Science Tribune - Article - March 1997
Scientific publishing : Paper or perish
David Atherton1, David Steffen2, Sinai Yarus3
1. Department of Pathology, University of South Florida, Tampa, Florida. 33612 USA
2. Biomedical Computing, Inc., 6626 Westchester, Houston, Texas 77005-3756. USA
3. Baylor College of Medicine, Dept. of Cell Biology, Houston TX 77030. USA Fax: (713) 798 8012.
E-mail : firstname.lastname@example.org.
Scientists publish articles for two main reasons, to disseminate information, and to advance professionally. Publishers have often agreed to help scientists disseminate ideas in order to generate revenue. Traditionally, publication has been done in weekly or monthly paper journals and the reputation of the journal in which an article appears has had a great impact on the professional advancement of the authors. This reputation is dependent upon the circulation of the paper journal. The possibility of publishing one's work electronically via the WWW, or "weblication", has changed ideas about disseminating information. In this position paper, we advocate the use of electronic journals to facilitate faster dissemination of information, without sacrificing career advancement possibilities. E-journals can retain the best aspects of traditional paper publications while offering opportunities to distribute data in myriad formats to an unprecedented number of potential readers. We also predict the effect of e-journals on authors, readers, publishers, advertisers, libraries, the concept of copyright, and even laymen trying to find information.
Discussion between referee and author and call for readers' comments
The reasons why a scientist publishes an article
Reason 1: To disseminate information or ideas
The scientist is interested in fast distribution to as large a group of interested readers as possible. Publication in a paper journal guarantees that a specific number of subscribers will receive printed copies of the article after a delay that may range from a few weeks to several months. This number varies depending upon the circulation of the journal. In addition, many paper journals are indexed in Medline (1) or a similar collective index. This system was instituted in recognition of the fact that there are so many journals that no individual can keep track of all the information being published (2). In fact, there is currently (at the time of this writing) a backlog of data entry for Medline owing to the tremendous number of publications to be indexed. Collective indexes offer a partial solution by electronic dissemination of full bibliographic information and an abstract. Readers wishing to read the article in detail must request a reprint or find the appropriate journal in a library.
Since weblication does not rely primarily upon distribution of physical copies of an article, the number of potential readers is theoretically as large as the number of people with Internet access. For the same reason, the distribution time is zero. The moment an article is posted, it is available to all although, in practice, there may be a delay of several days between the end of review and posting of an article in an e-journal. While weblicated articles may not appear in Medline, they are catalogued more quickly by a variety of search engines (alta vista, hot bot, etc) (3). A scientist may expedite cataloging of a weblication by direct submission of a URL to one or more search engines or by making use of <META> tags in HTML. Paper publication facilitates distribution of information only as text or 2-D pictures. Weblication is amenable to many media not distributed by conventional journals including sound, animation, video and Java scripted figures/image maps. Additionally, since there is minimal restriction on file size or type, there could be an improvement in the "graphical" nature of articles as it is possible to provide higher resolution images upon request and greater numbers of images per se.
In summary, weblication allows a scientist to disseminate information faster than paper publication, to a larger number of people than paper distribution, and in formats not available in printed journals. In theory, weblication requires no journal but, in practice, many scientists may choose to weblicate in e-journals.
Reason 2: For professional advancement
While e-self-publishing can easily help establish contact with individuals sharing your interest, paper publishing is generally looked upon more favorably by faculty search committees and grant review boards. These panels are often composed of senior scientists who may be less receptive to new methods. Although one might argue that senior scientists were innovative in their early careers - facilitating their advancement - there is no guarantee that they will admire innovation for the sake of innovation. Members of these panels will need to be convinced that a weblicated article meets the traditional criteria for a journal publication. E-journals, such as Journal of Computer-Mediated Communication (4) and Frontiers in Bioscience (5), are ideally suited to this function, while allowing the scientist to enjoy all the advantages of weblication.
The reasons why scientific journals are published
Paper journals are published either to generate revenue from advertising, subscription fees, and sale of reprints (e.g. Nature, Science) or to serve the needs of the scientific community (e.g. Development, PNAS). In the second case, no profit is generated but significant resources are expended on overhead expenses. These expenses must be offset by philanthropy, professional society dues, subscription fees or page charges. While many journals published commercially have "web sites", most of these are little more than advertising designed to encourage purchase of a paper subscription to the journal (e.g. CJC Overview). Those commercial journals that do offer full text versions on-line generally require a subscription to the paper version and levy an additional fee for the on-line version.
