A Guide to Electronic Archives

The Ohio Library and Information Network (OhioLINK) is a consortium of the libraries at more than 70 universities and colleges in Ohio, including Ohio's university and teaching hospitals. Funded by the State of Ohio, OhioLINK serves these universities and colleges by purchasing, storing, and providing access to information resources. In pursuit of this mission, OhioLINK has developed one of the world's largest electronic journal archives, the Electronic Journal Center (EJC). To date, OhioLINK has entered into licenses with the following publishers: Springer-Verlag, Elsevier Science, Wiley, American Physical Society, Project Muse, Kluwer, and Academic Press. Its journal archive, the EJC, contains more than 1.3 million articles, representing more than 78,000 issues from over 2400 journals. By building this archive, OhioLINK has learned many lessons about electronic journal archives.

Electronic Journal Archives and their Content

Who Archives Electronic Journals?

There are two types of electronic journal archives: a) the archives of journal publishers, and b) the archives of aggregators. Aggregation is a term used to refer to the collection of journals of multiple publishers on one web site. Aggregators get journals from publishers and store them in their own archives. Instead of accessing journals through a publisher's archive, users at an institution with a library that uses an aggregator will access journals through the aggregator's archive.

Aggreation is a term used to refer to the collection of journals of multiple publishers on one web site.

There are two types of aggregators: a) commercial organizations that sell access to their archives to libraries for a fee; and b) libraries or library consortia, such as OhioLINK, that aggregate and archive journals for their library members or users.

OCLC's ECO, www.oclc.org/oclc/menu/eco.htm, and Blackwell's Navigator, http://navigator.blackwell.com/BisJnlListFrames.htm, are two examples of archives created by commercial aggregators. OhioLINK, the University of Michigan, Los Alamos, and the University of Toronto, are four non-commercial organizations that have served as aggregators.

Aggregators are popular with libraries because they simplify the process that library users experience when they go to look for journal articles. Their web sites reduce the number of interfaces that must be navigated in order to find an article. The aggregator's archive creates one system for browsing and searching all articles found in a given publisher's journals. Without an aggregator, users would have to go to each of the publisher's archives for those journals; and each archive would likely have a different interface. Without an aggregator it would be impossible to search all of the journals at once. To find an article on a particular topic each publisher's site would need to be searched individually, a time consuming and inefficient process. Elsevier Science, www.sciencedirect.com, Academic Press, www.idealibrary.com, and the American Medical Association, http://pubs.ama-assn.org, are three examples of publishers that have developed their own archives. The web site of this journal, www.nursingworld.org/ojin/, is another example of this kind of archive. Within each of these sites the publisher controls the archive, its content and its features. At the present time, most users will gain access to journals through publisher archives; not all publishers have their journals available through the services of an aggregator. Many large publishers have decided not to sell their journals to commercial aggregators, fearing competition with their own archives. There are some publishers, such as Academic Press, that both maintain their own sites and sell to aggregators. Because most users gain access to journals through a mixture of publisher web sites and aggregator web sites, it may be difficult for the users to ascertain what journals are available to them and where.

These virtual archives are web sites that contain databases or lists of the journals to which a library subscribes and links to those journals.

This situation has led many libraries to create "virtual archives" for their users. These virtual archives are web sites that contain databases or lists of the journals to which a library subscribes and links to those journals. Two examples of these virtual archives are the Cleveland State University virtual archive, http://html.ulib.csuohio.edu/ej/, and the California Digital Library directory of journal resources, www.cdlib.org/directory/ .

What is the content of an archive?

Most electronic journal archives today contain electronic versions of journals that are issued in print. When someone says that a journal is in an archive, it is important to be aware of a large caveat attached to that statement. The electronic version of a print periodical in a journal archive will typically not include all of the content included in the print version. Electronic journal readers will lose access to some of the contents they are accustomed to seeing in print.

The electronic version of a print periodical in a journal archive will typically not include all of the content included in the print version.

