In 1989, the University of Houston Libraries decided to experiment with the new electronic medium and learn about the efficacy of electronic publishing by becoming electronic publishers. The first volume of Public-Access Computer Systems Review (PACS Review) appeared in 1990. At that time not much was known about how an electronic publication should be structured (a great deal is still not known); how an audience might respond; what experience from print publishing would cross over; if — or what — effect formatting would have on the publication, its readership, and its production; and whether a serious scholarly publication could survive in electronic form. It seemed at the time to be an excellent opportunity to learn firsthand the answer, or set of answers, to those important questions. That experiment was to be a proof of concept.

PACS Review was sent out by e-mail at first, using the BITNET listserv commands for requesting files. We added Gopher availability when that software became widely implemented. When Web browsers became popular in academic environments, the files were put on the Web, although again just as ASCII text without HTML markup. In 1995, PACS Review began to be produced in HTML format, in addition to ASCII. (For more about the early history of PACS Review, see "Electronic (Online) Publishing in Action . . . The Public Access Computer Systems Review and Other Electronic Serials," by Charles W. Bailey, Jr., in Online, January 1991, 28-35.)

In general, PACS Review has unquestionably been successful. Librarians, its primary audience, have been early strugglers and pioneers in and with electronic publications — as users, managers, and producers. That audience has been very responsive to the publication in its electronic form both as readers and as contributors to the content, some of which focuses on electronic publishing issues.

There was once a print version. In the early 1990s, the Net was a much less certain proposition than it is now, and there was strong feeling that for accessibility, PACS Review needed to be available in print. Volumes 1 through 5 were subsequently published in print in partnership with the Library and Information Technology Association. The interest in that format relative to the electronic form was less enthusiastic. Only about 500 copies of all volumes have sold through the years. Of course, a number of libraries do print the electronic copy and make it available in that format. Presumably, performing that task would suggest that at least some readers prefer to read copy in print form; it may also mean that some libraries prefer to archive it in that form.

Scope and Audience

Within the last two years, PACS Review's focus has been on scholarly electronic publishing and digital libraries. While maintaining an interest in that area, we are also promoting a move to feature more work on Web access and general public-service aspects of computing.

We have always aimed for the same audience: those interested in aspects of computing applied in service to needs of library users. They tend to be librarians, information scientists, and scholars and faculty interested in information, and library school faculty. All of those people tend to have access to the Web. Our librarian readers have tended to come from academia; it will be interesting to see the effect on the journal as more public, special, and school librarians have an opportunity to hear about and access the journal than in the past. We expect to cover more topics that should be of interest to those other types of librarians, and will advertise to appropriate lists that will reach them.

Because PACS Review does not charge for subscriptions, it seems to be more freely available than a print journal that does have a cost. Therefore it is more likely to be read than journals that cost money. However, it is only freely available to those with the right equipment. Just because access to such equipment has spread does not mean that we should not keep that fact in mind.

In building our readership, we took some specific actions to take advantage of the global nature of the Internet, and there has been a steady international audience for PACS Review. Some members of the Editorial Board are from outside the United States, and submissions are solicited and received from other countries, especially Canada and Australia.

How Articles Differ: Electronic vs. Print

Because PACS Review was intended to be like a printed scholarly journal in format and coverage, the content of the publication varies little from what it might have been if it had been created as a print publication. PACS Review covers all aspects of end user computer systems in libraries. The tone, types of articles, and the quality of the material matches print journals in most respects. It is our impression that most of the articles in PACS Review are not significantly longer, for instance. However, since there is no direct cost increase or other constraint keeping us from running long articles, we do take the opportunity to include them when the subject demands it.

The functional layout of the articles was specifically designed to mirror print articles. For example, the table of contents lists articles as if there were some true linear order to them; in the ASCII version, pages are separated and numbered, and footnotes are at the end of the article. That layout was chosen consciously to ease the transition to electronic publication and to offer a way to cite the journal. Some of those decisions were revised when we went to the HTML format. In HTML we don't separate and number pages, and we add links to footnotes at the end of the article.

One of the greatest variances in an electronic publication is periodicity. While it is important to maintain a presence and identity as an ongoing publication through regular issuances, with an electronic journal the editors have a greater publishing flexibility. That flexibility, in conjunction with the lack of space constraints, lends itself well to being responsive to the flow of material, currency of ideas, and editorial workload.

