Few publications can claim the kind of undisputed preeminence in their fields that the IEEE Transactions on Information Theory enjoys. Since the appearance of IT-1 in March 1955 (succeeding a periodical started in 1953 called the "Transactions of the IRE Professional Group on Information Theory"), the Transactions has grown to more than a meter and a half of shelf space. It now receives an average of around 500 submissions per year (roughly half of which are accepted), and it has published in excess of 2000 pages per year since 1994. A quarterly publication at its inception, the Transactions turned bimonthly in 1968, and, starting in 1998, it will publish seven issues per year.
The impact of the many landmark papers published in the IT Transactions is legendary. Indeed, in the last forty years, most major advances in information transmission/compression have appeared in its pages. The high standards of the Transactions continue to attract exceptionally talented authors from around the world. Those high standards have been achieved thanks to the unselfish efforts of a distinguished group of Editors, Associate Editors and reviewers.
All is not well with the IT Transactions, however. The average delay from submission to publication date for papers appearing in 1995-1996 was 22 months for regular papers and 19 months for correspondence items. It is not at all unusual for authors to have to wait for more than one year to receive the first round of reviews. We have grown accustomed to view the Transactions as a ``slow'' publication. Reviewers are authors: long publication delays are not good incentives for quick reviewing. Unresponsive reviewers are a major source of increased workload for associate editors, thereby causing longer publication delays for other papers. Positive feedback.
The status quo does not affect everyone equally. Papers written by authors new to the field are unlikely to induce more responsiveness than papers written by famous researchers. Moreover, publication delays may have a big impact at the beginning of a researcher's career. Not only they affect hiring and promotion decisions negatively, but the sluggishness of the IT Transactions is not the trademark of a vibrant field that attracts top young talents.
Some argue that this situation is an inevitable consequence of the ``mathematical'' nature of the IT Transactions and that a long reviewing delay is a worthy price to pay for an insightful review. I disagree. While it is hard to argue against the notion that, in general, IT papers take longer to review than those of other IEEE publications, it is also true that it takes the same amount of time to review a paper promptly than tardily. And unlike top wines and researchers, the evidence suggests that reviews do not improve with age.
As authors, we can do our share in order to expedite our own papers. It is hard to overemphasize the importance of elegance in writings. Granted that not all subjects are equally reader-friendly, and the last thing we want to do is to discourage abstraction. However, the IT Transactions is not the best forum for authors who give up hope to be understood.
I think the time is ripe for the IT Society to take effective measures to correct the current situation. The Transactions generates most of the revenues of the Society, yet it has been operating on a shoestring budget that relies heavily on the goodwill of volunteers and their institutions. The Transactions has become too big of an operation to continue that way. As I mentioned in my March column, an ad-hoc committee on Transactions Policy (Rob Calderbank, Jerry Gibson, Bob Gray, Bruce Hajek, Steve McLaughlin and myself) is considering various ways to decrease publication delays. Some of the avenues the committee is currently exploring are the following:
1. Incentives for referees to do a prompt and high-quality job. These can take the form of monetary rewards and/or public citations.
2. An administrator in charge of organizing and tracking the flow of papers and reviews. Some IEEE Societies have adopted this system and are satisfied with the reduction in publication delays they have achieved.
3. An increased number of Associate Editors.
4. Strict deadlines for Associate Editors to communicate with authors.
5. A Web site where authors can track the status of their paper from submission to publication.
I would like to hear your feedback on these proposals and on any other ideas you may have to improve the operation of the IT Transactions. We are fortunate that our journal enjoys a fine reputation, and no effort should be spared to continue attracting the best papers in the many fields that lie within the purview of the IEEE Transactions on Information Theory.