Electronic publication will be adopted by the scientific and engineering community only when it meets the needs of that community, overcomes the major problems with conventional publication, and offers new and worthwhile capabilities that exceed those of traditional journals. When it does that, electronic publication will deserve to be adopted by the professional community. Failing those tests, however, it will remain a marginal medium, professional pop art, appealing primarily to those with an interest in electronic publication prowess as an end in itself.

As a scientist with decades of experience reading, contributing to, and editing professional journals, I think electronic publication is unlikely to pass those tests for scientific journals. In response to apparent pressure to adopt electronic publication per se, I feel more angst than enthusiasm.

There is an intermediate path, however. Certain electronic features emerge as potentially essential adjuncts to more-conventional scholarly resources. Increasingly, those tools are being sharpened, implemented, and eagerly adopted by users. This paper looks at the implementation of those features that would meet professional needs (quality and longevity), solve certain problems of paper (publication delay and lack of linkages), and bring added value through innovative citation and search techniques.

As in Darwinian evolution, a principle of adaptive selection is at work as the various forms of electronic publication are reinvented and implemented. Critical mass will build around those attributes that are responsive to users' needs. The way forward is not to promote e-journals either as replacements or as alternatives to conventional journals, but rather to further develop those features of the electronic publication environment that are most helpful to the professional community.

Scholarly Publication

If electronic publication is to be adopted as a legitimate scholarly medium on a par with conventional journals, then it will have to meet the same standards. What is expected of a professional journal? The essential attributes are captured by the concepts of quality and longevity.

Quality

Recently the American Geophysical Union, the largest professional society devoted to understanding our planet and its environment in space, conducted a small survey [1] to determine the relative importance of quality factors for its principal publication, the Journal of Geophysical Research. The results are illuminating. Scientific content is the highest ranked factor. There is no factor in second place. Third place is shared by attributes of content: good writing and timeliness. New publishing technology places near the bottom of the list, number twelve of the fourteen factors cited.

The order of quality factors determined from the survey is especially interesting, given that the American Geophysical Union is a professional society that has asserted leadership in electronic-publication activities. In a typical year, it publishes thousands of scientific papers in its ten journals for a membership of about 35,000. By 2000, the society plans to make all of its articles available in both electronic and paper format. The AGU was one of the first professional organizations to provide e-mail for its members, to accept electronic format for journal and conference papers, to select and publish a standard electronic format, and to offer an online professional index, the Earth and Space Index (EASI), on its Web site. Further, the AGU publishes its totally electronic journal, Earth Interactions, in partnership with the American Meteorological Society and the Association of American Geographers.

The AGU's journal priorities are generalizable to other first-class professional societies. The memberships tend to rank the quality of the message above the technology of the medium, contradicting Marshall McLuhan's mantra. New publishing technology ranks below such relatively trivial factors as the use of color figures and the inclusion of titles in the society's referencing style. The medium is NOT the message.

The message should be clear: No matter the medium, quality cannot be sacrificed for the sake of lesser factors. Taken by itself, new publishing technology misses the point. There is room for innovation, but little willingness by the stakeholders to compromise on fundamental purpose. The corollary is that technological innovations are more likely to be adopted if they serve the primary purpose of scientific publication.

Longevity

Although the intellectual half-life of most journal articles may be relatively brief, some survive the test of time, retaining their value for many years. Content aside, the key longevity questions concern the mechanics (electronics?) of accessibility and legibility; in short, archival robustness.

The archival medium may support or suppress the message. The figure below shows that the longer-lived media are not electronic. The seven professional societies with which this author is affiliated are planning accordingly. The AGU, for example, has said that its policy for the foreseeable future is to rely on conventional paper format for its archival record.

Figure 1. Relative media longevity in years [2]

Note that actual lifetimes depend, among other factors, on the storage environment. In the case of electronic media, lifetime estimates are a function of the predictive model and its initial conditions.

