Steven M. Bachrach and Stephen R. Heller

Contribution from the Department of Chemistry, Trinity University, 715 Stadium Drive, San Antonio TX 78212 and NIST/SRD, Mail Stop: 820/113, 100 Bureau Drive, Gaithersburg, MD 20899-2310

For the past two years, the Internet Journal of Chemistry has published, exclusively on the Web, peer-reviewed articles in all areas of chemistry. The publication model employed by the journal is unique in the discipline of chemistry. Emphasizing the significant publication advantages afforded by the electronic medium, the journal publishes enhanced articles in a user-customizable format. This article describes how the journal came into existence, examines how many of the publication choices were made, and ends with a brief look at the journal’s current status and future plans.

Print publication has been the dominant (if not sole) medium for the dissemination of chemical research for the past three centuries. Print has served chemists very well, allowing for widespread access and long-term archival retrieval.

The advent of the Internet and the World Wide Web provides a new medium that may eventually usurp print’s stranglehold on chemical information. To explore the advantages and disadvantages of this medium, to test its ability to incorporate more data in better ways, to see if costs might be contained (or perhaps even diminished), we created the Internet Journal of Chemistry. This article describes the myriad of choices made in deciding to create yet another chemistry journal, how to construct it, and how to manage it. While it is still too early to determine the net effectiveness of the journal or the Internet as a publication medium, the two-year history of the journal provides some insight and guidance to others wishing to explore electronic pub lication. This article builds on two previously published articles.

In order to properly answer the question of why to launch an all-electronic journal, one must first recall the Web environment of mid-1996. By today’s standards, this is like peering back into history about 25 years! However, enough critical technologies were in place at that time to contemplate creating a true electronic journal.

By 1996 Web browsers finally possessed three features critical to a chemical journal. First, subscripts and superscripts could be properly imaged onto the screen. While this might seem a minor point for most disciplines, where the superscript might be solely utilized for referencing footnotes, in chemistry, subscripts and superscripts are essential for relating chemical formulas. A single example will suffice: the chemical formula for the dichromate ion is Cr₂O₇^2-. With out subscripts and superscripts, chemical formulas, the lingua franca of chemistry, are absent!

Second, Greek characters can be included directly within the text. Again, this may seem a minor point, but chemical names, relationships between atoms in a molecule, and mathematical expressions frequently make use of Greek characters. Prior to the inc orporation of the FONT tag with the FACE option, chemists were forced to include Greek letters as embedded gif images, a solution that looked ugly at best.

Lastly, tables could be properly displayed as text, numerals and images within columns and rows. Tabular data are an essential component of most scientific articles. A gif image will not suffice; it lacks the ability for a reader to cut-and-paste any e ntry into some other application, such as taking a table of numbers and importing them into a spreadsheet for further analysis.

With these stylistic additions to browsers, the vast majority of complaints about the limitations of the Web as a display tool of chemistry were obviated. We were convinced that Web technology had reached a state such that authors could write an articl e and have all of the essential components completely transmitted to the reader.

What the Web enables is chemistry- and information-enhanced publication. Articles can contain extensive use of multimedia, i.e. images, audio and movies. Tabular data can be easily cut-and-pasted into other documents or programs for further treatment. Plots can be published as spreadsheet files, enabling readers to refit the curves or directly compare their own data to the published set. Molecules can be published as fully three-dimensional objects (as specified in a variety of molecular representation formats) so that readers can manipulate the structure, rotate to a new orientation, or zoom in on a particular feature of interest to them. The key technological feature that allows for this seamless transfer of data across the Web to the appropriate vis ualization tool is the definition of Chemical MIME (Multipurpose Internet Mail Exchange), which provides a tag uniquely identifying the nature of the file’s content. Using browser extension technologies, such as plugins and Java applets, electronic publis hing allows for the complete transformability of data by author and reader alike.

Besides technology, a cultural feature was emerging that we felt needed to be addressed. The widespread notoriety of the Web had raised awareness among chemists that the Web was an emerging communications medium. One of the authors (SMB) had organized electronic conferences for a couple of years with great enthusiasm and success within pockets of the chemical community. We were sure that the major publishing houses would quickly introduce new electronic products and migrate their print products over to the Web.

Unfortunately, by mid-1996 only a handful of electronic journals were available. To make matters worse, most of them were fraught with technological and aesthetic errors. The HTML code was miserable, often not taking advantage of the features of the br owser, let alone the Web. Hyperlinks were rare; large tracts of text were inserted as images. Thumbnail graphics were unintelligible and their full-blown counterparts were scans of very poor quality.

Most critically, none of the early pioneers were fully committed to the medium. One journal required that articles be submitted on paper. Another published a paper compendium at the end of the year to serve as the archival source.

