Magic words these days, "Electronic Publishing." What does the phrase mean? The author, the publisher, the librarian and the reader all have different ideas of what that means.
Author - "No more proofreading. Immediate publication of my work!"
Publisher - "Production tasks can be automated."
Librarian - "Cheaper serials, more complex operations."
Reader - "Access to the full literature for free!"
Archivist - "Everything will be lost!"
These caricatures are, of course, exaggerated for effect, but serve to emphasize that each of us has special concerns when we think about electronic publishing. Without taking a broad view, it is impossible to know what electronic publishing really is today, much less what it will mean to all of us in three years.
BACKGROUND
[Back to Contents]
The electronic publishing situation is changing so rapidly that it is virtually impossible to keep up. Treloar(2) gives a recent summary which provides a good starting point. The Internet and the World Wide Web are growing by leaps and bounds. New tools are introduced every week. Java and other browser application packages are introduced with great fanfare. Microsoft and Netscape are vying for dominance(3) of the network market by adding extensions to their browsers, extensions which appear to be designed to increase incompatibilities. Finally, new electronic journals are hitting the screens in ever increasing numbers.(4) So what are serious users and producers of scholarly material to do?
The American Astronomical Society (AAS), a small professional society and The University of Chicago Press, a not-for-profit scholarly publisher, have teamed up to exploit the Internet to develop and to publish the electronic Astrophysical Journal Letters (eApJL),(5) one of the most fully featured scholarly electronic journals yet produced.(6),(7) After 13 months of production experience we have learned a great deal about the differences between traditional paper publishing and the new electronic version. Comments from authors, readers and librarians have provided a strong base of experience from which we can see the future a little more clearly. We are beginning to understand the great advantage of weaving various sources of information together with threads which extend around the world.(8) Perhaps our experience and conclusions will be of some help in understanding how electronic scholarly journals will evolve in the next few years.
SUCCESSFUL ELECTRONIC
PUBLISHING[Back to Contents]
As we have pointed out previously,(9) there are several keys to producing an effective electronic journal. The first key in this rapidly evolving area is to use well-established, open standards, to design the journal to be accessible with widely available browsing and printing tools. As the recent demise of OCLC's Guidon system for electronic publishing exemplifies, it is critical to avoid becoming dependent upon a single set of proprietary software tools. The fortunes of even the large organizations are changing so rapidly that depending upon proprietary systems is simply not a responsible solution for the presentation of material which has a shelf life measured in years or decades.
The second key is to take small steps, to be flexible, and to base future steps on the feedback from users. Very few people have sufficient imagination to envision the world five years from now and most people could not design a journal which, unchanged, would still satisfy users. The only hope is to design electronic publishing systems with an eye toward modularity and flexibility, as was done for the Hubble Space Telescope. A sufficiently modular design will allow upgrades and enhancements to be installed to various parts of the system as they are needed, as the astronauts did for the Space Telescope. Implicit in this statement is the assumption that publishers, and to some extent users, will maintain their enthusiasm for trying new approaches and adopting what works.
The Complete Publishing Process
[Back to Contents]
Another key is to understand the complete publishing process. We have previously described six
components of this process.(10) Five of them have evolved through the years to ensure effective
and accurate communication on paper. The need for these components does not disappear in the
electronic era. On the contrary, the need for valid content, clarity of expression, and effectiveness
of presentation increases as more of the journal is read on the screen. In addition, the new
electronic era requires another step in the middle, namely, the preparation of the electronic
manuscript database which constitutes the central core of an electronic journal.
The components
At the heart of the electronic journal is a well-constructed database of electronic manuscripts richly tagged in such a form that various representations can be derived from the central database automatically. Contrast this with the paper journal where the production run of printed copies forms the core of the journal. For reasons of access and archiving, printed copies are stored in libraries around the world. The electronic journal archival database, on the other hand exists in a very small number of places and is accessed over the network. Each manuscript must be well enough defined, through tagging of the material, to make it possible to reproduce the article, complete with mathematics, graphics, type fonts, headings, etc. SGML is the standard for electronic manuscript tagging which has been adopted in practice by the physical science community.