To be robust, an electronic archive must regularly rewrite all files onto fresh media. The preferred formats and media probably will change over the years. Thus, in the electronic context, the very concept of "archive" becomes dynamic, rather than static. Further, there is the implied obligation to maintain a meta-archive, a collection of functional reading and transcription devices whose capabilities together span all archived media and their several versions. Maintaining an active electronic archive implies expense, risk, and most likely missed opportunities. (If discovered after a few centuries in the desert, for example, a "Dead Sea-D-ROM" might well be useless.)

Such requirements effectively constrain, if they do not eliminate, the possibility that an individual researcher can maintain his or her own specialized library. Many institutional libraries, especially smaller ones, will face equal or greater difficulties. Rather than expand access to research materials, as the electronic-publication protagonists promise, an archival system based exclusively on electronic means could inhibit access.

Absent a vastly improved electronic alternative, those trends suggest that the most reliable and least expensive long-run archival medium is low-acid, buffered paper. Unless extensive means are provided to offset their disadvantages, electronic archives should not be relied upon.

Principal Problems

Maintaining scientific quality and timeliness is a challenge for any form of publication. Scientific quality comes at a price. Suitable standards have to be diligently maintained by authors, editors, peer reviewers, and readers. All that takes time. Although the authors and readers have some control over their time investment, the publication process is largely outside of their control. It can be lengthy. But for users, timeliness and speed of publication are important and pertinent factors. Assuming that scientific quality is maintained, they are the two dominant problems with conventional journal publication.

Timeliness

It is unusual for a paper to include references to pertinent articles that may have appeared shortly before its publication date. It is impossible to cite articles that may appear after the final manuscript is submitted (unless the author has inside information). Articles that appear after the paper is written are unlikely to be referenced, no matter how important they are. In many disciplines, change is sufficiently rapid that the citations in a conventional journal article are usually out of date by the time the journal is published. That is one of the reasons that researchers seeking the most recent work in their field often turn to conference proceedings and the like. Citation timeliness is a virtually insurmountable problem for conventional journals.

Figure 2. Submission-to-publication time, IEEE Transactions on Geoscience and Remote Sensing, 1997

Speed of Publication

Each journal article requires manuscript submission, review, revision, acceptance, scheduling, composition, and publication. The last time that the author can amend the work is at the galley stage, at which point extensive changes are discouraged (if not absolutely forbidden). The minimum elapsed time between galley proofs and publication is a few weeks, with six to ten weeks more the norm. If there is a backlog, then the delay between final acceptance and publication may be longer. Typical elapsed time between submission of the final accepted manuscript and its publication is several months or more.

Authors and readers alike have a vested interest in short publication schedules, under the proviso that quality is sustained. For most scholarly journals, the elapsed time from the first submission of a manuscript to its eventual publication is longer than one year, often much longer. Consider one example, the IEEE Transactions of the Geoscience and Remote Sensing Society (GRSS), which publishes approximately 1,500 pages per year in six issues: The submission-to-publication time for the 151 GRSS papers that appeared in 1997 is charted above. The chart is based on the original submission dates cited on each paper. Excluding special issues, the mean is 21.8 months, of which fewer than five were required for scheduling, composition, and actual publication. The remainder, more than sixteen months, is consumed by review and revision, and waiting time for the next available publication opportunity. The average waiting time for the 1997 papers was about four months.

For 1995 the GRSS average submission-to-publication delay was about fifteen months, of which more than ten months were required for review and revision. As a scan of representative journals will verify, those numbers are relatively typical of current publication delays. The review and revision process dominates the submission-to-publication time delay.

The GRSS special sections include invited papers that previously had been presented at symposia. Special-section guest editors take the responsibility for badgering their referees and pleading with their contributors, unscholarly tasks that generally lead to guest-editor burnout after only one issue!

Good writing is the second highest quality factor in the AGU survey. Correct and effective use of the English language is essential, especially given the international makeup of most technical societies' stakeholders. Most submitted manuscripts need considerable rewriting, and reviewers and editors often make detailed emendations. Authors may go through several rounds of rewrite and review.