Design Considerations

The one area of presentation that does currently vary from print practices is in the use and display of graphs, charts, and other such material. The editorial staff has been concerned about creating documents that are graphics intensive because of bandwidth and performance issues related to the Internet in general and to certain locations in particular. Earlier in PACS Review's history, length of articles and use of graphics caused problems for users in countries where access charges are high and based on bandwidth, and communications performance may be a significant problem; that factor influenced design. Over the years, PACS Review has had fewer graphics than comparable print journals.

When the journal was produced solely in ASCII format, all charts were produced using ASCII characters alone. If a chart did not lend itself to such presentation, authors were asked to reconsider the format of the data. Now that the PACS Review is also published in HTML, other options for presentation are available (e.g., graphics files, tables, etc.).

Graphics continue to provide challenges for the editorial staff, but recent issues have been produced with some more advanced graphic options. Certainly more use of graphics is expected in the future.

Changing Formats

PACS Review started as e-mail, then became a Gopher service, and now is a Web service in HTML. It is no longer efficient to send new articles out by e-mail upon request, although the Table of Contents will continue to be distributed as an issue announcement to several e-mail lists. The changes illustrate a fact of life in electronic publishing — electronic formats change rapidly, and, as yet, standards are not well enough established that would allow us to assume that one format or distribution mechanism will be good for all time.

Structural Considerations

When PACS Review began, there was a great concern about not wanting to overload readers' e-mail boxes with material that they had not specifically requested. Nor did we wish to generate undue network traffic in sending out entire issues via the listserv software. Thus, we designed a separate table of contents to facilitate notifying readers of new issues and indicating how they could retrieve the individual articles. That practice has continued. It also allows us to announce the journal more widely when an issue focuses on topics of interest to other disciplines.

Currently, the journal may include three types of articles: columns, communications, and refereed pieces. In the past, some reviews were published. We began refereeing in November 1991, in volume two.

The Nitty-Gritty Operation of an Electronic Journal

Producing an issue of an electronic journal is a new kind of enterprise in comparison to the long history of print journals, and we have is much less expertise to draw upon. The editor-in-chief currently involved more in issue production, Pat Ensor, has relied very heavily on training by the founding editor, Charles Bailey. It seems likely that for a while production will addressed by passing down "lore"!

Since we became editors-in-chief in January, 1997, we have made a deliberate decision to stay within the traditions established in the history of PACS Review. The constant challenge is how to stay relevant while not bending to every trend of the moment. Heavy use of graphics, constant redesign of the site, and linking to many external resources are certainly common on the Web, but can present a variety of problems for readers. The issue of link maintenance, for example, has kept us from providing many external links. An ongoing directive is to show that an electronic journal can be as stable and dependable as print journals generally are, while still relying on the advantages to be found in electronic publication.

Where's the Product?

Since electronic journals are by their nature not as established — or perhaps as respected — as print journals, and the rewards of publishing are not as clear, prospecting for articles is a major part of production of PACS Review. This journal is comparatively well established, and manuscripts do come in on a fairly regular basis; however, the flow of manuscripts would not begin to compare to that of a similar print journal. Producing an electronic journal, however, does make one more inclined to use the electronic communication networks to solicit and locate material. The editorial team of PACS Review sees any form of marketing of the journal as potentially drawing authors as well as readers. Marketing to seek readers should also bring authors; advantages to those authors are greater if there are more readers. We have published a call for papers on Web organization and access through discussion lists; that call for papers is also available on the Web site. We also use the Web and discussion lists heavily to "prospect" for potential papers.

That can include a variety of methods:

• We keep an eye out for contributions to lists that might indicate that the writer is involved in a project that might become a paper or that the writer might be qualified to write about a topic of interest to PACS Review readers.

• We look at electronic journals that cover related areas and tend to publish project reports. A brief project report might be followed up later with a lengthier paper more suitable to a refereed journal.

• We look at conference papers mounted on the Web and see if any might be suitable for expansion as a PACS Review article.

• We prospect for papers on topics, and by authors, of interest to PACS Review that the authors may have put up on their own Web page. Those prospects are usually identified through listserv messages and announcements of conferences.

