Sunday, July 20, 2008

Journal Policies -- IEEE/ACM Transactions on Networking

The Transactions on Networking, or ToN, is changing its policy somewhat on papers. Before, it used to be the page limit was nominally 12 pages, but you could pay $200 per extra page for up to 2 extra pages. Now, the page limit will nominally be 10 pages, and you can pay $220 per extra page up to 4 extra pages.

As I've mentioned before with Transactions on Information Theory (which is, I know, already changing its policy on correspondences), I don't understand page limits for journal articles. An article should take the space it needs for the authors to adequately express their ideas. Admittedly, for ToN, this is less of a problem than for ToIT; 14 pages is generally enough for most any networking paper. I suppose the page limit helps prevent papers with a seemingly endless series of graphs each presenting minimal information. (There are still plenty of networking papers like that, but page limits at least cut down the number of graphs that can appear.) Still, there must be some high-quality papers that authors either send elsewhere or artificially cut down to the page limit.

This change in policy, though, seems just to be a way for them to pull in some extra money. I know IEEE and its societies have had money problems in the past; maybe it's getting worse.

ToN is a high-quality journal, and has been a good outlet for some of my work. I haven't minded paying $400 in the past for the two extra pages when needed. At some point, though, there's a limit. Time for me to look closer at how Open Access journals like Theory of Computing are faring monetarily these days -- is there a networking equivalent yet?


Anonymous said...

I hope they start charging authors from the FIRST page. That's the simplest way to stop the IEEE nonsense on this and other matters.

On a second thought, did you ask the editor in chief what is their opinion and reasoning for this policy?

JeffE said...

Networking people must have money to burn. Why should authors pay anyone $200 for the privilege of selling even two pages of their work, the referees' work, and their editors' work?

The first rule of academic publishing should be "Get the hell out of the authors' way." (If not publishing in general.)

Anonymous said...

"Open Access" may not exactly be the right model either. For example the "Public Library of Science" open access journals have significant page charges for authors - for every page not just the first.

Mark Wilson said...

The tren toward "open access" (= author pays, reader doesn't) is discussed by John Ewing (Executive Director of the American Amthematical Society) here.

I find some of his conclusions disturbing. Why can't every field have journals that are basically free to submit to, and free to read, like the Electronic Journal of Combinatorics? Why do we need for-profit publishers involved at all?

On a related note, my old blog post

may be of interest (or maybe not!)

Anonymous said...

Hi Michael,

Theory of Computing is doing fine so far on a zero-cash-flow model ..


Michael Mitzenmacher said...

Anonymous 1 -- If I have time Monday, I'll find out who the current EiC is and point them to the blog. :)

Jeff -- I excused the $200 per extra page as, essentially, paying for paper. There are paper publishing costs still, for better or worse. Roughly $200 to pay for the cost of and shipping off the extra paper for a long article seemed OK to me.

Regarding open access -- I think the model of "author pays" could be a fine model -- if everyone got on board, it would then be a grant cost, and exceptions could be made for cash-poor authors. With either author-pays or reader-pays, I wouldn't want huge profits to be made at my expense, obviously. On the other hand, I can imagine there are some transaction costs that require someone to pay something at some point. Those transaction costs should be getting lower, not higher, so this new policy that increases page costs seems strange to me.

Anonymous said...

One nice version of the open access model is Journal of Artificial Intelligence Research (JAIR) published by the AI Access Foundation, Inc which is a non-profit set up expressly for the purpose of producing JAIR. JAIR is available electronically for free but printed copies are also produced by AAAI Press for a subscription fee. Authors do not pay page charges. Being set up as a non-profit also allows companies to donate people's time as well as money and write off the expense. It would be great if ToC would also follow this approach.