June 1996 € Volume 23 € Number 3

Journals Are DOA

Can a radical special-issue scheme cure the insufferably long publishing delays that are killing the value of scientific journals?

By Hans G. Daellenbach

Have you ever been frustrated when you followed up a promising reference only to discover that it said, "to be published in The Respectable OR Journal"? Have you been somewhat peeved when a paper accepted by Fast-Track OR Letters took 18 months to finally appear in print, a paper that could have saved you several months of work had you seen it sooner?

Such delays are not uncommon. Papers in the September 1995 issue of Management Science, for example, were published 26 to 37 months (including an average five months with the authors for revisions) after the manuscripts were first received by the journal (not including one that was initially submitted in 1990!). The Journal of the Operational Research Society (JORS) of the United Kingdom does better, but even it tends to publish papers 16 to 22 months after receipt and six to nine months after final acceptance.

I subscribe to or receive a dozen scientific journals. I make an effort to at least read the titles of the papers published in all of them and more often read the abstracts. There is the occasional abstract that leads me to actually read parts and, in rare instances, the whole paper, but most issues offer little of interest to either my teaching or my research. If I collected all the papers that I find relevant for my work in these 12 journals, in most years I would hardly get more than the equivalent of a single issue of Operations Research.

So why do I collect all these journals? They certainly look impressive on my bookshelves (although after a while their spines tend to fall apart). But that is hardly a good enough reason to spend close to $2,000 a year on subscriptions.

Graduate students frequently borrow stacks of issues. The lucky ones may find a special issue on a topic that has several relevant papers. But giving them access to my personal copies is simply a convenience that saves them a visit to the university library.

There is a far more disturbing phenomenon happening here. All too often the material presented in the journals may already be out of date by the time it reaches the printing press, due to the slow refereeing and clogged-up publishing process. I receive more useful and up-to-date stuff from colleagues at other institutions in the form of discussion papers or preprints. Some of my colleagues tell me that they get their latest up-to-date information about ongoing research through the Internet.

There is little doubt that OR/MS is a fast moving discipline. New approaches, new algorithms, improved versions of old algorithms, new computer implementations, and novel applications occur at a fast rate -- too fast for the journals to keep up. In fact, as one of my colleagues recently remarked, the up-to-date researcher and practitioner are miles ahead of the material published in journals.

Losing their relevance
So, are the journals losing their relevance, except for the bean counters at the annual promotion exercises at our universities? Should we be looking at a completely new form of dissemination of research results, perhaps via the Internet or the WWW? How can we still make sure that some of the chaff gets separated from the corn -- some form of refereeing -- so that we are not drowning in an overload of mainly worthless information?

Computer science is another field where new developments occur at a faster and faster rate. The discipline has already made a partial switch away from journals to publishing new developments in the form of minimally refereed conference proceedings, available at the conference itself or shortly afterwards.

A colleague pointed out that some form of printed journals could still be useful. Rather than publish regular issues covering a hodgepodge of papers, journals could only publish special issues, each issue dedicated to a given topic. Some of that is already going on. In particular, EJOR regularly features special issues under the editorship of recognized experts on the topic.

Published by Topic
Why not go a step further? Rather than the hodgepodge approach, related papers received and accepted over a given time span on one general topic (or possibly two) could all be collected in a single issue and published as soon as such an issue can be filled, without sticking to a strict periodic schedule of one issue per publishing period. There could even be an explicit understanding that no more than "x" issues would be published per year. Under such a plan, topic areas where lots of advances are made would fill their issues faster, while areas with slower rates of output would have fewer issues.

Under such a radical scheme, it would be desirable for journals to limit the topics they consider for publication, or that several journals cooperate. For instance, Management Science, Operations Research, Transportation Science, Mathematics of OR, Journal on Computing, ISR, Organization Science and Marketing Science could pool their papers and publish them as special issues.

I am convinced that this would speed up the rate at which hot topics get published. This would make it less important to publish in prestige journals, since those cooperating would all become prestige journals. (Naturally, Management Science and some of the other journals would still need to drastically speed up their refereeing procedures.) The form of journal subscriptions would also need to be changed, e.g., the subscription to INFORMS journals could be based on selected topics instead of specific journals.

Hans G. Daellenbach is a professor in the Department of Management at the University of Canterbury, Christchurch, New Zealand.

E-mail to the Editorial Department of OR/MS Today: orms@lionhrtpub.com

Lionheart Publishing, Inc.
2555 Cumberland Parkway, Suite 299, Atlanta, GA 30339 USA
Phone: 770-431-0867 | Fax: 770-432-6969
E-mail: lpi@lionhrtpub.com