How electronic publishing is related to standard publishing
Definition of weblication
Weblication is used in this article to describe any dissemination of information which occurs primarily via the Internet or World Wide Web. While weblications may be downloaded to local computers or magnetic media and/or printed onto paper by individual readers, this is secondary to their initial distribution. Such physical copies may be considered analogous to photocopies of articles from paper journals. While e-journals may be distributed as physical copies, on CD-ROM or similar media, this is for archival purposes or to facilitate additional distribution on the Internet or through a local computer network. This type of re-distribution would reduce the need for large computer hard-drives for the original publisher and, at the same time, reduce Internet traffic as well as bring the library one step closer to the development of full on-line archives.
Who can weblicate?
Anybody with permission to write files on a computer functioning as an Internet server, plus a rudimentary knowledge of HTML, can weblicate any document they choose. This type of e-self publication can disseminate ideas quickly but lacks scientific credibility, much as a self-published paper article does. This is because, in many instances, an article is self-published because it has been rejected by peer review; the author[s] refuses to accept the comments of his/her peers. Therefore, such e-self publication is unlikely to impress faculty review committees or grant review boards. However, e-self publications containing mailto links do facilitate communication, contributing indirectly to professional advancement.
The standards of E-journals
E-journals attempt to combine the power of the Internet with the professional standards associated with print journals. These standards include peer review, distribution, indexing, future availability, proof of priority and "impact score".
Peer review : The main professional standard is peer review. Most print journals in biological sciences use anonymous peer review. In this system the author's name is known to the two (or three) independent reviewers whereas the reviewers remain anonymous to the author. While this system works well in general, it may allow an unscrupulous reviewer to indulge personal grievances against competitors and to hold up a competitor's paper until theirs comes out, or it may cause a feeling on the part of the author that a review is biased. Whichever the case, the editor is expected to serve as intermediary and resolve the differences. Some journals in the physical sciences use an open review system in which the identity of the reviewers is known to the author. Editors of these journals claim that complaints from authors about the fairness of reviews are less frequent. Either open or anonymous review is suited to e-journals and it is the quality of the expected review, not the method, that an author should consider when selecting an e-journal as a venue for weblication.
Distribution and publication evolution : In Charles Darwin's "Origin of the species" (6), the hypothesis is that selective pressure causes living things to change their form. We maintain that the cost of traditional paper publication, together with its relative inefficiency, are creating a selective pressure which will favour the evolution of e-journals. Many of our colleagues in science share a feeling that information doesn't exist if it is not written down on paper. Those that have given the matter more thought realize that it is the persistence of a record through time, and not the media upon which that record relies, that is important. We respectfully submit that were Darwin alive today, he would be pleased to see that his works have evolved from printed to digitally distributed form owing to their long-standing popularity.
When considering weblication, an author should therefore seek an e-journal which can promise a stable URL for a reasonable period of time. In biological sciences today, that probably means 5 to 10 years. During that time, both the journal and the author should make archival copies of the weblicated article for re-posting at new URLs in the future, should the need arise. E-journals should also provide stability over time to weblicated articles by Medline indexing, mirror sites and CD-ROM deposits in major libraries.
Indexing and future availibility : While Medline indexing is not strictly necessary for retrieval of weblicated articles, it will call attention to these articles to our more traditional colleagues. Mirror sites are duplications of computer files on a second server, usually at a distant location. They serve to reduce traffic at the main site, allowing more efficient retrieval of weblicated articles, and provide a back up in case of irreparable hardware damage at the main site. In Internet terms, voluntary mirroring is one of the highest forms of recognition a web site can get. Not only is the 'webmaster' offering valuable time and energy in keeping the files up-to-date, but is also providing significant disk space and Internet bandwidth to allow the additional accesses. Mirror sites go beyond the standards of citation in that the whole article can be hyper-linked, ergo is available, from that site rather than just the author name[s] and précis/abstract. An author considering weblication should look at the number of mirrors an E-journal has as an indicator of credibility and stability.
Proof of priority : Chronology is important in scientific publishing. Therefore it makes sense for e-journals to adopt the traditions of volume numbers although they are not really necessary. For reasons of personal advancement, publishing first is important. If you are going to weblicate, you should insist on both the date of submission and the date of posting being clearly marked on the article. Although it is easy to make corrections or additions to a weblicated article after publication, the date of these corrections, as well as the original publication date, should be clear to the viewer. This documentation of chronology also allows e-journals to mesh with Medline, providing continuity in the archival process.
