I suspect many smaller departments would fall under more generic by-laws).

Though many CS departments (esp. the more prestigious ones) have won the battle of giving conference papers their due weight, it is still my understanding that other departments have not, and must fall in-tow with faculty/collegial guidelines.

So to answer this question:

Journal vs conference publications when looking for a job in in computer science

It depends. Though I may not have fully substantiated this here with only two anecdotes, I still hold by the position that in many CS departments (particularly smaller departments), journal articles often (not always) count for more than conference papers by merit of simply being journal articles.

(Some related discussion here.)

3
  • 3
    The letter of the bylaws is not the same as what actually happens 'on the ground'. I'd argue that if a department (as you suggest) is preferring journals to conference, they are going contrary to the actual practice of research in many CS subdisciplines.
    – Suresh
    Mar 22, 2014 at 23:12
  • 2
    What Suresh said. Also, you're conflating hiring policies with promotion and tenure policies. Your original claim was about the former; your later evidence addresses the latter.
    – JeffE
    Mar 23, 2014 at 2:45
  • Suresh, I don't disagree. @JeffE, the question is indeed about hiring, but my original claim was not solely about the former; verbatim: Still, even in the CS area, metric-wise (for hiring, positions, funding, etc.) .... The quotes I provide refer to positions, but also refer directly to how research output is reviewed in Computer Science.
    – badroit
    Mar 23, 2014 at 16:26
5

I don't think this will be a problem for your first faculty position. In TCS people will appreciate it if you publish journal versions of your conference proceedings. However nobody will expect, that you will do this immediately. Also, imho, having SODA/STOC and FOCS publications is what counts. Other things (teaching experience, grants, making a journal publications out of your conference contributions) are important but secondary criteria.

It is likely that many non-theory CS people are in the hiring committee. In other fields of CS journal publications count even less.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .