From the Transactions Editor-in-Chief

Alexander Vardy

As the incoming Editor-in-Chief of the IEEE Transactions on Information Theory, I would like to express my appreciation to my predecessor Robert Calderbank for his service to the Transactionsfor the past three years. During this period of time, as always, the Transactions published the best papers in our field, and maintain edits remarkable dominance in the core areas of information theory. In my role as Editor-in-Chief, I aim to keep the IEEE Transactions on Information Theory in its place as the premier journal covering all aspects of information transmission, processing, and utilization.

In keeping up with this goal, I am fortunate to have the advice of an
outstanding Editorial Board. In addition to:

 Alexander Barg  Coding Theory
 Ian F. Blake  Coding Theory
 Michael Honig  Communications
 Torleiv Kl/ove  Coding Theory
 Frank Kschischang  Coding Theory
 Neri Merhav  Source Coding
 Pierre Moulin  Nonparametric Estimation, Classification, and Neural Networks
 Shlomo Shamai  Shannon Theory
 Emina Soljanin  Coding Techniques
 Douglas R. Stinson  Complexity and Cryptography

eight new Associate Editors are joining the Editorial Board this year:

Venkat Anantharam Communication Networks
Philip A. Chou Source Coding
Imre Csisz'ar Shannon Theory
Thomas E. Fuja At Large
Sanjeev Kulkarni Nonparametric Estimation, Classification, and Neural Networks
Upamanyu Madhow Detection and Estimation
Joseph A. O'Sullivan Detection and Estimation
Ron M. Roth Coding Theory


These individuals bring broad scholarship and expert technical judgment in their respective areas to our Transactions. Working with each of them is a pleasure and a privilege.

I am inheriting from my predecessor a smoothly running operation. Extended backlog is ancient history --- papers accepted for publication are now forwarded to the IEEE without further delay, and almost every issue of the Transactions is produced promptly on time. This is due in no small part to the devoted effort of Publications Editor Steve McLaughlin. I am delighted to announce that Steve has agreed to serve the Transactions as Publications Editor for one more year, through July 1, 1999. His successor has been already identified, and the details of this transition will be announced in the next issue of this Newsletter.

Later in this article, I would like to discuss the general state of the Transactions. But first, let me describe some specific developments that will affect the Transactions in the near future.

Electronic submission: Although the backlog that was present a number of years ago is now gone, the overall delay from submission to publication continues to be a major concern. The current average delay of about 22 months for regular papers and 19 months for correspondence items is unacceptably high, and we have already taken concrete steps to reduce this delay. In the June 1996 issue of the Newsletter, my predecessor wrote, "It is not hard to imagine the day when all steps in the editorial process (submission of original manuscript, reviews, revisions, and final manuscript) are done electronically."
That day is today. Elsewhere in this Newsletter, you will find detailed instructions for optional electronic submission of manuscripts to the Transactions. For papers submitted electronically, the intent is to have all communication between authors, editors, and referees by e-mail, thereby expediting the review process. We estimate that electronic submission and review will reduce the delay to publication by up to three months. Furthermore, additional measures to reduce the time to publication will be implemented in the future. Notice that the electronic submission option is just that --- optional. Conventional submission of manuscripts -- on paper -- will certainly still be supported. Nevertheless, I expect that most authors will soon embrace this option, and the majority of manuscripts submitted to the Transactions will be submitted and reviewed electronically.

Special issues: This is an exciting time for information theory, distinguished by the 50-year anniversary of Shannon's seminal paper A mathematical theory of communication, which gave birth to our field. As an appropriate tribute, we will publish a special 50-th Anniversary Commemorative Issue in October 1998, the month of publication of Shannon's work. This special issue will contain survey articles by leading experts in diverse areas of information theory, and will undoubtedly become an essential reference for anyone interested in our field. Another special issue, devoted to the subject of multi scale statistical analysis, is scheduled for early 1999.


Contributor biographies: As of January 1999, each issue of the Transactions will contain the biographies of all contributors to the issue, regardless of whether the contribution is a regular paper or a correspondence item. David Forney, during his term as Editor-in-Chief twenty-five years ago, gave a precise definition of the distinction between regular papers and correspondence items. He wrote: "The distinction is not one of quality, but of nature; normally a correspondence item will have one point to make briefly and without adornment, whereas a regular paper will be a more well-rounded treatment of a problem area. "These words have appeared on the inside back cover of our Transactions ever since, and some of the best research reported in the Transactions was reported in the form of correspondence items (indeed, two of the 15 papers selected for the Golden Jubilee Paper Award by our Society are correspondence items). I wholeheartedly subscribe to the point of view that the quality of a correspondence should be just as high as the quality of a regular paper. As always, both forms of submission will be reviewed by theTransactions using the same criteria. Thus the disparity between regular papers and correspondence items in the Contributors section sends the wrong message to our authors and our readership.

The change of Editors might be an appropriate time to reflect on the state of the Transactions in the past and present, and identify the trends that will shape its future. The table below contains some numerical data that I gathered by counting the number of regular papers and correspondence items published in the Transactions over the past 10 years.

TABLE 1. Statistics on the state of the Transactions  
---------------------------------------------------
YEAR   I    #R   Rpa      #C  Cpa    #S   #A   Tp
---------------------------------------------------
1989   6    100  10.3     76  3.6    351  176  1396
1990   6     71  14.1    109  4.1    356  180  1560
1991   7     96  11.0    141  4.4    386  237  1804
1992   7     94  13.3    123  4.3    434  217  1908
1993   6    121  11.6    119  4.4    481  240  2056
1994   6    116  12.2    134  4.8    420  250  2176
1995   7    122  12.3    119  4.4    496  241  2232  
1996   7    128  12.9    120  4.7    467  248  2344
1997   6    107  12.4    133  5.2    554  240  2144
1998   5     89  14.6    126  5.6    482  215  2040
--------------------------------------------------- 
 Legend:
     I   number of issues
    #R   number of regular papers published
   Rpa   average length of a regular paper
    #C   number of correspondence items published
   Cpa   average length of a correspondence item
    #S   total number of manuscripts submitted  two years prior
    #A   total number of papers/correspondences published
    Tp   total number of pages published 

The data is here for you to peruse, and you can draw your own conclusions. For example, you can compute from Table 1 the average acceptance rate during the past 10 years, which stands at about 52%. Notice, however, that because the acceptance rate for papers submitted to special issues is usually higher, the average acceptance rate for ordinary submissions is only about 50%.


You may also notice certain trends that particularly stand out in Table 1. While the number of submissions has fluctuated, it has not increased significantly since 1993 (the total number of manuscripts submitted in 1997 was 423, and the number of submissions during the first six months of 1998 is about 240). Furthermore, the total number of papers published in the Transactions remained remarkably steady since 1993, although a substantial increase is expected this year. What has been steadily growing over the years is the average length of both correspondence items and regular papers. As a result, our Transactions has reached a formidable size, and significant further growth is expected this year. Extrapolating from the first five issues, we anticipate publishing close to 3000 pages in 1998, and our estimated page budget for 1999 exceeds 3000 pages for the first time in the history of the Transactions. Only two or three IEEE journals publish more pages, and it costs us money to publish each page. Thus the growth of the Transactions places a financial burden on our Society.


Nevertheless, I am strongly opposed to page limits on papers. I believe that each paper should be just as long as it needs to be in order to present its results in a lucid and succinct manner. However, as the average length of papers grows, the Transactions is clearly in a position to place very high demands on the quality of papers it accepts. A manuscript submitted to the Transactions has to be novel, significant, and of interest to our readership, in order to be publishable. Furthermore, papers in publishable form must be understandable without undue effort by their intended audience, which should be as broad as possible. As such, they have to be reasonably self-contained, well-polished, and absolutely clear. Writing such papers necessarily requires more effort, but not necessarily more pages.


Although the Transactions will continue to apply stringent selection criteria to the papers it accepts for publication, we will publish the best of the papers we can find in the field of information theory, very broadly interpreted. In fact, I would like to see the scope of the Transactions broaden as our discipline broadens and attracts contributions from diverse areas of scholarship. A good example of this are key contributions to coding theory that have come from the computer science community in recent years. Such papers often provide a different and valuable perspective that enriches our community, and they are most welcome in the Transactions. As always, the Transactions will also welcome papers in such areas as signal processing and applied probability, with relevance to information theory, but only if these are the best papers in their fields.

Ensuring the status of IEEE Transactions on Information Theory as the premier journal in our field has to be a team effort. I welcome your suggestions regarding all aspects of the Transactions. Please send your comments to it@csl.uiuc.edu.

The reputation and the vibrancy of our Transactions are the legacy of editors past, and continuing this tradition of excellence is a challenge. We are fortunate that information theory is a field in which profound theoretical investigation often has immediate and far-reaching impact on technology and practice. This spirit of truly fundamental innovation has pervaded the Transactions throughout the years, and will keep advancing the frontiers of information theory. It is also fortunate that our field has inherent intellectual beauty that has always attracted, and continues to attract, some of the best minds to information theory. At the core, it is this ceaseless flow of exciting first-rate research that makes our Transactions what they are, and makes my job as Editor-in-Chief most rewarding.