Publishers in the electronic versions of print journals do not distribute many items such as letters to the editor, advertisements, calls for papers, and other small items. Other items, such as journal issue supplements including those that may accompany a journal issue on a CD-ROM, may also be eliminated. In addition, when journals are printed higher resolution images for pictures, figures and illustrations are used than are later included with the electronic versions of the same journals. The electronic journals include lower resolution versions of the image files. The rationale given by the publishers is that since the images typically used for printing may be a number of megabytes in size they would make the article file too large for electronic distribution to most users. What are the formats of journal articles? For the content that is provided through a journal archive, the two most common file formats used are HTML (Hypertext Markup Language) and PDF (Portable Document Format). Each has its own particular merits and limitations for disseminating electronic journal articles. PDF is probably still the more common format used to publish articles on the web. It is still the primary format of the journals delivered to OhioLINK. PDF has the merit of being able to produce articles that appear very similar to articles in a printed journal. This has made it very popular especially early on when libraries were beginning to buy access to electronic journals on the web and some users were uncertain about electronic access as a replacement for print. HTML is becoming more common as a format for journal articles. As an example, the articles of this journal, Online Journal of Issues in Nursing, use HTML. Growing familiarity with the web and HTML is leading to its increased use. HTML lacks the control over appearance that PDF provides, but it can be viewed, searched, and printed much like any other web document. No special viewer is required. As most electronic journal readers are now very familiar with the Web, there is a growing expectation that all electronic documents should be just like a web site. HTML web pages are increasingly becoming the paradigmatic form of communication. The gap between HTML and PDF should disappear in the near future because of the development of XML. XML will allow much more precise control over the appearance of web pages than HTML. Some anticipate that it will be close to the control provided by PDF.

How Archive Web Sites Provide Access to Journals

To fully realize the potential value of electronic journals through an archive web site, the organization that maintains the archive has to create the means by which its users can find the articles they need. To do this, archive web sites have been designed with three primary capabilities: browsing, searching, and linking. The first two capabilities will be familiar to anyone who has ever browsed through journals in a library or searched a database. The third capability, linking, provides the ability to link to the articles from the other articles in the archive or from periodical indexes like CINAHL. Linking adds a new and important capability to journal archives and significantly enhances the users' potential to find the articles they need. Linking uses the hypertext linking capability of the World Wide Web (WWW) to bring together disparate pieces of the periodical literature.

Retrieving Articles through Searching and Browsing

An archive's web site must enable users to browse through its contents, as users like to be able to browse for journals online just as they might browse for journals in a library. Browsing can be hierarchical in which the user goes from a journal menu to a journal issue menu and then to a table of contents for that issue. Browsing can also be lateral. Lateral browsing allows the user to go from one issue, article, or volume to the next or previous issue, article or volume. The OhioLINK EJC and other archive web sites with large numbers of journals also provide menus of journals in specific subject areas and an alphabetical menu of journal names as additional means of browsing.

In addition to browsing capabilities, most archives also make it possible to search for articles. Archives can permit two types of searching: searching of full-text and searching of article metadata. With full-text searching, users are able to locate articles by searching through the text of articles using words or phrases. When searching with article metadata, the bibliographic data for an article, for example its title, authors, publication date and abstract, are searched.

While it might seem that the ability to search the full-text of articles would be of great benefit, it can be of questionable value with large archives. OhioLINK's EJC can index the full-text of the journals and originally OhioLINK allowed this type of searching. The decision to stop full-text indexing and only index the bibliographic information for the EJC articles was made because of the poor retrieval performance that occurred when the full-text was searched. With full-text searching of more than one million articles from journals in many subject disciplines, it was found that all but the most arcane scientific terms would occur too frequently for useful results. Even searches including technical terms like "gene therapy" would return too many results to be useful for the searcher. The search results were rarely specific enough for OhioLINK's users.

There had been initial enthusiasm about full-text searching as a partial replacement for the use of traditional citation databases, for example the Cumulative Index to Nursing and Allied Health Literature (CINAHL). Based on users' experiences with full-text searching, it was determined that these databases had important characteristics that were not replaced by searching electronic journal archives.

There had been initial enthusiasm about full-text searching as a partial replacement for the use of traditional citation databases, for example the Cumulative Index to Nursing and Allied Health Literature (CINAHL). Based on users' experiences with full-text searching, it was determined that these databases had important characteristics that were not replaced by searching electronic journal archives. The content of citation databases is selective: the citation database publisher only includes journals on specific subjects. The database publisher also supplements the information about articles in these databases: terms from a discipline specific vocabulary are usually added to these records. These attributes make citation databases a better tool for searching for journal literature than electronic journal archives.