If, a decade from now, a new standard emerges, it will be possible to migrate from today's richly marked up SGML to the new standard automatically, without losing information or altering the article. Without the complete information as coded in the SGML, electronic articles would be doomed to an unacceptably short lifetime. Even with SGML tagging, long term archiving (to 100 years and beyond) will require concern and attention on the part of the archiving organization to ensure survival of electronic articles.
ELECTRONIC LINKS
Back to Contents]
Electronic publishing is about content, presentation, links and services as adapted to the Web capabilities and limitations. The key to content is editorial control of quality and a production stream aimed at today's electronic journal and tomorrow's needs. In the electronic medium, a successful editorial style is very different than we have grown used to in paper.
The traditional paper journal article or book is a linear presentation, one path from start to finish. This is not necessarily the best presentation for computerized use. Electronic articles are best presented in a non-linear, matrix form: short, concise nuggets with a map and tools to allow easy navigation among them. The secret is to present material one or two screenfuls at a time. Most articles are not yet written for effective electronic presentation. The transition will be difficult. In fact, the most successful practitioners of writing and reading the linear article may have the greatest mismatch with the electronic medium.
The Red Sage project(11) spent a major effort trying to make the computer screen look like the pages of a book. The Adobe PDF format for electronic transmission, display and printing of pages is geared to the same approach. Yet, the readers of our electronic journal, who have the choice of either PDF format or an HTML presentation with links, navigation aids and a continuous flow of the article on the screen prefer the HTML presentation four to one. We believe our readers are using the PDF format only to print out the articles they actually want to save in paper form.
For these reason we consider journals delivered in PDF and PostScript®, which only provide electronic delivery of paper page images, to be electronic document delivery systems, but not true electronic journals. Just as we now recognize that an ASCII-only journal on architecture or astronomy would fail for lack of graphics,(12) so in a short while we will come to realize that PDF-only journals, without links to references, without tools to navigate within the article, and without a smooth flow of information on the screen, simply can not compete for an electronic readership.
Electronic capabilities far outstrip what can be done with print and paper. Readers of our eApJL rate the links to references as the most desirable feature of our journal. An electronic journal can, and should, provide links to original data, intermediate results and finished articles. The ability to roam the electronic library from desktop computer terminals is changing the readers' approach to research and scholarship. The capability to link outside the document and outside the journal is changing the way researchers in our field work. Although links to referenced articles should now be quite common, this is often not yet the case. However, reader pressure will soon force other publishers to provide this service.
An improvement on linking the reference list items to the cited resources is "forward referencing," links to electronic articles published later that refer to the paper being read. This capability - an immediate electronic citation service for each article - is in its infancy, but will become more and more useful as more journals move to electronic versions.
DISTRIBUTION
AND ARCHIVING[Back to Contents]
Once written, articles printed on paper are unchanging. As mentioned above, the electronic journal exists as a central electronic database. Access to an electronic journal has to be controlled. The author's rights and the integrity of the article have to be preserved. Readers must have confidence that the copy of the article they read has not been tampered with and that it represents what the author wrote and what the editorial process has vetted. The demands of scholarship require that the electronic article delivered to the researcher is the original one, dated and verified. Payments, reputations, patents, and even Nobel prizes may ride on the ability of an electronic system to remain tamper proof.
Nevertheless, authors and readers want to be able to change and update articles as new information becomes available. The challenge is how to accomplish it without undermining the scholarship and intellectual integrity of academic publishing. The concept of evolving articles books and monographs is something the electronic medium makes possible. In fact, readers have become very accustomed to looking on the Web for the latest updates to all sorts of information. It is simply accepted that the version on the Web will contain the latest information.
Fortunately, we have a means for updating which preserves the historical documents. In the AAS journals, we attach errata to the articles as they are approved by the editor. We require that updates and new results are published as new work, a common practice in modern scholarly publishing. An exciting benefit of the forward referencing system is that the new articles will be automatically attached to the previous articles. Although this may seem cumbersome to some of today's Internet surfers who have grown used to the free-for-all attitude of various newsgroups, such a system preserves the scholarship while adding flexibility and life to an article in ways undreamed of in the days of paper publishing. We see the publisher's role evolving to include authenticating the content; certifying the contents as being the paper at the date of publication.