Of course, we have no interest in soliciting anything that can be considered to have been published elsewhere. But we have no hesitation about receiving and refereeing publications that already have been mounted on the Web. The refereeing and editing process makes them into different works, and PACS Review publication gives them an imprimatur they would not otherwise have had.

The Process of Considering a Paper

The use of electronic communication also facilitates the processing of a paper. We prefer to receive submissions as e-mail, in ASCII format, or to retrieve them in ASCII format from the Web. ASCII is the simplest common denominator in the electronic world, and nothing else has been found to be as reliable. That means the editors must do more "scut work" with the file than if there were some truly standard word-processing format available. Graphics that cannot be represented in ASCII format may require some conversion to be usable.

When a paper is submitted, the editors first clarify whether it is being submitted as a communication or a refereed paper. If it is a communication, the communications editor reviews the paper, considers whether it is in scope, and whether it meets our quality standards. The editors-in-chief also review the paper before it is accepted or rejected. If the paper is accepted, the communications editor edits the paper, working with the author via e-mail.

If the paper is to be refereed, one of the editors-in-chief provides a paragraph describing the submission and sends it to an e-mail list composed of the editorial board of the journal. They constitute the corps of referees, and ask two of them to look at each paper. If we do not get two volunteers, we contact specific board members, based on subject expertise.

The referees are sent directions and an ASCII copy of the paper with distinguishing information removed; referees tell us whether the paper should be accepted, accepted with changes, or rejected. They look at content as the primary factor. Referees make comments to be sent to the author, and comments directed at the editor. They are asked to review the paper within three weeks, but this is an instance where print and electronic journals are similar — the work that relies on human attention and effort depends on when the human can fit it into his or her schedule. E-mail communication speeds up at least some parts of the process.

If the paper is accepted with changes, the author is given the appropriate input from the referees and asked to change and re-submit the paper. If it is accepted with no major changes recommended, the editor begins to work on the paper.

The Electronic Issue Production Process

Most of the PACS Review editorial and production team is composed of librarians at the University of Houston. One might think that the ease of electronic communication would mean that it doesn't matter where the team members are, but we have found that frequent face-to-face communication still greatly facilitates ease of production. One former University of Houston librarian now at New York Public Library, the associate editor for production, does the HTML mark-up, and that has worked quite well, although it does bring another potential delaying factor into the electronic production process. (The column editor could also, practically speaking, reside elsewhere.)

PACS Review is currently published in HTML and ASCII formats; the appropriate editor (either for a communication, a column, or a refereed paper) produces the final edited text of an article after numerous editorial passes. The editorial practices are quite traditional and are little affected by the fact that the journal is electronic.

Formatting is done with electronic viewing in mind, with block paragraphs and numbered sections. In the ASCII version, traditional page numbering is done, since from the beginning of PACS Review, it was decided to facilitate a traditional form of citation to articles, including page numbers. Due to arbitrary local needs, where page numbers fall is determined by the layout of the single-spaced ASCII file in WordPerfect 5.1. The page numbering convention is not maintained for the HTML version. (That has not been a problem, but if the decision is made to discontinue the ASCII version, page numbers would no longer be possible, which would be a big change for the journal.) The author is sent an electronic copy of the final text for his or her approval; the process is usually very quick, since it is electronic.

All the necessary files are made available in a secure directory accessible over the Internet to the production editor who is in New York; she does the HTML mark-up on the files. That includes linking to outside sites where appropriate, as well as doing internal linking to end notes. She places the finished files in a secure directory she has created on her server, also accessible on the Internet.

The associate editor for technical support actually produces the issue by placing the marked-up files on the server and by doing any appropriate indexing on the site. Once that is done, one of the co-editors-in-chief sends a copy of the table of contents of the issue (which includes abstracts of the papers included) to appropriate listservs, including two listservs that are maintained at the host institution, the University of Houston Libraries. One of those lists, pacs-l@listserv.uh.edu is a full-fledged discussion list. The other is pacs-plistserv.uh.edu; it consists of announcements and issues of some library-related electronic serials. Announcements are sent to other appropriate listservs, based on the subjects covered in the issue. Once that announcement has gone out, we consider the issue published.