Yet the Web held much hope as a new publishing medium that could truly revolutionize the way in which chemists communicated their results. We were greatly concerned that this tremendous opportunity for change would be missed and that the community woul d likely reject these first offerings as woefully lacking in features and too much of a bother to change from the status quo. We feared that this would taint the entire field of e-publication and thereby set back any serious attempt at creating a chemistr y e-journal for many years to come. We, therefore, decided that we needed to create an electronic journal that was completely committed to the technology—an electronic journal that was inherently capable of fundamentally exploiting the Web and the Interne t, a journal that was richer in content and more useful to author and reader alike than the status quo (i.e., print on paper).

Our first task was to decide on the scope of the journal. An obvious first choice was to limit the journal to computational chemistry. Many of the people involved in the journal project were practitioners in this field. Further, computationa l chemists were clearly committed users of computers and adept at using them to create documents. We would not have to spend too much time and effort in training this group on how to use Web technology or how to author Web documents. However, there were t wo major disadvantages with launching a new journal in computational chemistry. First, the discipline was already well served with journals, and the two most recent journals were electronic! Secondly, in a sense, launching a journal in computational chemi stry would be reminiscent of preaching to the choir. We felt that e-journals would be beneficial to all chemists—a computational chemistry e-journal would likely have little impact on the broader chemistry community and, therefore, would likely fail since it would not produce a strong enough subscription base.

This last argument convinced us to reverse our first inclination. For the e-journal project to succeed, the journal must have the widest potential impact. The journal had to attract authors and readers of all types in order to build a large subscriptio n base. We recognized a broadly focused journal might seem fragmented, with a smattering of articles in a variety of chemistry sub-disciplines. However, the potential payoff outweighed this concern; we decided to develop a journal that would cover all asp ects of chemistry.

The next issue was fairly easy to solve. Once we settled on the scope, the name of the journal came readily—the Internet Journal of Chemistry, which we abbreviate to IJC for simplicity’s sake. We registered the domain name ijc.com and the Web server is located at http://www.ijc.com.

We articulated a series of aims and goals that we wished to achieve with the IJC. First off, we intended to create a journal of high scientific repute. A strong editorial board and scientific advisory board would ensure editorial quality. We discussed the role of peer review, and while there were many differing opinions of its value, in the end we decided that the advantages afforded by peer review clearly outweigh its shortcomings. The chemistry community greatly prizes peer revi ew as an essential component of the editorial process, perhaps the most important role of the journal itself. Peer review provides a screening of the literature, which is especially useful when one reads articles outside one’s own area of expertise. It wo uld be fruitless to create a chemistry journal at this point in time that lacked peer review; the editors and publishers of such a journal would be constantly defending this position instead of working to make the journal a success.

We were completely committed to the total electronic nature of the journal. This commitment actually simplified many further choices. In an electronic journal, we knew the role that paper would play—none! — or rather as limited a role as we could possi bly achieve. A paper compendium is not published at the end of the year, thereby avoiding the question of what is the authoritative version. Authors would submit their articles solely by electronic means. Peer review would be handled via email and Web for ms.

We wished to exploit the Internet as fully as possible. Authors would be encouraged to incorporate multimedia and interactive tools and submit large data sets as supplementary materials. The IJC would be customizable so that readers could contro l the look of the journal and even some of its content. These features are inherently impossible to achieve with a paper journal. We believed the ultimate success or failure of the journal would rest on our ability to position the IJC as a unique p ublication, with features that so improved the communication of chemistry between an author and the readers that it would serve as the model for all future e-journals.

Resolving economic issues also ranks among our goals. We believed that there are economic benefits to electronic publishing and that these savings could be passed along to subscribers. A low subscription fee was therefore a significant aim. We also wis hed to address copyright and licensing issues in progressive and novel ways, perhaps in ways not possible with print journals.

Taken as a whole, we envisaged the IJC as a revolutionary model of the chemical journal. Technology would be utilized to create a rich communication process that would also be more economical and liberal in terms of use policy. These goals clear ly set IJC apart from its competitors.

While many of our goals are beneficial to the community at large, we never felt that the journal could succeed as a free journal. Again, consider the Internet landscape of mid-1996. Most people still considered the Internet to be free , and therefore all Web pages should be available at no cost. While this view was naïve even back then, there were a significant proportion of chemistry Web users who strongly supported this notion. In chemistry, there has been a spirit of free excha nge of software and Web pages. In 2000, the age of maturing e-commerce, the altruistic nature of the Web has gone the way of the dinosaur, and now there are a number of fee-based chemistry Web sites, as well as many which remain free, but are supported by advertising.

