Friday, March 11, 2011

Conferences, Journals and Computer Science


In my field (and also in most physical and biological sciences), journals are the preferred destination for our research output. There was a time when conference proceedings had a lot of prestige, but they started losing their sheen in the 1970s.

While conferences themselves are doing well, it is the proceedings whose reputation has suffered: the decline and fall have been so precipitous, someone joked recently, that proceedings now rank barely above e-mail spam!

Things are so completely different in Computer Science and allied fields, where conferences rule.

This difference in disciplinary cultures has always intrigued and fascinated me. So I was glad to find this paper by Jonathan Grudin who narrates the story of "why computer science in the U.S. shifted [away from journal articles] to conference publication in the first place."

A key quote:

Technology and a Professional Organization Drove the Shift to Conference Publication

By the early 1980s, the availability of text editing or word processing among computer scientists enabled the relatively inexpensive production of decent-looking proceedings prior to a conference. This was something new. Anticipating that libraries might shelve proceedings, ACM printed many more copies than conferences needed, at a low incremental cost.

ACM also made them available by mail order after a conference at a very low price. Papers in ACM conferences were thus widely distributed and effectively archival. These are the two features that motivated the creation of journals centuries earlier. [Bold and italic texts are from the original]

Are there other explanations for the shift [which, Grudin says, was largely US-centric, and didn't spread to Europe]?

* * *

Thanks to Suresh at The Geomblog for the pointer.

7 Comments:

  1. ajitjadhav said...

    As many intellectuals at the Ayn Rand Institute have noted, the computing industry, esp. that in the USA, has historically been relatively freer of coercive government controls, which in turn has led to the relatively astouding growth and pace of innovation which we witness in it. Also, a sense of adventurism, if you will, Abi!

    In short, computing industry in USA is relatively more capitalistic.

    As in any capitalistic economy, the inventors are left free to protect their intellectual property via patents. For freshly minted ideas, this means: *not* fully disclosing the idea---not until the time its *commercial* rights are secured (even if the IP strategy of the company is only defensive, in the sense, "we will not wield IP as an offensive tool to crush competition, but let's at least follow that "first, 'CUA'" advice).

    The best work in computing fields has been done not in the "pure university/academic" mode, but primarily in those dynamic start-ups (and also sometimes in big cos/labs which succeed in at least simulating that environment).

    If you want to protect your property but also wish to "disseminate" (i.e. advertise) your ideas, conferences are ideal---you don't have to give out all the details, and yet, you can create a citable reference which would eventually be useful while filing your patent claims.

    The usual journal model fails for such objectives. So does the "pure" university model. For instance, I do oftentimes find myself laughing at the PhD theses dealing with not only object-oriented paradigm, or modeling paradigms, but even compiler technology, yes, even the theses done at Stanford/MIT/UIUC/Berkeley/CMU/etc. And, I often feel almost like a dumbo when I read conference papers presented at those games developers' conferences, where the norm is not to have a PhD (and yet, none insults you if you do have a PhD). For example, check out Euler-based CFD, or nano-simulations etc., presented in nVidia conferences or similar.

    The reason this culture or mode of publication didn't spread to Europe is because USA, by and large, and esp. in the CS field, still is more capitalistic than Europe. An Indian need not belabor this point while talking to another Indian; suffice to ask how many of his talented friends work in SF Bay Area or other places in USA, vs. those working in Europe/elsewhere.

    The reason the culture didn't spread to other engineering fields is because the coercive government regulations are far tighter in those fields.

    QED.

    --Ajit
    I will really, really try to observe brevity the next time.

    [E&OE]

  2. Arun said...

    Publishing in a conference is a lot faster than the processes associated with a journal, right? I think this is the most important reason. In Computer Science, things change very fast, especially because of the easiness to create applications (you only need to code, no need to make changes in your flow shop) and also theory (once you get an idea, it may be easier to realize it and test it in Computer Science than in other fields) - so your competitor can catch up real fast. Also, the presence major high technology companies and start-ups who respect new technology and is interested in making money out of an innovative algorithm, rather than depending more on marketing and politics as in other industries. All of this related to the fact that the Computer industry grew from nothing to this ubiquity in such a short time.

  3. Arun said...

    And, IMHO, even in Europe, computer scientists prefer conferences over journals - a comment I often hear from my friends who are working in the field.

  4. Arun said...

    Another interesting thing across disciplines are the differences in order of authorship - in some alphabetical, in some the exact order of work done or in others the first author who did the most work and the last one who advised most/lab head.

  5. Desi Babu said...

    Professor Abinandan,

    Sorry for posting something out of context, but I thought you might enjoy this news story!

    http://www.telegraphindia.com/1110311/jsp/frontpage/story_13698981.jsp

    I found the HRD ministry's comment rather hilariuos: “His name was selected on merit. We never expected that he would refuse the job,”

    Since all sorts of questionable characters queue up to get a VCs job, they could never dream that someone would actually decline the job for a lack of courtesy on the appointing authority's part.

    Peace!

  6. WebMiner said...

    Even inside IIXs, it is common to get a letter in one's mailbox saying that one has been nominated by the Director to serve on some committee or another, while forgetting to get consent first. Most of these committees never meet and have no discernible purpose, and I know many people who forget that they got such a letter or that they have any meeting to attend. I guess it is a little more difficult to live such a life as a VC.

  7. Genghis Khan said...

    WebMiner,

    Your appointment letter probably says that you are supposed to perform duties assigned by the director (mine does). And, no mention of getting any consent. Those duties are just assigned.