Helping Users find Articles by Linking

Nurses, like other researchers, usually do not want just one article. They are looking for literature in a specific subject area. To find this literature a number of tools are used, including citation databases like CINAHL or MEDLINE and eventually nursing journals. Each of these sources contains part of the information needed by a researcher. Unfortunately, the access to these different resources has not been integrated. In the print world, it is necessary to search one source then another, and keep track of which articles have been found. One must eliminate duplicates, take citations found in CINAHL and MEDLINE and then look for the journals in a library catalog. In the online world, the same separations still occur. It is necessary to take the citations from a citation database, then go to journal web sites to find out whether they are available online. It may be necessary to check one source first to find out which archive contains the desired journal. Then that archive can be browsed to retrieve that article. It would be ideal if users could do all of this work quickly, easily, and efficiently online. Linking different pieces of content on the WWW facilitates this. Links turn a number of divided and separated research tools into an integrated search system that increases the probability of users discovering the articles they need. Given the opportunity to integrate the periodical literature, OhioLINK and other archives have adopted the strategy of linking their archives to related journal resources.

Types of Links There are two key types of links that go to or from electronic journal archives. The first provides links to and from citation databases. These links enable users who are searching a citation database to immediately retrieve the article they find from the archive. The second type provides links from one article to the next. This type of link lets users follow the connections within the literature by going from footnote or bibliography entry to full-text of the cited article. Links from Citation Databases to and from Electronic Journal Archives Citation databases and electronic journal archives can be interconnected so that whenever an article is indexed in a citation database, if it is available in an electronic journal archive, the user is given an immediate link from the citation to the full-text of the article. There is no need for the user to record the citation and then go looking for the article in an archive.

One mouse click can take the citation database searcher straight from the citation database to the full-text.

One mouse click can take the citation database searcher straight from the citation database to the full-text. Links can go from citation databases to electronic journal archives, and they can also go in the other direction. OhioLINK has made a link between its archive and its copy of the ISI (Institute for Scientific Information) citation indexes, e.g. Science Citation Index. These indexes show the articles cited by other articles. For the articles in the OhioLINK EJC, it is possible, using the ISI indexes, to find other articles that cite the article in the EJC. Since OhioLINK also has links from the ISI citation indexes to its EJC archive, the process can actually be circular: the user can find articles that cite an article in the EJC and then find the full-text of those articles in the OhioLINK archive. Citation database publishers and aggregators have both sought to develop their linking capabilities. Citation database to archive full-text links have been made by citation database publishers like Silver Platter through its SilverLinker project, www.silverplatter.com/silverlinker/index.htm. OVID is both an example of a citation database publisher promoting full-text linkages through its Journals@Ovid project, www.ovid.com/products/fulltext/journals/journals.cfm, and an aggregator. Ovid locally mounts some journals and links them to its citation databases as well as linking those databases to journals at the web sites of other publishers. Citation database publishers have entered into agreements with publishers and aggregators to enable these linkages. They have been similarly promoted by aggregators who want to promote the use of their full-text collections. The developer of OhioLINK's archive software entered into an agreement with SilverPlatter so OhioLINK libraries that are SilverPlatter customers could link from their SilverPlatter databases to the OhioLINK archive. Links from Article to Article Another important form of linking that can be enabled by electronic journal archives are links from one article to another. When one article cites another article, a link can be made from the footnote in the first article to the full-text of the article cited.

Another important form of linking that can be enabled by electronic journal archives are links from one article to another.

As these articles are going to be on related themes, this linking makes it possible for users to wander among related articles. This kind of linking is not yet supported by OhioLINK's Electronic Journal Center, but the journal archives of selected publishers support it. These publishers have all of the citation data they need to link among their own articles. They know which of their articles has footnotes citing other articles they have published. The American Physical Society, http://publish.aps.org/, is one archive that supports this capability. Article-to-article linking will grow significantly in the coming years. A multi-publisher project, CrossRef, www.crossref.org, has been started to enable publishers to have a standard way of linking from one article to the next. Most of the major journal publishers are a part of the CrossRef project. They expect to have an infrastructure supporting linking between millions of articles during the near future. CrossRef should make article linking a standard capability of web sites. It will create the ability for publishers to create links between footnotes or citations for multiple publishers and the archives where the full-text of those articles are available. These links would provide the advantage of linking to a broader community. The CrossRef system will make it possible for any journal publisher or author on the web to embed links to the articles they footnote or include in their bibliographies. While this broader use of linking is good, the efforts to create these standard linking systems will have some negative consequences for libraries and their users. The standard linking systems that have been developed to date always point back to one server that only retrieves the article from a publisher's site. Therefore these systems do not provide a means of linking from an article to aggregator web sites. Thus, they do not have a means of recognizing that a student at an Ohio university can retrieve the full-text of an article from an aggregator like OhioLINK EJC, but not from the publisher web site. As a result, the student seeing a link from a footnote in an article will be taken to the publisher's web site which will deny access to the cited article. The link will not connect to the archive where the article would be available to the student. It will be difficult for a researcher to take full advantage of all the electronic text available if the CrossRef system does not know about local arrangements made with journal aggregators.