We were quite sympathetic to the concept of a free journal but rejected it for two reasons. First, development of the kind of journal we envisioned requires skilled programmers, good hardware, quality maintenance, and robust Internet connection. All of these require real money and institutionalization. Second, for scientific journals there is a sense that free services have less quality, are less professional. This was an attitude we could not allow to taint the journal. IJC had to be perceived as a professionally run project, one that was not a fly-by-night operation or the passing fancy of some graduate student. We needed to show that there was a stable funding source to support the journal, maintain its Web server, archive its holdings, and c ontinue to develop new features.

We explored two funding strategies: work with an established publisher or create the journal as a start-up venture. We discussed the IJC with a number of publishing houses, and quickly realized that partnering with an established publisher would entail changing or foregoing many of our goals. Some of the conflicting attitudes included the desire on our part to create a low cost journal, the desire by most publishers to produce a print version simultaneously with the electronic version, and who w ould control the Web server. Maintaining cost control appeared unlikely when one publisher’s overhead to simply process one check was $27.50. Most publishers believed that a complete transition to electronic journals was doomed to failure, that an accompa nying print version was mandatory. Building a journal under these conditions would either severely restrict what features could be enabled on the Web side (so that the print and e-versions would be identical) or raise serious concerns regarding which vers ion was the authoritative one. The intended design of the journal would require significant development of the Web server, with root system access privileges. The publishers were understandably concerned about allowing this sort of access to their Web ser vers.

More than just a matter of settling the issues stated above, there was also a misfit in terms of trying to place this project within the framework of a print-based organization. In many instances, their corporate management culture and legacy were comp letely antithetical to the aims and design of the IJC. When it became clear that partnering with an established publisher was an unworkable situation, we next sought out private capitalization of the project. InfoTrust Ltd offered a partnership agr eement that allowed us to maintain the editorial control we needed while fully supporting all of our goals. InfoTrust provided capitalization, subscription and payment management, and marketing responsibilities. With financial backing in place, we next tu rned our attention toward designing the journal.

From its inception, IJC was designed as a completely electronic journal. We chose the Web as the display mechanism given its acceptance and reach within the chemistry community. Before we could build the appropriate Web server, we nee ded to determine how authors would submit their manuscripts and how readers would interact with the journal.

A key consideration, and one we kept in mind with all design decisions, was the question of how much burden would the proposed implementation place upon the author or reader. We firmly believed (and still do!) that for an electronic journal to succeed, the demand upon the author must be no greater than the status quo, i.e., authoring an article for an electronic journal must be no more difficult than authoring for a print journal. Since the IJC venture was so non-traditional, we felt that we cou ld not place any further impediment upon authors, otherwise they would choose to continue to submit to print journals.

We decided that authors would have to submit articles written in HTML, and we provided a brief template for assisting with this production. Since all major word processors had provided for over two years the "Save as HTML" option, we believed that requiring HTML would place no undue burden on any author. A Web form, wherein author names, addresses, article title and location of the article are supplied, would handle the actual submission of the article. The article itself would be placed on a n author’s local server or transferred to the IJC ftp server.

With authors providing HTML, we could have handled articles as if they were "camera-ready." In other words, the authors’ HTML would be published untouched; each author would define whatever style he or she wanted, what fonts to use, how to de lineate sections of the article, and so forth. Even with a full template, authors would have a great deal of latitude in their HTML style. We adopted a system whereby authors submit their articles in HTML using a template as a guide, but with no formal li mitations or mandatory conditions imposed. This not only gave the authors some control, but also transferred some of the cost and burden of proofreading, copy-editing, and typesetting to the authors and their institution, thereby lowering the cost of prod ucing IJC.

While this "camera-ready" approach is exceedingly simple to implement, it prohibits implementation of any reader control over the journal’s appearance or content. While the advantages of Web publishing for authors are pretty obvious—enhanced publication by incorporation of multimedia and interactive tools—the Web provides a truly unique facility allowing readers to determine the look and content of the pages presented to them. Web browsers allow for the selection of the font and its size, the size of the window, and the default colors of the text and background; these are simple tools for controlling the look of a Web page.

We wished to extend this user customization even further. A reader should be able to control not only the aesthetic issues (size of windows, frame layout, colors of text) but also the content. Just because the authors intended for the reader to view a three-dimensional representation of a molecule with the java applet they provided does not mean that the reader should be forced to use this tool. Perhaps the reader is already fluent with an installed molecular viewer; the reader should be able to visual ize a molecule with this (or any) tool. Another customization example is the choice of units, a purely arbitrary construct, though some units may be more suitable or logical than others. Users should be able to decide in what units an energy or distance m easurement will be reported, not some editor or author. One further example of the flexibility (and political correctness) of IJC is the ability of the reader to select the color scheme for an article. This allows for those who are color blind to e asily read an article.

Such a degree of end-user control requires a consistent formatting of all articles coupled with a Web server that can customize articles to the specification of each user. Our solution involves the creation of meta-files that are XML-like. These are de livered to readers via a customized Web server, which converts the meta-files to HTML according to the specific user’s preferences.