Impact scoring : Since current technology is unable to accurately demonstrate levels of readership for print journals, "impact scores" are currently assigned to journals based on the number of times that articles published in those journals are cited in other articles. In contrast, e-journals can demonstrate the real-time readership of any particular article courtesy of access/request log files. While e-journals today have low (or no) impact scores, we predict that these scores will rise drastically in the next several years as more and more young scientists come to view the Internet as an information resource (The Network Nation Revisited).
How weblication via e-journals affects those involved
Having concluded that scientists achieve their goals of information dissemination and career advancement by weblicating, we examine the effect of e-journals on authors, readers, publishers, advertisers, libraries, the concept of copyright, and even laymen trying to find information.
The author of a weblicated article can expect to save money and time. There should be significant savings in page charges, especially for colour photos, video or animation. Since e-journals don't have the delays associated with physical printing and regularly scheduled distribution, a short turnaround time should be the norm. The reviewers will probably receive the paper electronically, and it may be posted within days after review is completed. Similarly, since there are no limits on "space", the work should be published if it is scientifically sound. There is a huge potential circulation, assuming the journal has no subscription fee since actual readership may be increased by good use of <META> tags and direct submission of URLs to indexing services.
A weblication brings no reprint requests since readers can download a copy or bookmark the page in their browser. In addition, inclusion of mailto links expedites communication with readers facilitating future collaborations, exchange of ideas and a sense of scientific community. The author of a weblicated article may even add a note at the end of an article giving a hypertext link to a subsequent publication on the same subject, provided that such an addition is dated.
The reader of weblicated articles can also expect to save time and money. Time is saved by eliminating trips to the library and by searching all the literature from one index, or several central indexes, without regard to what is available at ones home institution. In addition, there is never any time spent looking for "..last month's issue of...", or any concern that ".....this week's issue of X hasn't arrived yet", or "...X is in, but somebody else has already got the only instance of it, and has removed it from the shelf". In the subsequent generations of weblication, references will be hypertext links so that primary sources are available without additional searching. Additional monetary savings come from reduced journal subscription fees.
The publisher of an e-journal is more likely to be an individual (or group of individuals) than a company and should be interested primarily in disseminating information. This is possible because of the reduced capital outlay. In general, no building, print shop, and furniture are required. In addition, the number of paid staff is greatly reduced. The major capital outlay is a computer to act as a server and the cost of Internet connectivity/accessibility. Assuming the publisher is affiliated with an academic institution, these costs may be absorbed to some extent by the institution. Reduced capital outlay reduces the need to generate revenue allowing increased focus on content and scientific integrity. Since staff members need not be in geographic proximity, it is possible to recruit many part-time experts on a volunteer basis. Opportunities for advertising eliminate need for subscription fees. E-journals should take commercial television and radio in the U.S.A. where revenue is generated solely from advertising and free access is available to anyone with the required hardware to receive the broadcast signal as their model, and not printed journals. Distribution expenses are drastically reduced, only CD-ROMs for deposit in libraries need to be distributed physically. Since each CD can be used to serve the entire journal to the Internet, a relatively small number of copies is sufficient.
The editors are probably the publishers and, in many cases, may be active research scientists themselves. They still fulfill the traditional roles of authenticating the identity of the submitting parties and arranging timely review. In contrast to traditional editors/publishers, they should be far more focused on disseminating information than making money, making the author/editor relationship more co-operative. Since many editors will be working part-time without pay, they may be interested in reducing the amount of time they spend resolving disagreements between reviewers and authors by adopting an open review policy. This should simplify the editor's job of getting authors to agree to reviewers suggestions. Weblication offers the possibility of a public open review policy as well. In public open review, the reviewers' names would appear in a note at the end of the article. Any suggestions from the reviewers not addressed by the author during the revision process could appear as part of this note with a response from the authors. This would allow the reader to get a feeling for the issues surrounding the research being reported.
The advertiser can expect drastic changes as publishers shift their emphasis from generating revenue to disseminating information. The modularity of HTML makes it easy to insert the same ad in many places simultaneously. Since every ad serves as a link to a corporate web site, a small banner turns to a multipage ad in a single mouse click. This means that advertisements don't need to contain as much information in writing. The advertisers' corporate web site facilitates requests for information, customer service enquiries, and on-line ordering. The possibility of profiling readers of a specific article by their IP addresses might allow purchase of advertising space on a page after publication. The possibility of adding advertisements retroactively means that an e-journal can publish volume one with no advertising and sell advertising space in that volume after they have established a readership. Advertisers should think seriously about corporate sponsorships of new e-journals as a means of reducing costly distribution of paper catalogs by educating readers to rely on their corporate web site as a source of product information. Concurrently, corporations may consider using the peer review capacity of the e-journals to lend credibility to their technical notes instead of costly paper advertising of specific products.