Circulation and Marketing

Free electronic journals are different from print and/or for-profit journals in ways that affect their visibility to a wide audience. They are not located in familiar places in a tangible format; they are not available through "jobbers"; and they tend not to have budgets for marketing. On the other hand, electronic-journal producers may be more likely to make use of freely available electronic channels to advertise their journals. The crucial point is making the audience aware of the journal's existence and location. Then the electronic journal has the advantage: It is freely available at the desktop to anyone who has the appropriate equipment. In addition, an electronic journal can easily make its full back file just as available as a current issue, ensuring more extensive use of the journal over time.

The primary form of marketing for PACS Review is listserv announcements of new issues, a free marketing avenue. One drawback to use of a listserv for marketing is that list readers do not tend to be patient with repeated advertisements of a journal that add nothing to the message. (In contrast, paid ads aren't required to impart "new" information.) It is important to keep issues flowing steadily enough to keep potential readers aware of the journal's existence, but not to have it appear so frequently that individual articles do not get an adequate airing.

That points out, incidentally, another advantage of electronic journals: There is no economic need to publish issues that are packed with articles. In fact, for visibility, the inclination is in the opposite direction, so that an individual article may compose an entire issue. That focuses attention on that one article — more than it would if get it were published along with lots of other articles. On the other hand, that approach narrows the appeal of any single issue ; the fewer the number of articles, the less likely it is that an issue will have "something for everyone."

Statistics on article access definitely rise steeply when a new issue is announced. A critical point with PACS Review is to get that crucial first access by a user, who may then browse the site and read more articles. (In fact, most of the articles with the heaviest current access are several years old.) In one sense, the indexing on the site is a form of marketing for the individual articles, and is a crucial part of the "package." Achieving the credibility of being indexed in seminal print indexes, Library Literature and Current Index to Journals in Education, is another form of marketing, since that may lead new users to the site.

It is more difficult to determine circulation of an electronic journal based on a widely accepted definition, and it is an ever-changing number. The computer can tell with precision how many times an article was accessed, but is it accurate to add all the accesses of all the files to get circulation? Print journals would not usually count (they would not be able to count!) how many times each article is viewed to show their circulation. Should one count accesses to the site's initial page? But not everyone may go through that avenue.

The Top Ten and Author Reporting

One feature of the PACS Review Web site that the founding editor established is the "Top Ten" — the ten most accessed articles on the Web site [Editor's note: link removed August 2001 because it was no longer active. PACS Review ceased publication of new issues in 1998.] Even the figures on which that is based are not complete, but they are indicative, and people seem to like that feature, based on the extent to which they access it. Of course, the existence of the top ten list influences what the most accessed articles are, and it tends to become a self-perpetuating list, but there are variations in it over time.

Recently, one of the editors-in-chief has begun to contact the authors in the top ten, telling them how many accesses their articles have received. We would like to do that for more authors, as it helps us to market the journal and encourage former authors and perhaps their colleagues to submit articles, but the process would have to be automated. Individualized statistics are certainly a strong point of electronic journals.

Conclusion

Electronic journals are still in their formative, pioneering stages, so their producers must establish procedures and practices in the absence of many precedents. The producers of PACS Review fervently believe in the promise of electronic journals, and are willing to work out the issues involved. As noted in a recently published PACS Review article by Bruce Morton, "Is the Journal as We Know It an Article of Faith? An Open Letter to the Faculty" (The Public-Access Computer Systems Review 8:2 (1997): 6-17), the journal-publishing enterprise must have a new paradigm, and PACS Review is part of it.

Pat Ensor (PLEnsor@uh.edu) is the Head of Information Services at the University of Houston Libraries; she has worked with electronic information since 1981 at California State University-Long Beach and Indiana State University. She is the editor of ALA Editions' Cybrarian's Manual, co-editor-in-chief of Public-Access Computer Systems Review, and book review editor of the Library & Information Technology Association's Telecommunications Electronic Reviews.

Thomas C. Wilson (TWilson@uh.edu), Head of Systems at the University of Houston Libraries, is the founding Editor-in-Chief of Telecommunications Electronic Reviews, the Library and Information Technology Association's first electronic serial. He was guest editor of a special issue of Resource Sharing and Information Networks. He has served on LITA's Publications Committee. He is currently working on a book on systems librarianship for ALA Editions.