Changes in Electronic Journals

Much of the current online environment for viewing journal articles is based on electronic journals that are the online equivalent of a print journal. As time passes and the WWW matures as a communications mechanism, electronic publishing will undoubtably change and journal archives will need to change at the same time.

...electronic publishing will undoubtably change and journal archives will need to change at the same time.

While these movements will provide opportunities to further increase the value of electronic journals, new problems will be created for researchers and authors. At present, there are two key trends underway. The first is the development of electronic-only journals and the exploitation of the abilities that are created by the Web as a medium of communication. The second trend is the decentralization of publishing. The Web reduces the barriers to becoming a publisher and makes it possible for organizations that cannot afford to be print publishers to become involved in the publishing process. Electronic Only Journals and Embedded Multimedia Printed journals include text and static images. Their electronic equivalents maintain these same limitations. But on the WWW, the number of media that can be used for communication is much larger. By publishing a journal primarily in electronic form and not just as a secondary version of a printed journal, a publisher can seize the opportunity to expand the ability of its authors to communicate to their audiences. Multimedia can be added to a journal, benefiting both readers and authors. There are many different kinds of new media that can be used with a journal on the web, such as a movie. The Heart Surgery Forum, www.hsforum.com/, is an example of a journal that uses movies as a supplement to the text of its journal. Other kinds of media are also available. Computer programs can be of value for a computer science or mathematics journal. Non-traditional media like virtual reality (VR) files or molecular models can be added to articles. These can be used to display in three dimensions structures that have had to be displayed in two-dimensional diagrams in print journals. Multimedia journal articles more effectively communicate concepts and ideas that are difficult to convey in print, but these media can also create problems for journals and their authors. The need to use new media adds to the work an author must undertake to develop an article. In many cases, authors do not have the expertise required to develop movies or other new types of media. They require expertise from media designers to produce an article. Either the author or the journal must supply a source of this expertise.

In many cases, authors do not have the expertise required to develop movies or other new types of media. Either the author or the journal must supply a source of this expertise.

Embedded multimedia in electronic journal articles will create new challenges for aggregators. Not only will they need to manage HTML or PDF text files, they will now also need to manage images, audio, video, and other media files. The most difficult problem this will produce for aggregators is how to preserve the new media. To preserve the media accompanying journal files it will be necessary to find ways to convert the formats of those media files to new formats as they develop. Given the number of different formats used for images and audio/video, this is going to be a significant task.

Summary

As OhioLINK continues to develop its Electronic Journal Center, it will do so in an environment that is increasing the benefits that can be offered to electronic journal readers and at the same time increasing the challenges faced by journal aggregators. The format of electronic journal articles is changing from PDF to HTML and those articles are becoming enriched by the ability to publish on the web with many different forms of media. The bodies of literature formed by these article form are becoming more tightly integrated and easily navigated by readers due to the ongoing development of links between articles and citation databases. While these developments enhance the value of electronic journal access to readers, the same developments will create new problems that electronic journal archives must overcome. The need to preserve complex multimedia journal articles will create a future archival problem. And the further development of standardized linking systems by publishers will challenge electronic journal archives' ability to function as an aggregator until such systems can direct users to both aggregator and publisher web sites.

Author

David Barber
david@ohiolink.edu

David Barber is Joint Technical Planning Director for four organizations: OhioLINK, a statewide consortium of university and college libraries; OLN (Ohio Learning Network), a consortium of universities and colleges promoting distance education opportunities for Ohioans; OARnet, ISP for Ohio's universities and colleges; and OSC, the Ohio Supercomputer Center. In that role, he works on technical planning for issues of concern to the four organizations, including video conferencing and authentication. Prior tothis position, he was Director of New Service Development for OhioLINK where he was involved in developing the technological infrastructure for new OhioLINK programs such as its Electronic Journal Center and Digital Media Center.

---