Here is the crunch: Quality requires thorough review and revision, which takes time. The associate editors and their panels of referees all are volunteers. Editors are over-worked (and too often under-appreciated). Submission rates are up, which implies increased load on the review process and growing backlogs. In the GRSS case, additional associate editors have been appointed, and there is increasing use of electronic communication between them and their reviewers. Most journals encourage electronic submission or camera-ready copy for the final manuscript, which helps to reduce the remaining time to publication. In spite of such efforts to improve the rate of manuscript flow, however, the review process takes a relatively long time.

Speed of Electronic Publication

What impact might electronic publication have on the speed of publication of a professional article? One might expect that electronic means would be faster; indeed, much faster is the oft-cited promise. On closer examination, the speed advantage attributable to electronic publication per se may not be very great. To expedite the following discussion, the publication process is construed as falling into two phases:

• from original submission to final acceptance of the revised and edited manuscript
• from final acceptance to eventual appearance of the published article.

If comparable quality is to be maintained in either form, then a comparable peer-review process must obtain for either genre. It follows for the first phase of the process that electronic media offer no intrinsic speed advantages over conventional journals, because any means used in the electronic world to accelerate the process would apply equally well to conventional publications. Here it is helpful to distinguish between electronic communication and electronic publication. Electronic communication, including e-mail, facsimile, and dedicated Web sites, deserves to be fully exploited in the process for all peer-reviewed publications, conventional and electronic alike. Of course, one might argue that use of electronic communications comes more naturally to those attuned to electronic publication. To the extent that is true, however, it is a matter of attitude, not attribute. Thus electronic media have no intrinsic speed advantage over conventional media in the first pre-publication phase.

During the second phase of the process, however, electronic publication can offer speed advantages. Electronic publication means that the final manuscript-preparation time can be reduced, the waiting time for the next-scheduled issue can be reduced if not eliminated, and article length and volume size limits become less constraining.

New Capabilities

Two great strengths of the online environment are interactive: search and linkage. They extend the online capabilities of electronic-publication resources significantly beyond the limitations of the conventional printed page. The potential invites innovation such as dynamic citations and dynamic abstracts.[3]

Dynamic Citations

Citation timeliness is a weakness of conventional journals. Online publications have no such inherent weakness. Timely citations can be assured even after publication. For those papers that appear on line, three citation lists could be appended: references, bibliography, and inverse citations. The reference list would consist of the citations called in the text, as is normally done. The bibliography would list all articles that are directly pertinent to the subject of the e-journal paper. The inverse citations would list all subsequent articles that reference the article itself.

A dynamic-citation list would consist of all three types of citations, updated as needed. For any online article a dynamic citation list could be maintained continuously and automatically. Each time the online article was accessed, its lists would be current. Concerns over citation timeliness would be a thing of the past.

There already is precedent for those capabilities. The Institute of Physics Publishing (http://www.iop.org) in 1997 launched its upgraded electronic-journals service with HyperCite™, which guides users to papers cited by, and papers that cite, their articles. Now those capabilities are limited to a subset of recent journals. In the future, similar citation power should be harnessed for almost all articles published on line.

Dynamic Abstracts

Several effective abstract-search engines are now available on line, such as the IEEE's Bibliographies On-Line [formerly http://www.ieee.org/web/developers/webthes/00002014.htm]. At that site, the abstracts of IEEE documents since 1994 are available — about 200,000 titles. The search on key words is fast, and yields titles and their source. One more click and the abstract appears. Then? Almost a dead end. Although the papers themselves are separately accessible, they are not available directly. The electronic medium can do better.

For many purposes, one needs only the information in the abstract AND the information that would be included in a dynamic citation list, including all articles that cite the article abstracted. It would be very useful to researchers to have access to a dynamic citation list for all papers published, not just for those published on line. Such dynamic citation lists could be appended to corresponding abstracts, creating a dynamic abstract. The implied online storage requirements would be modest, at least in comparison to the volumes implied by online access to the complete articles. To my knowledge, a dynamic abstract capability or its equivalent is not yet available.

Synthesis

Against the three criteria of needs, problems, and capabilities, how does electronic publishing fare? The question is best addressed beginning with the problems and ending with the capabilities of the medium.