There are two reasons to update documents. As noted above, there are salient reasons for pointing to an update of an article's content. In addition, the technology, tools, and links from an article will also need updating from time to time. The ability to make these kinds of enhancement must be part of the planning for any electronic publication as well. In the first nine months, we have reconstituted our public offerings three times from our richly tagged database, each time adding new features and more reader-friendly capabilities. Such updates are possible only if the database is designed from the start to be updated and enhanced in a systematic and automated fashion.
MIRROR SITES
[Back to Contents]
Experimental projects such as TULIP(13) were designed to provide multiple copies of journals for distribution by a number of university libraries. The feasibility of such wide distribution was based on delivering static, unchanging documents electronically. As soon as the concept shifts to allowing various changes and updates, as we believe any good electronic journal should, it does not become physically feasible to distribute more than a few copies. The TULIP project found it hard to distribute 43 journals to nine universities over the present form of the Internet.(14) In addition, anyone who has experienced problems with updates and incompatible versions of software can understand, all too well, the difficulty of maintaining multiple copies of updated articles. These are technical difficulties associated with transferring large files, but the underlying problem with TULIP was the page-oriented and link-free format that made the final product no different from the paper version. Despite attempts to do so, conclusions drawn from the TULIP program should not be extrapolated to predict user reaction to a fully functional electronic journal.(15)
Of more importance is that the electronic journal is already becoming closely tied to other journals and bibliographic reference services. Providing effective networking when there are a number of copies of the journal located in different sites is a complex problem, and solutions for it are at an early stage of development on the network. Standard addressing schemes, effective use of distributed caches, and inquiry and location systems which are transparent to the user must become part of the network as it evolves over the next few years.
The networks are also, at times, very slow. Wired Magazine says "WWW stands for World Wide Wait."(16) Despite the difficulties involved, there will certainly be pressure by larger libraries to negotiate for local copies of journal archives. To do so, libraries will have to take on more responsibility. They will have to provide the secure repositories for redistribution, ensure that access is limited to appropriate licensees, perhaps even maintain password lists and usage statistics. In such a scenario, libraries will, in essence, function as an arm of the publisher.
The development of closer cooperation between organizations providing separate links in the chain of electronic publishing would not be unexpected. As we have described(17), one characteristic of a successful electronic publishing program, which was discovered early on, is the need for increased collaboration and cooperation within the electronic journal production process. It is logical to expect that those institution who can make the same transformation from independence close collaboration will be the most successful at effectively distributing electronic journals.
THE ELECTRONIC
ASTROPHYSICAL JOURNAL LETTERS[Back to Contents]
The AAS and the University of Chicago Press, with partial support from the National Science
Foundation,(18) have been planning and developing the eApJL since 1992. The production version
of the journal, which comes out every ten days, debuted in September 1995. In the preceding
year, several prototypes of single issues had been presented to readers, evaluated, and revised.
The resulting journal incorporates many innovative and user-oriented features which are
enumerated below, but which are best investigated by trying the electronic version of the journal
at the URL <http://www.aas.org/ApJ/>.
Interlinking features
Internal features
THE INTERLINKED FUTURE
[Back to Contents]
Links are the central theme of our electronic journal. Reader feedback continues to emphasize the importance of links by which it is possible to retrieve referenced articles, navigate around the journal, and to download data tables from publicly available archives. Astronomy admittedly has an advantage in providing these links to live resources. NASA took a lead role for years in making data from space missions available over the Internet, and the agency sponsored the compilation of a bibliographic database of abstracts for use by space scientists.
That bibliographic database is now part of the NASA-supported Astrophysics Data System (ADS),(19) where it forms the core of a system by which page images of historical articles are made publicly available (Figure 1). Several publishers of astronomical journals have made arrangements to allow their current abstracts and older articles to be made available through this searchable database. In return, the ADS now links from their abstracts directly to the electronic versions of articles. The ADS also serves as a clearinghouse for links from the eApJL references, providing abstracts of the references and, when available, the page images of the historical articles as well.
In addition to enjoying the services of the ADS, astronomers have for years, been able to query a database of observations of stars, planets, and other objects at the Centre de Données astronomiques de Strasbourg(CDS)(20) in Strasbourg, France. This database is being used to distribute machine readable versions of data tables published in the world's astronomical literature, and is being incorporated into the electronic astronomical literature.