Articles are processed in two steps, graphically portrayed in Figure 1. First, authors’ HTML files are converted to meta-files by a custom-built intelligent parser. This parser identifies various logical components of an article, such as headings, figu res, references, units, and tables. These components are tagged using an XML-like markup. The parser creates a set of metafiles for each article, stored in its own directory.

[Figure 1 placed here]

We have previously described in much greater detail the process of creating the customization tags; a simple example will suffice here. At the top of Figure 2 is a sentence taken directly from an author’s manuscript written in HTML. The parser operate s on this line, noting the energy unit and tagging it as shown in the middle of Figure 2. This text is the metafile, which is not suitable for transmission to a browser.

The second step occurs when a reader opts to view an article. The IJC server receives the user’s request for a given page of a specific article. The server gathers the appropriate metafiles and converts them into HTML. When the server finds one of the IJC XML-like tags, it resolves the tags into HTML using the reader’s preferences and browser environment to create a unique customized file that meet the user’s specifications. Returning to our example, the server converts the metafile into proper HTML, incorporating the user’s preferences (bottom of Figure 2). In this example, the reader specified that all energies are to be reported in units of kJ/mol and to have on-the-fly unit conversions directed to the lower right frame (called "o ther") of the display window.

[Figure 2 placed here]

The HTML file sent to the user is not stored on the IJC server. Every page sent out by the server is created on demand, the definition of a truly dynamic Web server. In a sense, no journal article exists on the server—it is created on demand by the server and exists only on the display screen of the browser. A journal article is really an interaction of the metafiles, a user profile and the IJC server.

This two-step architecture achieves our goals in terms of meeting the needs of both authors and readers in a way that cannot be duplicated by print journals. Authors write their articles using the tools that they are familiar with, but, in addition, th ey can incorporate extensive use of color images, movies, and interactive molecules, spectra and plots. Authors can create a more content-rich presentation for the audience. Readers can customize the journal so that information is presented in a format of their own choosing, presumably one that will best enable them to manipulate the data and absorb the materials.

A completely electronic peer-reviewed journal requires a computerized system for managing the article submission/review/acceptance flow. We developed a custom program using a combination of perl scripts, Web forms and e-mail. The outline of this process is shown in Figure 3. An author uses a Web form to submit an article. This form is processed, information is added to the submission database, and an e-mail message is automatically sent to the editor notifying him of the new submission. The editor then selects appropriate referees and contacts them by e-mail; the message includes the URL of the article to be reviewed. Referees use another Web form to submit reviews. Upon receipt of reviews, the editor is sent an e-mail message. The editor th en uses yet another Web page to examine the referee report and to make the decision to accept or reject the manuscript or to ask for revisions. If revisions are required, the author is sent the referee report via e-mail, and then uses a revision submissio n Web form to return the revised manuscript. A critical point to note here is that the whole process occurs without any paper exchanging hands—no shipping, mailing, or faxing charges are accrued during the entire editorial process.

[Figure 3 placed here]

An important feature of e-journals is the ability to publish articles individually; there is no need to collect enough articles to fill an issue. Such flexibility leads to substantial reductions in publication time. For example, the American Chemical S ociety publishes articles on the Web following their ASAP (As Soon As Publishable) policy. They note that articles can appear on the Web from two to sixteen weeks prior to appearing in print. IJC adopted this policy from the outset. Articles are pu blished as soon as they are accepted and processed. Typical turnaround time from author submission to publication averages about six weeks, with the bulk of that time spent waiting for referees to submit their reviews.

IJC publishes special issues, at this point just Festschriften in honor of an individual’s achievements, but in the future special issues on an important topic may be published. Articles which comprise the special issue are published as they are completed an assigned sequential article numbers as usual. Articles comprising the special issue are highlights in the journal’s table of contents (TOC) by means of a special background color. There is also a link on the TOC to a description of the speci al issue, along with a list of its articles. The special issue is essentially an overlay selected from articles from the journal.

The intent of our project was to position IJC as both a technological leader and an innovator in policy issues. We have already discussed our goal to make the journal as cost-effective as possible; we will discuss financial issues in a later section.

The issue of copyright is beyond the scope of this article, and will limit our discussion to the IJC copyright policy. We firmly believe in the value of the copyright and instituted policies based upon this belief. Unlike most journals, IJC does not require a transfer of copyright from author to publisher. Rather, IJC exclusively licenses the right to commercial distribution of the article and in return pays the author $10 for this right. Non-commercial distribution rights remain w ith the author. No restrictions are placed on the non-commercial rights; for example, authors may post their IJC articles on their own Web site. Realize that this version will not have any of the customization features provided by the IJC. W e believe that the IJC offers significant added value to articles through its unique server, along with offering the authoritative peer-reviewed and abstracted version of the articles.