Libraries will change as e-journals replace paper ones. In the beginning, increased investment in computer hardware will be required to serve duplicate copies of e-journals either to the Internet or to local networks. Many libraries have already begun this process by putting their catalogs on-line and installing terminals for Medline searches. It should be stressed that not every library needs to serve every journal due to the nature of the Internet. This will lead to greater centralization with large library computers servicing very wide geographic areas. As currently existing paper journals switch to electronic formats, reduced costs from elimination of subscriptions, reduced size of physical plant required for storing paper, and smaller staff can be anticipated. The function of the library will remain archival, but electronic media will change the nature of the archive. At some point, a library will be a collection of computer servers and readers will no longer need to physically enter the library at all. At that point, access restriction will become pointless and everyone will be provided equal access to the libraries' contents. Some libraries, especially those affiliated with academic institutions, may add weblishing to their archival function by putting e-journals on-line.
Non-scientists seeking information about scientific developments operate at a distinct disadvantage in the current system of paper journals and academic libraries. As a result, most people rely on summaries of scientific developments that they read in the popular press or see on television. As a result, much of the information that they receive is actually disinformation either due to deliberate distortion of facts by special interest groups or due to poor reporting by the media. E-journals will make it easier to access source material by elimination of subscription fees and elimination of restrictions on library access (7). In addition, graphical presentation of data will improve comprehension, even among laymen. The tremendous demand for home Internet access shows that people are dissatisfied with the quality of information presented to them by the traditional media. Unfortunately, the same types of special interest groups that made such good use of print media to spread their message are strongly entrenched in the Internet (PeTA). The best way to reduce the impact of these groups on the public perception is better dissemination of factual information.
The concept of copyright as it applies to scientific articles will have to change as weblication supersedes traditional publication. Traditionally, journals have procured information for their subscribers in return for subscription fees. In order to protect this transaction, journals copyright articles at time of publication and authors waive their rights to the manuscript as a condition of publication. Since the size of a computer server is finite, e-journals will have to either continually expand their servers to accommodate new articles, or periodically archive old material. Authors, or other interested researchers, may wish to keep certain articles on-line beyond this point. Publication on the Internet implies the right to copy an article in its entirety since web browsers are designed to automatically store a copy of the file in a disk cache as soon as it is accessed. Our concept of weblication includes explicit permission to copy and to redistribute without any changes. E-journal copyright agreements should allow and encourage this practice with the condition that articles are reproduced in full including copyright information and all appropriate credit to authors and the e-journal which originally posted the article. Whenever possible a hypertext link to that e-journal's home page should also be included. An additional note explaining the circumstances of the re-posting and its date should be included in order to avoid confusion about chronology.
In summary, weblication in an appropriate e-journal disseminates information easily and quickly and should fulfill career advancement goals if only objective criteria are considered by those reviewing a scientist's publication record. If a scientist suspects that there is a prejudice against weblication at their institution, some work should be published in paper journals. Scientists with a sense of humour will cite their weblicated articles in their printed ones causing the eyebrows of more alert committee members to be raised. In a short time such artifice should no longer be necessary.
1. Medline: The National Library of Medicine offers Medline searches at [http://www4.ncbi.nlm.nih.gov/PubMed/] without charge. (Click here)
2. Steffen DL, Yarus S, Burch PE, Wheeler DA. A paradigm for electronic publication [http://mbcr.bcm.tmc.edu/ermb/citations/amia96.html]. 1996 Meeting of the American Medical Informatics Association. (Click here)
3. Search Engines [www.search.com]
4. On Line Journals: Burch PE [http://condor.bcm.tmc.edu/journals.html]. The Molecular Biology Computational Resource of Baylor College of Medicine. (Click here)
5. OnLine Journals: Atherton DJ [http://www.bioscience.org/urllists/jourlink.htm]. Frontiers in Bioscience; Virtual Library - List of Journals. (Click here)
6. Darwin C. Origin of the species [http://www.literature.org/Works/Charles-Darwin/origin/]. Online Literature Library.(Click here)
7. Virtual Library: Listings of on-line books at [http://www.ul.cs.cmu.edu/].(Click here)
Discussion between referee and author and call for readers' comments