Figure 1 sketches out an idea of the complex interactions surrounding the electronic journal of the future. Indeed, the power of the electronic medium is the ability to bring the whole world right onto your desktop. The challenge will be twofold: to manage the links and to find and retrieve the relevant information. Such challenges can be met only if everyone - authors, publishers, libraries and readers - works together. The rewards will be enormous.
COSTS
OF ELECTRONIC PUBLISHING[Back to Contents]
From the standpoint of a non-profit publisher whose role is to encourage and facilitate communication, the additional costs for an ongoing electronic publishing program are not enormous. The costs of the increased labor of adding appropriate hyperlink tags can be offset by the use of author-prepared electronic manuscripts. There is an additional cost of establishing a fund for maintenance of the electronic archives into the future.
It should be noted that the Astrophysical Journal is inherently complex and expensive to typeset. The potential savings by using author prepared manuscripts are much larger in this case than would be possible for simpler journals.
The major expenses we have encountered in bringing the electronic journal on line were for the experimental development and the re-engineering of the production and distribution systems so that they could take advantage of the electronic environment. Part of this work has been supported by the National Science Foundation, with the remaining development costs coming from the AAS and the University of Chicago Press.
For the first year, the costs for a combined electronic license and paper subscription appear to be about ten percent more than paper only. We expect some cost savings in future years as procedures improve and the percentage of electronically submitted manuscripts rises.
As outlined above, we believe that in the electronic era libraries and publishers will have to work more closely together. It is in everyone's interest to have an easily negotiated, standard site license which will provide unlimited use within an institutional Internet domain. Asking the library to administer individual passwords and access for many different publishers is unrealistic.
Our library customers have indicated a strong preference for predictable fees, not metered according to either time, simultaneous use, or volume of traffic. This is the way we are charging for access in 1997. However, this may pose too much of a burden on the smaller users. Only through experience, feedback and attention to market forces will publishers be able to develop an effective and equitable charging structure. with unlimited access. For the individual user or reader who travels often, we will be offering password protected, electronic-only subscriptions usable worldwide. An individually tailored alerting service will also be offered in the near future. We consider these charging schemes to be experimental and will adjust them as we gain experience. We will also share our experience with the library and publishing communities.
DESIRABLE FEATURES
IN AN ELECTRONIC JOURNALAlthough many publishers are claiming to have electronic journals, it pays to examine just what they mean. In looking at a number of scholarly journals on the Web, we notice a wide range of features, quality of presentation and convenience. We have developed a checklist for quality which may be of use for people contemplating purchasing an electronic subscription.
In our opinion a good electronic journal should start with excellent content. Users, after all, are interested in the content and the services (such as refereeing, copy editing and links to references) provided by the publisher.
Second, because it is yet to be seen how effectively an electronic archive will work, we suggest investigating the publisher's stability and commitment to maintaining an accessible archive for the indefinite future. If the technology changes, will the publisher migrate the archive to the new platform? Can they, in fact, make such a migration, or is the material stored in a way which makes this impossible? Most publishers have not yet addressed this question, and the cost could be substantial. Some we have talked to are thinking of keeping articles on-line for as short a time as two years. Such a policy will completely vitiate the usefulness of the electronic journal.
Look at the on-line presentation. Is it a full-featured electronic journal, or just electronic delivery of page images? Delivering PDF images is a quick and inexpensive way to provide electronic delivery. Some journals are loaded with gratuitous graphics which add nothing to the information content, but do add to the transmission time. Other journals are nothing but great square blobs of text which are not easy to read, or even to skim, on the screen. Electronic journals do not have to be ugly.
Does the journal have links to references? This feature is the single most important feature for readers. An electronic journal without such links should not be taken seriously.
Inspect the way the components of the journal are referred to. Can you get to the articles by an indirect name (URN) or must you use a precise URL? If URLs are used, any bookmarks and links to the journal that readers establish will break when the journal is moved. Does the journal have forward referencing installed? If so, how are the forward reference links derived?
Does the journal have links within the articles, e.g., from each reference to the citation within the text, from the table of contents to the sections, from the figures back to the text, and so forth?