The principal problem with professional publication centers on the delay between manuscript submission and finished publication. The review and revision process is by far the largest component of that delay. High quality, which all stakeholders desire, can be sustained only if the review and revision process is not compromised. Therefore, electronic publication is not a solution to the speed-to-publish problem, in spite of claims to the contrary. If the review process were to be downgraded, journal quality would suffer, and quality contributors and readers would no longer support those journals. The critical mass of contributors and readers would not be there. If the quality of the review process were maintained, but alternative means were developed to increase its rate of response, then the same methodology also would work for conventional publications.

Electronic publishing unquestionably offers capabilities that differ from those of conventional journals, particularly as experienced on screen and on line. However, those environments inherently imply disadvantages. Expanding on the attributes cited by Grenquist [4], those that are disadvantageous from the user's point of view include: limitations of the single-screen format, lack of user-friendliness in certain presentations, impossibility of making marginal notes, rate of reader fatigue, inefficiency of subsequent use and access, questionable authentification or even identification of authorship, lack of briefcase portability, and perishable citation. To me, it is the ready availability of the article that counts, not electronic acrobatics. Once I find the right electronic article, I prefer to render its form to be more user friendly. If an article is on line, and I choose to make more than a passing reference to it, I print it.

Although the great majority of articles that I need to read are not available on line, frequently I want only to find information about articles. Preferably that information will be in forms that facilitate rather than frustrate my efforts. Here the electronic environment comes into more user-friendly territory. The unique and compelling capability of the electronic environment is that it enables scholarly search and citation. The electronic interface gives online users unprecedented power to follow unanticipated paths to unpremeditated payoffs. Once a useful article is identified and its citation linkages exposed, the paper's ontology is irrelevant. Enhanced online resources, as exemplified by dynamic citations and dynamic abstracts, respond elegantly to the problem of citation timeliness.

The three primary criteria of acceptance — satisfying professional need, overcoming extant problems, and offering new possibilities — are unlikely all to be satisfied by electronic publication. E-journals will not displace p-journals regardless of the zeal of those who may promote the electronic alternative. However, new and powerful search and citation capabilities will be coming on line. Those possibilities will continue to generate enthusiastic acceptance within the professional research community. They will become essential research tools.

Acknowledgment: Preparation of this paper was partially underwritten by a Johns Hopkins University Applied Physics Laboratory Stuart S. Janney Fellowship.

Dr. R. Keith Raney is a member of the principal professional staff with the Space Department of the Johns Hopkins University Applied Physics Laboratory, and is the assistant supervisor of the Ocean Remote Sensing Group. Currently he is developing advanced radar altimeter and radar ice-sounding concepts. His contributions in remote sensing systems, theory, and applications are documented in more than 300 professional publications. He is on the founding Board of Associate Editors for the International Journal of Remote Sensing, serves as an associate editor (radar) for the IEEE Transactions on Geoscience and Remote Sensing, and is on the Editorial Board of the Canadian Journal of Remote Sensing. He is a past president of the IEEE Geoscience and Remote Sensing Society. You may contact him by e-mail at keith.raney@jhuapl.edu.

Notes

1. C. White, "Survey looks at meaning of 'quality' in AGU publications," EOS, Transactions, American Geophysical Union, 79 (10), 1998, 126. [doi: 10.1029/98EO00094]

2. L. Tangley, "Whoops, there goes another CD-ROM," U.S. News and World Report, Feb. 16, 1998, 67-68. For a glimpse of the ensuing controversy, and amendment of the cited CD-ROM data, go to http://www.cd-info.com.

3. Electronic media open up new dimensions for animated, interactive, or multimedia embellishments that go far beyond the limitations of the printed page. For those who would profit by their use, those features are sufficient to justify enthusiasm for electronic publication.

4. P. Grenquist, "Why I Don't Read Electronic Journals: An Iconoclast Speaks Out," The Journal of Electronic Publishing, 3 (1), ISSN 1080-2711, 1997. [doi: 10.3998/3336451.0003.110]

Links from this article:

IEEE's Bibliographies On-Line. [formerly http://www.ieee.org/web/developers/webthes/00002014.htm]