Finally, look for overall ease of use. Does it require a complex operation to sign on and access the journal or is it simple and easy for the user? Can you get multiple formats, print articles locally and move easily around the articles?
This is a time of development, experimentation and change. If you don't like the electronic products you see, now is the time to let the publishers know. After publishers' systems get firmly established, it will be more difficult to react to readers' and users' preferences.
THE FUTURE
[Back to Contents]The future for the astronomical literature indicates the pace of progress. The full Astrophysical Journal will come on line at the end of 1996, and will be restricted to paid licensees as of January 1997. Within six months after that it seems certain that the other important scholarly publications in astronomy will also be on line, perhaps with not so many features, but available, nevertheless. Discussions are now taking place among these journals with the goal of ensuring compatible formats, naming schemes and linking procedures. By the end of 1997, if not sooner, over 95 percent of the world's peer reviewed astronomical literature will be available on the Web with the capability of interlinked references. We have barely begun to understand what changes this will precipitate.
Figure 1. The better electronic journals will soon be serving as the hub for many links to other journals, bibliographic information, data and other types of information. In the electronic era journals can no longer be independent. Effective links require standard naming conventions, which, in turn, will promote the flow of information across disciplinary boundaries.
[Back to Contents]
1.
Boyce until recently was the Executive Officer of the American Astronomical Society. He now holds the position of Senior Associate in charge of electronic publishing. Owens is Electronic Publishing Manager for the Journals Division, The University of Chicago Press. Biemesderfer is a managing partner at ferberts associates, an electronic information consulting firm.2.
Andrew Treloar, "Electronic Scholarly Publishing and the WWW," Scholarly Publishing, 27, No. 3 (1966) 135-1473.
Michael Neubarth, "Microsoft Declares War," Internet World, March 1996: 36-424.
Ann S. Okerson and James J. O'Donnell, NewJour - Electronic Journals and Newsletters, URL <http://gort.ucsd.edu/newjour/ >5.
The Electronic Astrophysical Journal, URL <http://www.aas.org/ApJ/>6.
Steve Hitchcock, Leslie Carr and Wendy Hall, "A Survey of STM Online Journals 1990-95: the Calm Before the Storm," (1996), URL <http://journals.ecs.soton.ac.uk/survey/survey.html>7.
William H. Calvin, "The Future of Journals," Science Surf, 96.2, URL <http://weber.u.washington.edu/wcalvin/scisurf.html>8.
Peter B. Boyce, "Building a Peer Reviewed Scientific Journal on the Internet," Computers in Physics, (May/June, 1996)9.
Peter B. Boyce and Heather Dalterio, "Electronic Publishing of Scientific Journals," Physics Today, January 1996, 42-47 and URL <http://www.aas.org/~pboyce/epubs/pt-art.htm>10.
Peter B. Boyce, "A Successful Electronic Scholarly Journal from a Small Society," URL <http://www.aas.org/~pboyce/epubs/icsu-art.html>11.
The Red Sage Electronic Journal System, URL <http://redsage.ucsf.edu/>12.
Elwin C. Robinson, Architecture, Graphics and the Net: A Short History of Architronic, a Peer-Reviewed E-Journal, The Public Access Computer Systems Review, 7, no.3 (1996) URL <http://info.lib.uh.edu/pr/v7/n3/robi7n3.html>13.
Jaco Zylstra, "The University Licensing Program (TULIP): A Large Scale Experiment in Bringing Electronic Journals to the Desktop," Serials 7, no.2 (1994) 169-17214.
Ed Barnas, "Favorable and Unfavorable Experiments - What have we learned?," Scholarly Publishing Today, 5, No. 1 (1996) 3-815.
URL <http://www.elsevier.nl/homepage/about/resproj/trmenu.htm#ExecSummary>16.
Wired, (June, 1996): 7817.
Peter B. Boyce and Heather Dalterio, op.cit.18.
Under NSF grant AST94-03853 entitled "A Prototype Electronic Journal."19.
The NASA Astrophysics Data System, URL <http://adswww.harvard.edu>20.
Centre de Données astronomiques de Strasbourg, URL <http://cdsweb.u-strasbg.fr/>